1 00:00:05,664 --> 00:00:07,250 PROFESSOR: Hi everyone. 2 00:00:07,250 --> 00:00:08,060 Welcome back. 3 00:00:08,060 --> 00:00:09,560 So today, we're going to take a look 4 00:00:09,560 --> 00:00:12,750 at some forced oscillators and the exponential response 5 00:00:12,750 --> 00:00:14,150 formula. 6 00:00:14,150 --> 00:00:16,510 And the problem we're going to take a look at is first, 7 00:00:16,510 --> 00:00:20,570 for part one, to consider the equation x dot dot plus 8x 8 00:00:20,570 --> 00:00:23,060 equals, and then a forcing term on the right-hand side, 9 00:00:23,060 --> 00:00:24,287 cosine omega*t. 10 00:00:24,287 --> 00:00:26,870 And we're going to consider the case when omega squared is not 11 00:00:26,870 --> 00:00:28,010 equal to 8. 12 00:00:28,010 --> 00:00:34,470 So in the language of resonance, we're not on resonance. 13 00:00:34,470 --> 00:00:36,870 And we're also asked why is this called 14 00:00:36,870 --> 00:00:39,170 an undamped forced oscillator. 15 00:00:39,170 --> 00:00:41,970 And then for part two, to use the ERF 16 00:00:41,970 --> 00:00:44,002 to solve the differential equation x dot 17 00:00:44,002 --> 00:00:49,172 dot plus 2x dot plus 4x equals cosine 3t. 18 00:00:49,172 --> 00:00:51,130 And then, what is the natural angular frequency 19 00:00:51,130 --> 00:00:53,420 of this differential equation? 20 00:00:53,420 --> 00:00:55,820 So I'll let you take a look at these problems 21 00:00:55,820 --> 00:00:57,130 and try them out for yourself. 22 00:00:57,130 --> 00:00:58,338 And I'll be back in a moment. 23 00:01:09,487 --> 00:01:11,000 Hi everyone. 24 00:01:11,000 --> 00:01:13,400 Welcome back. 25 00:01:13,400 --> 00:01:15,810 So we're asked to find the general solution 26 00:01:15,810 --> 00:01:25,250 to the differential equation, x dot dot plus 8x equals cosine 27 00:01:25,250 --> 00:01:25,750 omega*t. 28 00:01:29,150 --> 00:01:31,180 And notice how this is a differential equation 29 00:01:31,180 --> 00:01:32,760 with constant coefficients. 30 00:01:32,760 --> 00:01:35,570 But it's being forced by a periodic function 31 00:01:35,570 --> 00:01:37,230 on the right-hand side. 32 00:01:37,230 --> 00:01:40,180 So the first thing to do is to write down 33 00:01:40,180 --> 00:01:41,850 the homogeneous solution. 34 00:01:44,904 --> 00:01:46,320 So the homogeneous solution, which 35 00:01:46,320 --> 00:01:51,770 I'll denote with a subscript h, solves the differential 36 00:01:51,770 --> 00:01:57,420 equation with the right-hand side of 0. 37 00:01:57,420 --> 00:02:02,590 This gives us a characteristic polynomial of s 38 00:02:02,590 --> 00:02:10,539 squared plus 8 equals 0, which then gives us roots 39 00:02:10,539 --> 00:02:13,100 of plus or minus root 8 i. 40 00:02:16,300 --> 00:02:23,870 So the homogeneous solution is some constant c_1. 41 00:02:23,870 --> 00:02:27,380 And when we have purely imaginary roots, 42 00:02:27,380 --> 00:02:34,250 we have cosine of the imaginary term times t, 43 00:02:34,250 --> 00:02:39,820 plus some constant c_2 times sine of the imaginary term, 44 00:02:39,820 --> 00:02:41,610 which in this case is root 8 t. 45 00:02:45,020 --> 00:02:49,516 So the homogeneous solution is always contains two constants. 46 00:02:49,516 --> 00:02:51,140 And it solves the differential equation 47 00:02:51,140 --> 00:02:53,720 with a zero right-hand side. 48 00:02:53,720 --> 00:02:56,990 The general solution, however, is the homogeneous solution 49 00:02:56,990 --> 00:03:00,060 plus one particular solution that solves the differential 50 00:03:00,060 --> 00:03:04,970 equation 8 dot dot plus 8x equals cosine omega*t. 51 00:03:04,970 --> 00:03:07,680 So we just need to find one solution to this differential 52 00:03:07,680 --> 00:03:10,670 equation. 53 00:03:10,670 --> 00:03:15,080 And in this case, I'll use the exponential response formula. 54 00:03:15,080 --> 00:03:18,090 But first note that when we use the exponential response 55 00:03:18,090 --> 00:03:22,090 formula, we need a forcing to be of the form e to s*t 56 00:03:22,090 --> 00:03:23,620 on the right-hand side. 57 00:03:23,620 --> 00:03:26,740 And cosine omega*t is not in that form. 58 00:03:26,740 --> 00:03:28,430 However, what we can do is there's 59 00:03:28,430 --> 00:03:36,629 a trick to complexify the right-hand side. 60 00:03:36,629 --> 00:03:38,920 So we note that the cosine omega*t is actually the real 61 00:03:38,920 --> 00:03:40,600 part of i*omega*t. 62 00:03:44,650 --> 00:03:53,910 So what this means is if we have a complex solution, zed, 63 00:03:53,910 --> 00:04:03,070 which solves z dot dot plus 8z equals e to the i*omega*t, 64 00:04:03,070 --> 00:04:12,250 then we can take x equals the real part of z, 65 00:04:12,250 --> 00:04:22,260 then solves the differential equation x dot dot plus 8x 66 00:04:22,260 --> 00:04:25,630 equals the real part of the right-hand side, 67 00:04:25,630 --> 00:04:26,380 cosine omega*t. 68 00:04:30,620 --> 00:04:33,380 And now we're in business because this equation 69 00:04:33,380 --> 00:04:37,500 has the form where we can use the exponential response 70 00:04:37,500 --> 00:04:39,300 formula. 71 00:04:39,300 --> 00:04:42,580 So solving this differential equation for z, 72 00:04:42,580 --> 00:04:45,470 we have a particular solution for z, 73 00:04:45,470 --> 00:04:50,980 is going to be e to the i*omega*t divided 74 00:04:50,980 --> 00:04:57,130 by the characteristic polynomial evaluated at the exponential. 75 00:04:57,130 --> 00:05:00,540 So in this case, the exponential is i*omega. 76 00:05:00,540 --> 00:05:04,960 So we must evaluate the polynomial at i*omega. 77 00:05:04,960 --> 00:05:07,640 And this is the solution provided that the polynomial 78 00:05:07,640 --> 00:05:11,352 evaluated at i*omega doesn't vanish. 79 00:05:11,352 --> 00:05:13,310 And in our case, the characteristic polynomial, 80 00:05:13,310 --> 00:05:22,020 p of s, we worked out already to be s squared plus 8. 81 00:05:22,020 --> 00:05:26,410 So this gives us a particular solution for z, 82 00:05:26,410 --> 00:05:33,810 which is 1 divided by i*omega squared plus 8 83 00:05:33,810 --> 00:05:35,730 on the denominator. 84 00:05:35,730 --> 00:05:37,720 We have e to i*omega on the numerator. 85 00:05:41,160 --> 00:05:47,300 And of course, 1 over i*omega squared plus 8 becomes 1 over 8 86 00:05:47,300 --> 00:05:48,350 minus omega squared. 87 00:05:53,880 --> 00:05:55,390 And we know we don't have a problem 88 00:05:55,390 --> 00:05:57,800 because omega, we were told in the problem, 89 00:05:57,800 --> 00:06:00,380 is not equal to the square root of 8. 90 00:06:00,380 --> 00:06:03,700 So in this case, we know that the denominator is not 91 00:06:03,700 --> 00:06:06,570 going to vanish. 92 00:06:06,570 --> 00:06:10,160 And now what we need to do is just take x, 93 00:06:10,160 --> 00:06:18,430 for the particular solution, to be the real part of z, 94 00:06:18,430 --> 00:06:24,490 which is going to be the real part-- I'll write it out-- 95 00:06:24,490 --> 00:06:41,950 8 minus omega squared cosine omega*t plus i sine omega*t, 96 00:06:41,950 --> 00:06:49,580 which gives us 1 over 8 minus omega squared cosine omega*t. 97 00:06:56,060 --> 00:07:00,440 So the total solution, the general solution, 98 00:07:00,440 --> 00:07:03,447 is going to be the sum of the homogeneous solution 99 00:07:03,447 --> 00:07:04,655 plus the particular solution. 100 00:07:07,160 --> 00:07:13,820 And in our case, that's going to be c_1 cosine omega*t plus c_2 101 00:07:13,820 --> 00:07:30,360 sine omega*t plus 1 over 8 minus omega squared cosine omega*t. 102 00:07:30,360 --> 00:07:31,489 Sorry. 103 00:07:31,489 --> 00:07:32,530 These should be root 8's. 104 00:07:38,600 --> 00:07:40,210 So this is the homogeneous part. 105 00:07:44,395 --> 00:07:46,110 And this is the particular solution. 106 00:07:49,850 --> 00:07:52,520 And this is the general solution. 107 00:07:52,520 --> 00:07:55,270 So we were asked, also, why this is sometimes called 108 00:07:55,270 --> 00:07:57,750 an undamped forced oscillator. 109 00:07:57,750 --> 00:08:00,360 Well, it's undamped, because in the differential equation, 110 00:08:00,360 --> 00:08:01,910 there is no damping term. 111 00:08:01,910 --> 00:08:04,745 There's no term proportional to x dot. 112 00:08:04,745 --> 00:08:07,120 And then, secondly, it's forced because we have a forcing 113 00:08:07,120 --> 00:08:08,078 on the right-hand side. 114 00:08:08,078 --> 00:08:11,610 We have an input function, f, which in this case is cosine 115 00:08:11,610 --> 00:08:15,200 omega*t which doesn't depend on x. 116 00:08:15,200 --> 00:08:18,420 And that forces the differential equation. 117 00:08:18,420 --> 00:08:20,640 Also note that the forcing term gives 118 00:08:20,640 --> 00:08:23,430 rise to part of the solution which is directly 119 00:08:23,430 --> 00:08:28,130 proportional to the forcing. 120 00:08:28,130 --> 00:08:31,000 In fact, it has the same frequency 121 00:08:31,000 --> 00:08:34,110 but a different amplitude. 122 00:08:34,110 --> 00:08:36,090 So this concludes part A. 123 00:08:36,090 --> 00:08:46,630 Now for part B, we're asked to use 124 00:08:46,630 --> 00:08:48,470 the ERF to solve the differential 125 00:08:48,470 --> 00:08:59,740 equation x dot dot plus 2x dot plus 4x equals cosine 3t. 126 00:09:06,950 --> 00:09:10,090 And again, we can use the same trick. 127 00:09:10,090 --> 00:09:15,510 The right-hand side isn't of the form e to the i*3t. 128 00:09:15,510 --> 00:09:21,660 But what we can do is we can write cosine 3t as the real 129 00:09:21,660 --> 00:09:30,870 part of e to the i*3t, and then solve the differential equation 130 00:09:30,870 --> 00:09:40,870 z dot dot plus 2z dot plus 4z equals e to the i*3t, 131 00:09:40,870 --> 00:09:45,110 and then take x equals the real part of z. 132 00:09:45,110 --> 00:09:46,610 And in this case, we're only looking 133 00:09:46,610 --> 00:10:00,130 for a particular solution, which we can compute using the ERF. 134 00:10:00,130 --> 00:10:02,420 So it's 1 over the characteristic polynomial 135 00:10:02,420 --> 00:10:07,127 evaluated at 3i times e to the i*3t. 136 00:10:07,127 --> 00:10:08,710 And what's a characteristic polynomial 137 00:10:08,710 --> 00:10:11,556 of this differential equation? 138 00:10:11,556 --> 00:10:21,680 Well, p of s is going to be s squared plus 2s plus 4, which 139 00:10:21,680 --> 00:10:31,110 means that p of 3i is going to be 3i squared plus 2 times 3i 140 00:10:31,110 --> 00:10:38,700 plus 4, which gives us negative 9 plus 4-- is negative 5-- 141 00:10:38,700 --> 00:10:39,370 plus 6i. 142 00:10:43,740 --> 00:10:47,530 And putting the pieces together, we end up getting that x is 143 00:10:47,530 --> 00:10:57,220 equal to the real part of 1 over negative 5 plus 6i e 144 00:10:57,220 --> 00:10:57,720 to the i*3t. 145 00:11:06,510 --> 00:11:11,050 And we can expand out the numerator using Euler's formula 146 00:11:11,050 --> 00:11:16,350 to get cosine 3t plus i sine 3t. 147 00:11:16,350 --> 00:11:22,360 And when the dust settles, I got 1 over 61-- 148 00:11:22,360 --> 00:11:35,480 let me check my notes-- minus 5 cosine 3t plus 6 sine 3t. 149 00:11:39,084 --> 00:11:41,250 So I'll let you work at this last step for yourself. 150 00:11:44,399 --> 00:11:46,690 And then, lastly, what is the natural angular frequency 151 00:11:46,690 --> 00:11:48,700 of the differential equation? 152 00:11:48,700 --> 00:11:50,530 Well, this is just some notation. 153 00:11:50,530 --> 00:11:53,430 We sometimes call the natural angular frequency 154 00:11:53,430 --> 00:11:58,830 to be the square root of the term proportional to x. 155 00:11:58,830 --> 00:12:00,870 So in this case, the term two here 156 00:12:00,870 --> 00:12:03,130 comes in as a damping term. 157 00:12:03,130 --> 00:12:07,420 The natural frequency, which is sometimes 158 00:12:07,420 --> 00:12:10,240 written as omega naught squared, is always 159 00:12:10,240 --> 00:12:13,300 equal to this term, which is 4. 160 00:12:13,300 --> 00:12:16,210 So the natural frequency, the natural angular frequency, 161 00:12:16,210 --> 00:12:17,930 omega naught, would be the square root 162 00:12:17,930 --> 00:12:19,030 of 4, which would be 2. 163 00:12:21,770 --> 00:12:23,380 So I just like to quickly recap. 164 00:12:23,380 --> 00:12:25,810 We've taken a look at forced oscillators. 165 00:12:25,810 --> 00:12:28,530 And we computed their solutions using the ERF. 166 00:12:28,530 --> 00:12:30,820 And one common trend, particularly when 167 00:12:30,820 --> 00:12:34,930 we have an oscillating input, a forcing term, 168 00:12:34,930 --> 00:12:37,400 is to first change the oscillating term 169 00:12:37,400 --> 00:12:41,330 into a complex exponential, then compute 170 00:12:41,330 --> 00:12:45,650 a particular solution using the ERF formula, 171 00:12:45,650 --> 00:12:49,880 and then take the real part of the ERF solution 172 00:12:49,880 --> 00:12:53,130 to recover a real solution to the ODE. 173 00:12:53,130 --> 00:12:55,009 That gives us the particular solution. 174 00:12:55,009 --> 00:12:56,550 And then, in some cases, we also have 175 00:12:56,550 --> 00:12:59,290 to add the homogeneous solution to it 176 00:12:59,290 --> 00:13:03,260 to get the full general solution to the ODE. 177 00:13:03,260 --> 00:13:04,800 So I'd like to conclude here. 178 00:13:04,800 --> 00:13:06,980 And I'll see you next time.