(3-8 “ Putting X = log x, so that x 2 -= » (@) takes the standard form (a + 3] =1 wali) Let =X + ay and put 3 = Sand 5 in (i), so that de)? | 42(dz)* Dede = (2 +a (2) 1 or fasa%2de=2du Integrating, Va+a%)2? =22u4b= 4X +ay)+b or 2 fd +a*) = £2 (logx+ay) +b which is the complete solution required. Form IIL. f(x, p) = F(y, q), i¢., equations in which z is absent and the terms containing x and p can be separated from those containing y and q. As a trial solution assume that f (x, p) = Fly, q) = a, say ‘Then solving for p,weget p= 4x) and solving for q, we get =v) , = dB dy = Since de = Fd + dy = pe + ady a dz = olx)dx + wordy Integrating, 2= foods + [yndy + which is the desired complete solution containing two constants a and b. Solution, Given equation isp? -x = y-q?=a, say ‘ pi-x=agivesp= Jats) and y—q?=agivesq= (@—a) Substituting these values of p and q in dz = pdx + qdy, we get dz = ax) de + (y~a) dy integrating gives, a= Zlasais 2y a+b which is the required complete solution. Solution, The equation can be reduced to the above form by writing it as fd) and putting(PARRADD FERENTIAL EUAN Eu = (i) becomes Pha Qraxtay? or PP x2 =y?—@ =a, say. P= Ja? +a) and Q= y?-a). dZ = Pdx + Qdy gives dZ = \ix* +a) de + Jy? ~ fre feo] = [: 4 pre [Jl re ax) a 4sHence the C.S. is Solution. Given equation i.e., in symbolic form (D? - ADD’ + 4D)z = e* +», Its AE. is (m—2)*= 0, whence m = 2,2. CF. =f, + 2x) + xf,y + 2c) week easy PL= Dope ‘The usual rule fails because (D — 2D’)? = 0 for D = 2 and D’ = 1. to obtain the P.L., we find from (D ~ 2D") = e**, the solution we J Playc—maddx= fe? de = x0 = 2019 Ie y=e—mx=e- 2x} and from (D ~ 2D’)z = u = xe***?, the solution ze fret Pay =k ef = Eatetety te y=e-mx=e-2r) Hence the CS. is z = f,(y + 2x) + xf,(y + 20) + Bre, Solution. Given equation in symbolic form is (D* + DD’ — 6D)z = cos (2x + y) tA. is m? + m—6 = 0 whence m = ~3, 2. CF. = f,y ~ 3x) + fy + 2x). Since D* + DD‘ 6D? = — 2° — (2X1) - 6-1)? = 0 It is a case of failure and we have to apply the general method. coe (ie +9)! eng (itz ay) 1 |= ey Dp ep? (D¥3D XD - 20° = pcap| fomeeee=BHas| = praplloset| [s y=e-me=0-2e = pdpresoran = [freee znd] = [frawrrerde] . (eee ie cotessa 5 [Integrating by parts] sige = Ssin (Gx + y—3x) + Leos (5x + y=3s) = Fsin x+y) + Leos (2x +9) 5 25 5 25 Hence the C.S. is 2= fly — 8x) + fly + Be) + : sin (2c + y) + Frey) 2 =f — 8x) + fly + 2x) + = sin (x+y) J eos 2x +9)_ Parma Dievenenmiat Eouarions oz Example 17,31. Solve 22 a or r+s-6t=y cos x, Solution. Its symbolic form is (D? + DD’ - 6D")z = y cos x and the A.E. is m? + m~6 = 0, whence m = 3, 2. CF. =f0- Rese ee) PL y 008 x = a-aa 1 . = oplet Sx)sin x +3 cos x]... 5, . 1 D-20" (sin x + 908 x) = [ J (c~2e)sin x3 cos x} dx] (Anna, 2005 S ; U.P-T, 2003), Bhopal, 2008: SW-T.U, 2008). DoaplJerseosxas| ls y=e-mx=e+2e] UIntegrating by parts} leay-2x = Mle = 2x) 008 x) ~(- 2X sin x) +3 sinx,,, 9, =~ycosx+sinx Hence the OS. is z = f,(y — 3x) + fy + 2x) + sin x —y cos x. ae ve Example 17,52. Solve 45> are = 16 log @& + 2y), Solution. Its symbolic form is 4D? - ADD’ + D’* = 16 log (x + 2y) and the A.E. is 4m? —4m + 1=0,m = V/2, 1/2, CF. fi(y+ds) xn (yds) as artay PL. = — > 16 log (x + 2) = 4- + ] tog (x + 2) (2D- Dy (> 1p XZ x tog {s+ ez ed ls y=e-mx=e~x/2] a ( 3) syed [ fiom eras] [x log (x + 2y)] DE te yD Ty = = = =4{f{s ww[era(e-S)fe] = iota 2e2 log (x + 2y) Hence the C.S. ise= A(y+ 3) +an(y+ 2) +22 tog e+ 29) a ‘Bove . following equations = ae, fe Ply 2 pe +35 tage er ag en eee (Biurdivan, 2003) 3. W®—2DD'+ D2 =e*% (Bhopal, 2007) a Fel Oe ee ay ois Ble andy”Ea Haver Enaneenine Martenanics: Oe 4 0 O25, , oy _jod’y & 55-27 +—F =sinx, (P.T.U., 2009S) 6 SFP F -E a Tay ae ae a TE ca SA a Bey 2ein ae 2p. (SV..U, 2007) 8, (DS = 7D)? — GD')z = cos (x + 2y) + 4. (Anna, 2008) 5 2, 9. = ae as =e? 40°79 + 008 (x +29). (UPQW.,2006) 2, 10. SFA = in x cos ty. (UPT.U.,2003) 11. (D* = DD w= e08 By (sin'r + cos 2). 12. DID %)2 = e-¥sin (+ 2y). Anna, 2009) (D? + SDD’ + 20% = Dixy. Of | Oe ote Me Sati te etree 15, (D®—-DD = 207 = G1) e% Bhopal, 2006) 16, WS + DAD! —DD2- Ds = e* cos 2y. 17. (D8+2DD'+ De =2e08y-xsiny, (PTU,, 2005) EYBER NON-HOMOGENEOUS LINEAR EQUATIONS If in the equation FD, Diz = Fox, y) o() the polynomial expression f(D, D’) is not homogeneous, then (1) is a non-homogeneous linear partial differential equation. As in the case of homogeneous linear partial differential equations, its complete sol P+ PLL ‘The methods to find P.I. are the same as those for homogeneous linear equations. To find the C-F., we factorize f(D, D’) into factors of the form D — mD’ ~c. To find the solution of (D— mD’ —clz = 0, we write it as p — mq = cz 2) ‘The subsidiary equations are dx_ dy _dz ‘Tom @ Its integrals are y + mx =a and z = be. Taking b = (a), we get z = e* oy + mx) as the solution of (2). The solution corresponding to various factors added up, give the C.F. of (1). ample 17.82. Solve (D* + 2DD' + D® - 2D = 2D'z = sin (x+y). Ww. Solution. Here f(D, D’) = (D + D'XD + D’~2) Since the solution corresponding to the factor D ~ mD — ¢ is known to be zZ=e* Oy + mx) CF. = 6,0 -2) + £9 -2) 1 D? +2DD'+ D? -2D-2D" 1 “Cao 2-2 PL sin (x +2y) sin tendigipans 2(D+D)-9 1 ee et) 8 sin (x + 2y) XD+D)+9 4D?+2DD'+D®) 81 = Fleas e+ 2) 3 sin & + 2yI1 Hence the complete solution is 2= 60-2) +e* Oy —x) + Fie eos tx + 29)~B sin e+ 2)NON-LINEAR EQUATIONS OF THE SECOND ORDER We now give a method due to Monge’, for integrating the equation Rr + Ss + Tt= V ) in which R, S, 7, V are functions of x, y, 2, p and q. Since dp = Pac Pay =rdx + tdy, and dq = sdx + tdy, we have r= (dp - tdy)idx and t = (dq ~ sdxV/dy. ‘Substituting these values ofr and ¢ in (1), and rearranging the terms, we get (Rdpdy + Tdqdx ~ Vdxdy) — s(Rdy? — Sdydx + Tdx?) = 0 w2) Let us consider the equations Rdy? — Sdydx + Tdx* = 0 wl) Rdpdy + Tdadx — Vdxdy = 0 Ad) which are known as Monge’s equations. Since (3) can be factorised, we obtain its integral first. In case the factors are different, we may get two distinct integrals of (8). Either of these together with (4) will give an integral of (4). If need be, we may also use the relation dz = pdx + qdy while solving (3) and (4). Let u(x, y, 2, p,q) = @ and v(x, y,z, p,q) = b be the integrals of (3) and (4) respectively. Then u = a, v = b evidently constitute a solution of (2) and therefore, of (1) also. Taking b = (a), we find a general solution of (1) to be v = (u), which should be further integrated by methods of first order equations. Solution. Monge’s equations are ady? + (x + yMdy dx + ydx? = 0 mo) sedpdy + ydgdx~ = *(p—q) dye = 0 fii) (i) may be factorised as (xdy + ydxXdx + dy) = 0 whose integrals are xy = c and x + y =c. ‘Taking xy = c and dividing each term of (ii) by xdy or its equivalent —ydx, we get dp-@ _dz-) 4 fre x-y This gives on integration (p — g(x —y) = ¢. Hence a first integral of the given equation is p —q = (x -y) @(xy) which is a Lagrange’s linear equation. Its subsidiary equations are dx_dy_ de (x—y)@xy) From the first two equations, we havex + y=a Using this, we have dz =~ (ax —x®).(a—2x)dx which gives z = 6,(ax—x?) +b Writing b= $,(a) and a=x+y, we get 2 = Oxy) + O42 +9)- + Named after Gaspard Monge (1746-1818), Professor at Paris.Ea Hionen Enanesina Marvenanis (Obs, Had we started with the integral x + y =e and divided each term of (ii) by x or dy, we would have arrived at the same solution. |) Bxample 17:36. Solve y'r—2ys +t =p + 6y: (Osmania, 2002) Solution. Monge's equations are y2dy? + 2ydydx + dx? = 0 and ydpdy + dgdx ~ (p + Gy\dydx = 0 fii) @ gives (ydy + dx)? = Oie. y?+ 2xe=e iii) Putting ydy = - dx in (ii), we get ydp - dq +(p + 6yldy =0 or (ydp + pdy)—dq + 6ydy =0 whose integral is Py + 89% Combining this with (iii), we get the integral py ~q + 8y? = Q(y? + 2x) ‘The subsidiary equations for this Lagrange’s linear equation are a YL oy? +2x)-3y? From the first two equations, we have y* + 2x =c Using this, we have dz + [9(c) - 3y*I dy = 0 whose solution is 2 +y@lc)-y2 =. Hence the required solution is 2 = y*~ y(y® + 2x) + y(y? + 2x). Ea Sae t Solve = | A qe Ds= +e 2 r= tcos?x pp tans =0. B. Brtr= Says + 2y*%t+ Apx + gy) =0. WN.T.U., 2006) 4. xylt— r+ (x*—y*) (s —2)= py - axe 6. @?r—2pas + p*t = pq? 6. (14g) 2 4 p +4 pals +(1 +p =0. OBJECTIVE TYPE OF QUESTIONS ea Sered Fill up the blanks or chiooge the correct answer in each of the following problems : ind 2 a) i og 1, The equation 5 120(%) +5 = Bis of onder. and degree..4 2, The complementary function of (D? —4DD’¥ 4D =x #y is 4, Asolution of (y —2)p +(2—x)g = 2 -y i... ‘The solution of S =sin Gy) is ‘The particularintegral of (D?-+ DD’ = sin (x yVis ‘The partial differential equation obtained from 2 = ax + by +b by eliminating and bis... Solution of Jp + /g = Lis... 8. Solution of py= + gfy =z is..... Solution of p ~g = log (x + y). 10, ‘Theorder of the partial differential equation obtained by eliminating from z = f(x? +y), is 11, ‘The solution of x defdx = 2e + is 12, By climinating a and b from 2 =. (x + y) + 6, the pve. formed i 13. The solution of (0° -8D°D' + 2DD7\z =0 is... 14, By eliminating the arbitrary constants from z = ax + ay? + b, the partial differential equation formed (Anna, 2008) eo pa gs 7 2 ou _ Ou. 6, ‘1fu =z? + isa solution of «> pgm then 15, A solution of u,, = 0 is of the form .. P ie.,dx 0, we have 22-0? ZZ, This is the partial differential equation giving the transverse vibrations of the string. It is also called the one dimensional wave equation. (2) Solution of the wave equation. Assume that a solution of (1) is of the form z = X(x)T(t) where X is a function of x and T is a function of ¢ only. where e? = 2 ol) ™Th SY =x. 7 and 2 =x. en and xv 17" 1 2 xX 2T > Clearly the left side of (2) is a function of x only and the right side is a function of f only. Since x and ¢ are independent variables, (2) can hold good if each side is equal to a constant k (say). Then (2) leads to the ordinary differential equations : Substituting these in (1), we get XT" oS -8X=0 (3) and oF ker 0 wd) Solving (3) and (4), we get (@ When k is positive and = p®, say X= ce" + ee™ ; T= cye™ + ce. (ii) When h is negative and =~ p® say X= c, cos px + ¢, sin px ; T= c, cos cpt + cg sin cpt. (ii) When k is zero. X= gx + ey 3 T= Cyt + Cy. Thus the various possible solutions of wave-equation (1) are (eye + cg6° (ese + ce) 45) (c c08 px + ¢,.sin px)(c, cos ept +g sin ept) (6) (gk # Cy MCyyt + C42) AD) Of these three solutions, we have to choose that solution which is consistent with the physical nature of the problem. As we will be dealing with problems on vibrations, must be a periodic function of x and t. Hence their solution must involve trigonometric terms. Accordingly the solution given by (6), i.e, of the form y= (C, cos px + C, sin px) (C, cos cpt + C, sin ept) (8) is the only suitable solution of the wave ek (Bhopal, 2008) .A string is stretched cdo ton nalts Lona Metla io yeasin cae, anti 2 -Show't eee ‘ in (x 608 (ne M0). “vau, Us, 201 58. ) ay _ 2ay Solution. The vibration of the string is given by 7 =e? 4) he ‘As the end points of the string are fixed, for all time, yO, )=0 ii) and y(l,t)=0 Since the initial transverse velocity of any point of the string is zero, therefore, (). =0 weft) Also ye, 0) =a sin (nx/l) lv) Now we have to solve (i) subject to the boundary conditions (ii) and (iii) and initial conditions (iv) and (v). Since the vibration of the string is periodic, therefore, the solution of (i) is of the form yx, t) = (C, cos px + C, sin px)(C, cos ept + C, sin ept) vi) By (ii), (0, t) = Cy(C; cos ept + C, sin ept) =0 For this to be true for all time, C, = 0. Hence yt, t) = ©, sin px(C, cos ept + C, sin ept) (vii) and 2 = C, sin px (Cg(—ep . sin ept) + C,(ep . cos cpt)) By (iv), (2) =, sin px . (C, cp) = 0, whence C,C,cp = 0. IC, = 0, (vii) will lead to the trivial solution y(x, 6) = 0, the only possibility is that C, = 0. Thus (vii) becomes y(x, t) = CC, sin px cos cpt elviti)By (iii), Since C, and C,#0, we have sin p= 0... pl=nn, ie.,p=nall, where n is an integer. Hence (i) reduces to yx, Csi 7 cos MA [These are the solutions of (i) satisfying the boundary conditions. These functions are called the eigen functions corresponding to the eigen values 2,, = cnvl/t of the vibrating string. The set of values i, hy, Ay, is called its spectrum] Finally, imposing the last condition (v), we have y(x, 0) = C,C, sin = =a sin which will be satisfied by taking C,C, = a and n = 1. net Hence the required solution is y(x,#) = @ sin * cost oli) a Solution. The equation of the vibrating string is Fee ‘The boundary conditions are y(0,t)=0, (J, ) = 0 Also the initial conditions are y(x, 0) = yg sin® (®) a =0 Gv) an ., Since the vibration of the string is periodic, therefore, the solution of (i) is of the form ‘ye, t) = (c, cos px + ¢, sin px) (c, cos cpt + ¢, sin ept) - By, 10, t) = e\(c, cos ept +c, sin ept) = 0 For this to be true for all time, c, = 0. & ‘yla, t) = ¢y sin px (cy cos ept + ¢, sin ept) Also by (ii), _y(l, ) = ¢, sin pl (, cos ept + c, sin cpt) = 0 for all t. This gives pl=nx or p=nwl,n being an integer. ‘Thus yx, =e, sin recy cos ee cy sin out) mO} By (iv), ‘Thus (v) becomes y(x, t) = ec, sin ‘Adding all such solutions the general solution of (i) is, ye, 0 = > 2, sin ™ cos ae (vi) nat"APPLICATIONS OF ParTiAL DirFeneNTIAL EQUATIONS: 60: from (iii), Yo sin’ =e 3sin _ or % to =hsin + b,sin 22 + bysin 2 4. Comparing both sides, we have by = By¢/4, by = 0, by Hence from (vi), the desired bie is ye 3 7 Yo gin BME. Srct SSeS . point x ofthe string at any time The boundary conditions are y(0, f) Also the initial conditions are y(x, 0 2) « “ (@).. ‘The solution of (i) is of the form Wx, £) = (c, Cos px + ¢, sin px) (c, cos cpt +¢, sin ept) By (ii), 40, £) = ¢\(c, cos ept +c, sin ept) = 0 For this to be true for all time, ¢, = 0. _y(a, t) = ¢y sin px (cy ¢0s ept + ¢, sin cpt) Ais by (i (l, 1) = ¢, sin plc, cos cpt + ¢, sin cpt) = 0 for all t. This gives pl=nn or p=nrl,n banat nbs ‘Thus Gx, 0) =¢,sin By cos ae ey sin 22) mC) 8 fo, win ME) BE (_ rot a Tyrese : 7) , nme) ne by (iv) (3) (ae) Fs ° Thus (v) becomes _y(x, £) = ec sin ™ eos 7 Adding all such solutions, the general solution of (i) is, net yx, D= x 8, sin a cos (vi) From (ii), lle 29) = 960, 0)= 9) 8, in mo 1 where 2 J H(le— x?) sin te, by Fourier half-range sine series378 [aan Sor zh a-ay Samat fie 2 mae ay 4 ay 4 og aye =m 2 sin Ae a-cant Hence from (vi), the desired solution is 2 ye = 12C¥ 5 =e A BIR IDK ogg B= Det 7? Sane 7 mt Solution. The equation of the vibrating string is ‘The boundary conditions are (0, t)=0, y(l, t) = 0 Also the initial conditions are y(x, 0 at 7 Since the vibration of the string is periodic, therefore, the solution of (i) is of the form YG, 0) = (€, C08 px +c, sin px) (cy cos ept +c, sin ept) By (ii), (0, t) = e\(c, cos cpt + ¢, sin ept) = 0 For this to be true for all time c, = 0. “ v(x, £) = ¢, sin px (c, cos ept + ¢, sin ept) and (2) =upsin® = 0 Also yl, 0 = ¢, sin pl (c, cos cpt + ¢, sin ept) = 0 for all t. This gives pl=nn or p= “Fm being an integer. ‘Thus sete, sin (0 os He + sin.) By (iii), O= cy sin for all x £e.,¢,¢,=0 Adding all such solutions, the general solution of (i) is mmx. enmt ye, = ~E 6, sin" sin TE 0) or ‘(asin 3 —sin 3) SAE sin EE sin 30 = 3 sin 0-4 sin? 6] 4 7 7 7 T © y gin ™ 4 2% gin ORE, Se, Sr = My sin 4 2 sin 4 BG, inEquating coefficients from both sides, we get Solution. The equation of the vibrating string is a = ‘The boundary conditions are (0, £) = 0, y(l, )= 0 Also the initial conditions are (x, 0)=0 and (3), = Axx) ih ‘As in example 18.6, the general solution of (é) satisfying the conditions (éi) and (iii) is ye, = x b,sin sin a “2 eT) oofv) By(iv), — dal-x) = (2 Je" rm in Rat L 2 Pcs 7 f son = 2h ae 32)(—- + cog 2™)_ Poi nme ane Flee 7 ¢ 20 (- ain) 0 £50) -S8a- ‘cos nx) = ae n-cy) ane Bae 1 (-1)"|=— "taking n = 2m -1. cal On an 11 emi ee Hence, from (v), the desired solution is But © 1 in (2 =U me, (Bm — Vnet 2 cat A Om —1F 7 i Solution. Let B and C be the points of the trisection of the string OA(= 1) (Fig. 18.2). Initially the string iis held in the form OB'C’A, where BB’ = CC’ = a(say).‘The displacement y(z, t) of any point of the string is given by es wo ae and the boundary conditions are (0, £) ube . ‘The remaining condition is that at ¢ = 0, the string rests in the form of the broken line OB'C’A. The equation of OB’ is y = (Ball) x ; the equation of B’C’ is sv) As in example 18.6, the solution of (i) satisfying the boundary conditions (ii), (iii) and (iv), is nM og tice TT Adding all such solutions, the most general solution of (i) is yx, t) =, sin nme Putting ¢ = 0, we have y(x, 0) = &. 6, sin n nt (Where b, = C,C;] salvi) (vii) In order that the condition (v) may be satisfied, (v) and (vii) must be same. This requires the expansion of yz, 0) into a Fourier half-range sine series in the interval (0, !). by (1) of § 10.7, 2[ p® 3ax nn 23 3a. < | mm Ba «nT af SP gin det [Ped —2e)sin "ae + fe Dsin ae 6a] ,{_costnmd| _ [_ sin (nee [* (ned) (nn? | |, (ore) in (nell) f° cos (mn sin (1 : a-20 (nnil) Je { (awl Hf. y (nna) in (nme) +] @—Dj- Sw |- ae _ { 0) (rll? J lay 6a] (_P ome, Pane), Pog tne 2? ane Pam OBL Bun 8 ne 3 ann 8B nn 8 oP +a sin a nn (0 3 \Bnx cos 2 5 in Bm 8 re 8sin 28% sin {im = 3 3 Hence (vi) gives yt, t) = cos (Take n= 2m) cos 27met (vii) Putting x = 1/2 in (vii), we find that the displacement of the mid-point of the string, i.e. y(l/2, t) = 0, because for all integral values of m. ‘This shows that the mid-point of the string is always at rest. (3) D’Alembert’s solution of the wave equation ay _ ay a-ak (1) Let us introduce the new independent variables u = + ct,v=x~ct so that,y becomes a function of u and v. By _ ay ay Then ed + e aay ) *y 9 ey ay and SR) aS) eae ae Similarly, ‘Substituting in (1), we get (2) Integrating (2) w.r-t. v, we get ® - pw) (8) where flu) is an arbitrary function ae Now integrating (3) w.r.t. u, we obtain y= J fwdu+ yo where y(v) is an arbitrary function of v. Since the integral is a function of u alone, we may denote it by 6(u). Thus y= Ou) + yoo) ie. yl, 2) = OG + et) + weet) wl) This is the general solution of the wave equation (1). Now to determine @ and y, suppose initially u(x, 0) = flx) and dy(x, OWAt = 0. Differentiating (4) w.r-t. t, we get 2 = c6'(x + ct) —ey'tx— et) Att=0, ¥@) = ye) AB) and (x, 0) = O(x) + whr) = fle) (6) (5) gives, 6x) = lx) + (6) becomes 2y(x) + k = fx) or yO) = $e and 6(x) = $e + AGo Hichen Enoneenina MatHeManics Hence the solution of (4) takes the form ey) = ifs et) + + Fleet) B= fla + et) + lr~ et) AD) which is the d’Alembert’s solution* of the wave equation (1) (V.2.U,, 2011 8) ‘Obs, The above solution gives a very useful method of solving partial differential equations by change of variables, Example 18,9. Find the deflection of a vibrating string of unit length having fixed ends with initial velocity zero and initial deflection fx) = b(sin x~ sin 2x). (WU, 2011) Solution. By d’Alembert's method, the solution is ya, = Fin + fle—cb)) = Glblsin (x + t)—sin 26¢ + et) + Aisin bx et) ~sin 2x ~c0 = h{sin x cos ct — sin 2x cos 2ct] Also yG¢, 0) = R(sin x ~ sin 2x) = fix) and y(x, Oot = k (—c sin x sin ct + 2c sin 2x sin 2ct),_=0 ie., the given boundary conditions are satisfied. 1, Solve completely the equation 3/0? = c92y/éx*, representing the vibrations of a string of 1, fixedat both ends, given that y(0, £)=0 jy, 0) = 0; ylx, 0) =/lx) and dylx, OVAL =0,0 0, is given by (4n—3P (an? 6. Ifa string of length is initially at rest in equilibrium position and each of its points is given a velocity o sueh that vex for) “= *See footnote of p. 373.| ieiiTiONS OF Pais Drrenenti. Eauarions 611 (1) ONE-DIMENSIONAL HEAT FLOW Consider a homogeneous bar of uniform cross-section o.(cm*). Suppose that the sides are covered with a material impervious to heat so that the stream lines of heat-flow are all parallel and perpendicular to the area a. Take one end of the bar as the origin and the direction of flow as the positive x-axis (Fig. 18.3). Let p be the density (gr/em®), s the specific heat (cal /gr. deg.) and k the thermal conductivity (cal/em. deg. sec.). Let u(x, ~ t) be the temperature at a distance x from O. If 8u be the temperature change in a slab of thickness Gx of the bar, then by § 12.7 (ii) p. 466, the quantity of heat in this slab = spa éx6u. Hence the rate of increase of heat in this slab, ie., spate =R,—Ry, where R, and R, are respectively the rate (cal./sec.) of inflow and outflow of heat. wot ho) the negative sign appearing as a result of (i) on p. 466. (du/ax.) — (du/dke) Hence spose & - #o() 4 wo( 2) ie, oH [erg Oe le bees Now by (A) of p. 466, R, = ~ta(%4) ax ae ot sp or Writing k/sp = c, called the diffusivity of the substance (em*/sec.), and taking the limit as &x > 0, we get du _ 2a S=FleS AL ra e we (1) This is the one-dimensional heat-flow equation. (V.T.U., 2011) (2) Solution of the heat equation. Assume that a solution of (1) is of the form u(x, t) = Xx). Tie) where X is a function of x alone and T'is a function of t only. Substituting this in (1), we get XT’ = OX"T, ie, XK = Mer (2) Clearly the left side of (2) is a function of x only and the right side is a function of ¢ alone. Since x and f are independent variables, (2) can hold good if each side is equal to a constant k (say). Then (2) leads to the ordinary differential equations 2 oe ix =0 ..3) and ocr 0 4) Solving (3) and (4), we get (i) When k is positive and = p?, say : Xaceh + ce, T= ce"; (ii) When kis negative and =~p?, say : X =e, cos px +c, sin px, T= og e°?* ; (iii) When kis zero : X= cpt +O T= Cy. ‘Thus the various possible solutions of the heat-equation (1) are UW = (oye? + coe) cgeP* 5) w= (c, cos px + ¢, sin px) ge" 46) = (6, + Cg)ey AD) Of these three solutions, we have to choose that solution which is consistent with the physical nature of the problem. As we are dealing with problems on heat conduction, it must be a transient solution, ie., uv is to decrease with the increase of time ¢. Accordingly, the solution given by (6), i.c., of the form w= (C, cos px + C, sin pr)e*?* is the only suitable solution of the heat equation. (8)“OF is u(x, t) = (c; €08 px +c, sin px) e-F** ‘When x = 0, u0,t)=c,e%" =0 ie, 0, =0. (ii) becomes u(x, £) = ey sin pret ealiié) When x = 1, u(1, f)=e, sinp. e- =Oorsinp=0 ie, p=nn. (iii) reducesto u(x, t)=b, e°"""' sin nx where b, ‘Thus the general solution of i) is w(x, = J) be°" sin nee ealiv) When ¢=0, 3 sin nnx=u(0,t)= > bye" sin nme at Comparing both sides, 6, = 3 Hence from (iv), the desired solution is Solution. Substituting u=XGYNO in the given equation, we get XT =X'T i.e, x = 7 2-42 (say) oT @WX ao aT as, TT yet =0 ad Geet X20 and + a! a) ‘Their solutions are X =e, cos kx + cy sin kx, T= eye 2) Af k? is changed to — k°, the solutions are X= 00 ogc, T= eget (8) If#®=0, the solutions are X=c,x +e, In (3), T > & for ¢ — © therefore, u also — solutions (3) while (2) and (4), satisfy this condition. Applying the condition G10 (8), we get =0. iS XT = Cy6¢=%, (say) AB) From (2), 7 2 ey sin he +c, c08 kx) heer Pt Applying the condition (ii), we get c, = 0 and —c, sin kl +c, cos Al =0 ie, ¢)=0 and Rl =nn (nan integer) wl) 9 the given condition (i) is not satisfied. So we reject the nr u=eqcoshe eye a,c a 6)“RveariONs.GF Parra. Direnewna Eouanons Ea Thus the general solution being the sum of (5) and (6), is u = a, + Ea, cos (nll) "ee LAD) Now using the condition (ii), we get lx — 22 = ag + Za,,c0s (neil) ‘This being the expansion of lx —x? as a half-range cosine series in (0, !), we get ie ar : is = 210-5 (cos nm +1) +0} =~, when n is even, otherwise 0. rr nt? Hence taking n = 2m, the required solution is Peay at A Solution. (a) Let the equation for the conduction of heat be au _ 2 Fu , eS ef Fed oD Prior to the temperature change at the end B, when t = 0, the heat flow was independent of time (steady state condition). When u depends only on x, (i) reduces to d°u/dx® = 0. Its general solution is wu = ax +d ii) Since u = 0 for x= 0 and w= 100 for x = J, therefore, (ii) gives b = 0 and a= 10/1. ‘Thus the initial condition is expressed by w(x, 0) = es Also the boundary conditions for the subsequent flow are u(0, t) = 0 for all values of ¢ iv) and u(l, t) = 0 for all values oft Av) ‘Thus we have to find a temperature function u(x, t) satisfying the differential equation (i) subject to the initial condition (iii) and the boundary conditions (iv) and (v). Now the solution of (i) is of the form u(x, £) = (C, cos px + C, sin px)e“?* (vi) u(0, t)=C,e°°** = 0, for all values of t. = 0 and (vi) reduces to u(x, #) = C, sin px. oP" vii) Applying (0), (vii) gives u(l, t) = C, sin pl. e~“? = 0, for all values of t. This requires sin pl = Oie., pl =nn as C, #0. = nvil, where n is any integer. b, sin ME oH Where b, = Cy. Hence (vii) reduces to u(x, #) T2 {These are the solutions of (i) satisfying the boundary conditions (iv) and (v). These are the eigen functions corresponding to the eigen values 2, = cnz/l,of the problem] Adding all such solutions, the most general solution of (i) satisfying the boundary conditions (iv) and (v) is > «Amex ze sin enteue ut, t , uce,0)= 5 b,sin Putting t = In order that the condition (iii) may be satisfied, (iii) and (ix) must be same. This requires the expansion of 100x/I as a half-range Fourier sine series in (0, !). Thus 100x 1 S = MEX 2 7) 100x . nm 2 in where b, = > if ap sin td _ 200 cos (nxx/ 1) sin(anx/)|] _ 200(_ 200 + = 20 of sere gy {sine T 200 sy) - 20 ay ar Hence (viii) gives u(x, t) = 200 5 ED" in BEE, grteonstfe Pa (©) Here the initial condition remains the same as (i) above, and the boundary conditions are u(0, t) = 20 for all values of t lx) u(l, t) = 80 for all values of t xi) In part (a), the boundary values (i.., the temperature at the ends) being zero, we were able to find the desired solution easily, Now the boundary values being non-zero, we have to modify the procedure, We split up the temperature function u(x, ¢) into two parts as u(x, t) =u, (x) + u, (x,t) (xii) where u, (x) is a solution of (i) involving x only and satisfying the boundary conditions (x) and (xi) ; u, (x, t) is then a function defined by (xii). Thus u, (x) is a steady state solution of the form (ii) and u, (x, ) may be regarded as a transient part of the solution which decreases with increase of ¢. Since u, (0) = 20 and u,(l) = 80, therefore, using (ii) we get u,(0) = 20 + (60/)x saci) Putting x = 0 in (xii), we have by (x), u{0, €) = w(0, #)— u,(0) = 20-20 = liv) Putting x = / in (xii), we have by (xi), ul, t) = ull, )—u) = 80-80 = 0 (xv) Also tue, 0) = ul, 0) ~ w,(x) = 2D — (02+ 2) {by (ii) and (ii) = 40% a9 sacvi) 7 Hence (xiv) and (xv) give the boundary conditions and (xvi) gives the initial condition relative to the transient solution. Since the boundary values given by (xiv) and (xv) are both zero, therefore, as in part (a), we have uj(x, t) = (C, cos px + C, sin px) °°? By (xiv), u(0, 0) = Ce" = 0, for all values of t. Hence C, = 0 and u(x, t) = Cy sin px. Pt ealavii) Applying (xv), it gives u,(l, t) = C, sin ple” FH 6 for all values of t This requires sin pl = 0, ic. pl = nm as C, #0. p = nmull, when n is any integer. Hence (xvii) reduces to u(x, ¢ nme ent! Where b, = Cy.| Rerucien oF Paemat Derenexrin. Eauarons | 615] Adding all such solutions, the most general solution of (xvii) satisfying the boundary conditions (xiv) and (av) is (a, t)= ¥ sin Be ore viii) i Putting ¢=0,wehave ujx,0)= 5b, sin" xix) a In order that the condition (xvi) may be satisfied, (xvi) and (xix) must be same. This requires the expan- sion of (40/2) x ~ 20 as a haif-range Fourier sine series in (0, 2). Thus =20= ¥ 6, 5in 2 where ,, 2! (40% _ 90) gin BR dy = - 40 is J, (295 - 20} sin 2 ae =- 22 (1 + c08 ne) mt i.e.,b, = 0, when n is odd ; = 80/nm, when n is even Hence (xviii) becomes u(x, t) = = () sin IEE on rt [ake n = 2m] sin Eg ttn es Finally combining (xifi) and (x2), the required solution is uty, = 49 + 29-49 x Jsin 20 gto In steady state condition, uv is independent of time and depends on x only, (i) reduces to Ausax* = 0. Gi) Its solution is u=a + bx Since u = 30 for and u = 80 for x = 20, therefore a = 30, b = (80 ~ 30)/20 = 5/2 ‘Thus the initial conditions are expressed by 4 (,0)=304 Sx ‘The boundary conditions are u (0, £) = 40, u (20, ¢) = 60 Using (i), the steady state temperature is u (x, 0) =40 + ov) To find the temperature u in the intermediate period, ue, )=u,(@)+u,(x, 0) where u, (x) is the steady state temperature distribution of the form (iv) and u, (x, #) is the transient tempera- ture distribution which decreases to zero as ¢ increases. Since u, (x, £) satisfies one dimensional heat equation u(x, = 40424 5 (a, cos px +b, sin px)? wv) W0,)=46=40+ Ya,e* whence a, my