Tensor Analysis
Tensor Analysis
(AM + A®) = B® which is symmetric 1 and C= 2 (AM AM) = (Am AM) =~ CP which is skew-symmetric, Similarly, it ean be shown for Apy Hence the given statement is true for any tensor. Example 9. /f A," and. B,P are tensors, prove that their sum and. ‘difference are also tensors. (Delhi ME 1976) Sol. Since A,?? and B,P4 are tensors. Zn de! aR! ax Agee wl 0 oe ot age wy Fin de! aR! axe "7 = Ha ae 2) = BO 8 aa at 2 Now oR! ae axe er Arts BP) Now (A, +B,P?) and (A,!"—B,P7) both are tensors ofthe same rank and type as A,P# and B,P7, Example 10, Prove that the contraction ofthe outer product of the tensors A? and By isin variant. (Delhi ME 1975) Sol. We know that Now (1) 1 AP B, = A” B, which is in variant. Example 11, Let A,ft bea tensor: choose p = tand q = s and show that Af4 isa tensor. What ists rank. (Roorkee ME 1971) Sol. ven A,B? isa tensor ai oe! a! ax? ac ax im = 5 oa SE ap Ain oP aah BF al am A Al) Now we have to show that A,£f is a tensor eee Let i= m,j=1in(1),then Ayf= S52pensoR ANALYSIS 1383 oe Agi . ‘Aff isa tensor of rank one and can be denoted as B, simp! Here by contracting twice the rank was reduced by 4 simply. He 12. i . Example 12. Show that any inner product of tensors A,? and B,* isa tensor of rank three. Sol. Since A,? and B,* are tensors (Delhi ME 1976) Zico! av Aj = P Wa agi wl) ae! oe" axl at ad 2) the outer multiplication is Ay Byit = OE OF OH AEM BY 4 n gor ax? dad ax? axe ax Now contracting by letting i = 1 summing, we have zi ax’ ag! aR” A! ax" AE! AE" A 4 np os ‘aa? Ox) at ov aI” (ate me AU, oF axt at 7" i => A,? B,® isa tensor of rank three. 21.17. REIMANNIAN SPACE AND METRIC TENSOR In rectangular cartesian co-ordinates the differential of are length ds between (x, y, 2) and (x + de, ay y + dy, x + dz) is obtained from ds? = de? + dy? + dz? which becomes > > sya dl by transforming to mi called three dimensional Euclidian space. general curvilinear co-ordinates. Such space is 2"), the differential of mal space with co-ordinates (x! N, we get Faw (GPE giadsl de? + gis de! de + + gu det del + Generalising this concept to N-dimensio ‘arc length between x’ and. a+ de = gy (de + 6") + « NON yet l= ye ds sing suman convention i i (1) ations of) As tis extension is due to Refmannian, the quadratic (1) ee he are length given by (1) is called Relmannian space, ___ The co-efficients gy 3° - is called Reimannian metric and any space in sP2Sy In particular if (1) can be written a, dst = (det + (Pte ay, then tis called a Euclidian metric and such SP2°8 © yw the di Frthe N-dimensional sp2e Now the di a ar displacements de dl are the ‘components of a contravariant vector or a tensor called Buclidian space of N-dimensions. wig scalar invariant £e., a tensor of order zero. In the equation of order one. de de is a tensor of order1384 ‘A TEXT BOOK OF ENGINEERING MATHEMATICS = _gy ust be components of a covariant tensor of order two socalled metric tensor or fundamental tensor, which is always symmetric. 27.18. CONJUGATE TENSOR OR RECIPROCAL TENSOR Let g =| gy | denote the determinant with the elements gi, where g 0. Let Gy be the co-factors of 8. Define g/ so that y _Se-factor of gy G, gi = Sector of gy Gy a) & Is] As gy and Gare symmetric in the subscripts, is also symmetric and Gi Now, B18 = gy tah and gya=e, tt kei y= 5" ie, sya itk O ifkni Similarly, sye! = 3! Again let a be an arbitrary contravariant tensor 4; = gy is an arbitrary covariant tensor a aj= gl gyal =8;!al giaj=al Since ajis an arbitrary vector, g! s a contravariant tensor of rank two (By qu nt law). ‘Therefore g' isa symmetric contravariant tensor of rank two so called ‘conjugate tensor or reciprocal tensor or second fundamental tensor. 27.19. ASSOCIATED TENSORS ‘The tensors that can be obtained from a given tensor by forming inner products with the metric tensor and its conjugate are called associated tensors of the given tensor. For example. The inner product of the tensor @ with the conjugate tensor, ‘8 iS another vector b; which is called the associated tensor of a! and can be written as a By = bs wl) Similarly in at. Noo “2 Biis called associated tensor of a4, ‘Thusiinagiven ensorby loweringor raising the indices we can obtain the associated tensorsby forming inner product with either of the fundamental tensors gy or gas in (1) or (2). Similarly, the tensors a//and aj can be obtained from the tensor a! by lowering the indices so called associated tensors of ay. 27.20, LENGTH AND ANGLE BETWEEN TWO VECTORS ~The vector can be written asa = a' g;or a= a; g! and the associated vectors are as If L in the length g vector a, then L By ai ora’ = gay, a)? =(algy. alg)! = (gdh! = (aa')i? a) a) 88)wurst aot 1385 Again vas V i nA (oe ag) gg); ‘d= 2. Dey goa; od which isinvariss ie, the length ofa vectors ivan. 18s the angle between two vectors a andb Letasag, and b= bg, a.b=|a|]b|o0s8 one St Nt U Vel Vegi? ab cas = Tae interns of associated vectors. Example 1 Determine the mic ens cnagte meric ser in () clinical (i) spherical coordinates (i) ds? = 3(de') + (de) + arte + Bx'e + 82x, Sol () Cylindrical co-ordinates, In this coordinate system he ferential ofa lengths given by j edhe da +d? wl)! Ifa! =r, = 0,2 =2, then ‘Comparing (1) with ds*= gy dt di! = 1gn=Peo7hgn= 670) 8" g.g=8) vege Bu Si Bs 100 andthe meri tensor in mati formis. |g $2 &|=|0 7 0 a1 ool Now 4 Coteorotan 1)? ©) is p-sientlt Beale 8 Cofactor of, Cofwtorofgn I]t Ofer and oe g Pio 6 100 iz 1 Simi, = f= 8 cnet mati firm ofernvsttesor gis) 0 50 oo1 (ii) Spherical co-ordinates iy sgn? sit and gy Ofori Hered? a4 Par +P od S08 +o 9 He sevsors °)9 7 aa ec, theme ewsoris 61°] to1386 A TEXT BOOK OF ENGINEERING MATHEMATICS ae oO Cofactor of gy The conjugate metric tensor gf = PSE; g «| gy |x is given by 0 eo 0 1 oo a P sin? @ since andg/=0 ¥inj, Pasi? (i) ds? =30P 430 442 48S eecd ‘Comparing with gi, de, dul, we get gy, sin? @ 3, 822 = 0, 33 = 3, 812 = 2 = B31 823 = 4 = Bam Bi: 324 © 8y=|2 0 4 | which is metric tensor in matrix form a ayes 8=|gi|=3(0- 16) -2(6- 16) + 4(8-0) = 40. Now the conjugate metric tensor gl = eo where G(i, j) is the cofactor of gy. vonin Example 2. Write the relation between the associated tensors B# and Byais. Sol. BIH = gi? git gg Beary and Byars = Sip Sin Sir Bly BM. Example3. Prove that the angles 8;2,033 and between the co-ordinate curves ina three dimensional coordinate system are given by §n 823 =82e, cos0y = : = 83 605 82 = ea 23 = Te gay? CSO = lesa’ (G.N.T.U. MLE. 1984) Sol. Let the co-ordinates be (x', 27, 2°). ‘Along x! co-ordinate curve both x” and x? are constants = po dh = ds? = gu(dr'y & ES is coordinate curve a unit tangent vector is a 7 ‘And along this coor ent vector isai'= 78 Similary, unit tangent vectors i £ i along x? and x* co-ordinate curves are az! = Veg and a=or ANALYSIS. 1387 ‘Therefore, the angle between a! and ays given by Cos B12 = gyay' a = gy Similarly, cos O35, 71. CHRISTOFFEL SYMBOLS k ‘The symbols [ij, k] and {i i| are called the Christoffel symbols of the first and second kind and they aredefined as ag B6u Bi warn} (+ raed wl) and {ifs (2) where gare the components of the metric tensor. Note. 1. The other notations for the Christoffel symbol of first kind are {ij k} or Ty, 2. There is no summation in (1), but summation is thereon s in 2). WHAL=bi Hanif }. {i}: 2722. TRANSFORMATION LAWS OF CHRISTOFFEL’S SYMBOLS The aim of these transformation laws is to show that both the Christoffel symbols are not tensors. Ge , 38a _ 984 | ax” ax! art 3, Both the symbols are symmetric with i,j i Now in. co-ordinate system [ij, 4] = In¥ co-ordinate system [ijk] = Fi { 2722.1. Transformation Law of Christoffel Symbol of First Kind Consider an wn(2) : 3) By cyclic cate of fj, Kand py Gor WE wees ~@) and (5)1388 ‘A TEXT BOOK OF ENGINEERING MATHEMATICS Substituting (3), (4) and (5) in equation (1), we get av? ax! ax’ ax? _axt we ae oF PO) og ae oe (6) Which isthe required transformation law. Due to the presence of the second order derivative in the right hand side of (6), the Christoffel symbol of the first kind is not a tensor. a= Since gis a tensor & £ gf (0) Inner multiplying (6) by (7), we get Pi ee orn eee, ten fi BS Se evan ZZ ore om $E8(5)-268 © Since 8,” #bar=stboan={ 5} and 8,78 Bq = 8 By = 8 Equation (8) is the law of transformation of Christoffel symbol of the second kind. Due to the presence of a second order partial derivative on R.H.S. of (8), the Christoffel symbol of the second kind is also not tensor. Example 1. Prove that © ral + tarp) = “8 and f-e (ah. Sot. (0 ton ann 3 { Sea | and trotn3{ eo te ‘Adding both the equations, we get [pr, q]+ [ar, p] = Pe : ) o%)-Fo=0 = gh 2 28 6, 20 or 5 2 Matiplyingby 2, gy SE =— ot gt Et ie, on a A UA stinay or cs at { J 22 tng, Example 2. Prove th Lihat Ze. Sol. We know that 8 = 8 Gh.D oe (sum over !only) “+ GUD dogs not contain gyexplicily FE — G4, 4ANALYSIS rae 1389 summing over jandk 28. 5G. 28k Bey det = 00 BS = gg 2 “ee inalstinn-e[ {Alf} 2 {hy a eEhogiaes, {im} eae ag OVE Hence the result. Example 3. For the metric AS = (de!) + (2627 + 3(e!)9) (a2, sna 1 } find} 95 \. Sol. Here, g11 = 1, 822 = 2) + 3(x')* and all others are zero, 1 1 san {22} 8" (622) = 510,22] = [1,22] =1(_ 282) _ =2(-3?]- Example 4, Find the values aff 2} : {2 } { 2 } {i} in (i) rectangular (ii) cylindrical (iii) spherical co-ordinate systems. Sol. For orthogonal co-ordinates gay =Oifpeq the system is orthogonal) a 2 pat AP +30 =-3 ee =-$ YP ; oleae (In rectangular co-ordinates gy=1 {oo =o (ii) In cylindrical co-ordinates. x! = 7,7 =, =z; ds?=(dr'P + Pd? + (ae = B= 1, g2=77,833= Land gg=0 Ypeg {a (iid) in spherical coordinate system x! =.= 0,7 = ds? = (de! + P(d2)? + 7 sin? 0 (de? gy =r sin’ @ ==sin 8 cos 6 ay Hh Sainte a0 Sin" 0) = cot 0,1390 ‘A ‘TEXT BOOK OF ENGINEERING MATHEMATICS 27.23. GEODESICS In Riemannian space, the distance s between two points f; and f on a curve x = x(t) is given by and the curve in the space which makes the distance a minimum is called a geodesic of the space. For example the geodesics on a plane are straight lines, 27.24, COVARIANT DERIVATIVES ‘We know that for an invariant 4, x is the partial derivative w.r. 2, which is a covariant vector, In general system of co-ordinates the partial derivatives of a vector or a tensor are not tensors. [Ref. Ex. 7]. Now Wwe study certain expressions involving the first partial derivatives of covariant and contravariant vectors which are also tensors. Such expressions are called covariant derivatives. ‘The covariant differentiation of a covariant vector LetA;andA, be the components of covariant tensor of order one in the.x’ and co-ordinate respectively. We have A=, BY? Ay axe ae av ast eae aA, Due to the 2nd term on RS. of (1) “Fis not a tensor. Differentiating if w.r.t, 3, we get 4 ine oY. Fi Replacing 5-5 using the expression In equation (1), we get or mq, 28% [Me (11, {ic} ae | ax’ | pr Here 2 i Ja isa covariant tensor of order two called the covariant derivative of A, wrt. and av Pr is denoted as Ap, 21.28, THE COVARIANT DIFFERENTIATION OF A CONTRAVARIANT VECTOR “The law of transformation between components of vontravariant vectors A! and Alin x! and ¥ system a =" a?(LYSIS ig0R AN 1H 1391 Differentiating it war.t.3, we get aa! BR aA? ax! | OR ad a ad art * aad ae 0) Now Substituting in (1), we get aAl _ axl ax! oar (nt3 or Here S54) to rand is denoted as AP, (. aa? eo zie {i yeu is.a mixed tensor of second rank, called covariant derivative of A? with respect 2726, COVARIANT DERIVATIVES OF HIGHER RANK TENSORS k at fi ae {i} amt 4 Lin] sions for covariant derivatives of tensors Ay, A% Now from the above article A, Extending these definitions we can write the exPre Aj‘etc. They are In general, int derivative of RE.C.W. 1984) Gi) gue Example 1. Write the covarial Ad Blt wr1392 ‘A TEXT BOOK OF ENGINEERING MATHEMATICS {in} {Lado eB, { 7 fava + { = rent ioe | i 28 38; - w ene {hfe [ Bal sem ia, ®} ~[ka, j] = 0 [By Example 15] 27.27. PERMUTATION TENSORS. The contravariant tensor Ej, = —= ej and the covariant tensor Et = Vg jg, are called permutation a tensors in three dimensional space. Note. (0) ej4= Of any two or all the three indices are equal. G e433 = e931 = e312 = + 15 €2 (id) In the similar way e!*is defined. 27.28. TENSOR FORM OF GRADIENT, DIVERGENCE, CURL AND LAPLACIAN 27.28.1. Tensor Form of Gradient. If isa scalar function (or in: int) then gradient of 6 is defined by a Vo= 2 @= ou Here 6, is the covariant derivative of @ w.r.t. x*. 27.28.2. Tensor Form of Divergence. The divergence of the contravariant vector A? is defined by wv arn 2 atl) which i ar . div ar = A al {nf rwhich is denoted as Ay? ie, itthe contraction A. Plat inve But {ro} ae Invg aa ga [fo aaP 1 avg iv ar = 2 a! Sinvg |=, (1 ave) 4 div A! wtf ae & a Ve ad A’ 7 ¥ 3 (Vg A!) which is the other form, Note. The divergence of covariant vectorAy i defined by divAy = MA, , 27.283. Tensor Form of Curl. Let Ap be a covariant vector. Then the curt of ‘Ay in tensor form is aA, Ady defined as Ap, q~Aa p= 39 aye L Proof. Ana™ 3a {oa 4 ~()-exs0R ANALYSIS 1393 - @-Oae A422 Gxt ~ aye tis tensor of rank two. Note, The cul is also defined ase" 4, aA, and Arse, Ja 27.28.4. Tensor Form of Laplacian, The Laplacian of 6 is the divergence of grad ¢ “PhaaiS)an(et3) 12 aed). 2b = hey Vie= ax) "Ve at Example 1. Express V? W in (a) cylindrical and (b) spherical coordinates. (Roorkee M.E. 1971) Sol. (@) In cylindrical co-ordinates g!! =1, g2=_, g ve oy) 1/9 ( dy), 2 Pere eS) a(S) 212/(, 0), 1 Fe, ey : (Shae which is the Laplacian of wp ie. Vp in cylindrical co-ordinates. 1 1 2p Fao (In spherical co-ordinates g!' = 1, g” Which isthe required equation. 2129. THE INTRINSIC DERIVATIVE 5 ‘The intrinsic derivative of A; pede oo derivative or A, and = and is denoted by = - 2x (pis defined asthe inner product of the covariant the vectors move parallel along the curve. The tty. we can define the intrinsic derivatives for IF the intrinsic derivatives ini oh Mtrinsic, derivative is also »alled abst “Bher rank tensors.2 ‘A TEXT BOOK OF ENGINEERING MATHEMATICS 27.30, RELATIVE TENSORS gi RIE Bl, IF the components of tensor AP transform Hy RR " then APY, is called a relative tensor of weight m. Here is the Jacobian of the transformation. Note. If m= 0, then the tensor APE, is called absolute tensor. If m= 1, then the A, is called tensor density Example 1, Find the intrinsic derivative the tensor of By) assuming that i, differentiable w.r.t.t aa Sol. rear Bl (1) ys fi] gs] abe awh (4 2 {fob de de TEST YOUR KNOWLEDGE 1. Write the following using summation convention Wal bP pt 4 a" bP es 4a"? bP ol 4 a2 bf of Gy +8? Gx +87 Gy +8 Ga. 2. Expand the following: OAGN=2 (ZR ANN =A 3. Evaluate (BP LAT (ip 6 8,48, 4. Write the law of transformation forthe tensors 7 OA Ble 5. A covariant tensor has components 1), 2y ~2, xz in rectangular co-ordinates, Find its covariant components in spherical co-ordinates. (Kakatiya 1988) 6 A quantity AC j h.m) which is a function of the co-ordinates 2 transforms to another co-ordinate system according to the law = de” de! a ax! ax’ (7,550) = AG km), a Dad ad del ax" age Ohh bm Is this quantity a tensor ? Ifso express it suitably and stte its rank (Panjab 1982) ax’ dx! 5s rove that 2 A 2a, cae ae ae? 8, ie components of wo tensors in on coordinate system, show hal hey are eqal inal erie systems 9, Prove hat there sno dstnetion between convarint and covariant vets ifthe vansfomation aw i ofthe la +b), where sand b's te constants such that! ca 7 (Allahabad M.E. 1970) formTENSOR ANALYSIS 10. a 2. 13. 14. 15. 16. 1. 18, 19 20. a 24, 25, 26, 27. 28. 29. 30. 2 iene ‘Show that Ay isa tensor if its inne; a Prove that 2* is a symmetric cont * Product with an arbittary mixed tensor By! is a tensor. For ds? = (dx! + 3¢ae2)2 ravariant tensor of rank two, +402 ote (tedras Me. 1972) +427 prove that gi ate vectors, show that —San” 8? Var4, 0°, Prove that for an orthogonal system gj=0 IfAP and BT " 30 Ming and gy=t ving Fi Find g and g'/ corresponding to the metric ds? ae + P(d0? + sin? @ deg), (Aligarh M.E. 1971) Prove that : © ar = lap, 7) wm) hale 9 wart = surf 5} telclal 2, Provethat 22" [fr] av" _ a ax | m Bel x! [It Get” axl ax" Derive Christoffel symbots of first and second kinds for spaces where g, = Oif p= q- Derive the taw of transformation of Christoffel symbol of (i) First kind (Ji) Second kind Find AM hn. g i a(* D8 jg Prove that (8) An!" )q = Sk A iva Express divergence of A” in terms of () Cylindrical (ji) Spherical coordinates If Ay, gis the curl of a covariant vector, prove that Ay, e+Ajs,: + Aki j= 04 (RECW 1983) Prove that the instrinsic derivatives of gf 8! are zero. Show that (D div curl A’ = 0 and G0 curt grad @ Prove that () Ap gr Ap. ra = Rar Am where Apis an arbitrary covariant ensor of rank one (i) Also prove that yy I 160808388 Ry = Bn pyr using tensor notation. Answers GS GyN=d al bP oN Al Al? 4 ART! + Ay WH eas19, 24. ~ A TEXT BOOK OF ENGINEERING MATHEMATIC, at of EH av ax? ay 1 0.cos ) (7 sin? 0 sin @ cos ) + (sin 6 cos q) (2rsin sin gr? cos? 0) + (cos 6) ( sin 8.05 8 cos 9) Mh i) Begg = (r-c0s 8 cos «) (7 sin? 0 sin pcos q) + (rcos 0 sin g(2r sin 8 sin p--1? cos” 8) + (~r sin 8) (7 sin 8 C08 8 cos 9) £ (sin @ sing) °? sin? @ sin g cos @) + (rsin cos q) (2rsin 0 sin g— 7? cos? 8) Yes; AYA}; rank = 5 due to contravariant of order 3 and covariant of order 2 Alt, j= Bin 15 ym [5 Ve gi Lat gf Laie Lat " ma fe {ag fee {J pat i te fg le oe 1 ain, 8 avant { is Cond a +t} 1_aAg (maw h2eayes a 3 39 (820g) +