You are on page 1of 190
ne oes 450 9 | Earth Conduction Effects in Transmission Systems By ERLING D. SUNDE Member of the Technical taf Bou Teuermoxe Lasozarons, Tre. siving sto 1 of IVER PUBLICATIONS, INC, NEW YORK curren Contents Tateoduetion I, BASIC ELECTROMAGNETIC CONCEPTS AND EQUATIONS 2 22 23 a 2s 26 27 28 29 210 2 212 IIL RESISTANCE OF GROUNDING ARRANGEMENTS: ey 32 33 3a 35 36 37 Fundamental Concepts and Units. ‘Basic Fleccrical Phenomena and Cheracresistis, Fundamental Electromagretic Postalate, ‘Methods of Solution of Electromagnetic Problems,» ‘Paani Line Pramenss. ‘Teansmission Line Eau Sexi and Veevr Ponts of Stationary Hil. Hlecromagnotie Frergy Relations ‘Wave Equati in a Homogensous Medium. Basic Solution of Wave Equation. Internal Impedances of Cylindrical Conductors. - ‘Seif and Muwual Impedances of Cylindrical Conductors ‘Fourier Integral Transforms and Operational Calculus... IL FARTH RESISTIVITY TESTING AND ANALYSIS Tntroduetion Basic Equacions Electrodes at Suriuse of Uniferm Barth... cts iow Sura of Uifarm Ear. ‘TwoLayer Stratification “Thrve Layer Stratification... Arbitrary Stratification ae Exponential Variation in Resisivicy-- Deter Analytical Apprenizati Earth fy Measaremenss,. arch Resistivity Data... Dara Seat Hariznet Wis. Buried Radial Wires, - ios of Earth Resistivity 2nd Structure from Measure. of Meayuted Resisivities. pace SSERRESE Senne eesshees aaes suxeee 3 comarren f CONTENTS CONTENTS ix ‘VI. D.C BARTH CONDUCTION AND CORROSION PROTECTION 64 Istraduetion. 62 Corrsian Charters 7 cr 3:10 Equivalent Racias of Flat Conductors...» z 54) Inversey of Heterogeretic Cutrents 181 SUV Bifet of Vanvion in Resistivity with Depth. eee ae ist 312 Effect of Chemical Treatment of Soil... 6.3 Stray Conduction in Proximity Zones... i” 188 3.13 Earth Potentials Near Grounds. 6.6 Stray Between Point Electrode and Long Conductor. . 189 ‘L14 Hesting Eset. 6.7 Stray Between Parallel Conductors. . . eee ‘192 3.15 Greund Reslerance Measurements. 6.8 Parallel Conductors in Close Proximity. . 7 194 a EARIELRETURN fo EettefVarmion n Rev exc Gave 98 of INSULATED 60. Ecce of Variation in Resistivity on Stray Coirent. 198 1V. MUTUAL, IMPEDANCE G11 ce Coins and Baritene, 2 = 6:12 Protestve Sectonalzation — Insulating Joints... 200 4.1 Introduction B (613 Protestive Drainage — Cathodic Protection » SI og 33. Veruca and Witton Dipole Pies, eee {S.L4 Measarements of Conductor Constants, --...0c0.... 205 43° Bacth Return Mutsal Impedances in Geeta a ‘LE Menautements of Current and Potentials... 00... ‘or £4 Wires of Infinite Length at Serface of Barc ge 6:16 Cathodic Protection Test8.- secs. b at £5 Wat nce Lh te Sls oP ri E47 Steven Cando Survey. con 25 Wine of Infinite Length below Susface of Far 4.7 Infinite Wire on Surface of Two Layer Ext i ‘Vil, POWER SYSTEM EARTH CONDUCTION AND INDUCTIVE 4h Tednite Wire soows Sartre of Two-Layer Earch Eee Sere 49 Infinite Wire — Arbitrary Horizantal Stratification co 7 4.20 Tafsite Wire — Exponential Variation of Resistivity. - TA Introduction....... x fs ms TL Finite Wires, Uniform Earth and Low Frequency... 13 72. Coupling and Circit Terminology Se noee al als 412 Finite Wires, Two-Layer Earth and Low Frequency. «0-0 +--+ 126 73. Current Disribution Between Reta Conductors 413 Resistive Coupling at High Frequencies. «<3: r ZA Return System Propagation Characteristics... -s..0-- toe et eerretc ck Woranecal and Vertical Conductors... 129 33 Earth-Gurrent and Barth Potential Distribution. 14. Maraal impede of Yercal Conductor, a 1 Betery Sipkan Potentalns spotter over Te Transient Coupling of BaeicReraen Cities 0.000000 ey 2. Shel fran see Gro Cec 2 fa Radiation. .cocsvssersesec ors 1B Shielding Effect of Concent Shields and Tape Arco Ee Geol Elis in Bo ' eh ee of Terminal to Longitudinal Voltages... .. 197 10 Transmission System Admiteances snd Impedances Bs Y, PROPAGATION CHARACTERISTICS OF EARTH-RETURN 71. Metheds of Fault Current Calcalation....... me ‘CONDUCTORS ‘Li2 Protective Devices for Communication Systems. 148 $1 Invrdoction 5 0 713 Protective Measures in Power Systests..- VL at 52. Fundamental Egaasons ee ww 74 Basic Factors im Nove-Frequency Induction... 000S. 282 $3 Rigorous Solution for Single Conductor... us a Approximate Solution for Single Conductor.» i ‘VII. SURGE CHARACTERISTICS OF EARTH. RETURN CONDUCTORS $5 Bees Appronmains fr Single Concucnr. +e 81 Bawaba nes 284 56 Pacallel Conduetors «++ -++0+ ie 82 Relation of High-Brequeney sad Surge Cheractristica. 00.00) 258 52 ena sessnesasattte ass 33 Propagation Constant for Bored Conductors, sv-s-ccsvvvsscee 258 8 Shunt Enegzation is 54 Sune Current Propagation in Long Buried Conductors... 259 £5. Conductive Enema 5 i 5-Sunme CarenePropsatn in Short Bar Condor. 2H $40 Tective Enerpzation. 2 sss sons isa TT gh 86. Surge Impedance Characerstice of Long Buried Candictor.. 263 SAL Effect of Farth Resistivity Variation with Dept ee 8.7 Surge Impedance Characteristic of Short Buried Conductors -.. 266 512 inte Conductor in Pre Space onion Ta 8. Efect of Soil lnization on Surge Characterietisg.- os sesss. 26? $13 Horizontal Conductor over a Perfectly Conduct OR: <<-*" yy 8.9 Propagation Characteristics of Insulated Buried Conductors,.,, 269 524 Hover oat ee eee a i $40 Propegaon Contato Innate Atl Contos «... mm $15 Impedance of oe = 8.11 Sune Current ition along Aerial Conductors .. 4 E46 Amtema Ground Lowes... un ae 172 ure Propagation along hut ” CONTENTS 8.12. Sune Impedance Characteristic of Aerial Conducta. 813 Effect of Carona on Propagation Characteristics... Rid Reflection Points ia Apseal Cond nctow.«-sosec0 8:13 Aerial Conductors with Uniformly Spaced Grounds, Inreanverion, ‘Aspe from the numerous transmission problems encountered in the develop- ment and operation of extensive communication and power systems, number of interference and protection problems also arise because of the unavoidable conflict of sac systems wich nature and with each other. This conilice resides in the circumstance that the earth is iavalved, it sre measure, 38 8 seturn canductor for both types af systems, either curing normal operation ot fr at the time of faults. The earth also serves as a return conductor for lightaing currents, which may occasion disturbances in communication and power systems. Since the soil is electrolytic, corrasion may be experienced ‘on buried conductors where an interchange of current takes place beeween the conductor and the earth. Problems thus arise both in communication and power engineering concerning the protection of transnission Ynes and ‘associated equipment against breakdown or interference due to excessive voltages. Protection mast also be provided against personal hazard due to abnormal voltages imposed on communication eircuits or on accessible ‘youncied metalic structores ofthe telephone and power plant. In addition, buried metallic seructores must be protected against excesive corrasiost Strokes 0 Atal Conlucers. 94 Stokes to Grune. ie , 3S Methane of Stokes co Buried Cordnctors Be Ghar Chorens for Stokes to Buried Conductors 27 peta Sree Bad Gnas ee Valeage Equations for Long Conductors... 38 Be VEE rhe used Cable for Direct Strokes Sio Veltpes in Buried Cable for Strokes to Ground... SAL Velases in Cable fer Discharges Beeween Clouds. .- 9.12 Provesve Mixsues (er Buried Cable. : 315 Lighting Trouble Expectancy for Buel Cable. S14 Voleages for Stokes to Aerial Cable ". Ss Prometive Measures for Long-Distance Amrial Ca 3g Carrere and Voltages in Local Aerial Cable 39 Bere Meee Leal Agila 318 Listening Voltages Imposed on Power Transmission Lines Bre rencctne Measures or Transmisnan Lines.e-v-e-o-ss 2-000 APPENDIX: DEFINITIONS AND TABLES OF FUNCTTONS 1, Eapooenti spre and Leguithmie Functions a resulting from stray current ot other causes, 2, Bxpoventia Inegr! Paretion. +++ x “To deal adequately with such questions, ic is necessary to consider theo- 3 ee ae - retical solutions to a number of basic problems in which the earth, as well 2s $ Rema functions. s--s buried eonductors or grounded serial conductors, is involved. ‘The analysis 6. The Functions © 9 and @ es 5. Paired Frequescy and Tine Response Funedians, References. 38 of such problems is inherently more complex chan that af completely metallic circuits, for several reasons. The first difficulty is the great extent of the earth, which necessitates the use of electromagnetic field theory rather than conventional transmission line and circuit theory in the solution of mast of the problems, Moreover, electromagnetic field theory alone does not suffice for the solution of all problems, sinee ionization effects due to high voltages ‘or to electrolytic action may also be involved. ‘The heterogencous character of the earth as a conductor and an electrolyte must be considered in some instances. Furthermore, the phenomena under consideration may be of anexrremely variable nature, asin the case of lightning earrents and voltages, necessitating statistical sreatment. Finally, in problems of inductive inter- ference, acquaintance may be required with the performance and operation of circuits and equipment of both power and communication systems. iit INTRODUCTION ‘Te mighe appear that with the many ead variable factors which may be involved, theoretical analysis may not serve any purpose, since the idealized conditions chat must of necessity be assumed will not be met with odinatily jn actual sitaations. On the contrary, this very situation necessitates theorctical derivation of relationships between the various basic factors. Sach relations cannot be established from experimental studies or observa. tions in particalar situations, since the various factors are aot under control and may not even be adequately known, On the other hanc, experimental studies staged under known conditions are essential and have beck used to establish whether or not excessive deviations fran theoretical resules may be expected because of departures from idealized basic physical assumptions, particularly as a resule of the heterogeneous character of the earth, Lightning disturbances are deale with extensively in the literarore, largely 1 atmospheric phenomena governed by the electric field in the air. How- lever, as regards the behavior and the effects of lightning near the surface of the ground and in communication and power systems they are primarily earth conduetion problems and should be so treated for proper understanding, and interpretation of a variety of observations. For instance, flashes that sometimes occur to objects within the protective zone of vertical conductors struck by lightning, as well as lightning damage to buried cable land the prevalence of lightaing outages of power transmission lines in high resistiviey trritory, are readily explained when account is taken of che finite conductivity of the earth and resultant earth potential gradients ‘Ina general analysis as presented here, advantage may be taken of the common nature of many of the problems to broaden the aspects of both theoretical and experimental results, For instance, the formulas for the conductive coupling becween a drainage ground and an excensive buried cable, which are of interest in covrasion protection, are basically the same as for the conductive coupling through the earth between a lightning channel to ground and an adjacent cable, excepe that in the first case the resist- ance of the cable is of main importance whereas in the other case the inductance is. The same basic effect is again involved as regards the earth potential near an electrified railway track, except that the important cizeuit parameters are again somewhat different. In adition to such similarities kn the physics of various problems, there may be other interrelations. Thus, the protection of power transmission linea against lightning outages provices, at the same time, a reduetion in the frequency of occurrence and magnitude oFiinductive disturbances in telephone cireuits. Protective measures against lightning damage to the telephone plant may also be effective in relation 1 inductive disturbances and, ia some instances, may affect corrosion aspects. "Although it is necessary torestrit the analysis to fairly simple fundamental INTRODUCTION xii cases on account of the complexities that would otherwise arise, in man; case eis possible by ingpection and analysis of theoretial reals to meas, nize fairly general Principles: ‘or rules which may be approximate but will serve as a useful guide in complex situations. To obtain engineering for- lay, it has been in this presentation sometimes expedient t9 make tae of mathematical approximations not afecting the fndamental aspects of the comet oy ewe etrors ae ‘usually be estimated and, where it has n sible to make experimental compar is, val tain tf make ceil company, the formulas thus obtained ong the application of carch-rerurn circuit analysis is here confined ‘transmission circuit protection problems, some of the devfvations ore base also to the study: of ground-wave radiation fields and antenna impedances. However, radiation phenomena are well covered in literature, with the cx- ception of antenna ground loses dissssed ely herein, _ Communication circuits are used frequently to investigate earth potenti diferences beeneen remote point Ppl fom ear eurenes on a vol, wide scale. ‘These currents are due ty terrestrial phenomena associatéd with the motion of the earth and with disturbances of solar origin in its magnetic field and im the ionosphere. Under abnormal conditions, these effects may be associated with interference in radio communication and in earth-retarn telegraph or control circuits. On long telephone circuits, the resultant earth potectial differences may also cause breakdown of equipment unless adequate Protection is Provided. The theory of terrestrial electricity and magnetism is dealt with extensively in literature and not considered herein. CHAPTER T Basto Exzetromacyetio Coxcerrs axp Equanioxs ‘This chapter is a brief review of basic electromagnetic concepts and equations and of conventional methods of solution of electromagnetic field problems, The presentation, which is arranged primarily for orientation and convenient reference to problems dealt with in later chapters, presumes acquaintance with differential and integral calculus and with furdemental clectrical terms and notions. For a more complete exposition and. for other aspects of the subject matter than aze of primary concen here, reference is mnade to the literature on electromagnetic theory, transmission Tine and circuit theory. 1 Fundamental concepts and units Blectricity appears to be of an atomic nature, the smallest particle being the electron, which, by conventions established before iis discovery, has been assigned 2 negative charge Since charges and their motion ace responsible for electric, magnetic and cleexrodynamic phenomena, it is natural that some unit of charge should be at the bese of an electrodynamic system of units. The additional units required to specify che motion of charges and the forces between them 2re provided in the international ‘meter-Hlogram-second system used in dynamics. For practical purposes, the coulomb is preferable as 2 unit of charge to the elementary electron charge 1.591 X 10 coulomb, and is employed here together with the other units mentioned above in the MKSC system advocated by Giorgi and Campbell, ‘The basic derived electrical concepes are as fellows: Quantity Defrion Symtol Designation Curent Coulanb/Sec J Ampere Porential Fote/Ceulms 7,7 Yale Power Tonle See PP Wate Plate Intensity Vole/Merer 56 Vae/Merr Magecsie Inceniy Arpere/Mexe HH Ampere Meter Revstance, Tnpedance Volt/Ampre RZ Ohm Conductance, Admiteence” Arpere/Volt GY Mh Inductance bum Second 2 EARTH CONDUCTION EFFECTS cue. In formal relations between the above quantities, such as elementary laws established from basic experiments or more: complicated relations established therefrom by mathematical derivation, the units are as specified above. To comply with engineering practice, however, other units are tused at timer in illustrative examples or as coordinates on curves, Thus, 2 length may be designated in millimeters, centimeters, Kilometers, inches, feet or miles, although the lengths used in formulas are in meters. Simi- larly, it may be convenient to express a voltage gradient in volts per Kilometer or per mile or in kilovotts per centimeter, although the voltage gradient or fcld intensity used in formulas is in volts per meter. Tt alsois convenient at times to designate conductor constants in ohms, rhes, henries or farads per kilometer or mile, rather than per meter ag in the formulas. 1.2 Basic electrical phenomena and chavatterisics ‘Two kinds of forces between electrical charges are responsible for the physical recognition of electricity and its utilization for energy. transtnis- sion. For a given separation between the charges, one of these depends on the quantity of charge only and is referred to as the electric force. The ‘other depends on the product of charges and their relative velocity, or the electric current, and is referred to a8 the magnetic force. Associated with a configuration of charges may be visualized an electric feld of force, the intensity of which is specified by the force & exerted on a probe of unit charge at the point under consideration. Similarly, a current may be con- czived to have a magnetic field of force, the intensity of which is specified by the force IK that would be exerted on a fictitious isolated unit magnetic pole, actually nonexistent. Although the forces are of an electrical nature, they are measured in the same units as gravitational and mechanical forces and, like the latter, 2 vectorial quantities, as is the electric current. ‘A given electric or magnetic field produces a certain reaction within a medium which depends on three basic electrical characteristics. These are referred to a cube of the medium 1 meter on the side and are designated as follows: Conductivity: #, mho/meter, or resistivity p, meter-ohm Capacitiviey: 1, farad/meter, 8.854 + 10~* for free space Induetivity: », henry/meter, 1.257 - 10°* for free space “The conductivity of a medium is a measure ofits ability co transfer charge jn response to a given impressed electric force. ‘The capacitivity and Inductivity relative to free space are designated the dielectrie constant, fe and the permeability, respectively. ‘The departure of these quantities from unity, ¢— 1 and w— 1, referred to as the electric and the magnetic ELECTROMAGNETIC CONCEPTS AND EQUATIONS — 3 susceptibility, may he thought of as measures of the ability of a medium to betome peasize hy displncments in i seructursio estore eek send magnetic forces. Permanent magnets retain a partial or remacens ‘ragnetizaton upon removal ofan applied magnetic feld and « corrspene, ing but less pronounced phenomenon may be observed in diclectics, which ray have a epocary ral plareatn 7 ‘onductors are generally classified into’ metals and cleciralytes, the conductiviy of metals being due tothe presence of free clecerons, that of electrolytes ta she dissociation of molecules into positive and negative fone Eleccrelytes are subject to the same clectromagetic laws as metals although the current in an electrolyte is sometimes regarded as a convection rates then « conduction cutrent and is associated with chemical phenomena that are outside the scope of eicctromagnetie feld theory. The electrom: asic laws also apply co convection earrents arising from che motion of electrons and ions in gases, although explicit consideration of thelr mesa may then be recuired to determine their paths and distribution in space In the case of conductors, any effects of the mass of the particles are in. cluded in the empirically determined condnctivisy, inductvity and capac, {sot actually encountered except in free space, in many problems « median of aabetantal conduevity my be regarded as an insta. In ooher robles metium of low eonduetvty may behave virally a «perfect, Residing in the atomic structere of a substance are other clectrk chest ttn the ested on, whch sero a a effects observed upon variation in the state of a medium or upon a ance Py nen et a no belong in the feld of electronics and sleetrothewiatey. tod ne caened briefly hete, the motion of electrifed particles oven in obedience to bor ot predictable from, the electromagnetic laws aling, since othes ects are invaved then the imeracton of else and magne fa a 5,0 these phenomena beiong electrifeation by friction between different inslators of witablesufae texture, resulting in tanner of surface lee {rom oe instr othe eter, the decree conduct of metal eh Increasing wemperarte aya result ofa decrease inthe numberof cleans, end the existence of contact potential between diferent reel, which fs inherent in the fact that there isa difference berween metale ie the 18y required to remove an electron from a metal in free spaces Other Bhenomens ae the cert proce na cloned chain of aiferet metalic conductors as a result of conversion of thermal into elecerieal energy whens 8 junction i heated ar by conversion of chemie! into local rere an electrolytic conductor is inserted in such a chain heeween ifr 4 EARTH CONDUCTION EFFECTS Coane _metallic conductors or if « metallic conductor is inserted between different electrolytes in a closed chain of electrolytic conductors or cells. Within electrolytic cells, metalic ions, which are always positive, travel toward and are deposited on the metal at the negative boundary or cathode, while the negative ions travel toward the positive boundary or anode, where they neutralize metal ions and thereby cause dissolution of the metal surface. Other phenomena are the removal of free electrons from metals by appli- cation of an electric feld of high intensity, known as high-feld emission, by heating the metal in a highly evacuated space, known as thermionic emis sion, or by illumination of certein metallic eurfaces, referred to as photo- clectric emission. If electrons released into 2 gas by any of these processes attain sufficient velocity, by virtue of an applied electric eld of high in- tensity, 2 phenomenon known as jonization may take place, in which some of the ators lose an electron in colliding with fast electrons to become tive ions. Under certain conditions this process may become cumul and result in an arc between the field electrodes, with both che initiat electrons and those emitted from the atoms as well as the positive ions colliding with other gas particles to release additional electons. At ex. ceedingly high temperatures, che electrons of neutral gas aroms may acquire sufficient velocity 10 become deteched, so thac a gas may also become ionized by thermal emission, although this cffect is ordinarily of secondary importance. In certain crystals, such as quartz, the pitzo-clectric effect is observed, in which charges appear on certain surfaces of the crystal as a result of me- chanical pressure and deformation; in addition, a pyro-clectric effect, in Which electromotive forces are developed berween certain faces of the crysl by application of heat, may be noted. The piezo-electric effect is revers- ible so that when the eryses is placed in an electric field, i is deformed by variation in the field intensity. Aside from the above phenomena in which charges or their external electric forces are observed, there are the very important internal electric Phenomena, resulting form the electrical structure of the atoms, that manifest themselves by chemical affinity between the elements and are thus responsible for chemical combinations. 1.3 Fundamental eeiromagnetic postuisses ‘The electric and magnetic fields at any point in a medium satisfy cereain physical postulates abstracted from basic experiments’. ‘These may de expressed symbolically without reference to any particular coordinate system and provide the hasis of mathematical formulation and solution of electromagnetic problems?-**, First, the following relations obtain between the field intensities and the ELECTROMAGNETIC CONCEPTS AND EQUATIONS 5 density J ofthe electric current, ® of the dielectric x or disp. 4 of the magnetic Bux ar displacement™, 7 — 01 1.02) . (2.03) pane ae ‘and 9, cand» are the characteritis of the medium at the point and in the direction under consideration. Th tity xb =O the displacement eutrent which, like the foresees fhe diplacemenc current which, lke the conduction curene «6, produces Second, the quantitice 9 § and B behave as ini ible and i stmuetible Puids so that the net fux over a neds ee yfhen it include a soure, in which case the total outward fux equals that from the saurce. The surface integral of the fx per unit volume as the enclosing volume becomes infinitesimal is defined as the divergence of the Yeetor. Because there are no free magnetic poles and since displacement Guzen i include with conduction eure, the divergences of and are Iways zero. This is not the case, however, with the ic displace ment 9 sad che faliowing relations obtains SCet dl 9-2 (104) divg -0 (1.05) divB a0 (1.06) ‘hare is the charge density at the point under consideration, change in the magnetic feld gives rise to an electric field, and converecly, the fed ineensces at 4 point being intercl wal clectromagnetic equations ce oe curl 6 = —s3ie (07) ene K = 66 + ab (2.08) here : and IC are time derivatives of 6 and frst equation i an expression of Faraday’s law regarding the voltage induced in a metalic loop by the variations in che fax treading the loss, {In a plane normal to the direction of the magnetic fax as shown in Fig, 11 a closed path sis assumed to include an area 2 of such small dimensions that the fx density may bereganted as constant over dd. By Faraday's Jam; the integral feof the tngential electric force 6 around the path * Scie letesaro used hereto dasa nstantancus values of quanta wh ‘vary arbitrarily with time, Sass 6 EARTH CONDUCTION EFFECTS rar tis equal to the time derivative of the flux vid within the path. The curl is, by definition, the integral per unit area, hence (1.07). “The second equation is a corresponding expression of Ampére’s law re- garding the line integral of the magnetic field strength around a path en~ losing @ conduction current, augmented by Maxwell to include displace- ‘ment current. F Kats taken in plane normal ro the dtection ofthe tor current (o& + xb) a included by she path, as shown in Fig. 1.2, ‘The absence of magnetic conductivity will be noted in the frst equation and is dae co the lack of isolated magretic poles in nature, Actually, the 4 ti Jeattat fester Fon 1. right-hand sides of the equations should be ~ @= —»¢— 7, and «6 +9~ 06 + 6 +46, but for media or circuits with fixed characteristics, as dealt with here, time variations in » and x are zero. ‘The electric and magnctic field intensities in (1,07) and (1.08) are not rovessarly the totel eld incensieies ata point. ‘There may be, for example, Potationary electric field due to galvanic or therraal action, or a stationary magnetic eld due to 2 permanent magnet of to current in 2 loop insulated Jom the medium. ‘The oxtl of any such feld vanishes, however, 36 the leetzie fed is not accompanied by variation in the magnecic (hux, and the Imagnetie Held is not accompanied ky current at the point under considera. tion. Such fields may chus be subtracted from the !eft-hand sides of the equations without affecting their validity, as is sometimes done to indicate that the equations apply to the fields arising from the vasiation in the magnetic fux or due to current at the point under consideration “The field equations may also be written in integral form applying to any path, rather than in the above differential form spplying to a point. The ELECTROMAGNETIC CONCEPTS AND EQUATIONS 7 latter form is convenient in formulating differential equati ttl eens Che ‘lowing integal frm is ors metic Eomseriog with clocui Sous=—f faa 1.09) Shute = f f Gras 2.10) _ bs ins eter on sheen ses are eke aun the boundary of any surface f over which the normal nents of & areir freed on the ghee ssa cn a in Fig. 13. The integral form follows ce fom the diferential Tor by application & cof Stokes’ theorem, which seates that the surface integral of the component of the curl of the veetor normal to the surface fexual the line integral of the tangential component of the veeror alongs the bound ary of the surface. The field equations also apply at dis- continuities such as the interface of two media of diferent electrical characteris, fies. Tn order to eatsty the equations at 4 boundary, itis necessary for the tangential components of the electric and magnetic field intensities to be equal on the two sides of the boundary: ds Ftsds=- [finda Bre, 13. Pa Bu Gey Hay = Tae a) ring fh v us 2 wee tant ten ie ee ae Fla 14 Boondary conditions fo magnetic felds Hag = Hay ‘This flows when she Fld intensities are de f ld intensities are decomposed into normal an asgental components, at shown for Jin ig Lin Whe the mente Sarge clang x natrow angen loop which incdes rhe boundary the width of the loop is made to vanish, the area did under the loop and the incearal $C, de = dA vanishes. This condition can be stised only if 8 EARTH CONDUCTION EBPECTS cans the feld insensities are the same on beth sides of the boundary, with the Pie erception: if a conducting shee is placed between the two media and the conductivity of the sheet approaches infinity at the same time as the sheet becomes infinitesimally chin, the product of sheet thickness a Currene density may reach a finite value <. ‘The pach will then include eh current and the equations are then satisfied with Bes Gx Has Hae = E 12) here’ is the sheet current density measured in amperes per meter width of the sheet LA Methods of solution of electromagnetic problems aati rreated here generally deal wit Flectromagnetic problems of the type treated ‘ the resultant current, field intensity or pec in ee —— rimary field. ‘The latter may be a voltage applied to te erminals ee eras generally, a distribution of electric force along @ conductor crogithin a oeediam. ‘The primary Geld may be arbitrarily assigned although actually t2 specify the distribution of charges and Sa iat will provide a primary field as assumed may be an electromagnetic problem init Pit, For the purposes of analysis, the primary field is assumed to give rise ro a secondary distribution ‘of charges and currents and ree Tecondary field. ‘The resultant field is the sum of the primary and # secondary fields, ito, pa eae" aa) i Each of eh corresponding relations for the potential and the current. aie C19) met saesty gern (17) ed (1.08) wal = see i) atthe boundary of eo mets te = problems: jnvolving fields varying with the time, the snilyiie ee (aa fied considerably by the introduction of the familiar soidal" ime sett ‘sed in the symbolic solution of alternating-current problems & = BAt = Bicosat + isin wt) & - whe (ula) is li f cosine functions, which is related where £ is the amplitude of the sine and. h = he rootmeanaguare value E by #=V2E. ‘The amplitudes. and Feotmean square valucs are vectors like the instantaneous quantities, Inserting (114) and a corresponding relation for 3 in (1.07) and (1.08), shin name of the function, Which i la refered toa & hasmonic: fstion, i aeived focr the relation off = cost fsinw = 8x. ELECTROMAGNETIC CONCEPTS AND EQUATIONS 9 and eliminating on both sides 2, the latter equations become curl B= — fav (11s) cart H = oF + ioe (1.16) Variation with time thus has been climicared. The solutions may be extended, moreover, to arbitrary nonsinucoidal variations with time by methods described lator. Generally, the problems dealt with may be classified into circuit or net. work problems, conventional transmission Tine problems and two or three dimensional fica problems. fn cxcait problems it is possible to, isolate elements each having separately the familiar electrical characteristics — capacitance, conductance and inductance. These elements may be in- terconnected in a variety of ways into networks or circuits, and the currents or potentials in various branches in response to an impressed or electro- motive force may be found from a set of simaleaneous equations. These equations must be set up to satisfy (1.05), which in circuit theory is known 1s the first Kirchhoff law, requiring the sum of the currents at a branch point to be zero. Furthermore, to satiafy the field postulates or the second Kirchhoff law, the voltage around each closed loop or mesh, cluding electromotive Forces, must be 2010. Tn conventional transmission line theory, the field at any paint in a plane perpendicular to the transmission line is assumed to depend on the current and the charge at the point of intersection with the plane, but not on che ‘current at other peints along the line. With chis assumption it is possible to assign to a transmission line element the characteristics or parameters of longitudinal impedance and transverse admittance, and to formulate the current and the potential at any point by a differential equation known as the transmission Line, or the telegraph, equation. Ix deriving the transmission line equation, the distrfbution of the current or the field over the cross section of the conductors or in the surround medium is left out of consideration, and its effect, as well as that of dis- tributed charge, is assumed co be included in the longitudinal impedance and the transverse admittance. These parameters include the effect of currents and charges induced within or upon the transmission line con- ductors and surrounding or adjacent media, such as the earth. Derivation of formulas for the transmission line parameters necessitates the applica- tion of the eleecromagnetic equations to the electric andl magnetic field in the plane perpendicular to the line, a particular case of field theory. More gmnerally, field theory is concerned with the transmission of electromag- netic waves in three dimensions through space or media having specified clectric and magnetic properties, usually in response to current in 2 short conthietor, as én the ease of radiation, or to current in s long eonductor, as 0 RARTH CONDUCTION EFFECTS cant in connection ith nmutusl inductive effects between earth zerorr ductors. ; The solution of feld problems generally is based om 2 three dimensional “wave cqustion derived from the general electromagnetic postulates wit the concepts of scalar and yeetor potentials introduced to facilitate the olution, By suitable choice of a coordinate system, depending on the type of symmetry involved in the problem at hand, itis possible to secure Pe ematical solutions to a pismber of problems involving homogeneous rnedie or combinations of such medi han ‘On account of their application in later chapters, the transmiss and wave equations will be desived here, togecher with foumulas for the transmission Une parameters LS Transmission line parameters a “Assume that a potential 7 is applied berween the inner and outer con- duceors of 4 coaxial conductor arrangement a shown in Fig. 1.5. The eS © eurcTRI HELD Mucxenc FIELD io, 15 Comin condor arrangement f suctor and, at some electric Sed will then be radial around the central conductor and, at sme GFatance rom the ens will have no component in the axial direction. coary (1.05), the radial cverent density at che distance 7 most equa Jin) = Jolt oan fe the total current between the inner and outer conductors per ware Jos th 2 ore arg thine fo oh dec fre the distance eis ther by (1.02) 20 =e th ag) GF tos) Oar ‘The applied potential muse equal the integral or (7) between r= a and 1 dyhence. he : fe Ver Sie btajds 7 Dale + ue) ELECTROMAGNETIC CONCEPTS AND EQUATIONS 11 ‘The unit length admittance defined by Y = Ja/V is therefore yw OS ot tee ais) $ kee bre ee eas. oa 2, (1.20) é log= jogs where G and C are the unit length leakage conductance and capacitance. Assume next a current J slong che inner conductor wich retum in the cuter, and let the conductivity and capacitivity of the medium between the condtuctors be negligible in comparison with chat of the conductors, 30 that longitudinal current in the medium may be neglected. The longitadina’ clectric force due to the finite conductivity of the conductors is then related to the longitudinal current density by (1.02). Nopleeting the capacitance of the conductors $n comparison with their conductance, and assuming @ ‘uniform current distribution over the cross section of the inner and outer conductors, the applied potential per unit length required to sustain the cearrent against the resistance of the circuit becomes Vi, = Ra Ry) = IR (1.21) Ra and Rp are the unit length resisvances of the inner and outer conductors: which equal Ry = 1/y4q and Xy = Lend, where og and oy are the con- duetivities, da and 4, the cross-section areas of the inner and outer con- ductors. ‘When the current varies with time, an additions! impressed force ia required to overcome the voltage induced in the circuit by the variation in the magnetic foe borween the innex and outer conductors. ‘The lines of magnetic Aux ate conceutzic circles about che inner conductors, and it Fallows from (1.10), with [f° Ju 2d = 1,0 more dineetly from Ampire's law that H= Ltr a2) By application of (1.07) or (1.09) to a path consisting of the inner can- ductor with return in the owter, or by direct use of Faraday’s law of mag- netic induction, i follows thar the applied voleage per unit lengeh required to sustain the current against the inductive voltage equals W=1T te pion 8 ta loge (1.23) 2 EARTH CONDUCTION EFFECTS cnet “The unit length inductance is defined by L = F/T or Vial and eavals Pig’ 1.243 Lopes (128) “The vn length longicodinal impedance is the impressed electric fore we, quired, per unit Fength and for nie corent wo overeome the resistive and inductive voltages ancl equals La Rt ak 2s) “To derive corresponding expressions for & pair of parallel nos-coaxial conducts ies convenient to regard the fe as consisting of two radial cePspomente due 20 appli potentials of opposite sins givin:rse toa Zada) cerent Jo in one and —Jo in the other conductor, both with setur jo @ camote Cendctar a the distance D. If the conductors are separated by ihe distance d, the potentials of conductors, of a rains @ and, #7 bb (pre pre mo xml ° i ) Jo yy? le + tx) tee -f ya dale ff) e “The potential difference betweck the conductors Py — Py equals Ing Y= ale pie) Bab (4.26) % aay «.28) ‘The admittance is accordingly a The + aed 7 (1.29) ae “phe inductance ofthe constr exanaerneat ay be devel in a siilaz nmaer to thet ased in deriving (24) and equals 2 = pes 120) baal Consider a hypothetical plane midway berween the two conductors, 6 caeceesi. 16, and assume that the two conductors have equal rai om asthe potentiel and che elsctric fore inthe tangential dines, Tener a neo. For this rezon, Hf ee plant were condcting, Fe would ELECTROMAGNETIC CONCEPTS AND EQUATIONS 18 not carry current. IF the conductivity of che plane is asoured infinite, i ie farthermore permissible to remove the conductor on one side without dis trbing the feld on the other sie of the plane. The field equations will then be satisfied, since the tangential components of the field intensity slong the plane will be zero, as mast be the case for a plane of infinite condvetiv= ity, both for the static condition and when the feld varies with time, Fic, LS FleewicSeld of « pair of oppaitely chard conductors, Since the conductor potestial and the longitudinal voltage drop will be cnly half as great as with two similar wires, che admit inde ca ha as eat wires, ittance and induetance asi) L- pe (1.32) ‘The magnetic eld of a stationary current is not affected by che presence of a conducting sheet, and the sbove formula for L applies only for 2 vari- able Geld of & stationary field over a plane of infinite permeal Aper- fectly conducting plane acts as 2 reflector or mirrof, and the method of images employed in the above derivation is generally applicable to con- ductors over such a plane, 4 EARTH CONDUCTION EFFECTS 1.6 Transnsission line equation om Ti in Fig. 1.7, and Consider a short segment ofa transmission Tine as shown in Fig. 1.7, assume that the potential ¥/= (x) and the longitudinal current J = Z(x) vary with time as e. Tn accordance with the definitions of ¥ and Z the following equations then apply, when the positive direction of current is along the positive direction of the s-axis: at few 1.33) z ¥) q = 134) = a a iat on wi 0.x and substicution in Differentiation of the frst equation with respect to.» and substitution i the second give a differential equation for the current, whereas differentia ee eee ee Tor Bio. 1.7 Transmison line equations Us 442) ~ Ie) ~~ al = YFG Vig + ds) ~ Mp) = ~6P = Bld tion of the second equation and substitution in che first give « differential equation for the potential. These equations are ar apy 7. (1.35) ‘where Tis the propagation constant: P= ZY = UR + al)G + wor” (1.36) ‘The solutions of these equations are Ts pete Tg) = ae — Be mi Ve) = Kae + Ke ELECTROMAGNETIC CONCEPTS AND EQUATIONS 15 where the second equation is obtained from the first by use of (1.33) and K/=1/¥ is the characteristic impedance: Km 2/YPF ARE WEYG + ioe? (1.38) The integration constants and B are determined from boundary con- ditions which require continuity of both conductor current and potential at series or shunt discontinuities, except at a source point. At the open end cof a conductor the current mast thus be 2ero, while at the junction of sev eral condvetors the sam of the conductor currents must be zero. At very high frequencies there may actually be a displacement current fram the open end of a conductor which may need consideration, but transmission line equations as considered here axe not applicable in that case. The dis. continuity at a source point is equal to che electromotive force of the source less the impedance drop through the latter. Atthe junction of two conduc tors of different material there may be an electromotive force due to galvanic or thermal action, which must be taken into account in some problems, The current and potential distribution for a conductor of finite length energized at one end x = 0 and terminated at the other end x= s, through aan impedance Z, is frequently required and is obtained as follows: Assume that the current (0) impressed at x~ 0 is known, When (1.37) is applied to x = 0 and + = s, the boundary conditions to be satisfied are Ate=0: 4-B=1(0) Ate mas (4eT — Bel) Z,= KideP + Bel The solution of these equations is 1 4=1OT AW: B= 10) where jis the reflection Factor: os— K)/ (Ze K) (139) ‘The current and potential at « are accordingly ee) (1.40 ve) as) The impeiance Z = ¥(0)/F(0) looking into the energized end of the condictoris Let CLS ae 1.42) 16 EARTH CONDUCTION EFFECTS aan lerivation of the transmission line ‘equation the current was Tn the abr ei to one or more impressed electromotive forces be ‘at certain points, When the ate ix cual brs dis lectri¢ Aeld intensity £°(x) in the axial direction along the con witued fs ‘will appear on the right-hand side of (1.34) and with this snedification the current equation ts ar 1.49) -r = —¥P@) a we i. “The general solution of this equasion may be written 2 Toe) = [4 + Pe ™ BE Qe (1.44) (eq) = Kid + PON + BT ae where ron kf rome (4.45) ow Ae [Boer er (1.46) ‘The constants 4 and B are determined by the boundary conlions at qnds of the conductors or at source points, as before. -_ ee (ad) may te derived a (llowet, Consider an infinite condctr slong which there i an impressed foree (9), ‘The caren at see ce the electric force B°(2) de along the conductor element ais the 21) = Pe) PK. Ths allo since she Garett Implant k ja connected t9 cach end of the element de and the al impede ice ofthe laters celyible by comparion. The curren at due te Iimproned fore at os then dZ(@) exp — 11x — 2] and the tral current rue to the impressed forces between —* and = equals Pe) = ef Ors a) ‘current may next be canceled at any two points by two series genesa- on 30 aS or shen ‘@ conductor of eet grey Esra eatin iy be wriien as CP — BPC, a0 that the to by the genera ycror assured to extend from O in the postive direction nay be writen e ates Te) = fio BE Le [BO de = Lf + POET — B+ 9G) ~ BC e™ ‘To obtain (1.44), the constant =~) is included in B. ELECTROMAGNETIC CONCEPTS AND EQUATIONS 17 When the impressed electric force along a conductor vaties as UG + of)", P(x) and Qe) may be expressed in terins of the functions ‘and ¥ defined ia Appendix I and, when J ~ 0, in terms of the exponential integral function. With several transmission Tines « namber of simultaneous differential equations must be satisfied unless physical conditions permit lumping into 2 single equivalent line. Although the formal solution for rmultiple lines may readily be obtained®* in a similar or corresponding, manner to that shown in Chapter V for two conductors, considerable difficulty may be encountered in the practical application to more than tes lines of different characteristics. Using the formulas for ¥ and Z. given in 1.5, the propagation constant for a transmission Fine of negligible resistance becomes, Pm Tarlo + fee) 0.48) ‘When it is permissible to assume plane wave transmission and to neglect the conductor resistance, the propagation constant is thus independent of the geometrical configuration. Expression (1.48) for the propagation constant recurs in the wave equavion and upplies to the medium in which the conductor is imbedded. LT Scaler and vector potentials of stationary fields To derive the particular form of the field equations in a rectilinear coordinate system, consider a small reccanghe ddan the y.% plane around the component H, of the magnetic field, as shown in Fig. 18. ‘The electric force around the loop is then [@Fx/8y — 3«/0s] dydz which, in accordance with Fig. 11, must equal —ioB, dydx. From this and corresponding ‘equalities for other components it follows that for time vatiation as eM! cutl E = 08,/ay ~ E,/0z = —iewH, curly E = aE,/0x ~ 8E,fax = ~inlly a9) curl, E = 0E (0x — 0B fay = —twrl, Similarly; curl, H = aH,/8y — aH,/0: = oF, tines curl, H = aHs/03 ~ OH ,/8% = oBy + inky (1.50) oul, Hl allyfar — 0/89 = cE, + iE ‘When applied to a cube in a rectilinear evordinate system, the expression for the divergence, as involved in (1.04), (1.08) and (1.06), becomes, in 18 EARTH CONDUCTION EFFECTS cnr. the case of J = oF + fanB av f= SE (st) 2 i irecti because fowis the rate at which current in the x direction changes the ae ‘of current through the four faces of the cube parallel to the cee ith a corresponding significance as regards the other two deriva- ives. ‘When the integral pendent of the paths of the field intensity between any two points is inde- “follows that che integral around a closed path — ovine me pA 2M pwted 2B ae a “The field intensity and each af cur ofthe field — is 20. ani ence the curt of the Se the negative gradient of a salar its components may then be expressed ax potential U: E, © —ade U = ~ Uae (GQ --mvr - [+ 2+ By the definitions of divergence snd gradient, the divergence of fels desived from a scalar potential becomes au eu, aU divend U= Fat Ge tg 7 40 «1.82 (1.53) which is referred to as the Laplacian of U. anu ie static condition, J = 0, and the stationary condition i = ob en hat cat! B= 0, ‘The field may be derived accordingly ELECTROMAGNETIC CONCEPTS AND EQUATIONS 19 from a scalar potential which must satisfy Poisson's equation obtained from (1.01), (1.04) and (1.52), ie, aU = -O/s (154) where @ is the charge density at the point under consideration. ‘When the charge density is zero, a8 in free space, the potential satisfies Laplace's equation au=0 (ass) which also, in view of (2.05), applies to the potentia! in a medium carrying a constant current and, furthermore, to a magnetostatic field not accom panied by current at che point under consideration. Such a ficld may be due to a permanent magnet or to constant cutrent in a metallic loop ulated from the medium. Where there ig current or electric force varying with the time in the medium, however, tis seen (earn (1.08) that the curl of the magnetic force does not vanish’so that the integral of the magnetic Geld intensity between cwo points depends on the path. ‘The magnetic field intensity may then he expressed s the curl of a vector po- tential 4, the equation for each component of which corresponds to (1.54) and is Ad =» 56) where Jis the current density in the dizection of the component of / under consideration and y the induceivity of the medium. ‘This equation is derived as follows: With Brawls as7 (1.58) equation (1.08) becomes curl curl 4 = 9 From expression (1.49) for the cor), it flloms that a a eu, (cul A) = Seve A — 5 early 2 (ede 240) _2 (oe Bae ~ ay) eae 3x ‘Adling and subtracting 0°4,/a2, the following relation is obtained: cat (carl A) = ad, + Sev 4 Comesponding relations are obtwined for the yan components, so that the fllowing relation cbrains forthe total field: curl (curl 4) = —Ad + grad div d (1.39) —= ~ EARTH CONDUCTION EFFECTS Cane. ‘Asmentioned before, the curlof the gradient of ascalar potential vanishes, so that (1.57) remains valid ifthe gradient of a scalar potential is added to ‘A. This acalar potential may be so selected that div = 0, in which ease (1.56) is obtained when (1.59) is inserted in (1.58). ‘When 2 point charge Qo is imbedded in a uniform dielectric, it follows from (1.04) that the dielectric flux density D at the distance R from Qs Qo/4rR?, Hence the electric ficld strength equals Qo/4anR? and the poten a the distance Rs dR Qo Om nde HE tee a) The capaciy of « charged sphere of radius Ris accordingly 4xR, Expres sion (1.60) # a solution of (1.54). The more general solution for charges Gistributed in a uniform dieleetrc ia obtained by superposizion of the po- ‘ential due to individual charges and equals UR z Sf asi) where @ is the charge density at the distance R from the point under con- Sideravion and the integration is extended over the entire volume. "These solutions also apply to the stationary condition obtaining in the ‘ease of aconducting medium. A current is then required to maintain the charge distribution, as when current supplied to a cylindrical metallic conductor is discharged along its surface into & surrounding mediam of Comparatively low csnductivity, asin problems dealt with in Chapter IIL ‘Thus, if Ip is the current radiating from a point having a charge Qo, the current density at the distance Ris J = Jo/4nR? and the electric intensity due to the current J9/4zoR®, The potential at R is accordingly jo (dR __Io Grade TE Gro In the stationary condition both (1,60) and (1.62) arc satisfied, but the solution is expressed ordinarily in terms of che current, which is usually of greater finportance and is moze readily observed than the charge: ‘Corresponding to (1.61) a3 & solution of (1.54), the following volume ‘neegral represents the general solution of (1.56) for the vector potential cof a stationary current distibution: aa ve (1.62) “The volume integral applies to each component of the vector potential, J being the carrent density in the direction of the component under con- sideration, ELECTROMAGNETIC CONCEPTS AND EQUATIONS 21 Assume a short conductor of length 25 imbedded in a conducting medium of infinite extent and insulated from it except at the ends. When the con- ductor is made to carry a current J, the return through the medium will ‘onsist of wo opposite radial flows ‘rom the ends ofthe conductor element. Evaluation of the vestor potential shows that the contribution fram the two radial return flows is zero. For distances that ate large compared to dS, the volume integral of the current density may then be taken as 14S/R, so that Ids TER (1.68) A short current element as considered al i :own as i J 48 as its moment. ie aie i ‘From (1.64) the following expression for the magnetic flux intensity di toa dipole, known as the Biot-Savart law, is obtained by use of st “e pati = WE sng 65) 4 where ys the angle included between the radius vector R from the i s [e incl from the ai {> the poine ae which Bir evaluated aad the dioteon of he dipole aa, rom ths expression, the ogni fd due wo erent ins cenductor of acirary path nay be obsined by inerton along he Ph of con. 18 Electromagnetic energy relasions Icis postulated that the energy CV#/2 required to charge a eapecita to a potential Vis dntributed inthe surrounding imediam with the den sity «6/2. Similarly, thatthe energy LI?/2 required to create a magnetic fild by a current J in a ciceuit of inductance L is distributed as +C2/2. ‘The instantaneous energy within a volume is then given by the followin volume integral = WH af es pra 66) When the fed varies with time, the energy lose per aie time equals &O o —D ~~ f (bb + 0) de 187) Fram (1.07) and (1.08), —sC = curl 6 and xb = curl K — 66 Hence, oD — OP x fea &~ Sean) de+ [ose (68) - fi ‘diy [60] do + f 08 do (1.69) a EARTH CONDUCTION EFFECTS Cu. 1 where (1.69) is derived from {1.68} as shown below, and 16H] = 65 sing (1.70) je the vector product of & and SK, and 0 is the angle incloded between these vectors. The vector product is a vector perpendicular co the plane through Pr A Fro, 19 Ratited power deity Ps — [EH] ~ EH sin and X, as shown in Fig. 1.9. In rectilinear coordinates, t [SIC], = 6H, — EH; [BK], = BK, — 6H ys ' (6901, = B.C, — 6,50. a7 ; a a div {69€] = 5 (GH. — 6G) +5, CMe — EH) + EGR, ~ 57) We OH) g (2s _ he ~~ [8G 2) + 6G - 32) iy ae ed ¢ ae By >) ‘a, a8," (aby 38," + [-(B- 3) +90 GF a6 (3-3) =e eens (72) ‘The integral throughout « closed volume of the divergence of a vestar is equal by Gauss’ theorem to he surface intégral of the component of the a ee BLECTROMAGNETIC CONCEPTS AND EQUATIONS 23 ‘vector normal in the outward direction ¢o the surface enclosing the volume, i.e, the total flux from sources within a volume must equal the flux through the surface. Hence, if 4 is the area of the surface enclosing the volume, (1.69) may be written as follows: - 2 = fiends fos? ao (7a) ‘The firs incegral represents the power radiated through the surface, the second integral the power dissipated in heat within the volume. The ‘vector , = (6H) a7) represents the instantaneous power radiated per unit area across « surface through & and 5 and is referred to as the Poynting vector. For a steady state Geld varying as cos ai, with a time phase displacement «9 between the electric and magnetic field intensities, the average dissipa- on P, per unit volume and the radiation P, across a unit area over a long period f are given by 7s) P= BR! [Ritemateos (ou beddy toe = YOR cos. sin = EH cos gin 6 76) where # and A are amplitudes and E = B/V2, H = A/x/2 are root- ‘moan-square values. ‘Near a power source, ¢ ney be nearly 90°, so that the eneray in the field ‘oscillates between the electric and the magnetic state, wich only « minor fraction being radiated. At remote points, however, approaches zero ‘0 that the field is a radiation field rather than aa induction field, 28 near ‘power source. In order to evehiate the reactive effect of the field on the primary power system and the radiation from such systems, i is necessary to derive the electromagnetic wave equation and to obtain an appropriate solution, as outlined in the following sections. 19 Woase equation in a hemogencous medium “Yo derive the equations for the scalar and vector potentials of fields varying with time, as e, the magnetic field intensity is again related to the vector potentia) by (1.57). When the latter relation is substituted in sz w EARTH CONDUCTION EFFECTS Cass. (07), curl (E+ iad) = 0 co that E+ isd is the negative gradient of a scalar potential U. Hence Em ~ied — grad U am which, when substituted in (1.58), gives, in view of (1.02), carl curl 4m =? = Fae u a.7ay here oP = favo + ie) (1.79) ‘Wieh (1.59) in (1.78), ~ad+enddivd yd — Tend 0 (1.80) ‘This cquation is satisfied when the following relations obtain: ad a7 asi) au = 70 1.82) diy Gad) = 7° (1.83) Equation (1.83) is obtained by observing that to satisfy (1.80) assuming (1.81), itis necessary that grad div 4 = ~ (7?/ia) grad U, which is identi- cal with (1.83), Equation (1.82) fellows from (1.81) and (1.83) when its obscrved that to satisfy (1.08) for a homogencous medium, in which there are no body charges, i is necessary chat div Z = 0, Heace, in view of (LIT), div (ied) = —div grad U = -AU 84) ‘which, when substituted in (1.83) gives (1.82). The solution of (1.81) and (1.82), together with (1.57) and (1.77), gives the magnetic and electri field intensities. “The potentials and U and the fické intensities snay also be expressed in terms of a single vector function IZ, usually referred to as the Hertzian ‘vector, each of the three components of which must satisfy the wave equation an= (1.85) ‘Equation (1.83) is satisfied with 4 U= -divt (1.86) ach component of the electric feld intensity and the magnetic flux density ELECTROMAGNETIC CONCEPTS AND EQUATIONS — 25 is then related ¢o the corresponding component of by B= 90+ grad div IE (1.87) a Post a Tn rectilinear coordinaces, E,= —7lle Dan (tle 2 oe a Oe A solution of the electromagnetic wave equation is represented by the following volume integrals: L per vp foye » on ~a lI ® (190) fay ee -plIge where Q and J are the charge and current densities in the elementary volume dv at the radial separation 2 from the point under consideration, and the integrals for 4 and II apply to the components of these vectors in the direction of the current component under consideration. These solv tiens are obtained by introducing polar coordinates and finding the current due to # single charge or current clement, from which the sclution follows by superposition. For # dipole in a uniform medium of infinite extent, the formula cor- responding to (1.64) is was oe =e ast iol as 1S 192) Integrals (1.90) presume a known charge and current i Ordinarily, however, the problem is, in effece, to determine the distribution, —s= 2% FARTH CONDUCTION EFFECTS cunt andy in turn, the potentials and field intessities. This is usually accom. shed seine manner outlined in following secions, without explice use of 19). 1.40 Haste solution of wave equation Jin dealing with problems involving cylindrical conductors and combina- tions of cylindrical conductors and media of great extent with plane boundaries such as the etrth, it is convenient to introduce cylindrical Fi 110. ‘elaon of efi and reir erat, coordinates as shown in Fig. 1.10. Equation (1.85) then takes the: form? * 16/4 1én, em 18g LT ax) CG)+tagta? To solve this equation, it is assumed that the solution may be written as the product of three functions, each involving only one of the variables: T= RP)Z@) (1.94) Substituting (1.94) in (1.93) the latter equation becomes 1 af am 1 1 eZ _ 11.95) ane) t Poet zee? a) Since the first two terms and the right-hand side are independent of 2, the third erm mast egal some constant independent of = which wil be designated m?, in which case the equation for Z and its solution is 22 a Dede + Be (1.96) a Equation (1.95) then becomes de £ 26,2) LER yp — my? as7) Sale) eae ot ELECTROMAGNETIC CONCEPTS AND EQUATIONS 27 in which the frst term and the right-hand side are independent of @, 20 that the second term may be represented by a constant X* not involving ¥. ‘The equation for & and ies solution is then Fe rt ee ae ae 98) Equation (1.97) sow becomes ra) +08 - ot aia <0 nh b? — G4 = myrIR 0 a.99) which, with ® = én and 4? — 1? = u2, is brought into che customary form ‘of Bessel’s equation for a function ® = R iru): aR cs e Fea tO Faas tured — aR = 0 4.100) The general solution of this equation is Rr) Aafulrd) + By¥alr2) (1.101) uy where Jo = Besse! function of the first kind, order » Y, = Bessel function of the second kind, order These and various associated Bessel fonctions are considered in more detail in Appendix 1. ‘The solution of (1.93) may now be written A,coune\ | Ae ArJalra) m=(+ “t+ {+ (1.302) By sin ne) Bet BY (ru) where a = (2 + 92)", In applications to specific problems for which this form of solution is valid, some of the conscants 4 and B are eliminated by physiesl considera tions of symmetry and of convergence of the solution for large and small distances. The temaining constants are determined from the boundary conditions of equal tangential eleceric and magnetic field intensities at the interfaces of different media. ‘The constant 2 will also be determined by such considerations and will be zero when the field is independent of g. When the field is independent of 5, w is eliminated since @ = 0 so that w= iy, When the fcid is independent of ¢ and r, = 0, 1 =0 and -—e 2B EARTH CONDUCTION EFFECTS Cann a= 7. Equation (1.93) zhen simplifies to the transmission line equation (1.35) and It = 4e-™ + Be, With the ficld depending on s, elimination of wis achieved by integrating (1.102) between appropriate Emits which depend on the integral representation of che impressed ficld, but in general may be written as Aycosme’\ (fee) J fw Inl) m=(+ 7 a2 lt de (1-103) By sin nel Jo \ > gitine JN galu¥ate Integration between finite limits may be secured by taking the f and functions zero outside the limits of the integration range. This solution jneludes (1.102) as a special case in which these functions are zero except for a single value of a at which they are equal to the constants 4 and B in (1.102). ‘Expression (1.92) for the I function of a dipole in a medium of infinite extent may be written in che integral form (1.103) es fend dS fu as See hem deira ae (1.108) wherer = (7 +72)! anda = ( + 94)". In problems involving a horizontal or vertical dipole above the earth, (1.104) is used as the II function of the primary field ta derive che resultant Tifunetion. Formulas for the Ul function of vertical and horizontal dipoles snd for che mutual impedance of horizontal earth return circuits, which may be regarded as composed of such dipoles, are derived in this manner jin Chapter IV. 111 Internal impedances of cylindrical conductors Tn later chapters reference will be made to the internal impedances of cylindrical conductors and shields when the frequency is such that skin ‘Hfect must be considered, which may be the case even at low frequencies when the permeability is high. For a long cylindrical conductor along the ‘axis with internal or external coaxial retorn, II must be independent of ¢ and zand will have a componentin the direction only. Hence, in (1.102), n= Oand m= 0, 0 that w= fy and W =H, = 4, Jolin) + B-¥olh7) = Aglolrr) + BoKoler) (1.105) where fp and Ky designate Bessel fonctions of the frst and second kind fer imaginary arguments and are defined in Appendix 1 ELECTROMAGNETIC CONCEPTS AND EQUATIONS 29 ‘The magnetic flux intensities derived from (1.89) are Bom Fldoli(or) ~ BoKi or] 4.106) Tf the conductor has an inside radius ¢ and an outside radius a, the mage netic feld intensity at r = & must be 0, provided that the conductor current T returns outside che conductor. At r= the magnetic flux density, fox the other hand, must equal vJ/2ea. Hence, with dly/dr = Jy and aKo/dr = —Ky, Acid) — BoKu8 D = In(ya)Ks(yb) — L1(8)K (va) ‘The electric force becomes Ey = —7"t1 or jel Es = Fp Hot Kilnd) + KolrrdvGr8)] (1107) ‘The negative ratio of the electric force along the outer surface r = a to the total curent is defined as the external surface impedance with external return Zr, and the corresponding ratio for r= 4 as che intemal surface impedance with external return, Ziq. Neglecting the capacitivity of che conductor so that 7 = (iava)''®, these impeciances become 1 fiw! Pee = Fad ®) Hoya) Ka (v2)-+ Ka(a)!a6)] (1.108) ‘aa\2 2u= 5 (® Mol) Kx G8) + Kol Li8)] (1.109) ‘The internal impedance with internal return is obtained by interchanging a and & in (1.108), while the extemal impedance with internal zeturn Zei = Zee: For solid conductors (6~ 0) Zee is usutlly referred to as the internal impedance of the conductor Z, a8 distinguished from the ex- -- 0 EARTH CONDUCTION EFFECTS Cure ternal impedance due to magnetic ux outside che conductors as considered in the next section. When the Bessel functions are replaced by their expansions for large valucs of the arguments 72 and 7, the following approximare formulas are obtained, which apply within a few per cent for tke type of cylindrical conductors ordinarily enoountered. Bug = Ry coth ao 110) Pag = Rrifsioh 8 qa) Zi = Rre/ts b= 0,1a>1 tia) where R isthe direct current resistance and y may be expressed in terms of the latter as 7 = (iur/2nt6R)"? when the thickness 8 = 0 — Fis jess than about 259% of the mean zadius = (@ + A)/2 112 Self and mutual impedances of eslindrical conductors Ta dealing wich the mutual impedance of two circuits, ft is convenient to designate one as the primary and the other as the secondary circuit. The ‘mutual impedance is the ratio of the voltage in the secondary to current in the primary, and it is he same regardless of which circuit is the primary. ‘The letter proposition, known as the reciprocal theorem, applies to linear cirouits oaly, in which the muvual impedance does not vary with current. ‘The external impedance of @ round conductor, which is the impedance due to external magnetic flux, is equal to the mutual impedance between the conductor and an infinitesimally thin insulsted wire at its surface. The self impedance of a conductor is the sum of the excernal impedance and the internal impedance, as considered in the preceding section. The external impedance depends on the characteristics of rhe medium svrrounding the conductor and, for a conductor in am insulating medium, includes 2 com- ponent referred to as the radiation resistance, as discussed in the following. With a dipole 28 along the z-axis, in a uniform medium of indnite extent in all dircotions, the electric field incensities in a plane through the dipole obtained from (1.92) and (1.89) are ras[-riw) +5 008)] (1.113) Ey a E, = 1a5 375, O(R) where r = (9 + 9?) and, with R= (+7), pay - 1 oe a@ R! "get ROM ELECTROMAGNETIC CONCEPTS AND EQUATIONS 31 ‘The voltage dV impressed along an element ds of a second wire lng 9 paths in the same plane as S, as shown in Fig, f,{1,is then Ey cas ¢-+ Ey sin ¢, where eis the angle between dS and ds. Hence, with dz = d9 oF = 148 ae ~PiReor eZ O6Reos e+ SX ORsin | =128a(-Pipenet 2 [2B ces Mone) asl a . = -1asai[ Peajeose+ 5% 008] (15) Pea 111 Manual impedance berwcen two wines S aes exeting Som 4 to B and 2 to Leia zane aa ‘The last relation is obtained by observing that the square bracket term é R 75 0(R)y where y is the angle between 28 and R, in the preceding expression equals 4 Q(R)[eas y cos « + sin ¥ sin = 2 Zotwem o-9= The mutual impedance between two wires and s of arbitrary path as shown in the latter figure may now be waitten Zan ff [Pm net A ow|ases = ff Pa coseds ds + dame» (att6) Qa—wus—» = Ga) ~ O(AB) + (Bd) — QtBa) (1.117) wy 32. EARTH CONDUCTION EFFECTS Cuan. t in which 4a, 42, Bb and Ba are the distances r between the four terminals of the wires. The doable integral in (1.116) is che mutual impedance between the wire paths alone, while the second term is independent of the wire paths and Tepresents the impedance becween the wire terminals through the surround. ing medium. For a straight conductor of length f and radius @, having no interchange of current with the outside medium execpt through the ends, the external impedance beoomes a Lf PUR) de de + 210(0) — O01 (tts) where R= {(¢— 2)? + 2°]! is the radial distance between the center of the conductor element de at x and the surface of conductor element dt at Tis assumed that the current at x may be regarded as concentrated at the center in calculating the electric force due to this current along the surface ‘atu, which is permissible except for short conductors of large eross section, which are not considered here- Expanding «7 as 1 — 7R + (PR)F/2! — (R}/31-+-+-, the follow ing formala ie obtained for the external impedance: Vat fod (2 3) _ iawn Z Met ee (ose 2) fern OAD) “The first term is the impedance of the conductor terminals, which, in the case of a nonconducting medium, is a capacitive reactance. The second term is the external inductive reactance of the wire with direct current return in the medium. ‘This inductance includes a term — tuslv/4a from “20(D. The third term is the first order approximation of the effec of ‘currents induced in the surrounding medium. For an insulating medium ‘the latter losses appear as radiation and the last term is, in this case, the ‘radiation resistance, which may also be written, with y = ioGox)"?, 9 = (Cre), w = 2rf and d= Hf, anieh) GQ) = (5) (1.20) where A is the wave length. This formula applies to 2 short: conductor Carrying the same current throughout its length. When a short conductor ELECTROMAGNETIC CONCEPTS AND EQUATIONS 33 is energized in free space, the current will Gecrease linearly from the poine of energization toward the ends of the conductor. The average current is then half as large as in te above case. Hence, ifthe radiated power is cxpressed in tema of the input current, rather ¢han the average curen the radiation resistance becomes a Ry = 20e? & (aay As the separation berween primary and secondary increases, the second term in (1.115) decreases as €7®/2° and becomes negligible in comparison with the fist term. The electric force normal ro Rin a plane through the axis of » dipole then becomes, with 48 = /, ¢ ye Hod ute amy where ¥ = 3 is the angle between /and R, of propagation, and d = o/f'is the wave length. ‘The magnetic ‘field intensity may be obtained from (1.92) and (1.89) in wy manner to that used in obtaining expression (1.113) for the intensities, With Z = HL, H, = 0, Hy = 0 in (1.89), the total magnetic field intensity H = (HE + Hy)*, i.e, the magnetic field in- vey normal to R and to the plane through the axis of the dipole, comes awl sin tans 4eR me E (Um)! is the velocity a iisiny oo ar ORR # when yR 2 1 .t23) Ig will be noticed that, when 7X2 1, as assumed in (1.122) and (1.123), the electric and magnetic field intensities are related by B-= (e/x)\2 ot Et = ol (1.124) ch shows that, under the conditions stated, the energies in the electric and magnetic Belds are equal. “The power radiation per unit area normal to Ris obtained from (1.76) with ¢ =@ and @ = 90°, since E and H given by (1.122) and (1.123) are me 34 EARTH CONDUCTION EFFECTS our in temporal phase but in spatial quadrature, and equals where Po is the total radiated power J°R, obtained by use of (1.120). 113. Fourier integral transforms and operational calewias Boundary value problems may frequently be solved by representation of an impressed field by an infinite integral, as ia the preceding sections. ‘This is also true of problems involving an arbitrary variation of the field with time, rather than in space. In the latter case the solution is accom- Wished ordinavly by finding the response characteristic co the simple time function ¢* assumed earlier. ‘The impressed field or current ie then sustained, and the steady state response, which slso contains the factor #4, may be characterized by # ratio F(a), a fanction of ie but independent foftime. The latter ratio may be the impedance or admittance at the point of transmission, the transfer impedance or admittance to some othet point, for the currenr or voltage transfer ratio, The response to an arbicrary: Jmpressed field or current may be obtained, in torn, by a Fourier integral sepresentation of the letter in terms of e#, as discussed. below. Tf a decaying exponential time fonction ¢* had teen assumed, the response characteristic as a function of # for a given value of a would have sufficed to describe che response 20 other impressed forces, provided thar the latter were represented by integrals involving ¢~**. Of particular interest is the unit step fanetion obrained with a = 0, which is assumed initiated at ¢ =0 and is unity thereafter. When the response cherae. teristic 5) for such a step function is known, she bekavior for an arbitrary cause may be evaluated by regarding ths lareer as composed of infinitesimal step functions. A third elementary function frequently used is the unit impulse, which i a highly concentrated field of infinitesimal width in time or space and of such intensity that the integeal or area of the impulse is unity. When the unit impulse response 5”) is known, the effect of any other cause may be evaluated by regarding the latter as composed of impulses. Various interrelations between the responses to the thros elementary time functions, and between the responses to these fonctions and to arbitrary causes, will bediscussed briefly here. ‘A Fourier integral representation of a Function is the limiting form of che series obtained by extending the integration rangs to infinity. As a result, an integral over & continuous frequency spectrum is obtained rather thar a summation of terms in discrete harmonic frequencies. Subject to eestain = _ ELECTROMAGNETIC CONCEPTS AND EQUATIONS 35 restrictions a funetion /(@) may thus be represented by the following, integrals in whick p = A: 1p WO = 5 fe [fe cng) de Efe Liwie + an =E fla flrme a (1.126) ‘The last integral is obtained by change of variable from ~p to p in the ¢ bracket term of che preceding integral. It is then observed that the gral of this component is identical to thar of ehe first bracket term. Let ¢¢p) zepresent che integral with respect to of (1.126). The double integral may then be written in the following paired form: Jone Lf" vee az) with eo fl yorta 128) sphere exch of the last two integrals is the Fourier transform of the other. ‘The function f(¢) nay have only a finite number of discontinuities and muse be such that the integral of the absolute value|/(?}| of the function be- tween —w and ++ exists, At points of discontinuiry, the integral will eve a single value equal to the average of the function. Assume next that the frequency response for the sinusoidal condition is F(p), with p = fw = 2rif. This response might be the impedance at the point of transmission, the transfer impedance to some other point, or che current or voltage transfer ratio, Then, if f(t} is an impressed force, to apy term 4 in (1.126) will correspond a response F(p)e2-", The regultant reapanse to f() is accordingly ow =2 fla [puree ae oo = 2 [le emrea (1.129) wich oor) = [" cme ar 90; _s 36 BARTH CONDUCTION EFFECTS Cuan In particular for a unit step cause, f(!) = 0 when ¢-<0 and 1 when £50. Ie this case #(p) obtained from (1.128) is U/p and the following relations obtain, when the response to a unit step cause is designated by 8. sood flet@a ast) ws FO) _ ft pn FO fo enswa aus where the lower limit —-» in the last integral has been replaced by zero Shee S(@) = Owhen? <0. ‘The function Fp) is known as the operational ‘equivaient of $(¢) for unit step current, and the relationship expressed by (1.132) as the infinite integral theorem.’ F(p)/p is the Laplacian trans- form of $(). For nee impulse, f(¢) = 0 except at ¢, where it is s0 defined that the area under the pulse is unity. Hence (1.128) becomes @(p) = 1. Between #(p) and the response characteristic to @ uait iinpulse the following rela- tions thus obtain: © Fp) a (1.133) with Foye [soe ast) For the above integrals to converge itis necessary to place certain restrie- tions on F(p). Thus, for a unit step voltage, (1.128) does not actually converge except as a limit of an applied voltage é~** when ¢ —+0,sinceother- ‘wise the integral of |/(@)| does not exist. For this reason, the integrals in the above form are confined to circuits in which the current in response to une sep volage ultimately decays to zoo, but sey sre adequate for fanctions F(p) of the type dealt with in later chapters ‘The response H1(¢) to an arbitrary cause C(¢) may be evaluated from the response to unit step current or unit impulse by application of one or the other of the following integrals, which are obtained by superposition of the responses to the elementary time functions: ue) = cos + [se- NOs acoso+ [ sace-na £ Bal = CH)50) + fe NEG) at =cnsot f stoee- ae Wee eee ELECTROMAGNETIC CONCEPTS AND EQUATIONS 37 For cettain functions G(@) it may be possible to find tie operational equivalent F\(p) from (1.132) or (1.134), in which case the corresponding time fanction is given by eh ft FOG) He = 5, fe FOR (1.136) where F(p)Fi(p) is now the operational equivalent of (4) when the cause fanetion is C@). For comparatively simple circuits the response may be found from (1.136) for other than unit step or impulse conditions. For frequency functions ax dealt with later, however, it is usually necessary to evaluate the response ftom (1.135) after Sor S/"has been derived. Even the latter response characteristics may not always be obtained from tables of paired Frequency and time functions,"° but they may be evaluated by approximate methods usually adequate for engineering purposes. These methods may consist ix finding the time functions Sp and Sw for small and large values cof time, which correspond to the frequeney functions F., and Fg for large and small values of 7. The approximate response may next be obtained by plotting Sp and S~ and drawing 2 transition curve between them. A check on the accuracy of the rime function S(¢) thus derived may be obtained by a numerical integration of F(p}/p in accordance with (1.132), for a value of pin the transition range. ‘The procedure outlined above is usually less laborious than a cumericel integration of (1.131) to obtaim St), on account of the complex quantities invelved in the latter inte- gration. ‘A uumber of associated responses to the three elementary functions are listed in Appendix I. - Ve CHAPTER Barri Resisriviry Tesino ano Ananysis 2A Introduction ‘Among the electrical characteristics of the earth, only the permeability may be regarded as constant and equal to unity. Alchough the dielecerie ‘constant varies within rather wide limits, between 1 and $0, depending on the nature of the soil or the rock, it is not ordinarily a factor of first im- portance in the kind of problem dealt with here, even at very high fre- ‘quencies. The dielectric strength of the soil is of importance in lightning phenomena but varies within fairly narrow limits, whereas in corrosion problems the chemical constitution of the soil must be considered. The resistivity of the earth, however, varies within extremely wide limits, be- ‘tween 1 and 10,000 meter-ohms, and is of decisive importance in all prob- Jems involving the earth asa return conductor, such as i studies of ground. ing, stray current conduction and corrosion, inductive interference and lightning protection. Because of this wide variation, it is usually neces- sary to measure the resistivity in the locality under consideration, and the ‘theoretical analysis presented here is of importance in the interpretation of the results of such measurements. Similar tests ere also employed in certain types of geophysical prospecting, but their interpretation often requires cansiderably more detailed anzlysis than is necessary in connec- tion with the problems dealt with here. "To measure earth resistivity, it is natural to pass current between ewo lectrodes in the earth and messure the voltage between two intermediate potential electrodes, at indicated in Fig. 2.1, in the same manner as for ordinary conductors. In most cases, the resistivity of a considerable ‘volume of earth is requized, so that the electrodes must be a good distance apart. ‘The dimensions of the electrodes therefore will be small compared t the distance betweer. them, so that the current distribution is practically the same as when the electrodes are considered as points, Since checurrent, paths are not parallel lines, the resistivity cannot be obtained from the ‘measurements of voltage between potential electrodes in she same simple ‘menner as for small cylindrical conduetors. For a uniform earth, however, the resistivity is obtained from « simple relationship involving mutual re- sistance and electrode spacing, When several measurements ere made for Gifferent electrode spacings, however, the resistivity obtained from this Fr EARTH RESISTIVITY TESTING AND ANALYSIS 39 simple relationship usually varies, sometimes to a very marked extent, because of increase in the volume and consequent variation in che character of the earth included in the messutements. Measurements have indicated chat such variations may often be accounted for if the earth is assumed to have a surface layer of different resistivity from that of the earth below. With such an earth structurc, the voltage between potential electrodes depends on the resistivity and che depth of che surface layer and en the resistivity of the earth below. It is evident that by proper choice of these three variables, @ great variety of observations may be approximately. represented. Although the theory fex such a twolayer earth is easily Mia, 21 Gienitarangement for mesmrement of eth resisiviey. extended to three or more layers, too many variables ate then involved to Justify its use except in connection with geophysical prospecting, where the detailed structure of the earth is the main objective, rather then an ‘overall approximation asin mest problems deale with here. Certain limit: ing cases of three-layer structures may serve, however, as a aseful guide in these problems. The cart resistivity may also be derived from measurements of mag. nitude and phase angie of the aiternating-curcenr mutual impedance of wires laid on the ground. Except when the wires are fairly long, the rea tive component of the impedance is not materially affected by earth resi tivity variations. ‘Thus, the additional information secured by measuring the reactive component is ordinarily of insufficient value to warrant the use of this method, particularly since it requires more complicated equip- ment and mathematical analysis. The resistivity near the surface may also be derived from measurements of the wave tilt of radiation fields, as discussed in 4.17, For a horizontally stratified medium consisting of two or more layers, the ome 0 EARTH CONDUCTION EFFECTS Cane. lectric ficld due to current entering at « point source has been stated by Maxwell! in terms of an infinite series of images, whereas the solution in the form of an inzegeal involving Bessel functions has been given by Oliendorf and by Stefancsco and Schlumberger.® Formulas in terms of images have been published by Hummel,‘ corves for the potential dis. sribution by Weaver, and numerical tables for the calculation of such curves by Roman®; Muskat! has given an analytical discussion of the formal solution and obtained certain expansion formulas. For the case of arbitrary variation of the resiscvity with depth a formal soketion has been. given by Schlichter,® who, together with Langer,'?° has considered the irect solution of the inverse problem of finding the variation in resistivity swich depth when the surface potential distribution is known. For uniform earth, 2 comprehensive treatment has been given by Carmpbel!” of the magnetic as well as the electric field due to divect current in wires berween point electrodes on the earth’s surface. In earth resis= tivity measurements by methods considered here, it is usual co make petiodie reversal of the current, ani zhe resulean¢ variation of the magnetic field gives rise to inductive effects which may influence the results in certain cases, The question of inductive effect, however, is deferred to a later chapter. As for electrolytes in general, che resissivisy of the soil decreases, with increasing temperature and is independent of field strength and fre- quency, except for extremely high values. (References 2, 3, Chapter VI.) 22 Basic equations ‘The eursent and potential distribution in the earth in response to ‘pressed electromotive forces must satisfy the electromagnetic feld equa. tions of Chapter IL. For the particular case of a stationary current ente.. ing a uniform or a horizontally stratified earth, over « spherical electrode cor & cylindrical electrode perpendicular to the surface of the earth, the gereral solution of thesc ciuations is represented by {1.103) with y = 0. ‘The solution for the scalar potential U, which for direct current may be wsed in place of I, is then o- LUMO HMA HMA — GoD ‘where d is used in place of 1 as a-variable of integration and ris the radial separation from the zaxis, which is perpendicular to the surface of the earth. The Bessel function fois appropriate in the present problem, since the potential remains finite for r = 0, except when z= O at the samme tie, while ¥p (AA) becomes infinite for r ~ 0. From the genera! sofucion, the povential may be obtained formally for any number of horizontal layers by determining the funcvions f and g from ioe EARTH RESISTIVITY TESTING AND ANALYSIS 41 the boundary conditions at the interface of adjacent layers and at the surface oF the earth. At the interface of two layers m and m +1, there must be continuity of the potential and of the current density: i, = — 0m dUm/ds normal to the boundary, ie, Un = Unt Uy ap Ong = om (2.02) where « = I/p is the conductivity and p the resistivity. At the surface of the earth the current density /, = 0, hence dU/dz = 0, The boundary equations are xecotdingly Fat MO BP = fant + Bust fapin fi Po hem Gam — Ftd pap, _ Soot Pm Pm Pm Pott ot Gy ‘These equations are identical with those at the junction of two transmis. sion Tine sections a and m + 1, when both sections have ‘the same propaga- tion constant A and characteristic impedances fim 2nd fj. The futctions fand z ‘are obtained accordingly in the same manner ag for the correspond- jing transmission line problem. This circumstance will be utilized to some sdyantage in later sections to extend the solution for two and three layers to an arbitrary number of layers and to continuous variation in ‘resistivity, with depth, ‘The formal solution for continuous variation in resistivity with depth may also be derived from che basic equations for this case. ‘These equa. tions and their solution, as given by Schlichter," are presented here for ‘completeness although they are not: essential to the development in later sections. Let it be assumed that the conductivity of the earth o = o(syy,t) is a fonction of the coordinates of the point in question and that the earth is isotropic, so that the conductivity is the same in all directions. The carrent densities may then be written as follows: Je=Es Jy=oEy Jam oBy (2.08) If Wis the potential at any poine in the earth, the electric intensities along the three axes are Fyn —dUjix By=—dUféy Ey= —dU/ds (2.08 ‘The current must satisfy the equation div J = 0 or 4s 4 de, dds ke + a + e o (2.06) 42 EARTH CONDUCTION EFFECTS With (2.04) and (2.05) in (2.06), the following equation is obtained: PU 2U UA [dlde | dU de #10 eon at a at olde dv dy dy de de Let the conductivity vary with depth only as ¢ ~ o(2), 80 that de/dx and do[dy = 0. The field will then be symmetrical about 2 and, introducing clindrical coordinates with r= (#7 +37)", the following equation is obtained: @U 1a PU, dole) Firat dk o@) where a! (z) = da(z}/d2. ‘The equation may be solved in the usual manner demonstrated in 1.10 by asouming U(ra) = u(r)o(2), and thus separating (2.08) into two equations: te Mee tye Ser G tuum 2, oe) do B+ TOS ve-0 f where d is an arbitrary constant. "The fst of these equations is that of Bessel fonctions of ero onder. AS { shown later on, the sosoad equation can be identified with the equa A general solution of (2.08) may new be writzen in the form T= [0 omar reae a where g(x) is a function of d only, at disposal to satisfy certain boundary conditions. ‘Equation (2.09) is satisfied by the Bessel func: case, since the potential remains finite forr = 0 (except if = time), while the function ¥y has a singularity atr = 0. "The boundary conditions require tht the electric fexce in the 5 direction must vanish at the surface of the earth, since the current in the 2 direction {s zeros hence, for = = 0, ZL emnoovoaa 0 where 0! Qa) + dv/ds. i ee ee Ee plane wave propagation in a medium where the conductivity changes along the dizeetion of propagation, propagation constant b remaining unchanged, because the longitesiaal and teansverse conductivities change in the same sof the frst and second kind, Jo(ar) and YoQ). Of these, the former is appropriate n the present EARTH RESISTIVITY TESTING AND ANALYSIS 43 Equation (2.12) is satisfied by ¢(} = AA/0"Q,0) where 4 is a constant, since tie integral vanishes when the integrand equals A/o(\r) dh. Hence v= af" poreaaa 2.13) where _ 0a) ane) = MBO eu) For uniform earth, a pair of solutions of (2.10) is ¢™, the negative sign being taken to make the potential vanish at infinity. For other cases, the general solution of (2.10) consists of two functions, the one vanishing 5 infinity being appropriate to present purposes. Fora uniform earth erefore, va af" non anserer® rs) ‘The solution for a uniform earth, as obtained in a simple manner later on, is U — (lp/2n}- (P+ BY, so that 4 = Io/2n, I being the current in the electrode and p = I/e the earth resistivity. ‘When the resistivity varies arbitrarily with depth, the ratio of potential sta pointton the surface of the earth to current entering the electrode, which i the mtu restane ofthe electrode withthe pent in question, may be written thus: 90) - 2 f emrsor) a 16) where po = p(0) is the resistivity at the surface and &(A) = £00). Equation (2.16) applies when the current is supplied chrough a point lectrode or through an elearode consisting of a conducting hemisphere, so that earth current is uniformly radial near the electrode. Another case of interest is thet of a iat circular dise electrode of negligible thickness and resistivity, which excails a modification of the solution for points near the clectrode. In this instance, g@}, in addition to satisfying the condition of zzr0 Vertical electric force at the surface of the earth, must also be such that the potential is constant when the radial distance rs equal to or less than the radius a of the dise, for z= 0, ‘At the same time, the potential must ish as To/2er when r becomes large compared to the radius of the disc. ‘These conditions are satisfied when the factor sin ah/an ie inserted in the integrand of (2.13) and 2.14), The potential at the surface of the earth is thens 210 = 2 f° 2 203 Joo) ar em where #0) 's determined as before, -_aea 4 EARTH CONDUCTION EFFECTS cunt For uniform etrth, the mutual resistance for a dise electrode is then, wit 4Q) = 1, when tS 2 when r 2a 28) when > a. by Weber's integral. The above results may also be obtained by introdie jing ellipsoidal coordinates tnd regarding the disc as very thin ellipsoid of revolution, ‘The formula for ra ie obtained in the latter manner in section 3.2. ‘When earth resistivity varies continuously with depth, the problem i solved formally by obtaining a corresponding solution of (2.10) and thence Ose) from (2.14). When there are Giscontinuities in the resistivity, as ix the ease of several distinct horizontal layers, (2.10) applies to each indi- vidual layer, o”fe being zero and tm = nf ™ + fine™ the solution for layer m. It is then, necessary to determine the functions fm and gm in the manner mentioned before ané thus obtain an explicit expression for cin the upper layer andi, in eurn, an expression for AQ). 23 Electrodes ab surface of uniform earth Let a current J enter the ground at @ point electrode and lot the other current electrode be sufficiently remote so that its presence may be nes: ected. ‘The current is then radial about the surface point electrode, Imagine a hemispherical surface with center at the electrode and radius s, ‘The area of this surface is 2xs¥, and the radial curzent density in the ground atthe distance sis then J = //2xs%. Ifp isthe earth resistivity, che electric intensity in the ground in the radial direction at the distance sis E(s) = Jy EQ) = Ipftas? (219) “The potential at the distance s from the electrode is the integral of electric force between 5 and an infinitely remote point: Um [BEd = Ipftes 2.2) ‘The ratio of potential to current; or the mutual resistance of the electrode and the point under consideration, is then Qt) = o/ Ins at) FARTH RESISTIVITY TESTING AND ANALYSIS 45 ‘Assome next two electrodes, | and 2, on the earth’s surface. Let current enter the earth at 1 and leave at 2, aad consider the anutual resistance with 4 point 3 in the ground or on the surface, a8 indicated in Fig. 2.2. The re asltane mutual resisranceis then the diference between (sya) and Qa) or, jn a different notation, e(i a On = 3, C-- 2.22) Qumws = Ore where sig end sop are the distances between points 1 and 3 and points 2 and 3,and the subscript (1-233 of Q denotes thas the current path is between points 1 and 2 while the potential is taken at 3, sti Bia 12 Metuelseatanee af ‘loctrodes | and wich. Fi 23 Four cleerece methods of ‘arth reiscvity measurements. Banas = Ors = Qos Measured mutual redaraneeiQ—v/. ei(Helh) Apparene earth redeceay p =2araQ. With an eddivional point, 4, in the ground or on the surface, as shown in Fig. 2.3, ic follows that the mutual resistance of @ circuit between 1 and 2 with a circuit between 3 and 4 is Gomes = Dis ss — et Ooo dee. 0 =EGraoat) o> In the above formulas, the distances s were between a point on the sur~ face of the carch and a point in the ground or at the surface. Let r denote 1 radial separation along the surface of the earth and consider four clec- trodes on a séraight line at equal spacing re, 1 and 2 being the outer elec- taodes. Then fig = fou = 7, x4 = fog = 2ry. The mutual resistance is then Gave 9-2 (2- or, ifthe resitviey is desired, p= Und (2.25) When 7p is taken in meters and (? in ohms, (2.25) gives the earth resistiv- ity in meter-ohms. ace rol ~ Per (228) 46 EARTH CONDUCTION EFFECTS cunr.2 Such an arrangement of four equally epaced electrodes, known as the 4 fpurelectrode” metinod, i frequently used to obtain the earth resistivity from the measured tmatual resistance. Te is easily verified chat all the above equations hold when current and potential leads are reversed, as is required by the reciprocal theorem. 2A Elecrades below surface of uniform earth Electrodes below the surface of the earth are rarely used to determine earth resistivity, but the mutual resistance of auch electrodes is required in connection with the calculation of the resistance of buried grounds, to be considered in a later chapter. ‘Consider wo point electrodes, 1 and 1’, in an infinite conducting medium, as indicated in Fig, 24, and let a current, J, enter each electrode. The ‘pve 24 Canent mating fom ono peintslecuvaes in an Saisie meio, Cament azarae to plane P— 2” in whch # — 3 potential due to current in each electrode is half as large as that considered Jn the preceding section, since currettt now radiates in all directions. The [potential due to current in both electrodes #s thus Joft 1h o- B(L+3) 25 where 5 is the distance from the polit within the earth to. electrode i, ands’ that to electrode 1. ‘Imagine a plane midway between the electrodes and perpendicular to the line between them. For reasons of symmetry, it is evident that in this plane there will be no current perpendicular to the plane. Therefore, che medium and the electrode on one sie of the plane may be removed without affecting the current distribution and the potential on the other side ‘When the imaginary plane ie identified with the surface of the earth, the potential due to an electrode below the surface is thus obrained from (2.26), F being the distance to the electrode and s’ that to the image of the elec trode above the surface of the earth. EARTH RESISTIVITY TESTING AND ANALYSIS 47 2.5 Teonlayer stratification Jo many cases the results of tests may be approximated by assum thatthe earth consists ofa surface layer ofa cortain depth d ‘oa rede x and a lower layer of infinite depth and resistivity ay. Ror such an earth structure, the mutual resistance function (@{r) between points an the surface of the carth, separated by distance r, may be expressed as a func- don M of cwo ratios, a = r/d and 8 = po/py, thus: a ler 2) = 7 Mw) 2.27) The fanction Mf, which will be determined shortly, is shown in Fig. 25. When r is small compared 20 4, the mutual resistance approaches py/2er and, when ris large compared *@ 4, it approaches p2/2er. ‘Thus, when the earth consists of rwo layers and measurements are made with increasing spacing, the earth resistivity derived from mutual impedance varies between limits g2 and fe. ‘The manner in which ic varies with r depends om the ratio r/d, and by comparison of the observed variation with char obtained from curves for various assumed values of d, an approximate value of the Inter may be obtained. Tn (227) the produc eas) may be roarded as the apparenc resistivity, ie, the resistivity that is used in connection with the formulz for uniform earth. ian When the four-clectrode method referred to before i resistance function becomes a acaaainnaaas Panes = 95, Malas) 228) where Molaf) = Mee) ~ M(20,8) Tn expression (2.28), the term piMo(as) is nt res Ta aed ean tai 0O7) ® HE ABE In the conventional method of solution of this and similar problems, the resultant field ig assumed to consist of two components, a primary fil, sometimes also referred to as the impreswod field, and a secondary field. ‘The primary field is usually the field for a known simple case, in the present case that of uniform sarth, whereas the secondary field is due to currents or charges induced on of in bodies placed in the primary field. ‘The resultant potentials Uy and U within the upper and lower layers may be written thse ave UO, = U4 UF 29) ivity. -_-* = EARTH RESISTIVITY TESTING AND ANALYSIS 49 where U" is the primary Beld, Ui! the secondary field in the upper layer and Uj/ that in the lower layer. Rach of these is a fanction of the coor: dinates, U’ being a known function applying in both the upper and the lower layers. Ifa single electrode is considered, and the axis is assurned to extend ats ‘the earth through the electro, it is evident that for either a uniform or horizontally stratified earth thote is circular symmetry about the = axis “The potential at a point in he etrth then depends upon and radial dis tance r, parallel to the surface of the carth, ta the point. Thus, in the east 8 EARTH CONDUCTION EFFECTS Cear. 2 i ey Pepa iom soo s pep son | | ~ : = : 6 — ‘9 | 50 1 2 ' | ° , 3 A & 4 & 2 Shee A £] & ; | a & nla 35 0 1 = 2 =? a a a 4 4 4 1 os to" = ae a in ai u os MS ons sl es Leal cal J a a a S ) & § I 20 50 100 200 09 0002000 S090 ard a-RId Fro. 25 Relation apparent resistivity jg to murface resistivity or for twolayer Fig. 26 Relation of apparent resisdvity pq to surface tesistivity py for twolayer ark suc tocar ith o-code metal, of a uniform earth of resistivity py equal to that of the upper layer, In - Perey a0) By means of a Fourier transformation, the term (6 + 72"? in chis ‘expression may also be written in che form @+AP (231) 50 EARTH CONDUCTION EFFECTS where For) = i fo ene + PH ae = sae) a in f° v-@ [Pron a By the above transformation, the primary potential has been brought into a form involving 2 and rin separace functions. Also, this form of solution may be generalized to other cases involving the samne type of symmetry, by multiplying the integrand by an arbierary function of A. By chis gen- eralization, and by the above separation of the variables, a type of solution is obtained which can be fitted to boundary conditions. Generaliaing the solution, the secondary potentials may be written uF = [7 Ore + w@VeeQ) 20 oy = [7 add + we YoQs) an which is of the same form as (2.01). By = —dUjdx =0 i is necessary that fi = 51. To have the potential vanish as = becomes infinite, its necessary that gy ~ 0. Functions fi0\) and /(X) are obtained from the boundary condition at 1 = d, where the resultant potentials Uy and Uys defined by (2.29) mus. be equal and the current densities Ja and Jy in the z direction must also be ‘(A/os) aU/ds and Jo =~ (I/os) dUa/ds, the fol equal. With J = lowing boundary equations are obtained: h a Mt panes e8) = Bes fie 2 fe ERO + 2 ab Je ay orem] ase Solution of chese equations gives In mre FO) ae = Bene deat Mp) where in = (e1 ~ 02)/(e1 + pa)- ar i In order to satisfy the boundary condition at the surface of the earth, EARTH RESISTIVITY TESTING AND ANALYSIS st ‘The mutual resistance U1/Z for points on the surface of the ground is then en) [ [2+ x09] pana 8 [Lae onan (=m)* 5 ab + he war] pea (2.38) where a = r/d. AA physical interpretation of this formula is az follows: the surface po- tential for unit current may be calculated by adding to the potential for uniforin earth of resistivity py the potentials due to an infinite number of images, —m, +l, -ai->+ below the surface of the earth, at distances a> 2d, Ad, 6d 26 Three-layer stratification The formulas and curves for two-layer earth may also be used to find the approximate mutual resistance or the apparent resistivity for cereain types of three-layer earth. Assume that the earth consists of « top layer of resistivity oy and depth dj, an intermediate layer of resistivity py and thick- ness dp much greater than dy, and a third layer of resistivity os extending to very great depth, as shown in Fig. 2.7. When the electrode spacing is less than da the earth may be approximated by a two-layer structure with the second layer of resistivity ps assumed to extend to infinite depth. For elsctrode spacings greater than ds, the earth may be approximated by 2 two-layer structure with the original top layer a: absent. A somewhat better approximation in this latter case is to combine the two upper layers inta a single layer of depth d, +d, and equivalent resistivity (dit da)/ Giles + da/p2). From the two curves of the mutual resistance ir) or the apparent resistivity thus obtained, approximate curves for the three-layer structure may be obtained es shown in Fig. 27. When, however, the thickness of the intermediate layer is synall com- pared to that of the upper layer, the above method fails. Even if the thickness of the intermediate layer is small, is effect may be noticeable provided that its resistivity is markedly diferent from that of the top or bottom layers, as is illustrated in an example shown in Fig. 2.8, which has been calculated from formulas derived below. The general three-layer i lered here, together with the special ease in which the inter~ yer is thin enough and has a sufficiently high conductivity 0 that it may be replaced by a thin conducting shee? or, if t has a very high 32 EARTH CONDUCTION EFRECTS cune resistivity, by a leakage conductance or insulating sheet between the uppe and lower layers. Proceeding 25 in the case of two layers, the potential inthe third layer é written as Us = + UY, The potential UY’ is expressed in the same ee ts CR REEL. Pa/Ps Wea as 3 tld Fic. 27 Redon of appurentrostvty to surface resistivity for» threeleyer earth acuetsr, Carve 1. Upper twe levers ankr, Cure 2, Laer swo lagers only Carve 3. Aoproximatetarceaver vaiation, pip. =-1 pier = 10 e/a = 10 ‘way os the potential UY’ before, with an unknown fonction fa. In th ‘expression for the potential UY’ itis now necessary wo retain the funetior f2 on account of the Bnite thickness of the second layer. Solving te ‘equations for /1s/x £2 and fy rom the four boundary conditions, the folles. Sng expression is obtained for the mutua! resistance between points on surface: 20) EARTH RESISTIVITY TESTING AND ANALYSIS 53 where pum ae 9/1 poe) ot woe PLE wee) a) = a ‘ af 34s (2 34s a a0 Tay Fic. 28 Relation of apparent seisvity 0 sata ‘arth structare witha Hn intermediate layer Carve 0. Upper and lower layers enly. Garve 1. Thin ftermciiate conducting leyer a4 = Of. Gave 2 Thin intermedia insolatig layer po — 1500p a= (ea~ Pa)/ (a + m0) 4; = Thickness of upper layer dy = Thickness of intermediate layer When the intermediate layer is assumed to be very thie, so that exp (—2ade) = 1 — Dada this expression for ah) simplifies to eat py oe = a 5 BE Nay 2Q) = an ome 54 EARTH CONDUCTION EFFECTS Cua. When the resistivity of the intermediate layer ix very high, compared eo ‘av and pa, there is « forther simplification as follows: pt = 6a — Road pi + on + Nonde where pads is the resistance of the intermediate layer in the vertical direc tion, referred to a unit of surface area. The intermeatiate layer or shes: will not carry, in this cese, appreciable current in the radial direction, and ita principal effec is to insulate partly the lower from the upper layer. Wher the resistivity of che intermediate layer is very low compared to 1 and pp, so that it carries appreciable current in the radial direction, the following expression is obtained: aa) = 082) 1 = P+ Ma rosin pa + pa — Ma papal a Naud 5 + 09 + Road #Q) = (45) where 61, 9 and of are the ecmductivities, Xez being the conductance of ths sheet in the horizontal direction referred to unit length. Tt will be noticed that the formula for such « conducting sheet in terms of conductivities i similar, except for sign, to ther for a resistance sheet expressed in terms of the resistivities. 27 Arbitrary stratification ‘The mutual resistance for two leyers, as given by (2.38), may be written ape ‘ 09 = 2 baad for) a cas lene 1+ mae was * (ax ~ a2)/€ox + pa) ‘The function & will be referred to as the kernel of the integral. For the case of three layers, the mutual resistance as given by (2.40) may be expressed in the form ha 20-2 [amon nona es fas = 1 ps 4/0 F sae 8) EARTH RESISTIVITY TESTING AND ANALYSIS 55 a1 = pats Pi + pokes fay = (1 — pase ™)/(1 + wage) saa = (on — eal/oa + 0} Tn the latter expression, éy is the kernel of zhe two lower layers alone. As shown later, for the general case of 2 layers the mutual resistance Detween points on the surface is maae = 200 -EE Aaa. cn Jo) ad (246) fa. aan waa. hemi bins Pact = pebncot) an eT ut amie) ne 1 hin yar kein = ; OTE penne mF Homie ~ = 7H In the above formulas kigiym is the two-layer kere! fanetion for an upper layer of resistivity ys and depth dy and « lower layer of resis tivity om and infinite depth. Thus, when curves are available for various twolayer kernel functions, the ratio saa... may be calculated and, in turn, the finetion Aig.» In chis manne= the rmutaal resistance for the 1 layer case may he evalusted by aumerical intetration in a systematic way, For instance Ja the case of four layers, the process would be as follows: ame, eb nee me eae fe pee BE tam oka nase ole MT uae = 2 — Phan ikaw oe ptt pak 1 TE wane ms 56 EARTH CONDUCTION EFFECTS Cuan. “The above form of solution, consisting of a serics of substisations, is similar to that employed in successively reducing an electrical ladder net- ‘work starting at the far end. I is also possible to express the solution by use of determinants, as has been done by Stefanesco and Schlumberger, which, in effect, corresponds to setting up a number of simultaneous equations for 2 network. However, as in the case of ladder networks, che latter form of solution appears to be less convenient. ‘The above formulas constitute a general solution for the case of arbitrary variation of resistivity with depth, since by variation of depths and resis- tivities any possible variation may be accounted for. The kernel kr appearing in (2.44) will be recognized as the factor by which the char- acteristic impedance of a transmission line must be multiplied to obtain the impedance when there is a point of reflection at the distance diy sna being the reflection factor and \ the propagation constant of the line. When the solution is extended to three or more layers, the kernel similarly cor- responds to two of more reflection points, provided that the characteristic jmpedance of various sections of the line is changed in such a manner that the propagation constant remains unchanged. ‘The correspondence with the transmission line problem is inherent in the identity of the boundary conditions at points of impedance discontinuity, as discussed in 2.2. The requirement of a constant propagation constant arises from the circom- stance that the earth is assumed isotropic, so that the ity in the longitudinal (2) and transverse (r) directions changes in the same manner with depth. ‘To obtain the general formulas (246), assume that « transmission line section m ~ 1 is terminated in an arbitrary impedance Zm at the end of i ‘The impedance looking into the beginning of section (248) at = Kena bemetpe Keantin VE encima 29} a knan Zn Bonn OEE Now Za may again be regarded as the impedance of a transmission line ‘composed of several sections and, in accordance with (248), be written in the form Zm = Key > knit)» Hence, : Kact = Kut him) oon Wem Kant Ke Rime) oe (2.50) pz |_| EARTH RESISTIVITY TESTING AND ANALYSIS 57 where in all expressions K is replaced by p for the case ofa stratified earth, by virene of the equivalence of p and K as shown by (2.03). Thus, when the kerne! Aniest) ...a is known for & given structure of resistivities pny wit «Px it may be found for a structure in which has been added surface layer of resistivity pi. _ To find the solution for the transmission line in the general ease of con tinwous variation, assume that the Tine has a tnit length longitudinal Impedance p and transverse admittance », both functions of 2, a coordinate parallel to the ms ofthe ine. The equations fr voltage oand current Z are then dofda= In, dfs = —09 si) Differentiating the frst equation and inserting the second, the following equation for the voltage is obtained: ede Boog =o (252) In the latter expression gv — 2%, 2 being the propagation constant, and —s'In may also be written o/c. A solution of the above equation is obtained by the substitution 9 = wexp (blog 2) 83) wrhere w is obvained fram the differential equation eu a 1d fo" fe ofe_(2¥_14(2)] ere -Gyte@)l-2 es The ratio of voltage to current at 7 = 0 is given by 9(0)/2(0) = =9(0) 2(0}/e'{0}, which is greater than the Spangler ps uniform Hine, —p(0)/A, by the factor HQ) = 900}/0"(0) (2.55) ‘The latter expression is the same ag that obtained from (2.14) for = = 0. For the general cae of consinoous vavston of rexivty with depths erefore, a) = 2 rarsotr an 2.56) Both expressions (2.47) and (2.55) for the kerne! of the integral may be regarded as general solutions for arbitrary variations in resistivity with depth, one for a finite number of discontinuities and the other for con- sinuous variation in resistivity. Expression (2.55) was obtained by ™_as CS 8 EARTH CONDUCTION EFFECTS oan? taking the limiting case of continuous variation by means of a differential equation for &. To carry out the reverse procedure and to find solutions for » distince layers when differential equation (2.52) is knows, it wonld be necessary to solve the latter for 2 uniform earth and then obtain (2.46) by satslying the various boundary conditions in the manner outlined in preceding sections. The circumstance that the kernels #(\) or fiz...» may be identified wich the variation in the input impedance of a transmission line havieg certain types of irzegulerities suggests the possibility of using an artificial network to find the Kernel experimentally in the fallowingmanner. Con- sider an artificiel tranemission line consisting of a great number of sections of series inductance and shunt capacitance. Assume that a given variation in conductivity with depth is to be sinralated. This corresponds to a similar variation in the shunt capacitance along the line, and an inverse variation in the series inductances (ie., the latter vary in the samme manner as the resistivity). The propagation constant 4 — sw(ZC)"? thea re mains the same for all sections along the artificial line. The variation of input impedance with frequency corresponds to the variation in # with X. 2,8 Exponential variation in resistivity In some cxses it may be possible to fit observations by assuming am ex ponential variation in resistivicy with depth. This case invoives only two varigbles, the surface resistivity oo and the rate at which the resistivity varies with depth, rather than the three variables of two-layer earth. Formulas and curves for inductive coupling are available both for ewe. layer earth and exponential variation in resistivity with depth, and the Jatter may be used, in some cases, in preference to the former if direct eur. rent measurements should indicate it is « better approximation, Let the resistivity of che earth be assumed to decrease a3, we as Equation (2.52) is then A solution of this equation is 7 4, seth ew ew( $204 d+ aver ) 2.58) ofa File dt ae EARTH RESISTIVITY TESTING AND ANALYSIS 59 and, since the positive sign ig required to make 9 = O when k= , #00 = 7 +1] x) leg)" 4] Inserting (2.59) in. (2,56) and differentiating both sides, 222 poo[fest)"- wales §]oa m= 8 [hewn i er 60) w = oar Neer) 61) aed wan -eF-fni(2), exo 2 2.62) Neer) FF 8(Z) + ortogar «<0 Here Hi is the exponential integral defined in Appendix I. When a <0, the sign of the last bracket term in (2.60) is reversed, bot rot that of the preceding term originating from a*/4 under the radical. ‘The two last terms in (2.60) then combine into 4/r and tence the logarieh- nic term in (ar). When the four-lectrode method is used, the rautual resistance is ob- tained in the same manner as for two-layer earch and is giver: by 2 % ae Nolan) 26) where ro isthe electrode spacing. and No(aro) = 2 (ar) — N(2ara) 261) 29 Determination of earth resistivity and soruetuye frome measurements In the preceding sections, the potential at the surface of the earth was determined for certain assumed variations in zesistivity with depth. The ‘curves thus obtained may be used to establish variations in earth resistiv. ema = £ # 3} |_| wo EARTS{ CONDUCTION EFFECTS cuss? ity with depth from measurements of the mutual resistance between points on the surface for various electrode spacings. This is usually done by comparing curves of measured mutual resistances with appropriate theoretical curves. In the case of actual two-layer structures, the approxi- ate resistivities and the depth of the upper layer ave readily obtained frcm dhe measurements after « few trial comparisons. Except in connection with geophysical prospecting, an approximate twodayer stratification is asually satisfactory, even when the measurements indicate a more com- plicatedearth structure. Should amore accurate determination of resistv- iy and earth structure be required, approximations as outlined in con- nection with three-layer stratification may be adequate in many instances. A direct solution of the inverse problem of finding the conductivity depth functions o(2) when the mutual resistance is known is formally possible. “Thus, by an inversion of (2.56) the following expression is obtained for the kernel: aah fe = ("ons 2.68 FQ) = [0 oenseorirdr (2.65) ‘This is consequence of Hankel’s inversion formula which states that Fo) = [0 010. , 10) = [0 Fos ovyrer are equivalent formulas, j_ being an nth order Bessel function of the first kind. “Assoming earth resistivity to vary with depth only, the kernel £(X) thos may be obtained from measurements of (7) over 4 range of separation sufficient to enable evaluation of (2.65) to the accuracy required. "The inverse boundary problem of finding the conductivity function o(@) from (2.52) when k()) is known have been solved focrnally byLanger. ‘The Kernels corresponding to various depth functions, a8 obtained by solution ‘oF (2.52) for o and use of (2.55), have becn gives by Schlichter and may be ‘ised a¢ an aid in the solution of the inverse problem. As noted above, this problem may also be solved using the simulative artificial line, ‘Langer's method consists of developing the kernel in inverse powers of ae AQ) = Lb ah ai? + 2.67) When £(A) is not given in analytical form but as a curve obtained by evaluating (2.55) from the resulss of measurements, as would be the cast in actual applications, the constants 2, a2, 5. --, etc. may be obtained by solution of the appropriate number of simultaneous equations. Thos, EARTH RESISTIVITY TESTING AND ANALYSIS 61 if m points were selected on such a curve, coefficients could be obtained from m simultancous equations. From these coeficients, a Function p(z) is calculated fom the following expression: POG) = dg + (mad + Laas + (al — Lares + dag ie? + $0Balag — Bai — aya — 2a + day) + dj — ahs + Seay + Teed — (2.68) Amey — 2ayay + Yaga ‘The conductivity as 2 fonction of depth is then of) = onexp [-2 9 «| 2.6) ‘The above method applies when ¢(=) is a continuous function and will formally give the resistivity to any depth, slthough the accuracy will de- grees the depch incense, When thee ate dssontimities in the con. juctivity, a different expression for the conciuctivity function would appl on eich side of the dicontinulty. In moot practical canes there are aie continuities in the conductivity, and the conductivity for large values of 3, a obtained by the above method, will then approach a value cifferent froin that derived directly from the matual resistance for large electrode spacings in the customary way, using (2.25). Although the method indicates the presence of discontinuities, which may usually be assumed in the frst place when there is 2 marked variation in the apparent resistivity as obtained from (2.25), it does not indicate the depth of the discontinuities. Proof of the above method may be found in the papers referzed to before,S® and a more elaborate method for the case of discontinuities has also been devieed by Langer." 210 Analytical approximation of measured resiuivities Jn problems involving earth potentisls due to distributed leakase of current from conductors, or distributed leakage of current into conduetors ve to earth potentials, a3 in the case of lightning enrrent or stray current in buried cable, itis sometimes necessary to conside: the efect of varistion i the resistivity with depth. In such problems, where an integration is involved, it is convenient to make use of the following expression for the ‘mutual resistance between points on the surface of the earth: 1 QO) = lm — Gor — one) (2.70) ‘The bracket term in the above expression approaches py and py for small and lange values of r respectively, corresponding to the resistivities near the ary [0 ee ee eee a a a EARTH CONDUCTION EFFECTS Cat surface and at great depths, By proper choice of a1, p2ard earth sesistiv- key measurements may usually be spproximated to the same degree of accuracy using the above expression as by the curves for ewo-layer earth, for the reason that ideal cwe-layer stratification is seldom encountered. Compazison with carves for two.layer structures indicates that the Larter may de approximated by (270) when =u an) being the depth of the top layer and 6 2 constane which depends on the resistiviny casio about as follows: ver = 10010 a 3b 02 my a te DAL Earth resistivity measurements Earth resistivity measurements by the four-clectrode method mentioned before may be made by use of a battery, an ammerct, and a sensitive volt. eter. Allowance ant be made for the volemeter resintanee in relation to the total resistance of the potential circuit, unites a meter of very high Jnipedance is used, the resistance of the wire and the two potential ele=- trodes in cries being measured by means of the equipment mentioned Wwe. This method, however, is not seliable when the electrode spacing ig large, since the voltage to be measured is likely to be sinali relative to extraneous Voltages caused by stray current in the earth or galvanic elec- trode pocential differences. By simuleaneoualy operated reversing switches in the curreht and potential cireits, the voltage due so the vest current nay be obtained by taking the average of two readings or, ifthe difference je large, Sy averaging several sets of two readings. Lf che reversals are rude at a rapid rate, by using a double exremutator rather than reversing Switches, extraneous vaitage will not be recorded so that s direct readings obtained of the voltage éne to the test current. One tyPe of instrament hhased on this principle contains a hand-criven generator ap an ohrameter, which gives a direct reading of the mutual resistance of the current and potential electrode circuits, “in ancther type of instrament the mutual resistance ia measared by means of a potentiometer. At the Ute of balence thers is, thesefore, 20 current in the pacential cirelt, so that che resistance ofthe poren ak circuit does not affect the reading. Even with such af instrument, i is desirable, however, to use clectredss of not 200 high resistance, since otherwise the all indication may be indefinite, the voltage across the sul detector being small even when balance is not good. In high resistivity soil, therefore, it js usvally advisable to drive the ground rods, which are commonly used at electrodes, two or three feet into the ground. EARTH RESISTIVITY TESTING AND ANALYSIS 63 When the current is rapidly reversed, as in the above types of inseru- ment, 2 voltage is induced in the potential circuit because of the current reversal. This induced voltage may be particularly noticeable for large clectrode spacing, since the mmatual resistance decreases while the inductive coupling inexeases as the electode spacing is increased. This inductive effect is also more noticeable for low than for high earth resistivities, since the mutual resistance is proportional to the resistivity, whereas the induc- tive coupling depends on the resistivity to a lesser extent. The inductive coupling may be reduced somewhat by separating the current and potential leads, although this is inconvenient when the leads are long and the j ductive eect is appreciable. From a practical standpoint it is also de- sirable in taying out the wires to keep them fairly close cogether. The dificulty of inductive coupling has been overcome in a potentiometer type of instrument by setting the commutator so that the potential contaers are closed after the current contacts. Sufficient delay is used so that the induced transient voltage has disappeared by the time contact is made.!? The measurements are usually made at various lectrode spacings, ranging from a few feet to perhaps a thousand feet and sometimes even more. ‘The range of electrode spacings required depends on the nacure of the problem involved and on ehe variation in the apparent earth resistivity with the spacing. If the resistivicy increases or decreases rapidly as the ‘electrode spacing increases, itis usually necessary to continue observations ‘ntl the curve of apparent resistivity asa function of spacing has fattened cut to such an extent that a reasonably accurate estimate may be made of the resistivity for larger spacings by extrapolation of the curve. When the for-lectrode method is used, the apparent earth resistivity js calculated from expression (2.22), or by the corresponding expression 2 = 192100 when the electrode spacing is taken in feet. The results of the teats may be analyzed by che methods outlined before. In cases where an average resistivity over a fairly long eistance is required, 25, for example, in ‘estimating inductive effects or the liability of eable to lightning damage, the results of observations at several locations are averaged. The average resistivity curve as % fanetion of spacing may then be analyzed to find representative earth scracture or equivalent resistivity along the section under consideration. ‘The analysis of surface potential measurements in connection feological prospecting is discussed in 4 wumber of publications.!#.14!5.18 212 Barth resiscsity data Depending on the nature ofthe soil or the rock, the resistivity may range from an unusvally low value of 1 aan vewsualiy high valve of 10,000 meter- Jotivity may somesimes be estimated from geological maps, ome “ EARTH CONDUCTION EFFECTS cman. based on the results of tests for similar geologic formations as indicated by the maps. A study of such maps may indicate what type of variation in the resistivity with depth is to be expected. If, for instance, there is a cop layer of sand with shales at greater depth, as along @ great part of the Atlantic coastal plane, the resistiviey is likely to decrease from a very high value near the surface 3 a low value for grester depths. On the other hand, ifthe surface materiel.is clay while the underlying rock is granite, che ‘opposite type of variation will be found. ‘The relation of earth resistivity to goological age and formation, xs obtained from a large number of mieasure- ‘ments in connection with such problems as mentioned before, has been given by Card!” As a rule, the resistivity inereases with the age of the geological formation. Measurements with very large electrode spacings, ‘about 330 km between the outer elecerades and from about 25 to 140 km berween an outer and the nearese intermediate electrode, indicate aa upper Jimit zo the resistivity of deep strate in the order of 10,000 meter-ohms, ‘with lower values at depths over 50 km or so.'® The resistivity rear the surface, to a depth of about a moter, may vary appreciably betwenn along dry and 2 long wet season, and thereby affect the resistance of grounding clectredes. Otherwise this variation in earth resistivity is of secondary importance in protection problems, as ix does not appreciably infiaence voltages in cable ot transmission lines due to lightning or low frequency ‘currents, ‘The classification in the following table gives some indication of the frequency of occurrence and the range of resistivities that may be expected for various geological formations. EARTH RESISTIVITY TESTING AND ANALYSIS 6 CHAPTER IIT Resissaxcr oF Gzounome Areawceuenns 3A Introduction Ground connections on electrical circuits may be provided for several purposes, depending on the nature of the cirenit. Grouncing is employed to prevent excessive potentials between metallic structures and groand at times of faulr or during normal operating conditions, to provide low fault impedence and thus prompt operation of fault clearing devices, ta conduct ghtning current to earth ané thus limi lightning voltages on transmission lines or in telephone cable or, ifthe earth is utilized as a return conductor, asin the case of radio transmitting antennas or forced drainage installations for corrosion prevention, effective grounding is desirable from the stand. joint of operating efficiency RSE sa nmieune Capand at the frequency in question, be small compared co the impedance of the Circuit in which itis connected. By this criterion, it may be perraisuble in some instances to have a ground of high resistance, several thousund ohms, a in the case of an “electrostatic” apparatus ground, get o ound of insulated apparatus cases being ordinarily quite high. In other ‘aves, heworer, «resistance of only # few ohma may be roired for eve tive grounding. ; “Anough calculations of ground redntance are rarely made before ie- stallacion, it may be advisable to make earth resistivity measurements and calculations in situations where a ground of low resistance is required, particularly if the installation site is such that « high resistance may be anticipated. Earth resistivity measurements may indicate that a ground rod or pipe driven to a certain depth is preferable to a ground wire extending at shallow depth along the surface, or vice versa. Again, should a surface ground be preferable, there may be a choice between the use of several ground rods or buried wires, between a single wire, several parallel wires, er several wire cadiating from a common point. Sach questions cenest easily be decided experimencally by comparing actual installations, the earth rersiey snd srucrareuiualy vary to such an extent from ont location to another, even for near-by ses thatthe elect of arin inte rounding atrangeinent may be small by comparison. The resistance for travious grounding uvangerca, howeves may be eudiy calculated whes & RESISTANCE OF GROUNDING ARRANGEMENTS 67 the earth i assumed to be uniform, and the effect of certain variations in the resistivity with depth may also be evaluated. In the following theoretical treatment, the grounds are assumed to be of such limited excent ther the voltage drop along the conductor may be neglected in comparison with the potential of the ground structure. In general, chis will be the case even for grounds consisting of a substantial Jength of buried wire, say a few hundred meters, at least wien the earth resistivity is high enough so chat @ ground of this length is required to ob- tain @ reasonably low resistance. ‘The resistance for a ground of such {ength that the drop along the wire must be considered may be obtained, however, by use of the formula hea (BV" RA w= (5) ” coer 2) a1} =e when R/R <4 (RR when R/R> 2 where & = resistance to ground for conductor having zero longitu. inal resistance R = longitudinal resistance of tota! length of wire. This formule is obtained by considering; the ground 2s a transmiscioa line of length open at the far end, of unit length leakance G = 1/3? and unit length resistance R/i The formula indicates that when the longitudinal resistance R is 40% of the resistance &, the actual resistance Q is about 5% higher than &. When the longitudinal resistance ie 2 times , the actual resistance is 4% higher than the characteristic resistance (RRY”. Ie is assumed in (3.01) that current enters the wire at one end. When there are m equal branches carrying cqual current, the combined longite inal resistance will be 1/m times the resiteance of one branch whereas the combined resistance to ground will be ereater than In times the resistance te ground of one branch, because of matual effects between branches. ‘The ect of the wire resistance is then less than indicated above. For a single wire or parallel wires, the formulas given below, though derived for direct current, may also be used for altereating current when consideration is taken, where necessary, of the alternating-current loogi- tudinal impedance Z of the wire, or of several wires in parallel, which re- places Fin the formula above. When the wires are not too long, say less than 100 meters, the formulas may also be used to calculate che surge im- pedance of wires to lightning currents except for unusually steep wave fronts. For shore wires or grounds consisting af only a few ground rods, cy EARTH CONDUCTION EFFECTS Cras however, the resistance to heavy currents, as encountered in lightning surokes, may be appreciably reduced beceuse of ionization of she soi! azound the ground rods, as discussed in Chapter VII. “The conduezors have been assumed to bein direct contact with the earth Measurements indicate that this is not strictly true, even after grounds hhave been installed for along time, and that, a a resale of contact resistance at the surface of the conductos, the resistance of grounds may be some 20% higher than calculated. At high frequencies, this contact resistance is bridged by capacitive admittance between conductor and ground, ani the resistence to ground is thus somewht reduced. The conductance to ground of « given conductor arrangement is derived in the same manner as the capacity of the same arrangement in free space and, except for constant factor, both are given by the same formula. Ex- cept for wites of certain shapes not used in actual installation, rigorom formulas arc quite complicated. However, formulas based on the s0- called average potential method used int calculating the capacity of wites are usually fairly simple and are known to be rather accurate. When the average potential method is employed, the caleulation of the self and mutual electric potential coeficients of conductors, or of self- and rmutual resistances of grounds consisting of wires or rods, involves evaluation of integrals identical with those encountered in. calculating the self and ‘mutual inductances of wires.’ Based on the average potential method, the capacity of various wire arrangements has been given by Howe* and Grover.® For various types of grounds of limited extent as considered here, formulas for the grounding resistance have been given by Ollendor! who has also considered the eect of heating; in addition, 2 comprehensive sumnmary has been published by Dwight.” Various practical considers. tions in connection with ground installation and testing are dealt with in x number of publications.®”°# ‘The effect of « bare interconncetion wire between ground reds connected in multiple, 2s well as the effect of variation in resistiviey with depth on the ance of grounds, have been considered here. Some tables and figures ace given of the esistance, assuming for convenience an earth resistivity of 100 meter-ohms. As is evident from data given in seetion 2.12, however, this may not be used as a representative earth resistivity, 3.2 Basic farmulas ‘The resistance of grounds having the form of = hemisphere ota plate at the surface of the earth or that of a buried sphere may be obtained in s direct manner from the formulas in Chapter HI, as shown Jater on. In stalled grounds are usually long cylindrical conductors or wires, and the general theoretical considerations outlined here are confined to such con. RESISTANCE OF GROUNDING ARRANGEMENTS 69 dactors. The resistance of a vertical or a horizontal eylindsiesl ground near the surface of the earth may be obtained in a simple manner, ws shown lates on, from the resistance of a cylindrical conductor imbedded in a con- dyeting medium of infinite extenc in all dircetions. When the mutual re sistance between any two such conductors is known, it is also possible to find the resistance of vertical or hotizoptal grounds in multiple. Con. siderations will be confined ia this section, therefore, to the self. snl utal resiotances of cylindrical conductors surrounded by 4 conducting medium of infinite extent. Assume that a conductor extends along the axis between « ~ —//2 and #= //2, and that a current Jp enters the latter at the mid-point. Lee the conductor current at the distance x = u be T(u). ‘The potential at a point xy in the surrounding medium due to current leaving @ conductor clement die at x = wis then Ble oF + yt au 3.02) a0 whore the term multiplying 2 au ig the mutual resistance between ewo points in a medium of infinite extent, as derived in the sections 2.3 and 2.4, In this expression, itis assumed that the potential due to current leaving ‘conductor element is the same as for a point source at the axis of the con- ductor, which is permissible in dealing with long conductors. The potential due to current leakage along the entire conductor is then Poona 2 fie ote ay Tf the resistance of the conductor is assumed to be negligible, so that che voleage drop along the conductor may be disregarded, the boundary con- dition to be satisfied at the surface of the conductor » = is that a (ea)/de = 0, Hence the current distribution along the conductor must satisfy the integral equation on fice wt tye ay By successive approximations, it is possible to find a distribution of leakage current 2 (i¢)/du along the conductor which satisfies this equa- tion. When the current distribution is obtained in this manner and the conduetor potential is next calculated from (3.03), ie turns out that the average value of the latter is, within a few per cent, equal to the average potential obvained from (3.03) when df (w)/d is taken equal to the average leakage current, 2U(0)/2=7,/2. ‘This has been done in connection with the (3.03) (3.04) reenact ee Ee a §— | EARTH CONDUCTION EFFECTS ca.9 » calculation of the capacitanee of conductors or, eather, the patential coefhi- Gients, where the charge distribution has to satisfy the same condition ax the leakage current in the present case. Ta the caleslecon of the capacitance of a conductor in 2 medium of eapacitivity «, p/ter is replaced by L/tre in @.03) and dI(u)/de by the charge density Q(w), which for uniform distribution of charge is taken as Qa/!, Qo being the total charge on the conductor. The potential coefficient ¥/0o is accordingly obtained by replacing p/4z by 1/4ne in the forrmulas for the resistance. In calculation of the inductance of a conduetorina snedinm of inductivity », o/4r is replaced by r/4x and 4f(u)/du by the longinudinal corrent fy, which is assumed constant slong the conductor, ‘The inductance 's accordingly obtained by replacing »/'tr! by vi/dacin the formulas for the resistance. constant current leakege dI(e)/du = 21(0)/2 along the con- ductor, evaluation of (3.03) gives UO)o ae! Feo) 305) where log = log and og a(syy) VEEP FS + e+) VEq TP + eI) ‘When the length of the conductor is very much greater than its diemetec, the potential at the mid-point and at the end points is, with y = 2 and (UD + aH? = 2m a, PO2) = (ey) = 3.08) Oe el é eas Gon Oe). 2 Pfra) =P eS ‘Thus, aseuming uniform current leakage, the potential at the mid-point is nearly twice as large as at the end points. It is evident, therefore, that the lenkage current will actually be greater towards the ends of the con. cuetor. When the average potential obtained by integrating (3.08) betwee 4 = Oanda = //2is divided by 2/(0), the following expression is obtained for the resistance of the conductor in & medium of infinite extent in all directions: = E (oe[La + vrF eA] +5 — VIF GI) (3.08) RESISTANCE OF GROUNDING ARRANGEMENTS — 71 or, when > 2, w a a = fi[ice™ — 1] ~ Pou (3.09) Another method yililiag much the same result is as follows: Returning 4 equation (2.05), jc can be shown that the equipotential lines /"(s,y) constant, corresponditg to a(#,3) constant, are confocal ellipses with foci ar the wire eermmals. When a is given, the major axis 4 and the minor semi-axis @ of these ellipses aze obtained from @.10) @ cequipotential surface may be replaced by a thin sheet of negligible ivity, without disturbing the potential distribation. If the material within such a shect is next replaced by material of negligible resistivity, the potential oatside the eqripotential surface is not altered and aground is obtained having the shape of an ellipsoid of revolution. Its resistance is obtained by substituting in (3.05} 7 as obtained from (3.10) and a as obtained by solution of (3.11), with che result pense 8 — FEI hope 3.12) a= ald + 4 V1= Fell 9 = ofa? — 2 When (3.12) is applied to = sphere of radius a @—+1 and 1 lim"; loga—1, so that the resistence becomes, with = 24, = p/4na, which is correct. When the major axis approaches 2e70, the ellipsoid of revolution becomes « Set, circular dise. In this case h/e 1 and a — 1 ~ tlyfa, 90 chat with log (—1) = ir, Rr = p/8e. This is the proper result for @ Fat dise in an infinite mediam, as may be verified by comparison with (2.18} for a dis at the surface of he earth. “The resistance of an clipsoid of revolution of length fq and maximam radius 7 — 4 much smaller than J, surronnded by a medium of infinite extent, may be taken as Pgh Ie Ea R= (313) EARTH CONDUCTION EFFECTS Coan n ‘The latter formula is the same 28 that obtained by taking the potential fat che mid-point of a cylindrical conductor and differs only slightly from that obtained when the average potential method is used. Let (2) designate the resistance of a conductor as given by (2.08) ‘The mutual resistance R, (2,2) of two parallel conductors of ajual lengih is then obcained from (08) when the separation giz between the com ductors is uved in piace of the radins ¢, Touse the simpler formula G09), the length of the conductors aust be muck greater than their separation. Tn addition to the nmtual resistarce of two parallel conductors, that between two conductors starting at a common point and making with each other an angle 6 will be required in order to obtain the resistance of several wires diverging from a point. Using the average potential method, this mutual resistance is obtained by integrating the potential given by (G.05) along a straight line y = kx = x sin®, with x taken fram one ené of the wire rather than from the mid-point. The mutual resistance thus obtained is Gs Ths formula i derived in the same manner a che mural ioducane between the wires, except that in the latter the factor cos @s4/2e replace p/2xl. With this modification, formulas given by Campbell for the sei ‘and mutual inductance of straight wires! for conditions considered abow ‘and for various other couditions, a8 for wites of diferent length aot staring at a common point, also apply to the self- and mutual resistance of wires ‘When 1 parallel conductors of the seme length are arranged on a sirch of diameter D, so that all the canduetors earry equal current, the combing, resistance of the conductors including all mutual effects is wer a {axe + Baws main] Gus where Ry(2) and RW vin m fn) ate obtained from (3.08), ax = Dain mx/n being the distance between grounds 1 and m. Instance of G.15) are ®y = pIRle) + RD] Ry + HIRO) + 2RAD sin «/3)] Bq = EIRa) + BRD sin 4/4) + K(D)] Ry = ER (e) + WD sin x/5) + WD sin 2x/'5)) Re = Riba) + RD sin £6} + RD cin 2/3) + KD} ees RESISTANCE OF GROUNDING ARRANGEMENTS 73 When » conductors of equal length radiate in the same plane from a com. ‘mon point, with equal angles ¢ = 2e/m between adjacent conductors 50 chat all carry the seme current, the combined resistance equals a ~ Paws Ea, rnin) where Ry is the mutusl resistance between wires 1 and m, the angles be. tween which are @ = 2em/n. Thus, Bo = FIR (2) + By (H)] Ry = FIs (0) + Re (2"/3}] Ry = EIB (a) + We (/2) + Re (I Bs = §IBRy (a) + Ay (2n/S) + Wo (44/51 Ba = § (Rs () + Bo (x/3) + Wo (2n/3) + Ro ()] Wher the conductor arrangement is not symmetrical, so that all eon- ductors do not carry the same current, the currents in the several conduc tors and their combined resistance may be obtained by solving a number of simultaneous equations with the conduezor self. and mutual resistances as coefficients. When all conductors are of qual size, however, and the con- ductor spacing is fairly uniform, it is usually permissible to assume equal currents ip all conductors in calculating ehe combined resistance. ‘Thus, the resistance Ry of m paralicl conductors in the same plane at equal spac ing s may be abtsined by calculating the resistance with respect to each conductor and taking the average value, which becomes G6) 8 = 2 (860) +2 ter — 180) FOR OD Fo — ) GUT) 33. Spherical grounds Although itis impracticable to construct a ground in the form of « kem- ‘sphere at the surface of the earth crin the form of a buried sphere, grounds ofsuch shape may sometimes be approached by the huried metal structure of transmission ling towers oF When the resistivity of the soil around a wound is reduced by salting. Spherical grounds may also be used for experimental purposes, when a uniform radial field is desiced. As determined in section 2.3, the mutual resistance between a point elec trode at the surface ofthe earth and a point at the distance s in the ground is ven by O(8) = o/2es. ssmay be assumed in the case of metal, is placed in the form of a hemisphere Ieisevident chat when material of zero resistivity, ” PARTH CONDUCTION EFFRCTS Cuar.5 ‘or radiuy 1 azourd ehe point electrode, the potential of the electrode will RESISTANCE OF GROUNDING ARRANGEMENTS 75 be the same as at the distance s, ar Baihak the effective distance tothe imagemay be taken as 2d; the resttance R= p/des G.18; ae Re at ied 2) If the ground consists of a hemispherical metal electrode of radius 4 szrromaced by mera of ressvty on Mp toa real tances the) To obtain the mstnce of a plate at pth small compared to te sistance of the ground is diameter, consider first the series tance fr Infinit li ; st the sevies resistence in an infinite mediom of £+e(t- 5) G.t9) plates separated by the distance 24-4: . To ealewlate the latter rece, dns Ue \o0 sy ance, only the medium between the plates needs to be considered, and the : 7 sistance is then 2ad/ra®. This resistance equals twice the di ‘The latter formula would apply when low resistance materia is placei Tistance is d quals twice the difference around she ground or when the soil resistivity is reduced by saleing ar by between the self-ané znutzal resistances of the plates, so that the mutual ionization hecause of high voltage gradients or increased by evaparation a pee ai pffa— ed/ra®. The resistance of a plate buried at the depth moisture as a sesult of heating. ‘When current enters 2 hemispherical ground #5 considered above, the Ra ep i “| (28 potential gradient in the vicinity of the ground som ‘be undesirably high te Fa, ) ‘The gradient at the surface, as well as the resistance of the ground, may 4.5 yrpiea es, reduced by using » buried sphere, although this may be accomplighed i 03, 7S! gotnd rods or pipe | tore practical manner bythe ase ofa diferent type of ground- When: , 8 Sotinon ype of made grotad conse of one or mare ground ros sphere is buried at a depth d large compared to che diameter of che sphere, 45 € J meters ( to {0 fe) long and 1.25 to 2.5 em (4 to 1”) in diameter, follows from considerations similar to those cutlined in section 2.4 that th: hen several reds are used in 4 permanent installation, they are uaually resistance of the ground is interconnected by a bate wire laid in a trench betwee the rods, although =. , tminsulsted wire may sumesimes be preferable ro prevenc it from cooling x B tal GO. as a result of galvanic porential difference with respect to the rote oe continued leakage current from other sources, A bare wire may in iselt «being the radius of the sphere and 2d the distance from the center of tu POvide a good ground which, however, is not taken ineo cecoune th the sphere to the center ofits image above the surface of the ground. Tai rmvlas of this section, where an insulated wire fe assumed The we formula is rather accurate when the depth of che center of the sphere ex stance of a buried bare wire, as well as that of ground rods interconnected ceeds its diameter. Tt is assumed that connection co the sphere is made 6 *Y 4 buried bare wire, is considered later. an insulated conductor. If a bare conductor of some length is used, tk Assume a plane perpendicular to the axis at the mid-point of the eon. resistance may be reduced appreciably. ductor of fengeh 7 considered in section 3.2. The medium, the conductor and the eu:zeat 1(0) to one side of this plane may then be removed without Bab Plate grounds Gstorbing the curent oF the potential on the other side. The porencial In some instances, wire nets are used to obtain a ground of faitly lor ® thes obtained for a vertical conductor of length //2 carrying a current resistance. When the meshes of the nets ae not ta0 large, the resisang J), Since the porzota! sche ame but the total cures? 90) hecend to ground will he practically the same as for 2 solid plate. of 21(0), the resistance of such a conductor is twice as large as that given From the discussion following (3-12) it is evident that the resistance oft 4” (0.08) when fis twice the length of the conductor. When /is the Jength flat, circular disc of radius @ at the surface of the earth is of the conductor, the resistance is thus: ae : and oan = (ue Fa 4 ve wR] viF@m) ea When the plate is buried ar a depth, d larger than the radius of the plate ol (iar! é. ) > when (ea (625) 6 EARTH CONDUCTION EFFECTS cm RESSTANCE OF SROUND OHM a a z—5 7 © LUeNGTW OF oRoUO- METERS Fic. Sad Vasiation of resistance of rerealgtound re with lenge far varius dametar 2 indicated co survey, for aa en7thresitiviey of 100 meters. a a) The resistance of rods or pipes of varlous lengths and diameters, a1 ‘calculated from the sbove formula for » = 100 meter-ohns, is given in the following table and is shown in Fig. 3.1: rlio2 5 © $0 100 « 36 US 3 6 SO ete 1 gs )ms 192 SDT 47 he as [is Ss MTS BO 43 Y xs list 95 47 SSS O72 40 2 so [us 7 wo B 16 «5 46 6 cs ee ee ee ce 30 he 30 | 4 3 2 A 90 46 26 RESISTANCE OF GROUNDING ARRANGEMENTS 77 in parallel, separated by the distance a2, the combined resistance obtained by use of (3.15) is 2 = o2)*s ai (3.26) 34h! G27} ‘The first of these formulas is the same as for a single conductor of cquivalent radius a’. In general, the combined resistance of equsl grounds arranged at equal spacing on a circle of diameter D small compared to the length of the round may be computed by use of an equivalent radius given by [ arte d=[ne 6) [". p< ‘When the # ground rods are arranged on a circle of diameter D and the spacing between adjacent rods is equal to or greater than the length of a 10d, as would more likely be the case in actual installations, the mutual sesistance may be taken as ® (aim) — 5” rain? as obtained by use of (9.15) is then (3.28) and the combined resistance Le Iggy tpt Be hile 1+ 5S, =a where D sin xm/m is the distance between grounds 1 and m, the radii which make the angle 2rm/n with each other. Thus, ° 9- Failed 1+ 5] 6.29) 1 a as~ 5ah[ve 1 pawl a woe ‘pgatol <1 ee ah ') Dine t Danae When a is replaced by aza in (3.24) and (3.25), these formulas give tt: _ For large values of , the summation in (3.29) may be replaced by the tual resistance bepween tito parallel grounds. With two ground rat integral between 1 and w — 1 with respect to a of (sin ome/n)", or by we ee ee kh 2 EARTH CONDUCTION EFFECTS Cun! (2a/r) log Qn/z). The combined resistance is then, GH) where sis the spacing between grounds and xD & ns when m is large. “When the spacing between rods is equal to or ess than the length of = fod, and the latter is smell compared t9 the diameter of the circle, the frst ‘mutual resistance terms in (3.29) should be calculated by the more accurate formula (324) until dja, > 4 alter which the remaining term is obtainet from (3.29) or (30). “The combined resistance of various numbers of typical rods on a circls ‘when the spacing is equal to ths ength of the rods, is given inthe following table in per cent of the resistance of a single rod, and is shown in Fig. 3.2 ee ee ee "i SF 62 AS IS 8S 48% Tn actuct installations the combined resistance may depart considerably from the values shown in this table beceuse of irregularities in the earth resistivity. Thus, the curve in Fig, 3.2 may be regarded only as aa expec- ancy curve, which might be obtsined by taking the average of a larg number of measurements on actual installations. The valnes in the table apply when the ratio of rod length to diameter 120, as for #”-5 fe of 1”=10 fe rods, although the values are not materially changed when the ratio varies in che range between 60 and 240. ‘When several ground rods are placed at equal spacing on a straight ling the current is not the same in all. To obtain the combined resistance, 3 would then be actessary to solve a number of simultaneous equations fo the currents. With soificient accuracy the current may be assumed, however, the same in all rods and the resistance calculated from (3.17), which is based on the average potential ofall the rods. RESISTANCE OF GROUNDING ARRANGEMENTS 79 Lid Porat nner afods (L412 poo?) appochs Coe erroter| (3.23) CONEANGD REMITANCE MH FER CENT OF SINGLE Rl RESIGTANCE E oa a) 5 2 suman OF noes Fin, 32 Variation in combined reetance of rods connected in mahiple when aranged ‘For rods ar a uniform spacing, 5, the combined resistance Rq thus be 00 8 stnight ine or « cca with spacing between rode equal mo length of ede, Dested Ri Bata) given by O38) and Sm ith a =m, | tae idles confined sedenae without maid oface. a cnr 8S by Ct al) br Gwe = feats temic ane nn nnn Sa Red Keg 20 cna et Re. 5 (Bitar Fe ARCS) + (DIR (Ds) Ra (ne ol) 31) When the separation between rods is less than the length of the rads, the first mutual resistance terms of (3.32) are calculated more accurately ‘fom (3.24) until the separation is greater than the length of the rods. ‘The routusl effect of the various ground rods is less when they ars on (6.32) a straight line than on a circle and for this reason the resistance of various When + J so that ons) & p/2ems, (3.1) becomes 8 = *a ee (he brit +4) 80 EARTH CONDUCTION EFFECTS Guat ambers of typical rods oa a straight line is slightly smaller than shows in Fig. 3.2. The difference ia insignificant, however, so that the curve shown in Fig, 3.2 may be used for both conditions. When s mumber of rods are arranged at equal spacing to fill «certain area, the combined resistance may be obtained by the wse of simultaneous equations or, with same sacrifice in accuracy whea equal currents art assnmed in all grounds, by methods similar to shose outlined above. Te the formulas above, an insulated connecting lend waa assumed betwees ground rods. Usually the rods are connected by a bare wire which, by itself, may have a resistance comparable to thet of the rods. Thus, wher ‘an extensive ground system is required, it may even be possible to dispens: ‘with the rads azd use « buried wire, a3 considered in the following, 3.6 Buried straigit horizontal wire In many situations where itis desirable to obtain 4 ground of low resist. ance, it may be convenient to install a buried wire ground, cither by plow ing or hand trenching. ‘To obeain the resistance of auch a ground, imagie «plane through the axis of the conductor considered in section 3.2. Th medium, the conductor and she conductor current to one side of the plare may be removed without disturbing the potencial, Since the total cos. ductor current is now 1(0) instead of 2/(0), the resistances of a condacte of length Z at the surface of the earth is cwice as large as thst given by 8.08) oF Ho) = Alcea vTE my +4- Vira] ox ge (he? ), when 1 6 G35 ‘The above fornmula assumes the conductor to be half buried, i.e. th ants of the conductor is assumed to be in the plane of the earth's surface When the conductor is buried at a depth d, the resistance is obtained by assuming two conductors a distance 2d apart in 2 medium of infinite exter in all dicections. By reasoning similar eo that of section 2.4 it then follom that the resistance of a buried wire ground equals. Ry = Kala) + Ria) a a S [lee aire 1], when de? 6x Thus, when a wire is buried at a depth d, its resistance is the same for a conductor at the surface having an equivaient radius (2ad)!!®, RESISTANCE OF GROUNDING ARRANGEMENTS 31 In the following tubie and in Fig. 3.3 are gi is nth nd in are given resistances of a No, 10 wie (@ = 125m) for different lengths, at the surface (d= 0) and buried at a depth of 2 = 30cm (12) in soil having a istivity of meter-ohms. Q eee") 4/1020 50100200 $0000 Fe a 306 S30 15000 meters 0 @ 49 4 We 56 24 129 Se ee 109 so ESISTANCE oF cnouM-OHMS ZS °° ae Uenors Or crouib-wetin 2 SP f0 Fg 33 Vasason of renee fra round with ne fr agnd ra ground athe sr face end at a depth of 30 cm, for an earth resistivity of maenckns a i dismeter of .25 em (410 wire). Shee eae ‘The moewal resistance of two parallel wites at depth « separ » at depth 1 separated by the siseance ai small compared to te length of the wire ic giees be flog 2 + tog 2 Blow) ~ F (062 + tH — 2) 627) where ia is the separation between wires, ale = (4d? + aly)! j dnsace fen one tte eng of ae ie Sa Gate fective separation. ‘The combined resistance ofthe two wires, 2 obtained ied se ee Cuan EARTH CONDUCTION EFFECTS ( where the equivalent radius a” of a wire at the surface is f= (ad) Cerra) OF With two wires, each of .125 cm (057) radius, 30 em 12") apart, 1 ‘a depth of 30 cm, the equivalent radius is: 10 cm as. compared with 2.7 cr for a single wire. When two wires are placed in parallel, the resistance! thus seduced about 22% when she wire is 300 meters a4. 30% when th vi ta eel wires are buried close together 0 that (3.57. may be used for their mutual resistance, an equivalent radius may b employed. Thus, for 3 wires of radii ay, 49 and og at different depths, ch equivalent radii with respect to conductors 1, 2 and 3 and the combing quivalens sadius for a wire ax the surface are oh = [(2a,d,)"? Carsatad sets) PS 2h ~ { (2a) (asgala) acta P = [(2agdg)" (arse Corea) 1 a= hab For a greater number of wires the equivalent radius mey be obtained a fairly obvious extension of (3.40). 37 Buried radial wires , Instead of using parallel wires, &t may be preferable to nin the wie radially fiom a conimon point. This arrangement is frequently employe for tadio uansmitting antennas and sometimes to provide a low resistant for transmission line towers. It has che advantage that the initial ug impedance to lightning currents is lower for wxiven direct current resstam than with 2 single or a few perellel wites. | For radial wires the following formula is obtained 82 by use of (3.15) is then Gat; G4 bby use of (5:16): m= Zoey 1+ veo at where, T+ sin m/s aE eg amen ox NO) = E00 sin cain & ( — 1) log (2- 1.707) — log when 7 >: n=? 3 4 6 8 12 100 N@)=7 1.53 245 442 65 TNO He RESISTANCE OF GROUNDING ARRANGEMENTS 83 When the wites are buried ar a depth a, che radius @ is replaced by the equivalent radius 2” = (2a2)"!?_ Tn the above formula it is assumed that the angle between any two adjacent wires is che same, 80 that in the case n= 2 the wires extend in opposite directions ftom a common point. The formula given shove i then the same as thar for a wire of length 2. TF the two wires were at right angles with each other, M(n) would have been 88 instead of .7. ‘As n becomes very large, the value of V(r) approaches 1.22% and the resistance then approaches R. = p/28T%, as compared with 9/42 obtained for a solid plate of zadius 4. The difference arises from the different dis- tribution ofleaksge current in the case of radial wires. Since the current is assumed to enser the earth uniformly along each wire, the density of the leakage current in a plate obtained by making m= < will be inversely propottional to the distance from the center of the plate, which differs ap- preciably feom the actual distribution in 2 solid plate. Actually, as the ruber of wires is increased, more corrent will enter che earth towards the tai of the wires. The mutval effect of wires will then be deszeaseé some- vehae and the 2ctual resistance will be somewhat emaller than calculaced from (BAT). For a very lari. number of wires, M¢n) should be taken there- fare 25 the limiting value 1/4 which would give a resistance more con- sinvent with the aetval current distribution. “The resistances of various numbers of wires of 125 em diameter buried at ‘adepth of 3 meter (12'”) are given in the following table and in Fig. 3.4 for an carth resistivity of 100 meter-ohms, 28 38 i ne 7139 Ohm 2 TA 89 Ds 4 3 sa 29 1s ” a as aa 13 « U 365 20 1.05 s a 33178 Ost “ 2 30157 at « Pate) Ls 083 0.92 “ a8 To obtain the resistance of a buried wire ring, itis necessary to modily che forratlas in section 3.2 and integrate along a circumference rather than along a straight Fine. Assume the center of the ring at the origin of the coordinares with the axis chrough the center of the ring and the x ad y axes in the plane of the ring, as shown in Fig.3.5. The distance from 3 EARTH CONDUCTION EFFECTS Cun RESISTANCE OF GROUNDING ARRANGEMENTS 35, wire elemene du toa point in the s-y plane is then (7? + x? — 2rx cos 6)¥2, where, is the radius of the ring and ¢ the angle between r and x. ‘The dis. ‘ance from a wire element to 2 point at the distance x from the xy plane ie (+ 38 +2" — Drv cos @)!. IF 7 is the current entering the wire, the leakage current in a wire element du = rdp is rdp(Z/2er). Ifthe ring is assumed at the surface of the ground, the potential at a point in the ground is then 8 2 do 0 -RL PaaT a iseare O) Bio Evaluation of this integral after substituting ¢ = x — 29 gives . Jo 2 Fir) 3 ven=Z Ties pee 4 ere 43) : where 7s ¢ -/—# > 3 ‘ ls PE a] 2 and FO)» [0 Baty tay ‘tthe complete elliptic integral of the fest ind, tabolated in Jahnke and Enide’s Tables of Functions, ‘The resistance is obcained by dividing by J in (3.43) and taking « = ¢ and 1 = a in the latter expression, a being the conductor radius; iy which ot Case, Exot oF whet - . ns Se canal esters (2 coun ee fe, fore sas aia 00 ethan We ame of ; ; cr ane Since @ Rar = RESISTANCE OF GROUNDING ARRANGEMENTS 87, Te will be noticed that, although the wire has a resistance compara Se 18 when the rode wre nterconsected by a bare wie. Alvi the ground consists of a iong buried wire, the installation of ground 1 ie redistance only some 25 to 30%. found rs reac th 310 Hpuivolent radius of flat conductors Ifa tape were used in place of a round wire, a somewhat lower res « used in oa wer resistance ‘woald he obtained with the same amount of copper. The equivalent zadlus of a zhin eape of width ag is 4 = af PP Day 6.48) ‘This formula is obtained by the average potential i a is 0 ¢ average potencial method, assumin tape to consist of a number of infinitesimal wires and to be long enought ‘that the potential varies logarithmically with distance. ‘The ratio of average potential co current is then proportional ro Sl Vim ig f° LP toe ce — wpa de = tog agi) By way of comparison, if the tape is bent into the form of a low cylinder, the sass of thelattes youll be an a= e4/2n% 1620 Bill Ect of seriaton im arth resstcty with depths As mentioned in section 2.10, when the earth resistivity vari 7 ity varies with depth, the mutual resistance between points on the surface may te approximated, in many instances, by the expression 26 Is has been shown previously chat, with appre eon, the resi ance of a wire may be based om the potential at the mid-paint. Tn the present case th esiatance this btaied fora wire a he seace a aah fot at da EE FR (oa en) exp = He + Sa Ts the latter expression it is permissible to take exp — (u? + a2)}%4 = 1 Deptt ~ Bs mE" (8.49) 35 cP — | 3, which cae the resistance in given by the following expres- G1 Ygsion when da <1 and af: aa eo 391s 63 ms wa 8k 52 26 eee as R= Biog! — [5 - #1072)] 3.51 6 8 68 B & Sa aT LM 0%) Ca ‘where Kis the expanential integral defined in Appendix I and y = 1.781" 88 EARTH CONDUCTION EFFECTS cnaes RESISTANCE OF GROUNDING ARRANGEMENTS 89 may be used: earth as the number of rods is increased. When the latter effect is abe |. When the latter effect is appre- Ciable, however, it is because of a fairly small thickness of the upper ver combined with a very low resistivity of the lower layer, in which case it When #//2 <2, the following approxim: = BiG/2) & log 281/2 — H/2 “Che resistance then becomes = Pye SPS 6) a When B1/2> 1, the Ei function in (51) may be neglected and He resistance fs thea sce Pt og WE 4 Pog t : Ro pen + Rb os Approximate values of } for two-layer strnctares were given in section 210. {a Fig. 3.6 ae shown caloulated resistances for a wire 150 meters (300 Jong of 2 em rudius (OB inch) at the surface of = horizonealiy stratifcs ‘Geolayercxcth, for vasious depths, dof the upper layer. Ie will be natiow Thats ohn the suzfree resistivity i low, the resistance af the ground may be appreciably higher shan for a uniform earth of the same sexistvity # at the surface, eves when the surface Jayer is of substantial thickness. I ie evident, then, that knovledge of the surface resistivity alone &s we faffcent fr sla eotimstes of the resistance to be expected for roun Tn the cue of a high surface resistivity up to a moderate depchy sy I smetuiy, ground of substandally smaller resistance may be obtained ly driving one or more rods or pipes into zhe lower Layee, ‘Thus, with S mete: thickness of the upper layer, the resistance of a pipe 10 meters long of Sex 34 Guameer iz ahoot 18 ofme as compared with 170 ohms for a 150-me' sree the surface, and about 130 dams for awe at S0-eM dep probably preferable to drive one ora few Barked cecresse in esti with depth sequen found in seilye, me 7 aga ne ee soilnao shit it may be possible te drive «pipe to considerable depeh withes . Ree oG te 3.12 Eifect of chemical wreament of soil ee eonsive ground fs constructed by driving a farge number Ifthe resistance ofa ground consisting of one or more i chee into the suslace layer, the resistance is aflested in substantalireduced, it may be okaille wise Staidig asst marghio oe Seen iene evarlation iy redtviey with depth as in che case ofesstivity of the soil around the rods by chemical treaxmert with certain the same: manner by vaveton ff TZigle eis determined eainly by thas such as sodium chloride (common salt), ealcum chloride or sod be en a of aasent monber af ros dspends ttnrte. When auch teatmen¢ is employed, is usually where the rei sure ey a ek ipvciotvity wich depen. Thus, whotivty is Very high so thatthe resist ofthe salted sol around the rod Fee ee paepi the ccesaf mural wsmuateebeewooma?be assumed to be negligible in comparison with the surrounding NUS. i pester than for uniform earth andas che number ofrodsisinercaxiearth. Consequently, dhe eect of salting isto increase che radius ries ea cnc ito rapidly. than for uniform cari round rods. ‘The sults have, ofcourse, «tendency to be abcorbed by th Fe oe a cae eeaiviny doors with dapth the oppssurounding each so thatthe redveton in fesistance may not be perma ect is observed, the resistance decreases more rapidly than for ubifornent, but may last a few years. inane at Dy ESISTANGE OF GRCUND- OHNE ae oak ‘iewvesd or urmen Lavee= meters Ec ova en incr esitvly wi depth onthe estan of ‘ground 150 meters long and of A cm diamecer buried at the surface. ae ee i a RESISTANCE OF GROUNDING ARRANGEMENTS 91 CONDUCTION EGFECTS Cmue! - re : | 100) = 2 pg UE + AP + 12 ‘The following table shows the resistances of various numbers of 1f iad i ay meter (5 ft) rods, when the radius is assumed to be increased by means 0 fae salting. The resistance ie shown in per cent of that abtained withor Big! when ac! ar Lo 12s » von 160 100 190100 % cy a uM we ag 2 By a log when d a - Tm cosme f° Ll sue" falre) de (402) where 1 is a numeric, cos = a/ra = (x48)! and Jy is an nth order Bessel function of the first kind. ane ‘As mentioned in 1.10, che primary TI fancticm of the dipole in the air is ‘taken as maf * mm pn) da 423) a where ay = (+ 8)", 40 = dolore)? and _ fel dS Ang The resultant II function will first be derived for a vertical dipole at the ‘eight A above a uniformly conducting earth. The resultant feld may be segarded as the suim of a primary and a sccondary eld, in the following manner: c Inthe sit: Ib, = 1+ Th, 2 >0 In the earth: Th, = 1%, z<0 The Primary field has for simplicity been omitted in the earth, although it sight have been included, asdiscussed later. The secondary and resultant fields must be independent of g, so that with m = 0 in (4012) the resultant fields become sae [eZ + em] 220 7 (408) Te = [het Ireydu <0 MUTUAL IMPEDANCE OF EARTH-RETURN CONDUCTORS 101 vwhere in the Brst integral exp — ag|2— | = expao(s— A) when <2< 4 and exp — ap(z— A) when = > 4. ‘The boundary conditions at = = 0 required that Fors Eig of Boy Buy How = Fhe ot Hayy = Fw Using the scomponents, i follows from (401) that, with » = 76: ls FH,» ae dz ax Oz aad (4.05) Mos 9 Ths ways Inserting (4.04) in (4.05), and omitting the integrals, the fellowing equa. ‘dons are obtained: Ca — age = afr [chem +o] =A (4.06) 07) where ae [0 BB ea “bo aolvieo Foal and RiaP+ GAs Rae + atay ‘The first two terms represent the I function for a perfectly conducting carth, due to the dipole and its image, whereas A represents the effect of the finite conductivity of the earth. If the primary field were assumed to permeate the earth, the integrand of (4.04) for Th, would have contained an additional serm CR eM Jo(ru) dhe. The function fy obtained from the boundary ss : equations would then dilfer by a corresponding amount from the fonction (Re e_—=Een 102 EARTH CONDUCTION EFFECTS curt ‘AA previously obtained, s0 that the resultant field would be the same by ‘both methods. ‘The same solution would also be obtained by assuming a resultant field in the tir and in the earth, without initially assiming any primary fald whatever. After the boundary conditions had been satisfied at 2 = 0, there Would then remain one function to be determined, which would be cbrained by observing that for 3» = 3, the solotion would have to agzez wich (4.03). The latter method will be used in the fallowing derivation for a horizontal dipole over a wwo-layer earth. “The dipole is assumed along, the s-avis at the height A above the earth's surface, and the earth is assumed to consist of two horizontal layers Subseripts 0, i and 2 sre used for constants of the air, the upper and the lower layers respectively, while d denores the depth of the upper layer. ‘The field is symmetrical with respect to thes ~ =plane, butnot with respect to the x — y plane, je,, the feld above the earth isdifferent from that below, For these reasons it is nccessary, in order to satisfy boundary conditions, to assurme a wave function with ewo components Il, and Hy. “The resuleant wave functions in the three media are taken as Ba Moxy Moss Th = Ua = Tar Has Boundary conditions at = and at = —d consist in the continuity of the tangential (7) components of E and H. Equations arising from these boundary conditions ean be simplified by differentiation or intcpration with respect to andy, which is permissible by virtue of (402) (i.e.,a diferentia- tion oF an integration of the Jo function gives # solution wich = 1). Using (4.01) with TL, ~ 0, and equating in order He, Hy, Fy, and Fay the following equations are obtzined from tie boundary conditions: 2-0 ues Tet od lor = Thy ith, — vibe {4.08 2 Aly: OTe athe 2 Alle ae Be = ge 409) Mog 5 Mos The _ Oy Alig hy _2Nlae | Tae eta ar as ae Tae ae 10) los ~ 1He vith, = Blse an) In (4.09) the 11, terms which arise on both sides have boen canceled by use of (4.08) and (4.02), and in (411) the 3U/ax terms have been canceled similarly by uve of (4.10) and (4.02). From the two pairs of equations (4.09) and (4-11), ehe s-comporents may be determined separately and then used in finding the =-components, MUTUAL IMPEDANCE OF EARTH-RETURN CONDUCTORS 103 For the x-components, (4.02) applies with » = 0, viz. oe =f" (fre + soe) Jolrw) dt O28 12) the =f" Hert mem pole) du, 4 S50 3) Mae = fo fart fiyeds, Sd (44) ‘The arbitrary functions are then determined from the following equa tions, obtained from (4.09) and (4.14): eo fo — go) = Hab ~ 21) Ue +) = At nd Har (ao ~ gre) = vie fae (4.18) Aer + mel) = The solution of (4.15) is anol 7 yh 2oolen + ae fs (4.16) if Paolo = 22) Loam np yn Ae eg te dlos + a) + (eo ~ ax)(ax — 9) where, ie may be recaled, = a 4 of. “The function fo inay now be determined by noting chat with sh = Yo = yo the selotion must be the same as for 4 woifrm mediam of Tnsnite extent in atl directions. In this case gi 0 and f= fo. To obeain the cate feld at z= —A as from (4.03), fy oust then equal a8 eu) tran. Thexefoe, fun sas BME) 5, an ” ) 4.17) = 14g itt Mle) simp at nares ee Nie may sow be ohttined from (413) and in torn Tye and Ips by use of @ll). “The boundary conditions forthe = components may be sstisfied by taking ae ee a es 104 EARTH CONDUCTION EFFECTS Court 7 = Lin 4.02) so that Boe = oo fm Jule) ae 18) The = cosy f (ret + ge) ir de (49) Mae = cos f~ peet™Jalru) du 4.20) where p and g have been used in place of f and x. From boundary conditions (4.08) and (4.10), together with the expres- sions for the » components as already determined, the following equations are obtained for the » and q functions: Hl. + a1) ~ ribo exile: — a2) = ~anripa + nah — DU + 0) Aare + get) = ype @21) aah re = ne) = ening + wo ~ eo + me) The solutions for p, and g, are y p= Baa A + 6) — OE = Dal hee + nn) Le Mae ae b= Wal Git wee + Gt — Dane + nw) ‘AGA + AG Aye Rae 4.22) an where Amani tan’, at = avi — ad As aritani, Ai = ant - ant components of the electric force at = 0 are by (401} Ben at Lit etre) de ‘The tangenti faa + Zoo [0 ui +) tale — alien) ae Fee Lah tan) + aalpy — ail flond ae cose [on tu = ~ 2 [PO an Since i ek _ MUTUAL IMPEDANCE OF EARTH-RETURN CONDUCTORS 105 these expressions may be rewritten as follows: Ba AL” Sit siddotre) ae -Bf[-usm+fe-moroa aan Bee Sf [itt So. - m|piea ee ‘These electric forces may be expressed in the form FO”) , POIrY Faby = 14s [—79) + 22D, 200) (424) where 14sP) = Af" + hein de PO = Etat eat (ou — endo Mole) dee (4.25) SL At Sos — afore ae 00) = [A] 408 - Praccte me EL el a ier ba) + Ger — deta Yolen) ae Am (ot ai)(a $29) + (eo ~ a:)fen ~ anh 1 = (oot + axrd)(or + ari) + Coot — avid) (air — aartde Ba = (aq + an} lerrd + agri) + (aa — 2) (ory — ernie In the same way as shown in Chapter I for a uniform medium, the ‘mutual impedance between wire elements 25 and dr lying in any horiznntal plane z = 4 and including an angle ¢ between their direction is = #00) 2, = ads [Pp eos + 8 2) “nn 43 Karthreurn mutual impedances in general Neglecting propagation effects, as discussed in 4.1, the mutual impedance of two wires J and sin a horizontal plane, extending from 4 to B and from 106, EARTH CONDUCTION EFFECTS Onan 410 6 respectively, is obtained by integration of (4.27) and equals “8 Td7O(r) Pam ff [ERE + Pee cose asar 428) iia = Cea + fols, (4.29) Quaren= ff 22 asa = Q(4a) — O(Ad) + O(BE) — O(Ba) (4.30) fata = ff" Povcoseds az 431) ‘The wires are assumed verminated ia che horizontal plane in which they are situated, the circuits being completed by radial flows fram the wire terminals, ‘This approximation is permissible for the usual type of circuit, where the mutual effects of vertical down leads may be neglected in com- parison with that of the horizonta! wire paths. The functions Qir) and P(r), given by (425) and (6.28) for the cate of two layers, are functions of che various propagation constants ae well 23, ofr. Ieis fairly evident that (428) applies to any horizontally stratified carth structure, since P(r} and Q(r) will then have fixed values for a given separation between wirt elements. It is noted that integration of Qfr) is accomplished by inserting the four distances r= 40, ft, Bb and Ba between the wite terminals, ix accordance with (4.30). For infinice dis. tances feten the einai of oe wire and chase he othr, Quansyime Both Qua s)ie») ard Lg, are generally complex-valued and thus do not represent resistance and induetance, as otdinarily defined, es might be implied by the similarity of (4.29) to che usual impedance expression. At zero frequency ial.g. vanishes and Qca_ajcea becomes QPesyiady # real number, the dc mutual resistance of the circuits as considered in Chapter II, with the superscript 0 used here for distinction. For frequen- cies sufficiently low, such that terms involving higher powers of the fre- quency in the expansions of the foxctions in powers of the frequency are negligible, the mutual impedance can be expressed jt the ordinary form 2s Bi = Pao + FoLa-ayu-n + LR] (4.32) Here Qfy2)io») 18 the direet-current mutual resistance while 224 is the ‘Neumann integral of the wires as defined in section 4.11, 2% pyi-a is the cocficlent of in che expansio® of Quayqao abd is equal to the sam of the Neumane integrals of the earth flows with the wites and with each i MUTUAL IMPEDANCE OF EARTH-RETURN CONDUCTORS 107 other, the exrth flows being those for direct cument. The bracketed terms ‘thus give the é-c mutual inductance "3 aoa of the wires with carth return, ‘Thesubscripts (§—) (se) indicates that the mutual inductance 4g beeween wire S with its earth return & and wire 5 with its earth return e, ‘The physical ¢istinction of ca—ayie-» and Lse may be illustrated by the following two cases: In che first, ove wire Js supposed straight and of arbitrary length; the second extends at right angles to it from two ground ing points and is closed at infinity (e., by a segment parallel to the frst wire and at such éistance chat its mutual impedance with the first wite is ‘negligibly small). In this case, in the perpendicular segments cos «= 0» anc in the parallel cegment PF) = 0, since r= =, g9 that Ze, = 0 and the mutual impedanee is given entirely bY Qra-syiet, ies the mutusl impedance depends only on the grownding points. Im the second cases the two perpendicular segments of the second wire extend away from parallel segmone to grounding points at infinity. Here the mutual pedance is given entirely by Lsy, since Q(°) and, therefore, Quayte-8 vvanishes for the limie7 = %. ‘In sections following, the rrutual impedas.cr of straight wires is examined farther, frst on the assumption that at least one wire is lorg enough to be regarded as infinite and then for wires of such lengeh that this condition is not satisfied. For a given allowable error in the mutval inductance, the Gistance « that one wire has to extend beyord the ends of the other to be rogarded a infinite is emaller the fonger the shorter wite. When one of ‘evo parallel wires is of infinicesimal length, anc great accuracy is required, fs in determining the earth resistivicy fom mutual inductance measure. ments, the distance should be such that y# 2 2 as shown in 4.11. Thus, Jor an carth resistivity of 1000 meter-ohms and a frequency of 60 cycles, the dissance « would equal 3300 meters, so that che cotal length of the Jonge wire would have to equal or exceed 6600 meters, abont 4 miles. For an earth resistivity of 100 meter-ohmns « total longt’ of 2100 meters would suffice, wherees for an carch resistivity of 10,000 meter-ohmns 21,000 meters would be required. When the earth resistviey is of the latter order of uagaitude, however, the mutual resistance may be more imporcant than the mutusl inductance, so thar a greater error in ¢he mutual inductance is permissible, Tn raost situations where magnetic induction is large enough to be important, one wire is long enough se thar i is permissible co use formulas for wires of infinite iength in calculating mutwal inductance, although the mutual resistance Qc4-aco- must also be considered. The Griteria mentioned above are important, however, it analyzing che results ‘of mutual impedance measurements, since a small departure in the mutual impedance may correspond to « large errar in she earth resistivity derived from such measurements,

You might also like