2.004 Problem Set #4 Solutions

MASSACHUSETTS INSTITUTE OF TECHNOLOGY Department of Mechanical Engineering 2.004 Dynamics and Control II Problem Set #4 Solution 1. (a) Problem 21(a) from Nise textbook, Chapter 2 (page 113). (b) After you find the transfer function, locate the zeros and poles, draw them on the s–plane, and determine the type of the response of the system (undamped, underdamped, critically damped, or overdamped). a. Answer: We choose current i1 (t) and i2 (t) as shown in the figure. I1 (s) = Vi − V1 6 5 × 105 + 10s I2 (s) = Vo − V1 6 105 + 10s and I2 (s) = −I1 (s). Also V1 = 0 because the positive terminal is connected to the ground. The negative terminal should have the same potential. Arranging the above equations, we obtain the transfer function given by 1 (s + 10) Vo (s) =− . Vi (s) 5 (s + 2) b. Answer: Pole: -2, Zero: -10. It is a first order system since has only one pole. It also has one zero. 1 Step Response Pole−Zero Map 1 −0.2 −0.3 −0.4 Amplitude Imaginary Axis 0.5 0 −0.5 −0.6 −0.7 −0.5 −0.8 −0.9 −1 −12 −10 −8 −6 −4 −2 −1 0 0 0.5 1 Real Axis 1.5 Time (sec) 2 2.5 3 (Drawing the step response is not required) 2. Problem 25 from Nise textbook, Chapter 4 (page 236). Answer: From the free body diagram, 3ẍ = f (t) − 33x − 15ẋ. In the Laplace domain, the equation of motion is (3s2 + 15s + 33)X(s) = F (s). The transfer function is X(s) 1/3 = 2 . F (s) s + 5s + 11 (For the following items, refer to the lecture note 7 p.26) √ • The natural frequency ωn = 11 rad/s. • The damping ratio ζ = 5/2/ωn = 0.7538. � � ζπ √ • %OS = exp − × 100 = 2.7% 2 1−ζ • Ts ≈ • Tp = 4 ζωn = 1.6sec. √π ωn 1−ζ 2 = 1.4415sec. • To compute the rise time approximately, we can use the normalized rise time method in section 4.5 of Nise textbook (pp.196). Since the damping 2 ratio ζ ≈ 0.75, the normalized rise time will be 2.2965 (s). If we divide it by ωn , the rise time Tr ≈ 0.6925 (s). If you want to compute the exact rise time,�you may use the exact step response. Since σd = ζωn = 2.5 and ωd = ωn 1 − ζ 2 = 2.1794, (by partial fraction expansion) the exact step response is � � f (t) = 0.0303 1 − e−2.5t cos(2.1794t) − 0.0681e−2.5t sin(2.1794t) u(t). The step response and the normalized one (steady state = 1) are plotted. Step Response Normalized Step Response OS 1.03 0.035 0.03 1 0.9 Amplitude Amplitude 0.025 0.02 0.015 0.01 Tr Tp 0.005 Ts 0.1 0 0 1 2 3 Time (sec) 4 0 5 0 0.15 0.84 1 Time (sec) 1.44 1.6 2 All specifications are marked and note that Tr ≈0.69 agrees with the ap­ proximated method very well. 3. Derive the transfer function of the electrical network shown below with the source voltage vs (t) as input and the voltage vc (t) across the capacitor as output, where R2 = 10Ω, L = 1 mH, C = 10μF and i) R1 = 5Ω; ii) R1 = 100Ω. L R1 vs + - R2 C + - vc For each case, a) locate the zeros and poles and draw them on the s–plane; b) determine the type of the response (undamped, underdamped, critically damped, or overdamped); c) if you find that the system is underdamped, determine the natural frequency, the damped frequency, the settling time and the percent overshoot (%OS); if, on the other hand, you find that the system is overdamped, determine the two time constants; 3 d) determine the steady–state value directly from the transfer function; e) plot the response to a step voltage of magnitude 5 V and verify that it matches your answers to the previous questions. Answer: From the voltage divider rule learned in lecture 11, we can think the following equivalent circuit. The combination of R2 , C and L can be considered as impedance Z1 and the voltage across Z2 is denoted as vz (t). From the two figures, we can derive the following two relations in the Laplace domain. 1/Cs Vz Ls + 1/Cs Z2 = Vs R 1 + Z2 Vc (s) = Vz The impedance Z2 is given by 1 1 1 = + . Z2 R2 Ls + 1/(Cs) Rearranging the above equation, we can derive the transfer function. R2 Vc (s) = 2 Vs (s) LC(R1 + R2 )s + R1 R2 Cs + (R1 + R2 ) R2 1 = × 2 R1 R2 1 (R1 + R2 )LC s + L(R +R ) s + LC 1 2 From KCL and KVL, you will end up with the same result. i) R1 = 5Ω; a) The transfer function is T (s) = (2/3) × 108 . s2 + (1/3) × 104 s + 108 � � √ � √ � Poles: p1 = −5/30 + j 635 × 104 and p2 = −5/30 − j 635 × 104 4 Pole−Zero Map 4 1.5 x 10 Imaginary Axis 1 0.5 0 −0.5 −1 −1.5 −1500 −1000 Real Axis −500 0 b) The system is underdamped because the two ples have imaginary com­ ponent and their real part is negative (i.e., on the left–hand plnae). c) ωn = 104 rad/s, ζ = 0.1667. � ωd = ωn 1 − ζ 2 = 9860 rad/s. Ts = 4 ζωn = 0.0024 (s). � � ζπ %OS = exp − √ 2 × 100 ≈ 58%. 1−ζ d) From the final value theorem, 1 2 f (t)|t→∞ = lim s T (s) = lim T (s) = . s→0 s s→0 3 e) The step response. Step Response Amplitude 1.06 0.667 T s 0 0 0.5 1 1.5 2 Time (sec) ii) R1 = 100Ω. 5 2.5 3 3.5 −3 x 10 a) The transfer function is T (s) = � Poles: p1 = −10/22 + √ � j 1196 � √ � × 10 and p2 = −10/22 − j 1196 × 104 4 Pole−Zero Map 4 1 (1/11) × 108 . s2 + (1/11) × 105 s + 108 x 10 Imaginary Axis 0.5 0 −0.5 −1 −5000 −4000 −3000 −2000 Real Axis −1000 0 b) The system is underdamped because the two ples have imaginary com­ ponent and their real part is negative (i.e., on the left–hand plnae). c) ωn = 104 rad/s. ζ = 0.4545. � ωd = ωn 1 − ζ 2 = 8907.5 rad/s. Ts = 4 ζωn = 0.00088 (s). � � ζπ %OS = exp − √ 2 × 100 ≈ 20%. 1−ζ d) From the final value theorem, 1 1 f (t)|t→∞ = lim s T (s) = lim T (s) = . s→0 s s→0 11 e) The step response. 6 Step Response 0.12 0.109 Amplitude 0.0909 T s 0 0 0.5 1 Time (sec) 1.5 −3 x 10 4. Consider again the system of a compliant mass driving a shaft that we modeled in PS01/4.a and derived the transfer function for in PS02/4. Substitute numerical values M = 1 kg, J = 1 kg · m2 , fv = 2 kg/ sec, r = 0.1 m, and two cases of translational compliance i) K = 0.1 N/ m; ii) K = 1.0 N/ m. For each case, a) find the poles of the system (you can use the roots function in Matlab , or any numerical analysis software of your choice); b) determine if the system has dominant poles; c) sketch by hand the step response of the system as accurately as you can from the transfer function but without solving for the exact solution. Do not use any numerical analysis tools for this part of the problem. Answer: The transfer function is fv r Θ(s) = . 3 2 F (s) JM s + fv (r M + 2J) s2 + (fv2 r2 + KJ) s + Kr2 fv Plugging in the numerical values, we end up with 0.2 Θ(s) = 3 . 2 F (s) s + 4.02s + (0.04 + K)s + 0.02K i) K = 0.1 N/m T (s) = 0.2 . s3 + 4.02s2 + 0.14s + 0.002 7 √ √ a) Poles: p1 = −3.9850, p2 = −0.0175 + j 0.014, p3 = −0.0174 − j 0.014 Since two of the poles have imaginary parts, we can expect this three– pole system to have oscillatory response; this will become even clearer immediately next when we examine the system’s dominant poles. b) Dominant poles: p2 and p3 . (|Re {p2 or p3 }| << p1 ) c) Neglecting the fast pole p1 , we can approximate the system transfer function as follows T (s) = 0.2/(−p1 ) 0.2 ≈ (s − p1 )(s − p2 )(s − p3 ) (s − p2 )(s − p3 ) 0.0502 = 2 . s − (p2 + p3 )s + p2 p3 (we divide by −p1 to ensure that the steady state value remains the same after approximating the three–pole system with its two dominant poles only.) From the denominator of the response, ωn = 0.0224 rad/s and ζ ≈ 0.78. This is an underdamped system. The estimated specifications are %OS≈ 2%, Ts = 228.5 (s), Tp = 224.7 (s). Step Response 120 OS Amplitude 100 3rd order sys. 2nd order sys. Tp 0 0 100 200 Time (sec) T s 300 350 (The exact response of the third order system and the response from the two dominant poles alone are plotted together for your information.) ii) K = 1.0 N/m T (s) = s3 + 4.02s2 0.2 . + 1.04s + 0.02 a) Poles: p1 = −3.7436, p2 = −0.255, p3 = −0.0209 b) Dominant poles: p3 = −0.0209. (|10|p3 < |p2 | and 10|p2 | < |p1 |). c) Its steady state value is 10. Neglecting the two fast poles p1 , p2 , we can 8 approximate the system’s transfer function as follows K0 K0 /(p1 p2 ) 0.2091 ≈ = . (s − p1 )(s − p2 )(s − p3 ) (s − p3 ) s + 0.0209 (we divide by p1 p2 to ensure that the steady state value remains the same after approximating the three–pole system with its single domi­ nant pole.). The approximated system is a first order system, with step response f (t) = 10(1 + K0 ep3 t )u(t). Step Response 12 10 Amplitude 8 6 4 2 0 0 3rd order sys. 1st order sys. 50 100 150 Time (sec) 200 250 300 (The exact response of the third order system and the response of the approximated first order system are plotted together for your informa­ tion.) 9