690 Chapter 15. Intocuction tothe Laplace Transform Both the initial and final values could be determined from h(t) if we Knew it, See Example 15.11, where h() is given. 15.7 Obtain the initial and the final values of 6st + Os Ss Ge) => s+ +9, Answer: 6, 0.4167 15. The Inverse Laplace Transform Given F(s), how do we transform it back to the time domain and obtain the corresponding f(1)? By matching entries in Table 15.2, we avoid using Eq, (15.5) to find fi, Suppose F(s) has the general form of «as.47) where M(5) is the numerator polynomial and D(s) is the denominator polynomial. The roots of N(s) = 0 are called the zeros of F(s), while the roots of D(s) = 0 are the poles of Fis). Although Eq. (15.47) is, similar in form to Eg. (14.3), here F(s) is the Laplace transform of a Software packages such as martag, _fnetion, which is not necessarily a transfer function, We use partial Mathcad, and Maple are capable of ‘fraction expansion to break F(s) down into simple terms whose inverse fincing partial frection expansions ‘transform we obtain from Table 15.2. Thus, finding the inverse Laplace ‘quite easiy. transform of F(s) involves two steps. Steps to Find the Inverse Laplace Transform: 1, Decompose F(s) into simple terms using partial fraction expansion. 2, Find the inverse of each term by matching entries in Table 15.2, Let us consider the three possible forms F(s) may take and how to apply the two steps to each form, 15.4.1 Simple Poles Recall from Chapter 14 that a simple pole is a first-order pole. If F(s) has only simple poles, then D(s) becomes a product of factors, so that Otherwise, we must fist apply long No SDS) = Fo =§ (as.48) vision so that f(s) = N(S)/D(5) = e+ ple = pd 6 tpn Q(s) + R()/D(S), where the degree ERNE PIE Po) Cf R(), the remainder of the long where 5 = pi, =p... —Py are the simple poles, and p; # p; for all civision, is less than the degiee of D(s). i # j (ie., the poles are distinct). Assuming that the degree of N(s) is154 The Inverse Laplace Tanstorm less than the degree of D(s), we use partial fraction expansion to decompose F(s) in Eg. (15.48) as kn + (15.49) s+ Pm ‘The expansion coefficients ky, ka, .., k, are known as the residues of F(s). There are many ways of finding the expansion coefficients. One way is using the residue method. If we multiply both sides of Eq, (15.49) by (6 + pp), we obtain (> puke stp + Pky s+ Pp (+ PDF) = + 50) Since p; # pp, setting s = —p, in Eq, (15,50) leaves only k, on the right-hand side of Eq, (15.50). Hence, (8 + PDF) |r=-p, = as.s1) Thus, in general, (s+ pDFts) » (45.52) This is known as Heaviside’s theorem. Once the values of kare known, wwe proceed to find the inverse of F(s) using Eq. (15.49), Since the inverse transform of each term in Eq. (15.49) is £~'(k/(s + a)] = ke u(0), then, from Table 15.2, FQ) = (Rye ge PE ooo hye DD | (5.53) 15.4.2 Repeated Poles ‘Suppose F(s) has m repeated poles at s = ~p, Then we may represent Fis) as ke Fo o + py + (s.54) where Fy(s) is the remaining part of F(s) that does not have a pole at 5 = =p, We determine the expansion coefficient ky as fey = (5 + DFS) [pany 45.85) as we did above. To determine k,—., we multiply each term in Eg, (15.54) by (s + p)" and differentiate to get rid of k,, then evaluate the result at 5 = —p to get rid of the other coefficients except k,1. Thus, we obtain ky (15.56) 4 a Gilet PFO L~ Repeating this gives 1a . Fy gle + PFO |e (as.s7) oot Historical note: Named after Oliver Heaviside (18501995), an English engineer, the pioneer of operational calculus.Chapter 15. Intocuction tothe Laplace Transform ‘The mth term becomes nom = oy Gyal (s + py"F)I |=» (15.58) where m = 1, 2,...,m — 1. One ean expect the differentiation to be dif ficult to handle as m increases. Once we obtain the values of ky, kas x by partial fraction expansion, we apply the inverse transform, 1 ment = - Cu 5.59) . la al on? 5S) to each term on the right-hand side of Eq, (15.54) and obsain - nog Baa LO= (he Ps kyle he k (45.60) @-b + no ‘ru + HO 15.4.3 Complex Poles A pair of complex poles is simple if it is not repeated; it is a double or multiple pole if repeated. Simple complex poles may be handled the same way as simple real poles, but because complex algebra is involved the result is always cumbersome, An easier approach is a method known as completing the square. The idea is to express each complex pole pair (or quadratic term) in D(s) as a complete square such as (5+ a)” + B* and then use Table 15.2 to find the inverse of the term. Since N(s) and D(s) always have real coelficients and we know that the complex roots of polynomials with real coefficients must occur in conjugate pairs, Fis) may have the general form Ay +A, P+assb Fe) + Fils) (45.61) where Fy(s) is the remaining part of F(s) that does not have this pair of complex poles. If we complete the square by letting Partha #4 2st art B= (s+) +B? (05.62) and we also let Ays + Ay = Aye + a) + ByB (15.63) then Eq. (15.61) becomes Ais + a) BiB Fis) = Tp Tho ase Gtatre Gray From Table 15.2, the inverse uansform is FO) = Aye™ cost + Bre™™ sinBu(t) + fi) | (15.68) ‘The sine and cosine terms can be combined using Eq. (9.11). Whether the pole is simple, repeated, or complex, a general approach that can always be used in finding the expansion coefficients154 The Inverse Laplace Tanstorm 693 is the method of algebra, illustrated in Examples 15.9 to 15.11. To apply the method, we first set F(s) = N(s)/D(B) equal to an expansion containing unknown constants. We multiply the result through by a common denominator, Then we determine the unknown constants by equating coefficients (ic., by algebraically solving a set of simultaneous equations for these coefficients at like powers of 3) Another general approach is to substitule specific, convenient val- ves of s to obtain as many simultaneous equations as the number of ‘unknown coefficients, and then solve for the unknown coefficients, We ‘must make sure that each selected value of s is not one of the poles of F(s). Example 15.1 illustrates this idea Find the inverse Laplace transform of ~~ Examplenisigi Solution: The inverse transform is given by )-eGa) eG) =G-SeT + 3sin2uy, =O f) = LF) = £ where Table 15.2 has been consulted for the inverse of each term, Determine the inverse Laplace transform of Pras 6 is FOSS * 4 Faas Answ BU) + (Ge — 7 cos(S1)) ule. Find f() given that -ExampleniSi9m P42 *O= G4 DGD Solution: Unlike in the previous example where the partial fractions have been provided, we frst need to determine the partial fractions. Since there are three poles, we let 241 A KetDG79 5 B sl c std as. where A, B, and C ate the constants to be determined. We can find the constants using two approaches,694 Chapter 15. Intocuction tothe Laplace Transform I METHOD 1 Residue method: +12 2 AOSD oo TEED l aa? +12 4+12 BH 6+ DFO) ea oy a e+i2 9+12 CHG FIFO) |r sis +2) |pn C3(-D I METHOD 2 Algebraic method: Multiplying both sides of Eq. (15.9.1) by s(s + 2)(s + 3) gives S412 = Als + 2 +3) + Bols +3) + Cols +2) PH AD= AG? 4 Ss + 6) + Bl? + 35) + Cle? + 25) Equating the coefficients of like powers of s gives Constant; 12-64 = A=2 : O=SA+3B+2C > 3B 42C=-10 &: 1=AtB4+C0 = BHC=-1 Thus, A = 2,8 = 8, € = 7, and Ba, (159.1) becomes 28 7 POSS Tea aes By finding the inverse transform of each term, we obtain Fi) = @~ 8e™™ + Toul) 15.9 Fina nie 6s + 2) FO" Geer DETD Answer: f() = (eT! + 3e7% = de" u (0. [ample 15.10 — Calculate v() given that los? 4 VW) = ne + D> 2} Solution: While the previous example is on simple roots, this example is on repeated roots. Let tot +4 ss Fs + 29 A,B c ye s stl @tap st2 va) a5.10.1)154 The Inverse Laplace Tanstorm I METHOD 1 Residue method: los? +4 4 A= (8) ico = — DG + 2 Wey los? +4 14 B= +1) ars 4 = or va 562 enay 7 _ los? +4 CHEF DMO bnna = ETD a oie) a (lor +4 oa gurrw| S| (B+ 920s) — 0s? +4 Qs +) Gre MH METHOD 2 Algebraic method: Multiplying Eg. (15.10.1) by s(s + Is + 2), we obtain 10P + 4= AG + Ie +2? + Bee +2? + Cs(s + 1) + Ds(s + Ils + 2) 108 +4 =A +58 + 8544) + BIS + 4s + 45) +O +9) + DEP +32 +28) Equating coefficients, Constant: 4=4A = A=1 # O=S+4B+C+2D + 4B+C+2D=-8 a W=SA+4B+C4+3D = 4B+C+3D re O=A+BtD = B+D=-1 Solving these simultaneous equations gives A = 1, B D = 13, so that -14,.c MB 2 SS Sse st2' Grey Ya) = ‘Taking the inverse transform of each term, we get w() =~ Me™! + 1307 + 22%) Obtain g(0) if, Practice ProblemASiOm PHI +6 a= s(s + Dis +3) Answer: (2 — 3.25e~" — LSte! + 2.25e- u(t). Find the inverse transform of the frequency-domain funetion in ~ Example 15.1 Example 15.7. 20 WO = ae +BChapter 15. Inteduction tothe Laplace Transform Solution: In this example, H(s) has a pair of complex poles at s* + 81 + 25 = 0 ors = 4 * 73. We let 20 A GIO + FI) SHS Bt He (+ 8s + 25) asi.) ‘We now determine the expansion coeflicients in two ways, I METHOD 1 Combination of methods: We can obtain A using the method of residue, A=(s+ DHS) |, Although B and C can be obtained using the method of residue, we vill not do so, to avoid complex algebra, Rather, we can substitute two specific values of s [say s = 0, 1, which are not poles of F(s)] into Eq. (15.11.1). This will give us two simultaneous equations from which to find B and C. If we let s = 0 in Eq, (15.1.1), we obtain wale Boss 20 = 25A + 3C (15.11,2) Since A = 2, Eq, (5.11.2) gives C = 10. Substituting s = 1 into Eq, (15.111) gives 20 A B+C OH 4° 4 20 = 344 + 4B + 4C (45.1.3) But A = 2, C = —10, so that Bq, (15.11.3) gives B= -2, HL METHOD 2 Algebraic method: Multiplying both sides of Eq, (5.11.1) by (s + 3)(s? + 8s + 25) yields 20 = Ale + 8s + 25) + Bs + CVs +3) asa. = AG? + 85 +25) + Bs +39) + Cle +3) Equating coefficients gives # O=A+B > A= F O=844+B+C=5A+0 > CH-5A Constant, 20=25A+3C= 254-154 A= That is, B= =2, C = =10. Ths 2s +10 2 +H +2 Hg) =, - eae 543 @ tet std Gaaree 2 G+) 23 SHS GFP HO Terao155 The Conoltion egal on Taking the inverse ofeach term, we obtain ne = (2 = 26M cos = Ze sin a ass) lis alright to leave the result this way. However, we can combine the cosine and sine terms as AU) = Qe = Re cos(3t = ul) (5.11.6) To obtain Eq, (15.11.6) from Eq. (15.11.5), we apply Eq, (9.11). Next, ‘we determine the coefficient R and the phase angle Ra VES @ = 2108, 9 wn = 18.43° Thus, Ud) = 2e~™ = 2,108e™* cos(3t — 18.43°)u() Find (0) given that Practice Problem 15/14 60. 60" eset = 6e™* cos 31 = 2e7 sin 3r, 1 = 0. 15. The Convolution Integral The term convolution means “folding.” Convolution is an invaluable tool to the engineer because it provides a means of viewing and characteriz- ing physical systems. For example, itis used in finding the response y() of a system to an excitation x0), knowing the system impulse response ‘h(). This is achieved through the convolution integral, defined as yO = f YH = A) AL (15.66) or simply HE) = 2 # H(0 (as.67) ‘where A is a dummy variable and the asterisk denotes convolution. Fqua- tion (15.66) or (15.67) states thatthe output is equal tothe input convolved with the unit impulse response. The convolution process is commutative: 3 =X * A = hO #09) (15.68a) or so [sane aan= [mane a00 ase) This impli tht the onder in which the two functions are convolved is immaterial, We will see sory how fo take advantage of hs com mutative property when performing graphical computation of the con- Solon ig