You are on page 1of 23

Horizontal Stratification of the Soil in Multi-Layer Using Non-Linear Optimization Estratificação Horizontal do Solo em Várias Camadas

Usando Otimização Não-Linear
Wesley Pacheco Calixto Electromagnetism and Electric Grounding Systems Nucleus Research and Development - Federal University of Uberlandia - Department Electrical Engineering, Av. Joao Naves de Avila 2160 Campus Santa Mônica Bloco 1E - Sala 1E20 CEP 38400-902 Uberlandia, Minas Gerais, Brazil wpcalixto@ieee.org Luciano Martins Neto Electromagnetism and Electric Grounding Systems Nucleus Research and Development - Federal University of Uberlandia - Department Electrical Engineering, Av. Joao Naves de Avila 2160 Campus Santa Mônica Bloco 1E - Sala 1E20 CEP 38400-902 Uberlandia, Minas Gerais, Brazil lmn@ufu.br Marcel Wu Electromagnetism and Electric Grounding Systems Nucleus Research and Development - Federal University of Uberlandia - Department Electrical Engineering, Av. Joao Naves de Avila 2160 Campus Santa Mônica Bloco 1E - Sala 1E20 CEP 38400-902 Uberlandia, Minas Gerais, Brazil marcel@mestrado.ufu.br Hipólito Barbosa Machado Filho Electromagnetism and Electric Grounding Systems Nucleus Research and Development - Federal University of Uberlandia - Department Electrical Engineering, Av. Joao Naves de Avila 2160 Campus Santa Mônica Bloco 1E - Sala 1E20 CEP 38400-902 Uberlandia, Minas Gerais, Brazil hipolito@ifgo.edu.br Resumo: O propósito principal deste trabalho é apresentar uma metodologia que utiliza um processo de otimização não-linear na obtenção dos parâmetros da estratificação horizontal do solo em várias camadas. O método utiliza uma

Revista Ciências Exatas e Naturais, Vol.11 nº 1, Jan/Jun 2009

curva de resistividade experimental proveniente de leituras feitas no solo. Esta curva experimental é comparada com uma outra curva de resistividade teórica produzida pelo processo de otimização. A curva teórica é embasada no algoritmo de Sunde que é exatamente o processo inverso da estratificação horizontal do solo em várias camadas. De posse das duas curvas, pode-se mensurar o erro produzido na estratificação horizontal do solo. Os parâmetros a serem estimados são os números de camadas N , a resistividade ρ i e a espessura hi de cada camada. Palavras-chave: Resistividade Elétrica do Solo; Estratificação do Solo; Otimização Não-Linear. Abstract: The main purpose of this paper is to present a methodology to obtain the stratification of the soil into horizontal multi layers from measuring results, obtained from the Wenner’s method. A mathematical modeling is presented combining the horizontal stratification of the soil method in several layers, with the non-linear optimization of Quasi-Newton method. The process initiates, calculating the resistivity of the first layer in a non-traditional way, and then calculating the resistivity and the thickness of the following layers, based on the Sunde’s algorithm, which is exactly the inverse process of soil stratification into multi layers. The parameters to be optimized in this methodology are: the resistivity ρ i , the thickness hi and the number of layers N . The results achieved point out advantages in comparison to other known methods. Key words: Electrical Resistivity of Soil; Stratification of Soil; Non-Linear Optimization.

1 Introduction
In an electric grounding project, it is essential to know the soil behavior, regarding to its electrical conduction properties. In metals, this property is represented by the value of its electric resistivity, which varies according to the

68

et al. Wenner’s Method Source: The authors 69 . justifying its wide use.e. one of them will fit best in the electric grounding project [3].CALIXTO. Wenner [2]. maintaining a great homogeneity along its dimensions. the electrical resistivity variations are not only extremely large. impurities etc. where each one of them has the same electrical resistivity throughout its points. In case of the soil. 1. (i) with how many layers its soil can be represented. Figure 1. carried out a measuring method. this model can be applied with a good accuracy in the major part of the electric grounding designs. However. and (ii) the values of thickness and resistivity of each soil layers. Since an electric grounding extension does not overtake more than a hundreds of meters. With a soil stratification model adopted. variations on the electrical resistivity of electric conductors are relatively small. a model that represents the soil by its electricity conduction behavior is developed. The method configuration is shown in fig. W. as well as it has a heterogeneous behavior along its dimensions [1]. it is necessary to know. Due to this difficulty. it is only necessary to define the area to be stratified i. The most common model is the one which represents the soil as horizontal layers. which the results contains the answers of those prior questions. Although this measuring method has some fluctuations. Frank A. P. metal temperature. to a certain area addressed to build an electric grounding. This characteristic represents a major obstacle for designers of electric grounding. where four align and equally spaced electrodes are inserted into the soil.

This curve stores information about the number of layers with their respective electrical resistivity and thickness. Knowing the soil’s layers. As it is presented throughout the paper. with their respective 70 . with a fixed reference point. Therefore. In fig. determinate the number of layers and its respective values of electrical resistivity and thickness. the resistivity of a homogeneous soil is given by (1). ρ a (a ) . 2 Soil Stratification The vast majority of electric grounding projects are a combination of vertical and horizontal ground conductors.. considering the Wenner´s method. V the relation R = becomes inversely proportional. 6]. If the soil is homogeneous.. varying the distance a .e. which characterizes a homogeneous soil with resistivity ρ . related to a . which is a function of a . the relation R = can I also varies. P is the depth in which the electrodes are immersed into the soil and R is the measured soil resistance. it is possible to rise the experimental curve ρ a (a ) .11 nº 1. From [5. i. 1. so ρ remains a I same value. using the Wenner’s experimental curve.e. i. advantages are found in using this method of soil stratification rather than other methods already known. V However. The main purpose of this paper is to show an efficient and accurate alternative way to stratify the soil. when a varies. This variable resistivity is going to be called apparent resistivity of the soil ρ a . when the distance a varies. ρ (a ) = a 4π aR 1+ 2a a + (2 P ) 2 2 - 2a 2 (1) (2a ) + (2 P ) 2 Where a is the spacing between electrodes. The methodology employed applies a non-linear optimization process based on the mathematical modeling deduced from the Sunde’s algorithm [7]. the voltage V can be measured through the internal electrodes [4]. the resistivity ρ has different values for each distance a .Revista Ciências Exatas e Naturais. Jan/Jun 2009 When a current I is forced into the external electrodes. in a heterogeneous soil. Vol. but not in the same way as in the prior case. ρ a (a ) .

producing superficial potentials that are reduced in small distances [12]. P. the current that flows from it tends to spread itself throughout the first layer. transforming the vertical conductor application less dangerous when related to step voltages[11]. resistivities and thicknesses. Another uncountable number of analyses can be done relating the grounding conductors with the soil’s structure in horizontal layers [14]. These analyses define extremely useful orientations criteria to the development of a good electric grounding project. producing surface potentials until great distances [9. it is probably that the grounding should be built superficial. the idea of comparison between the experimental Wenner’s curve ρaE (a ) and the theoretical 71 . 10]. i. assuring the best use of the soil. If the opposite occurs. A common case is the vertical conductors. However. Whether are known the soil structure values. mainly the superficial ones. When interpreted along with the values of soil’s layers. the depth that the vertical conductor should achieve can be defined. causing a abruptly decreasing in the potentials curves [13]. These phenomena smoothes the potentials curves. the current will flow deeper. the vertical conductor employ. those superficial potentials can define certain projects criteria.CALIXTO. In this case.. if this same superficial layer has high resistivity values. can become more dangerous regarding to step voltages. If there is a thick first layer with low resistivity in comparison to the others. W. so much regard with the grounding resistance. et al. 3 Mathematical Modeling The soil stratification proposed on this paper has on its basis. as well as in the potentials. leaving the horizontal conductors only with the role of electric linking. to work with vertical and horizontal conductors in a most efficient way. it is possible to model the grounding geometry. being more indicated the use of horizontal conductors [8].e. If it belongs to thick and low resistive layer. vertical ground conductors should be employed.

n 0 1 2 2 (4) It is possible to verify by Sunde’s algorithm that for E = 0 it is obtained N n = 1 and. which enables to express the characteristic function N n in E . 72 . equality (3) can be adopted. 2 n 1 2 (5) Assuming. created from a known soil structure. From [7]. i. number of layers N . the values of the soil structure that generates this theoretical curve correspond to the desired soil stratification. In this way. a polynomial expression described in (4) was chosen. we can choose j values for E . at first.. the basic idea is: when one theoretical curve equal to the experimental curve is obtained... using Sunde’s algorithm. considering a maximum admissible error. i. can be achieved. Vol. (2) can be written. therefore A0 = 1 . and N n is the Sunde’s characteristic function. of a generic layer i . So (4) becomes (5). aT ρ (a ) = 2 ρ a ∫ N (m )[J (ma ) − J (2ma )] ∞ a 1 0 n 0 0 (2) The first layer resistivity is ρ1 ..11 nº 1.e. If the soil structure is known. j . in this particular case. For each value E .. it is possible to obtain its characteristic function by Sunde’s algorithm [7]. thickness hi and resistivity ρ i . Therefore. ρ i and hi .e. N = 1+ A E + A E + . Jan/Jun 2009 Wenner’s curve ρ (a ). E=e-2mh1 (3) In (3) it is presented a variable change from m to E . a certain number of terms j to (5).2. This last one. from which the values for Ak with k = 1. In Sunde’s algorithm. equation (5) gives a system of j equations and j unknown variables. it is possible to obtain the corresponding value N n .. and.. remembering that 0 < E ≤ 1 . J 0 the Bessel function of type one and order zero.. N ..Revista Ciências Exatas e Naturais.. N = A + A E + A E + . m is the auxiliary integration variable.

S (a )   j a 1 1 k k 1 k =1 (11) 73 . To initiate the optimization process. i. W. 1+ 2∑ A . the function ρ a (a ) of (6) may be written assuming pairs of values of a . r (a )= r .k + 1 h1 a 2 2 (9) β= aT (10) To obtain the theoretical curve ρ (a ) equation (6) is used. minimizing numerical errors generated by the polynomial decomposition.b . ρ (a )= ρ . If the values of Ak are known. the electric resistivity of the first layer.S (a )   j a 1 k k k=1 (6) Where: S (a )= S + S k 1 2 (7) In which: S= 1 1 4. saving processing time. For this. Solving the integral in (2) and using some algebraic manipulation (6) is obtained. This aE became possible determining at first. et al. a1 and a2 . the equation (5) express N n as function of E . 2 1 b . 1+ 2∑ A . it is possible to solve the integral in (2) analytically. P.k + 1 2 2 (8) S = 2.e.. Considering this data.CALIXTO. it is interesting to obtain the initial values of the process coherent with the experimental Wenner’s curve ρ (a ) . It is possible by an iterative process to obtain the best values of j .

ρ (a ) = ρ (a ) 1+ 2 A .2 ) (16) Where: M (a k 1.2 )= S (a ).2 k 2 (17) B = ρ .S (a ) j (13) d ρ (a a 1. Vol.S (a )] a 1 a 2 1 k k 1 k 2 k =1 (14) From (13) and (14).Revista Ciências Exatas e Naturais.S (a )   j a 2 1 k k 2 k =1 (12) Dividing and subtracting (11) and (12). Jan/Jun 2009 ρ (a )= ρ .M (a k k k =1 j j 1. This process is explained in the next section. expressions (13) and (14). are respectively obtained.2 )= 1+ 2∑ A .2 k =1 d ρ (a 2 a 1. 74 . 1+ 2∑ A .2 ) = r ρ ( a ) -1 a 1.S (a ) k 1 a 1.ρ (a ) = 2.S (a ) k 1 k 2 (19) Thus.2 (15) ∑B .P (a )= k k 1.A k 1 k (18) P (a k 1.2 )= ρ (a ). equations (15) and (16) can be obtained.S (a ) ∑ k k 1 a 1 k =1 j a 2 k k 2 k =1 j r ρ (a a 1. respectively as: ∑ A . A [S (a ). ρ1 can be obtained from (15) and (16).∑ρ .2 )= S (a ).r ρ (a ).11 nº 1.

one can adopt the thickness of the first virtual layer as 10 cm. P. this inverse process is approached by finite iterations using the first order differentiations. the characteristic function N n can be obtained by Sunde’s algorithm. both with resistivity ρ1 . At the proposed methodology. Bk and h1 . from (15) and (16). ρ1 is obtained. Therefore. 4 Methodology When a1 and a2 are both modified by the algorithm iteration. the set of values N . From expression (6) the theoretical Wenner’s curve ρ (a ) can be draw. the Ak coefficients also can be obtained by using the polynomial (5). and thus.e. Considering that in the real world. This layer can be imagined as composed by two virtual horizontal layers. the number N of layers is a parameter to be optimized. W. as it is done in the gradient method [15]. 75 . Knowing the ρ1 value. a non-linear optimization method is employed. The optimization method adopted is called “Quasi-Newton’’ [15]. as it is done in the Newton’s method. regarding the number of inflections p presented at the experimental curve of apparent resistivity of the soil i. already known as the Secants Method its use is justified once it brings the simplicity of the Gradient method as much as the quickness of the the Newton’s method [16].CALIXTO. Rather than calculating the inverse of Hessian matrix. In order to generate this curve it is necessary to adopt values of N . ρ i and hi are being modified. remembering that ρ1 was definitely obtained. ρ i and hi represent a good stratification. expressions (15) and (16) represents two system of equations. N = p + 1 . h1 can be assumed to 10 cm. the goal is to get only the first layer resistivity ρ1 . The comparison between the experimental Wenner’s curve and the theoretical Wenner’s curve shows if the assumed values of N . aT At this point. so that by an iterative process. until a theoretical curve as close as possible of the experimental curve is obtained. ρ i and hi . Firstly. the thickness of the first layer is usually superior to 10 cm. The usual employed methodologies estimate the number of layer N . et al. for purposes of ρ1 calculation. So. where the unknown variables are Ak .

which are illustrated from 2 to 7 were obtained through an interpolation of the experimental curves points. Small deviation values are obtained between the experimental and theoretical curves of apparent resistivity of the soil. several measures are done in the field at different directions and. distinct from the other values of ρ i and hi . there is a different treatment for the N value optimization. Jan/Jun 2009 Assuming that ρ aEx is the experimental resistivity curve and ρ aTh the theoretical resistivity curve. imposing conditions for the approach function to be continuous and having continuous derivatives until a certain order [17]. two has resulted in a two layers stratification. from the Sunde’s algorithm. in general. This makes the curve to not have peaks and neither abruptly changes of curvature in its knots. is to demonstrate a new methodology to stratify the soil. proceeds from field measuring. obtained through a lower degree polinomial. The resistivity experimental curve. ρ i and hi to be optimized. Among them. the evaluation function f ( x) which is a metric is defined by: f (x ) = ρ -ρ ρ aEx aEx aTh (20) The Quasi-Newton algorithm updates the matrix with the parameters N . due to this. This interpolation is made with groups of few points. softening the curve. As the measure harvested in the field reflects the soil real configuration and. The N value. is a integer number. and the rest of them resulted in a four layers stratification. other one has resulted in a three layers.11 nº 1. and the theoretical curve. in 76 . some are discarded after analysis. Vol. 5 Results This section presents the results from six different soil stratifications process. minimizing the deviation value f ( x) found. The Wenner’s curves.Revista Ciências Exatas e Naturais. The experimental and theoretical curves were obtained from the interpolation of the soil apparent resistivity values. what we aim here. In the Wenner’s method application.

tab. fig. A larger amount of measured points. the tab. two distinct soil stratification methods were used. For stratifications numbers 5 and 6. with the Wenner’s curve represented in fig. 4 with the fig. 11. the interpolated curves are plotted on fig. 12. 14. At tab. being tab. One of them is the proposed method. 7 and tab. given at tab. heads to lowest deviations between the experimental and theoretical curves of apparent resistivity of the soil. The theoretical and experimental values of soil resistivity found by the proposed method are listed at tab. are presented at tab. for a larger amount of a values utilized to produce the experimental curve of apparent resistivity of the soil. i. P. For the process of Soil Stratification 1. and the results of this stratification. 3. 10 contains the theoretical and experimental values of apparent resistivity of the soil from the process of stratification number 5. tab. this work. W. and the other is known as the complex image method [18. found by the proposed method and the method of the complex image. 3 and tab. 2. 6 presents the curves interpolated with the points listed at tab. et al. 1. for the Wenner’s method application. 2 and the results of the stratification process (thickness and resistivity of each layer). related to the Soil Stratification 2.CALIXTO. found with the proposed method and the complex image method. using the method of the complex image.e. and the values found with the complex image method are listed at tab. 9 are listed the theoretical and experimental values of apparent resistivity of the soil from the the stratification process number 5. 9 and tab. 77 . are presented in tab. using the proposed method. are given at tab. 8 with the fig. The same sequence described above were made for the soil stratification process number 6. 13. 6 with the fig. 10. 4 to Soil Stratification 3. the data of the experimental and theoretical curves of apparent resistivity of the soil. the results of thickness and resistivity of each layer. 7. 5 to Soil Stratification 4. The same way is used for the others stratifications processes. the measures were harvested just in one direction. 5 and tab. 19].

08 2091.99 2257.62 1437.Revista Ciências Exatas e Naturais.57 Deviation (% ) 1.33 − 1.0 8. Jan/Jun 2009 5.0 2.62 1887.95 0.71 1850.Results Layer Layer Number Thickness ( m ) 1 d 2n s t Layer Resistivity ( Ωm ) 408.Experimental and Theoretical Curves a (m ) 1.95 0.72 inf 78 .35 Figure 2: Curve of Apparent Resistivity of the Soil from Stratification 1 Table 2. Stratification 1 . Stratification 1 .32 Theoretical Curve ( Ωm ) 653.81 1.0 16. Vol.11 nº 1.1 Soil Stratification 1 Table 1.0 Experimental Curve ( Ωm ) 641.27 2098.43 1014.77 1.29 1456.0 4.83 996.

47 Theoretical Curve ( Ωm ) 3663.08 1927.83 Figure 3. P.0 Experimental Curve ( Ωm ) 3582. Stratification 2 .2 Soil Stratification 2 Table 3. W.92 3354.0 16.CALIXTO.81 3258.0 2.73 2106.95% which corresponds to point 8 meters .0 4.Experimental and Theoretical Curves a (m ) 1.26 − 2. Through the tab. one can notice that the major deviation between the theoretical and the experimental curve is − 1.74 2210.58 2322.28 Deviation (% ) 2.97 0.57 1872. Curve of Apparent Resistivity of the Soil from Stratification 2 79 . 1. 5.88 2.05 2104.0 8. et al.10 − 4.

11 nº 1. one can notice that the major deviation between the theoretical and experimental curve is − 4.31 Through tab.67 9483.99 24970.31 2.Experimental and Theoretical Curves a (m ) 1.21 19966.35 1089.18 23716. Stratification 2 .3 Soil Stratification 3 Table 5.28 − 3.27 − 5. Stratification 3 .83 2007. Vol.15 Theoretical Curve ( Ωm ) 16120.15 Deviation (% ) − 4.21 2. 3.Revista Ciências Exatas e Naturais.0 Experimental Curve ( Ωm ) 16841.02 Figure 4.83% which corresponds to point 16 meters.0 4.75 inf Layer Resistivity ( Ωm ) 4120. Jan/Jun 2009 Table 4.Results Layer Number 1 d 2n 3rd 4th s t Layer Thickness ( m ) 1.0 8.61 0. 5. Curve of Apparent Resistivity of the Soil from Stratification 3 80 .11 − 5.26 15422.0 16.29 20715.58 8999.0 2.34 4220.20 15381.

4 Soil Stratification 4 Table 7.0 Experimental Curve ( Ωm ) 9560.17 11994. Stratification 4 .92 − 0. P.90 7328. 5.48 − 0.60 7073.CALIXTO.51 11105. one can notice that the major deviation between the theoretical and experimental curve is − 5.11% which corresponds to point 8 meters .37 Deviation (% ) − 2.06 0. Table 6. Stratification 3 .73 11990.Results Layer Number 1 d 2n 3rd s t Layer Thickness ( m ) 1.68 11135.09 21814.0 8. 5.26 3.78 inf Layer Resistivity ( Ωm ) 14315.38 Through tab.35 Theoretical Curve ( Ωm ) 9281.81 13535. Curve of Apparent Resistivity of the Soil from Stratification 4 81 .0 4.86 11591. et al.02 0. W.Experimental and Theoretical Curves a m ( ) 1.0 2.0 16.66 11589.03 Figure 5.

99 1078.46 82 .39 1073.11 nº 1.59 3.0 16. one can notice that the major deviation between the theoretical and experimental curve is 3.58 Layer Resistivity ( Ωm ) 8126.92 Deviation (% ) 2.0 4.0 8.30 4. Stratification 5 .59 1300.11 − 0.48% which corresponds to point 4 meters .34 1379.69 − 0. Stratification 4 .0 2. Jan/Jun 2009 Table 8.39 − 5.Experimental and Theoretical Curves (Complex image Method) a (m ) 1.90 Deviation (% ) − 1. Stratification 5 .0 Experimental Curve ( Ωm ) 1338.28 1459.Experimental and Theoretical Curves (Quasi-Newton Method) a (m ) 1.73 1301.78 1389.Revista Ciências Exatas e Naturais.86 1472.55 10257.07 − 0.57 9174.86 1472.76 1360.87 Theoretical Curve Quasi-Newton Method ( Ωm ) 1324.90 1075.0 16.58 1.0 8.0 4.87 0.Results Layer Number 1 d 2n 3rd 4th s t Layer Thickness ( m ) 0. 7.76 1360. 5.28 Table 10.15 1438.64 1302.87 Theoretical Curve Complex Image Method ( Ωm ) 1373.0 Experimental Curve ( Ωm ) 1338.32 inf Through tab.5 Soil Stratification 5 Table 9.0 2.99 1078.59 1300.63 0. Vol.09 5.68 17388.

ρ 3 and ρ 4 .78 7.52 137. h3 and h4 obtained from each stratification method. P. 11. Curve of Apparent Resistivity of the Soil from Stratification 5 Table 11.94 1538. Figure 6. ρ 2 . W. The values of each layer resistivities ρ1 .28 0.99 1866.52 7. et al.89 336.96 3. and the respectives thickness values h1 .CALIXTO.66 inf 83 .99 962.11 7.74 7.23 9208.31 inf Complex Image Method 1371. Stratification 5 .Results Layer ρ1 ρ2 ρ3 ρ4 h1 h2 h3 h4 Quasi-Newton Method 1137. h2 . are compared at tab.

38 690. Stratification 6 .82 Deviation (% ) 0. 5.0 Experimental Curve ( Ωm ) 3389.05 574.00 Theoretical Curve Complex Image Method ( Ωm ) 3226.22 585.0 32.49 − 22.0 16.6 Soil Stratification 6 Table 12.00 568.76 88. 6. 10 and fig.0 4.17 1609.0 8.46 2294. were found.72 Deviation (% ) − 4.01 − 15.0 4.0 32. However.00 585.00 568.00 Theoretical Curve Quasi-Newton Method ( Ωm ) 3389.00 1900.00 585. there is a great proximity between the results obtained by the stratifications made with the proposed method and the complex image method.Experimental and Theoretical Curves (Complex Image Method) a (m ) 2.00 823.95 21. smaller deviation between the curve of apparent resistivity of the soil obtained with the proposed method and the experimental one.15 − 6. Vol.48 1105.15 84 .0 16.22 Table 13.Experimental and Theoretical Curves (Quasi-Newton Method) a (m ) 2. Stratification 6 .11 nº 1.80 20.0 Experimental Curve ( Ωm ) 3389.01 1.00 823.00 1900. Jan/Jun 2009 As can be observed at the tab. tab.07 640.55 771.30 0.Revista Ciências Exatas e Naturais. 9.0 8.

10 inf inf The values of each layer resistivities ρ1 and ρ 2 . Figure 7.15 570. In the soil stratification process number 6. showing a improvement in the proposed stratification method. Stratification 6 . the number of layers N was forced into N = 2 . 14. et al. and the respectives thickness values h1 and h2 obtained from each stratification method.CALIXTO. when compared with the complex image method. However. in soil stratifications processes. there is a great proximity between the results obtained by the stratifications made with the proposed method and the complex image method. 7. Analyzing the curve of experimental apparent resistivity of the 85 . tab. were found. are compared at tab. W. which is a very usual employed method.Results Layer ρ1 ρ2 h1 h2 Quasi-Newton Method 99905.36 Complex Image Method 3550. Curve of Apparent Resistivity of the Soil from Stratification 6 Table 14.00 630. 13 and fig. smaller deviations between the curve of apparent resistivity of the soil obtained experimentally and through the proposed method. P. 12.77 0.00 3. As can be observed at the tab.

would be already enough to estimate a number of layers N > 2 .Revista Ciências Exatas e Naturais. invalidating the rest of the stratification. N = 2 . Nevertheless. 12 and tab. Jan/Jun 2009 soil. it is observed through tab. 6 Conclusion This paper has presented a methodology and a mathematical modeling for the soil stratification in multi layers. tab. even forcing the number of layers to N = 2 . This fact. using the described equation in section 4. At tab. Knowing that a exploitation method is purely mathematical. 86 .e. it can be notice that the number of curve inflections p = 2 . A such order error is unacceptable for a ρ1 value. when the curve ρ a (a ) does not have a good behavior. the results have presented lower or equal deviations to those found in the pertinent literature. Vol.11 nº 1. N = p + 1 . the resistivity and the thickness of each layer are calculated and optimized from the experimental curve of apparent resistivity of the soil. where the number of layers. In all cases in which the stratification was made by the proposed methodology. Some error at ρ1 can be tolerable. Whether this stratification was made considering a number of N = 3 layers. the obtained results can reach errors which overtakes 200% . the deviations would have a considerably decrease. but. 12. Several cases were processed. i. where ρ a (a ) is extrapolated until the point where a is equal to zero.: (i) The most common method employed to determinate the first layer resistivity is the extrapolation of Wenner’s experimental curve. 13 the listed deviations presented to be larger than the acceptable in grounding projects. fig. 7. evidences the necessity of a methodology that optimizes the number of layers N in the soil stratifications. This horizontal soil stratification is a typical case of stratification. in this case. something that frequently occurs. 7 that the proposed method achieved lower deviations values than that obtained through the complex image method. both to the two applied methods. 13 and fig. where the number of layers N is forced in a fixed number. This is explained by the facts that differs this methodology.

some soils behave differently. Although. the great amount of small errors that appear in this paper. By the proposed methodology. for example. ρ1 is obtained from the derivations expressions of (6). it is concluded that the soil model made by horizontal layers is not applicable to this type of soil. considering its electrical conductibility in horizontally and uniformly layers. et al. In electrical engineering. the set of unknown variables are: N . the stratification of the soil for electric grounding construction project. the number of layers is not a given date of the problem. turning its accuracy not reliable. opens up to new investigations possibilities. motivates further studies in order to obtain lowest errors on soil stratification. representing the physical signification in attendance with Wenner’s measuring method. a very accurate stratification process is important. Actually. a more appropriated study on the optimization process. Therefore. One parcel of this error. the accuracy obtained at this work. This explains why sometimes the accumulated error of stratification is not due only to the numeric process of curve adjustment. adopts the soil as a model. Still. In this situation. (ii) Soil stratifications considering a fixed number of layers are common.CALIXTO. resistivity and thickness of each layer. there is something more to be analyzed. If the stratification accuracy is tied up to the project and the electric grounding construction. there is no sense in using errors lower than one unit percentage. Finally. W. sometimes the major one. what justifies. but an unknown variable. this is enough to interfere on the results of Wenner’s method. In the proposed methodology. but a high error is still verified in the values obtained at the stratification. An error around 10% is perfectly feasible. 87 . P. constructions imperfections of electric grounding justify this order of magnitude. ρ i and hi . depending on its order. can be caused by the fact that the soil does not have uniformity in its horizontal layers. Restrictions on the proposed method can appear with the existence of a vertical or oblique rift in the layer. So. it can propagate throughout the stratification. soil stratification involves three unknown variables: number of layers. This ensures that ρ1 accuracy depends mostly on the adopted model to represent the soil. If it is assured that the error made at the curve adjustment is small.

J. FORTIN. Inc. Jan/Jun 2009 Acknowledgment This article results from the researches of a P&D project (Development of a Philosophy of Electric Grounding Project with Practical Extension) sponsored by Energetic Company of Goias .’’ Institute of Electrical and Electronics Engineers. n.11 nº 1. and Earth Surface Potentials of a Ground Systems.. [6] IEEE Std 1100-1992. A. Washington D. References [1] IEEE Std 81-1983. 12. ZONG. p. 37 n° 5 p. 1968. [3] H. 1916. W. 2006.’’ MacMillan & Company. 88 . YUAN. Inc. WENNER. [4] P. SUNDE.. v.Revista Ciências Exatas e Naturais. Power Deliver.’’ IEEE T. 21. 4. London. LAGACE. H. ``Recommended Practice for Grounding of Industrial and Commercial Power Systems. v. Magn. YANG.’’ IEEE T. ``Method of Measuring Earth Resistivity.’’ Institute of Electrical and Electronics Engineers. ``Guide for Measuring Earth Resistivity.C. M. [7] E. ``Recommended Practice for Powering and Grounding Sensitive Electronic Equipment. Inc. ``Multilayer resistivity interpretation and error estimation using electrostatic images. VUONG. D.’’ Bulletin of the National Bureau of Standards. J. J. ``Earth conduction effects in transmission systems.CELG jointly with The Brazilian National Agency of Electric Power . 1954-1960. Revision of IEEE Std 81-1962. LEFEBVRE.ANEEL. v. ground Impedance. [2] F. 36843687 2001. Vol. M. ``Determination of Three-Layer Earth Model from Wenner Four-Probe Test Data.’’ Institute of Electrical and Electronics Engineers. [5] IEEE Std 142-1991.

H.’’ IEEE T. P. 1974. Power Deliver. D. ``Optimum Design of Substation Grounding in a Two Layer Earth Structure. Phys. D. Syst. ``Earth Principles and Practice. 1992. W. Power Deliver. B. ``Nonlinear Programming: Theory and Algorithms. L. BAZARAA. 89 . ZHANG. ``Numerical Methods for Unconstrained Optimization and Nonlinear Equations. [19] B. [16] M. DAWALIBI. Power Ap. B. [17] E.’’ Pitman and Sons. Appl.’’ AIEE transactions. paper 59-806. X. nov 1959. ORELLANA. J.’’ J. L. TAKAHASHI. YANG.’’ IEEE T. [14] R. D.. L. [11] F. SHERALI. ``Grounding Resistance of Buried Electrodes in Multi-Layer Earth Predicted by Simple Voltage Measurements along Earth Surface: A Theoretical Discussion. 1996. C. R. D. RYDER. DAWALIBI et. [10] F. 1966. P. [9] Y. E. K. p. CUI. KINYON. J.. 2005. [13] A. ``Correlation of Measured and Calculated Substation Ground Grid Resistance.’’ Wiley. ISSACSON. H. CHOW. HE.’’ IEEE T.CALIXTO. SRIVASTAVA. ``Ground Electrode Resistance Measurements in Nonuniform Soils. Power Ap. 569-574. J. p. SRIVASTAVA.’’ SIAM. Part I-II-III vol. 7. KELLER. DENNIS. 378 p. 853 p. v. W. v. [18] Y. 20 n. L. J. ``Complex Images of a Ground Electrode in Layered Soil. 1991. al.. 10 1995. J. K. SCHNABEL. ``Prospeccion Geoelectrica en Corriente Continua. 1394-1401. [12] T.’’ John Wiley and Sons. M. CHOW. ``Calculation of Earth Resistance for a Deep-Driven Rod in a Multi-Layer Earth Structure. v. and Syst. 1993.. 1969. [15] J. SHETTY. 2. PAS-93. 1974. PAS-94 1994. Madrid.’’ IEEE T. v. ``Analysis of Numerical Methods. ``Parameter Estimation of Horizontal Multilayer Earth by Complex Image Method. [8] E. v. YANG.’’ Paraninfo. London. 71. et al. MUKHEDKAR. LI. S. P. Power Deliver.’’ IEEE T.