Laplace Transform and Di↵erential Equations HarishKumar hkumar@iitd.ac.in Department of Mathematics, Indian Institute of Technology, Delhi Harish Kumar Laplace Transform and Di↵erential Equations September 20, 2020 1 / 55 Definition Consider a real valued function f (t) on [0, 1). We define Laplace transform of f (denoted by L[f (t)](s)) as, Z 1 L[f (t)](s) = e st f (t)dt, (1) 0 provided the integral makes sense. Definition (Exponential Order:) A function f is said to be of exponential order if there exist constants M and ↵ such that, |f (t)| Me ↵t . (2) For example: Any polynomial is of exponential order. Similarly, any 2 bounded function is also of exponential order. However f (t) = e t is not of exponential order. Harish Kumar Laplace Transform and Di↵erential Equations September 20, 2020 3 / 55 Definition (cont.) Also Consider: Definition Piecewise Continuous: A function f is said to be piecewise continuous on domain D if f is continuous on D except on a countable set S on which it has jump discontinuity. Combining both the definition, we have the following result: Theorem Suppose f is piecewise continuous on [0, 1) and satisfies (2). Then L[f ](s) exists for s > ↵. ( Ga ) - [ stlfctydt.cm/e-stedtdt--Mf?e-CEdtdt = I f*e Stf - # de Harish Kumar ) s it e- s > a. Lcf] is ) converges Laplace Transform and Di↵erential Equations . September 20, 2020 4 / 55 Important Remark Remark In the following, we will not explicitly specify that f is exponential order but we will always assume it, whenever we need to define its Laplace transform. - Harish Kumar Laplace Transform and Di↵erential Equations September 20, 2020 5 / 55 Example: Consider f (t) = e ↵t which is exponential order. Laplace transform of it is calculated as follows: Z 1 1 ↵t L[e ](s) = e st e ↵t dt = , s > ↵. s ↵ 0 S > L For ↵ = 0 we have L[1](s) = 1 . s = Harish Kumar Laplace Transform and Di↵erential Equations 570 September 20, 2020 6 / 55 Linearity Theorem The operator L is linear i.e., - L[af (t) + bg (t)](s) = aL[f ](s) + bL[g ](s) where → LED & LEFT is defined for all s and constants a and b. - - Proof. L[af (t) + bg (t)](s) = =a Z 1 0 Harish Kumar e st f (t)dt + b Z 1 Z 1 e st (af (t) + bg (t))dt 0 e st g (t)dt = aL[f ](s) + bL[g ](s). 0 Laplace Transform and Di↵erential Equations September 20, 2020 7 / 55 Linearity (cont.) Example Laplace transform of hyperbolic functions cosh(at) and sinh(at): cosh(at) = 1 at (e + e at ), 2 sinh(at) = 1 1 (L[e at ] + L[e at ]) = 2 2 ✓ 1 1 L[sinh(at)] = (L[e at ]oL[e at ]) = 2 2 ✓ L[cosh(at)] = Harish Kumar II. 1 s a s a Laplace Transform and Di↵erential Equations e at ) 1 s +a ◆ = 1 s +a ◆ = + w 1 1 at (e 2 , a 17-0 s s2 2 ☒ a s2 a2 . September 20, 2020 8 / 55 Linearity (cont.) Example Laplace transform of Trigonometric function sin(at) and cos(at): Define Lc = L[cos(at)] and Ls = L[sin(at)], then using integration by parts, Z - - - - - - Lc = L[cos(at)] = - = = e st s cos(at) 1 e st cos(at)dt Z I a 1 st 1 e sin(at)dt = s 0 s 1 - 0 0 soo a Ls s ⑤ = Similarly, Ls = a Lc s Solving these two linear equations for Lc and Ls we get, T Lc = Harish Kumar s s 2 + a2 , Ls = a s 2 + a2 Laplace Transform and Di↵erential Equations . September 20, 2020 9 / 55 Di↵erentiation of Laplace transform Theorem ( Skl Suppose f is of exponential order, then in its region of convergence F (s) = L[f (t)](s) is di↵erentiable infinitely many times at each point and - = F (n) (s) = ( 1)n L[t n f (t)](s). "" Proof. F 0 (s) = = d ds = Z 1 0 e st f (t)dt = Z 1 0 ; f (t) = d (e st )dt = ds = L[tf (t)](s). = Then the result follow from induction argument. Harish Kumar Laplace Transform and Di↵erential Equations September 20, 2020 10 / 55 Integration of Laplace transform Theorem If F = L[f ] then, - - L f (t) = t Z 1 F (s1 )ds1 . s Proof. Z 1 F (s1 )ds1 = s - Z 1 Z 1 s f (t)dt ds1 = 0 - = Z 1 0 Harish Kumar e s1 t . e st Z 1 f (t) 0 f (t) f (t) dt = L . t t - = Z 1 e s1 t ds1 dt s . = Laplace Transform and Di↵erential Equations September 20, 2020 11 / 55 Laplace Transform of Derivatives of a Function Theorem Suppose f is di↵erentiable for t # well defined then, 0 and of exponential order. If L[f 0 ] is = I L[f 0 (t)](s) = sL[f (t)](s) Lcf ] us ' Proof. By definition we have, = S - - ? fest - f (0). = f- lol F- LEFT = = L[f 0 (t)](s) = = Z 1 f. Gifts ] e st f 0 (t)dt 0 : o - Moo) x④¥I=s= Integrating by parts we get, ⇥ ⇤1 L[f 0 ](s) = e sr f (r ) 0 + s - - Harish Kumar €-0 Z 1 0 e st f (t)dt = sL[f ](s) ④ = f (0). = . Laplace Transform and Di↵erential Equations September 20, 2020 12 / 55 Laplace Transform of Derivatives of a Function (cont.) Corollary Now repeating this process n times we get, L[f (n) (t)](s) = s n L[f ] = s n 1 f (0) s n 2 f 0 (0) ··· f (n 1) (0). O↵ course we have to assume that L[f (n) ] exists. Harish Kumar Laplace Transform and Di↵erential Equations September 20, 2020 13 / 55 Laplace Transform of Integration of a Function Theorem L Z t f (⌧ )d⌧ = 0 L[f ] s Proof. Rt Define g (t) = 0 f (⌧ )d⌧ , then g (0) = 0 and g is di↵erentiable with g 0 = f . Furthermore if f satisfies (2), then - = |g (t)| - . . Z t # 0 |f (⌧ )|d⌧ M Z t 0 e ↵⌧ d⌧ = == # M ↵t (e ↵ 1) M ↵t e . ↵ So, g is also exponential order. So, L[g ] exists. Now, L[g 0 ] = sL[g ] g (0), gives the desired equality. Harish Kumar Laplace Transform and Di↵erential Equations September 20, 2020 14 / 55 From these result we will now derive the following properties of Laplace transform: LCH Theorem - ↵t L[eO f ](s) = L[f ](s - Fcs ) L t fit D= Fcs - H ↵) O Proof. L[e ↵t f ](s) = Harish Kumar Z 1 0 e ↵t e st f (t)dt = Z 1 0 e (s ↵)t f (t)dt = L[f ](s Laplace Transform and Di↵erential Equations ↵). September 20, 2020 15 / 55 Laplace Transform of Various Functions S. No. 1 2 3 4 5 6 7 8 9 10 11 12 f (t) 1 t t2 tn a t (a positive ) e at cos(at) sin(at) cosh(at) sinh(at) at e sin(!t) e at sin(!t) - ↳ ( wt ) L[f ] 1/s 1/s 2 O 2! s3 n! s n+1 (a+1) s a+1 1 s a s s 2 +a2 a s 2 +a2 s s 2 a2 a s 2 a2 s a (s a)2 +! 2 ! (s a)2 +a2 Table 1: Laplace transform of some functions Harish Kumar Laplace Transform and Di↵erential Equations September 20, 2020 16 / 55 Heaviside function Consider the Heaviside function at a, ( 0, if t < a, Ha (t) = 1, if t a. = - - then L[Ha (t)] = - Z 1 0 ta e st Ha (t)dt = Z 1 A e st dt = a as e s - o Holt ) YI Hao CHI L Harish Kumar Laplace Transform and Di↵erential Equations September 20, 2020 Ho * 18 / 55 Characteristic function Similarly, consider the Characteristic function, 8 > <0, if t < a, (t) = 1, if a t b, [a,b] > := 0, if t > b. x# f- { to Is - . . . - na . - a c b Laplace transform: L[ [a,b] (t)](s) = = Harish Kumar Z 1 e st [a,b] dt = 0 - Z b i a e st dt = - Laplace Transform and Di↵erential Equations e bs e as + s s September 20, 2020 19 / 55 Shifted function f˜(t) = f (t a)Ha (t) = ( t.EE . f- It - !! 0, f (t = if t < a, if t a, = a) F- ⇐ Hi Ho ) e Harish Kumar Laplace Transform and Di↵erential Equations September 20, 2020 20 / 55 Shifted function (cont.) Theorem Let f (t) be an function which satisfied (2) and F (s) = L[f ](s). Consider the shifted function ( 0, if t < a, ˜ f (t) =° f (t a)Ha (t) = f (t a) if t a, has Laplace transform L[f˜] = e as F (s), i.e. L[f (t Harish Kumar a)Ha (t)] = e as F (s) Laplace Transform and Di↵erential Equations September 20, 2020 21 / 55 Shifted function (cont.) Proof. L[f (t a)Ha (t)] = Z 1 0 - =D e as f (t a)Ha (t)dt = Z 1 a o e st f (t - - a)dt. = Now introduce variable t1 = t a and change the variable in the integral. We get, Z 1 Z 1 s(t1 +a) as L[f (t a)Ha (t)] = e→ f (t1 )dt1 = e e st1 f (t1 )dt = e as F (s). = , 0 Harish Kumar = Laplace Transform and Di↵erential Equations 0 # September 20, 2020 = 22 / 55 Shifted function (cont.) This theorem is very important as this allows us to calculate Laplace transform of functions with jump discontinuities. Furthermore, from this we can also deduce that L[f (t)Ha (t)] = e as L[f (t + a)], which is more useful form of the above result. t Harish Kumar tf hit ✓ I Laplace Transform and Di↵erential Equations September 20, 2020 23 / 55 Example: Find the Laplace transform of the following function: 8 > <2, if 0 < t < 1, f (t) = 12 t 2 , if 1 < t < ⇡/2, > : cos(t), if t > ⇡/2. Using the Heaviside function we will rewrite f (t) as, XCTIK 'D Xa i 1 # H1 (t)) + t 2 (H1 (t) H⇡/2 (t)) + H⇡/2 (t) cos(t) 2 Xco , D f (t) = 2(1 = ¥ .÷a , = Harish Kumar Laplace Transform and Di↵erential Equations September 20, 2020 24 / 55 Now using the Linearity of Laplace transform, we get, L[f ] = 2L [1 = Now, ⇤ 1 ⇥ H1 (t)] + L t 2 H1 (t) 2 ⇤ ⇥ ⇤ 1 ⇥ 2 L t H⇡/2 (t) + L H⇡/2 cos(t) 2 =\, 2L [1 = H1 (t)] = - and ⇤ 1 ⇥ 2 L t H1 (t) =e TL 2 - Similarly, Now, I e s) 2(1 s ta 1 (t + 1)2 = e s 2 ✓ 1 1 1 + 2+ s2 s 2s Ox ◆ . I ①← Ice ⇤ 1 ⇥ L H⇡/2 t 2 = e ⇡s/2 2 O ✓ 1 ⇡ ⇡2 + + s3 2s 2 8s L[H⇡/2 (t) cos(t)] = e ⇡s/2 L[cos(t + ⇡/2)] = - - ◆ . e ⇡s/2 = 1 . 1 + s2 ' - ans - = Combing all the terms we get the Laplace transform of f . Harish Kumar Laplace Transform and Di↵erential Equations September 20, 2020 25 / 55 Dirac Delta ”Function”I = Consider the function, - fk (t a) = -0 ÷ . it ÷ . Harish Kumar - a' - - f: ( I 1 k 0, if a t a + k, otherwise. P% Laplace Transform and Di↵erential Equations September 20, 2020 26 / 55 Dirac Delta ”Function” (cont.) Note that, Ik = Z 1 0 fk (t a)dt = 1, - We define Dirac delta function (t - (t (t and a) as follows, a) = lim fk (t a) = Z 1 a), k!0 = = which gives, 8k ( 1, if t = a, 0, otherwise. (t a)dt = 1. g¥ 0 - - imf ful t Hdt - a) f- Mdt le Harish Kumar Laplace Transform and Di↵erential Equations September 20, 2020 27 / 55 Dirac Delta ”Function” (cont.) One of the important property of Dirac delta function is, H ful t af Z 1 II g (t) (t a)dt = g (a) - = = Ig 0 gca , , for any continuous function g . This follows from → Z 1 0 Harish Kumar g (t) (t a)dt = lim k!0 Z 1 fk (t a)g (t)dt. 0 Laplace Transform and Di↵erential Equations September 20, 2020 28 / 55 L. T. of Dirac ”Function” f. [ felt HI - § - stfu It a) dt a)] = e as = Jagathy - st at a Note that L[fk (t = - e = 1 e ks ks are #€ So, we define, L[ (t Harish Kumar a)] = lim L[fk (t k!0 a)] = e as . Laplace Transform and Di↵erential Equations September 20, 2020 29 / 55 Consider a IVP with constant coefficients, DO x 00 (t) + ax 0 (t) + bx(t) = g (t), x(0) = x0 . x 0 (0) = x1 . Let X (s) = L[x(t)](s), then s 2 X (s) sx0 x1 + a(sX (s) x0 ) + bX (s) = G (s). Solving it for X , we get, A X (s) = w¥d_ sx0 + x1 + ax0 G (s) . s 2 + as + b → 1 f- [✗ I ¥ = - if Solution x(t) is, Harish Kumar ✗ is) s⇐÷ ¥# gkfH1 x(t) = L = Lexx] = 1 [X (s)] = L 1 sx0 + x1 + Aax0 G (s) . s 2 + as + b Laplace Transform and Di↵erential Equations September 20, 2020 42 / 55 Inverse L. T.: Well Defined? Lcf ] = Icg] Important Question: L is well defined if true , f f- I G Let f and g are two functions such that there Laplace transform is same. What can be wait about these functions ? are they equal. This question is answered in the following result: - . Theorem Let f , g are two piecewise continuous on [0, 1) and satisfy (2). If L[f ] = L[g ] on their common region of convergence, then f (t) = g (t), for all t 0 except at countable number of points. Furthermore, if f and g are continuous then f (t) = g (t) for all t 0. assume e- Stf its adt we Sg e stay # It = § { ( fit Harish Kumar ) e- *' - = dt - o ⇐ - f- 9 , mummy Laplace Transform and Di↵erential Equations f. and gave continuous September 20, 2020 ) 31 / 55 Definition Let F (s) is a given function. If there exists a function f (t) such that L[f ](s) exists and F (s) = L[f ](s), we say Fofis inverse Laplace transform of f and denote, f (t) = L 1 [F (s)](t) or simply L 1 [F ] = f . . ¥ # Example: Itt ← f- Font As L[Ha (t)] = e as /s, implies L 1 [e as /s] = Ha (t). y Linearity of Inverse L.T. - sa - - , of discontinuity Note that as L is linear operator, it follows immediately that L 1 is also linear i.e. L 1 [aF (s) + bG (s)] = aL 1 [F (s)] + bL 1 [G ]. Harish Kumar Laplace Transform and Di↵erential Equations September 20, 2020 32 / 55 Also note the following results: I L 1 [F (s I L 1 ↵)] = e ↵t f (t) F (s) = s Z ⌧ f (⌧ )d⌧ 0 I L 1 [e ↵s F (s)] = H↵ (t)f (t Harish Kumar Laplace Transform and Di↵erential Equations a) September 20, 2020 33 / 55 Some useful Inverse Laplace Transform: S. No. 1 2 3 L 1 [F ] 1 1 ↵t cos(↵t)) (s 2 +↵2 )2✓ 2↵2 (sin(↵t) → s t sin(↵t) 2 2 2 (s +↵ ) 2↵ F (s) ¥-00 ← ( - - s2 (s 2 +↵2 )2 1 2↵ (sin(↵t) + ↵t cos(↵t)) Table 2: Inverse Laplace transform of some functions partial Harish Kumar traction Laplace Transform and Di↵erential Equations Exercise September 20, 2020 34 / 55 When solving for linear ODEs we will encounter several functions of the form P(s)/Q(s) where P and Q are polynomial such that degree of P is less then degree of Q. We need to take their inverse Laplace transform. This can be done via the following results: O Theorem Let P(s) and Q(s) be polynomial of degree n and m respectively such that n < m. If Q has m simple roots 1 , · · · m , then L 1 Harish Kumar m X P( i ) P(s) = e it 0( ) Q(s) Q i 1 Laplace Transform and Di↵erential Equations September 20, 2020 35 / 55 Convolution Definition The convolution of f (t) and g (t), denoted by f ⇤ g , is the function define as, Z - - - - t f ⇤ g (t) = f (t ✓)g (✓)d✓ O= 0 I Some properties of convolution are given here: 1. f ⇤g =g ⇤f - - 2. f ⇤ (g1 + g2 ) = f ⇤ g1 + f ⇤ g2 - - - 3. (f ⇤ g ) ⇤ h = f ⇤ (g ⇤ h) ← Harish Kumar Laplace Transform and Di↵erential Equations September 20, 2020 36 / 55 Convolution (cont.) 4. f ⇤0=0⇤f =0 - - - 5. 1 ⇤ f = f ⇤ 1 6= f Harish Kumar Laplace Transform and Di↵erential Equations September 20, 2020 37 / 55 Convolution and L.T. Theorem Let f , g be piecewise continuous function of exponential order and F (s) = L[f ] and G (s) = L[g ], then = = L[f ⇤ g (t)](s) = F (s) · G (s) - and this implies, - I L 1 [F (s)G (s)] = (f ⇤ g )(t). = Harish Kumar If Laplace Transform and Di↵erential Equations September 20, 2020 38 / 55 Convolution and L.T. (cont.) Proof: F (s)G (s) = Z 1 f ( )e - 0 As s d Z 1 0 f (⌧ )e 9 - s⌧ d⌧ and ⌧ are independent variables, we can rewrite F (s)G (s) as, Z 1 Z 1 F (s)G (s) = f ( )e s( +⌧ ) d g (⌧ )d⌧. 0 0 We fix ⌧ and introduce the change of variable for internal integral as t = + ⌧ , then dt = d and we get, Z 1 Z 1 F (s)G (s) = f (t ⌧ )e st dt g (⌧ )d⌧ 0 = Z 1 Z 1 ⌧ f (t ⌧ )g (⌧ )e st dt d⌧ 0 ⌧ a Harish Kumar Laplace Transform and Di↵erential Equations September 20, 2020 39 / 55 Convolution and L.T. (cont.) We reverse the order of integration to get, - t ÷÷t z F (s)G (s) = = Z 1 Z t 0 - Harish Kumar Z 1 Z t 0 f (t ⌧ )g (⌧ )e st d⌧ dt 0 A. 0 f (t - ⌧ )g (⌧ )d⌧ e st dt = L[f ⇤ g ] - Laplace Transform and Di↵erential Equations - September 20, 2020 40 / 55 Consider a IVP with constant coefficients, x 00 (t) + ax 0 (t) + bx(t) = g (t), x(0) = x0 . x 0 (0) = x1 . Let X (s) = L[x(t)](s), then s 2 X (s) sx0 x1 + a(sX (s) x0 ) + bX (s) = G (s). Solving it for X , we get, X (s) = sx0 + x1 + ax0 G (s) . s 2 + as + b Solution x(t) is, x(t) = L 1 [X (s)] = L - Harish Kumar 1 I sx0 + x1 + ax0 G (s) . s 2 + as + b Laplace Transform and Di↵erential Equations September 20, 2020 42 / 55 Example: Consider the IVP, x 00 (t) + x(t) = g (t), - - We get, -5¥ X (s) = So, x(t) = L 1 - x(0) = 0, x 0 (0) = k. - k + G (s) . s2 + 1 I → 8M k G (s) G (s) +L 1 2 = k sin(t) + L 1 2 . s2 + 1 sI +1 s +1 - . * - - Inverse Laplace transform of F (s) = 1/(s 2 + 1) is sin(t). Using the convolution property, Z t G (s) 1 GH Asin H L = sin(t ✓)g (✓)d✓, s2 + 1 0 # which gives, x(t) = k sin(t) + Z t sin(t ¥4 , ✓)g (✓)d✓. 0 Harish Kumar Laplace Transform and Di↵erential Equations September 20, 2020 43 / 55 Higher Order ODEs with Constant Coefficients x (n) +a1 x (n 1) · · ·+an 1 x 0 +an x = g (t), x(0) = x0 , · · · x (n 1) (0)xn 1 . Taking Laplace transform of the ODE we get, s n X (s) s n 1 x0 s n 2 x1 ··· xn 1 + · · · an X (s) = G (s), which can be written as, P(s)X (s) Q(s) = G (s), :-# where P(s) = s n + a1 s n 1 + · · · + an 1 s + an = and = Q(s) = s n 1 x0 + s n 2 x1 + · · · + an 2 (sx0 + x1 ) + an 1 x0 . Harish Kumar Laplace Transform and Di↵erential Equations September 20, 2020 44 / 55 Solving for X (s) gives, X (s) = G (s) + Q(s) , P(s) - - 9¥ . I . and so the solution will be, x(t) = L - 1 [X (s)] = L - 1 → Q(s) G (s) +L 1 . P(s) P(s) f- @tttpartialfratvntE.a . . Harish Kumar Laplace Transform and Di↵erential Equations September 20, 2020 45 / 55 Linear ODE with variable coefficients: Note that, d L[tx 0 (t)] = [sX x0 ] = ds Similarly, - ÷÷÷ X s dX . ds dX . ds In fact, if the coefficient are of the form at + b, taking Laplace transform results in first order ODE of X (s). Solving this ODE and taking inverse Laplace transform results in the solution. L(tx 00 ) = 2sX s2 Consider the IVP, x + tx = 0, ± 00 x(0) = 0, x 0 (0) = b. Taking Laplace transform we get, s 2 X (s) Harish Kumar b + L[tx(t)] = s 2 X b Laplace Transform and Di↵erential Equations X 0 (s) = 0 September 20, 2020 46 / 55 So, we get, X 0 (s) s 2 X (s) = b. T This a linear ODE which can be easily solved by method of integrating factor. Then taking the inverse Laplace transform results in the solution. Harish Kumar Laplace Transform and Di↵erential Equations September 20, 2020 47 / 55 Generalized Solutions I In all the theory discussed so far for ODE we have always assumed that coefficient and the forcing terms are atleast continuous. I However, in several practical problems we have to consider the forcing which are not continuous or not even function (e.g. Dirac delta). These equations are still need to be solved. - - I The method of Laplace transform is powerful enough to handle such situations. ✓ I However, the solution we get may not be of desired smoothness. For example solution of second order may not be C 2 . These solutions are called generalized solutions. - Here we will try to find ”Generalized Solutions” ✓ v e if ¥bIde , Ho Harish Kumar CK discontinuous emt Laplace Transform and Di↵erential Equations September 20, 2020 48 / 55 Generalized Solutions (cont.) Example 1: Consider the IVP, T.EE x 00 + 3x 0 + 2x = H1 (t) H2 (t), ¥7 x(0) = 0, x 0 (0) = 0. . Taking Laplace transform s 2 X + 3sX + 2X = - which gives, 1 (e s e 2s ) s = " E e÷÷ X (s) = e s e 2s . s(s 2 + 3s + 2) Now using partial fractions, 1 1 = s(s 2 + 3s + 2) 2s Harish Kumar 1 1 + . s + 1 2(s + 2) Laplace Transform and Di↵erential Equations September 20, 2020 49 / 55 Generalized Solutions (cont.) So, f (t) = L ① 1 1 s(s 2 + 3s + 2) = - Using the shift theorem, ⇥ ⇤ x = L 1 [ e s L[f ] e 2s L[f ] = f (t = 8 > <0, 1 >2 : e + 12 e 2(t 1) (t 1) + 12 e 2(t 1) + e (t 2) 1 1 e t + e 2t . 2 2 a 1)H1 (t) f (t 1 2(t 2) 2e if if if (t 1) e 2)H2 (t) 0 < t < 1, + . 1 < t < 2, t > 2. t & . = Note that the solution is not C 2 (Check !). Te Harish Kumar Laplace Transform and Di↵erential Equations September 20, 2020 50 / 55 Generalized Solutions (cont.) Example 2:✓ Instead of Heaviside function on right if we take (t 1) then we get, x 00 + 3x 0 + 2x = (t 1), x(0) = 0, x 0 (0) = 0. ES taking Laplace transform we get, X = e s =e s 2 s + 3s + 2 Inverse Laplace transform gives, ( 0, 1 x(t) = L [X ] = e (t 1) = ✓ 1 s +1 e 2(t 1) 1 s +2 . = = , Note that the solution x(t) is not even C 1 . = Harish Kumar ◆ Laplace Transform and Di↵erential Equations if if 0<t<1 t > 1. ¥: S s September 20, 2020 51 / 55 System of Equations Consider a system of ODE, ¥8,8 x10 = a11 x1 + a12 x2 + g1 (t), x20 = a21 x1 + a22 x2 + g2 (t). ) - Let us define X1 = L[x1 ] and X2 = L[x2 ], then assume G1 = L[g1 ] and G2 = L[g2 ] exists, we get, = a- = (a11 = a s)X1 + a12 X2 = - - a21 X1 + (a22 [D - s)X2 = - x1 (0) G1 (s), x2 (0) G2 (s). I = which can be written as, ~ = sI )X I (A - ~ (0) X - G~ . - This is a linear system for X1 and X2 . Solving this we get expression for X1 and X2 is the form of s. Then inverse Laplace transform gives the solution x1 (t) and x2 (t). .. I - Harish Kumar - Laplace Transform and Di↵erential Equations September 20, 2020 52 / 55 System of Equations (cont.) Example 1: Mixing problem involving two tanks Consider the System: 8 2 x1 + x2 + 6. 100 100 x10 (t) = - Similarly for tank T2 we have, x20 (t) = 8 x1 100 8 x2 . 100 - with initial conditions x1 (0) = 0 and x2 (0) = 150. Taking Laplace transform of both the equations we get, ( 0.08 s)X1 + 0.02X2 = (0.08)X1 + ( 0.08 Harish Kumar s)X2 = 6 , s 150. Laplace Transform and Di↵erential Equations September 20, 2020 53 / 55 System of Equations (cont.) Solving for X1 and X2 we get, 100 62.5 s s + 0.12 100 125 X2 (s) = + s s + 0.12 X1 (s) = - - 375.5 , s + 0.04 75 . s + 0.04 Taking inverse Laplace transformation we get, I c x1 (t) = 100 62.5e 0.12t x2 (t) = 100 + 125e 0.12t Harish Kumar 37.5e 0.04t , 75e 0.04t . Laplace Transform and Di↵erential Equations September 20, 2020 54 / 55