Lecture – 2 First and Second Order Linear Differential Equations Dr. Radhakant Padhi Asst. Professor Dept. of Aerospace Engineering Indian Institute of Science - Bangalore Linear Differential Equations Principle of superposition: If x1 (t ) and x2 (t ) are solutions then α1 x1 (t ) + α 2 x2 (t ) is also a solution for any scalars α1 and α 2 Example: x + x = 1 (Non homogeneous) x1 = 1 + cos t and x2 = 1 + sin t both are solutions. But 2(1 + cos t ) or [(1 + cos t ) + (1 + sin t )] are not solutions. ADVANCED CONTROL SYSTEM DESIGN Dr. Radhakant Padhi, AE Dept., IISc-Bangalore 2 ADVANCED CONTROL SYSTEM DESIGN Dr. Radhakant Padhi, AE Dept., IISc-Bangalore 3 Linear Differential Equations However, if x + x = 0 (Homogeneous) then both of the above are solutions. Lesson: Principle of superposition holds good only for homogeneous linear differential equations. ADVANCED CONTROL SYSTEM DESIGN Dr. Radhakant Padhi, AE Dept., IISc-Bangalore 4 Homogeneous First-order Equations System dynamics: x + kx = 0, x (t0 ) = x0 , k = 1/ τ Solution: ( dx / x ) = −k dt ln x = −k t + ln c ln ( x / c ) = −k t Initial condition: Hence, x = e− k t c x0 = e−kt0 c, c = ekt0 x0 x(t ) = e − k (t −t0 ) x0 ADVANCED CONTROL SYSTEM DESIGN Dr. Radhakant Padhi, AE Dept., IISc-Bangalore 5 Homogeneous First-order Equations Note: 1) 2 2 3 3 a t a t at + +" e = 1 + at + 2! 3! 2) If t0 = 0, then the solution is x(t ) = e − k t x0 = e − (t / τ ) x0 x0 0.63 x0 (0, 0) τ t (no overshoot) ADVANCED CONTROL SYSTEM DESIGN Dr. Radhakant Padhi, AE Dept., IISc-Bangalore 6 Non-homogeneous First-order Equations: Response to Step Inputs Consider: x + 1 τ x=A (A: constant) Homogeneous solution: 1 xh + xh = 0 τ − t xh = e τ c Particular solution: x p = B Substitute: 1 τ B=A ∴ xp = B = τ A ADVANCED CONTROL SYSTEM DESIGN Dr. Radhakant Padhi, AE Dept., IISc-Bangalore 7 Non-homogeneous First-order Equations: Response to Step Inputs x(t ) = xh + x p − t x(t ) = e τ c + τ A x0 = c + τ A x(t ) = e − t τ ⇒ c = x0 − τ A ( x0 − τ A) + τ A xss = lim x(t ) = τ A t →∞ (Note: xss ≠ A) Complete solution: x(t ) = e − t τ ( x0 − τ A) + τ A Steady state error (wrt. input A): ess = A − [ x(t ) ]t →∞ = A − τ A = (1 − τ ) A ADVANCED CONTROL SYSTEM DESIGN Dr. Radhakant Padhi, AE Dept., IISc-Bangalore 8 First-order Equations in Special Form: Unit Step Response Consider: ⎛1⎞ ⎛1⎞ x + ⎜ ⎟ x = ⎜ ⎟ A ⎝τ ⎠ ⎝τ ⎠ Homogeneous: xh + (Special case in text books: A = 1) 1 τ xh = 0 − t xh = e τ c Particular solution: x p = B Substitute: 1 ⎞ ⎛1 = B A ⎟ ⇒ ( B = A) ⎜ τ ⎠ ⎝τ ∴ xp = B = A ADVANCED CONTROL SYSTEM DESIGN Dr. Radhakant Padhi, AE Dept., IISc-Bangalore 9 First-order Equations in Special Form: Unit Step Response x(t ) = xh + x p − t − t x(t ) = e τ c + A x0 = c + A ⇒ c = x0 − A x(t ) = e τ ( x0 − A) + A xss = lim x(t ) = A t →∞ Complete solution: x(t ) = e − t τ ( x0 − A) + A Steady state error (wrt. input A): ess = A − xss = A − A = 0 ADVANCED CONTROL SYSTEM DESIGN Dr. Radhakant Padhi, AE Dept., IISc-Bangalore 10 Unit Step Response of a First-Order System x (t ) T =τ A =1 Settling time: Ts = 4T = 4τ ADVANCED CONTROL SYSTEM DESIGN Dr. Radhakant Padhi, AE Dept., IISc-Bangalore 11 Second Order System Homogeneous: x + ax + bx = 0, Guess: (a = 2ξωn , b = ωn 2 ) x(t ) = eλt ( λ 2 + aλ + b ) e λ t = 0 eλt ≠ 0 ⇒ λ 2 + aλ + b = 0 : characteristic equation − a ± a 2 − 4b λ1,2 = 2 ADVANCED CONTROL SYSTEM DESIGN Dr. Radhakant Padhi, AE Dept., IISc-Bangalore 12 Second Order System Note: a, b ∈ R ⇒ Arises three cases Case 1: Distinct and real roots Case 2: Complex conjugate roots Case 3: Real and repeated roots ADVANCED CONTROL SYSTEM DESIGN Dr. Radhakant Padhi, AE Dept., IISc-Bangalore 13 Case-1: Distinct and real roots x1 = e λ1 t x2 = e , λ2t x ( t ) = c1 x1 ( t ) + c 2 x 2 ( t ) x ( t ) = c1e λ1 t + c2e λ2t ADVANCED CONTROL SYSTEM DESIGN Dr. Radhakant Padhi, AE Dept., IISc-Bangalore 14 Case-2: Complex conjugate roots λ 1 = σ + jω , λ 2 = σ − jω x 1 = e ( σ + j ω ) t = e σ t (c o s ω t + j s in ω t ) x 2 = e ( σ − j ω ) t = e σ t (c o s ω t − j s in ω t ) x ( t ) = c 1 e ( σ + jω ) t + c 2 e ( σ − jω ) t However notice that: Linearly independent terms x1 + x 2 = eσt cosω t 2 x1 − x 2 = e σ t s in ω t 2 j ADVANCED CONTROL SYSTEM DESIGN Dr. Radhakant Padhi, AE Dept., IISc-Bangalore 15 Case-2: Complex conjugate roots So, we can also write the solution as : ⎛ x1 + x2 x(t ) = A ⎜ ⎝ 2 ⎛ x1 − x2 ⎞ ⎞ ⎟ ⎟ + B⎜ ⎠ ⎝ 2j ⎠ x(t ) = eσ t ( A cos ωt + B sin ωt ) ADVANCED CONTROL SYSTEM DESIGN Dr. Radhakant Padhi, AE Dept., IISc-Bangalore 16 Complex conjugate roots: Typical response plot x (t ) ADVANCED CONTROL SYSTEM DESIGN Dr. Radhakant Padhi, AE Dept., IISc-Bangalore 17 Case-3: Real and repeated roots λ1 = λ 2 = λ x1 (t ) = e λt How to find x 2 (t ) Method of variation of parameters Assume: x2 (t ) = u (t ) x1 (t ) Then x2 = u x1 + u x1 x2 = u x1 + u x1 + u x1 + u x1 ADVANCED CONTROL SYSTEM DESIGN Dr. Radhakant Padhi, AE Dept., IISc-Bangalore 18 Case-3: Real and repeated roots Substituting back and rearranging the terms, it leads to u ( x1 + a x1 + b ) + u ( 2 x1 + a x1 ) + u x1 = 0 However, ( x1 + a x1 + b ) = 0 ( since x1 is a solution ) ⎛a⎞ Moreover, x1 = λ eλ t = λ x1 = − ⎜ ⎟ x1 ⎝2⎠ ( 2 x1 + a x1 ) = 0 a⎞ ⎛ since = − λ ⎜ ⎟ 2⎠ ⎝ This leads to u x1 = 0. Moreover x1 = eλt ≠ 0. Hence u = 0 ADVANCED CONTROL SYSTEM DESIGN Dr. Radhakant Padhi, AE Dept., IISc-Bangalore 19 Case-3: Real and repeated roots u (t ) = t One solution ∴ x (t ) = t x (t ) = te λ 2 t 1 x(t ) = c1 x1 (t ) + c 2 x 2 (t ) λt x (t) = (c1 + c2 t) e Corollary: When λ = 0 (double poles at origin) x(t ) = c + c t → ∞ as t → ∞ 1 2 Double pole (in general, multiple poles) at origin is de-stabilizing..! ADVANCED CONTROL SYSTEM DESIGN Dr. Radhakant Padhi, AE Dept., IISc-Bangalore 20 Summary Case 1 Roots of characteristic equation Distinct & Real λ1 , λ 2 Complex Conjugate 2 3 λ = σ + jω Real double Roots Basics General Solution e λ1t , e λ2t c1e λ1t + c 2 e λ2t eσ t cos ωt eσ t sin ωt λt e , te eσ t ( A cos ωt + B sin ωt ) λt ADVANCED CONTROL SYSTEM DESIGN Dr. Radhakant Padhi, AE Dept., IISc-Bangalore (c1 + c 2 t )e λt 21 Example x − 4 x + 4 x = 0 λ 2 − 4λ + 4 = 0 x(0) = 3, x (0) = 1 λ = 2, 2 x(t ) = (c1 + c2t ) e (Real, repeated) 2t x (t ) = 2 (c1 + c2 t ) e 2 t + c2 e 2 t Solving for the initial conditions c1 = 3 c 2 = −5 x(t ) = (3 − 5t ) e 2t ADVANCED CONTROL SYSTEM DESIGN Dr. Radhakant Padhi, AE Dept., IISc-Bangalore 22 Particular Solution: Method of Undetermined Coefficients x + ax + bx = f (t ) Terms in f(t) ke pt Choice of xp(t) ce pt k tn k cos ωt kn t n + kn −1 t n −1 + " + k1t + k0 k sin ωt A cos ωt + B sin ωt A cos ωt + B sin ωt Substitute these expressions back in the differential equation and determine the unknown coefficients ADVANCED CONTROL SYSTEM DESIGN Dr. Radhakant Padhi, AE Dept., IISc-Bangalore 23 Example Find x p (t ) for Choose x − 3 x + 2 x = 4t + e3t x p (t ) = (k1t + k0 ) + ce3t Substitute and equate the coefficients 1 k1 = 2, k 0 = 3, c = 2 1 3t ∴ x p (t ) = (2t + 3) + e 2 ADVANCED CONTROL SYSTEM DESIGN Dr. Radhakant Padhi, AE Dept., IISc-Bangalore 24 ADVANCED CONTROL SYSTEM DESIGN Dr. Radhakant Padhi, AE Dept., IISc-Bangalore 25 Second order system in Standard form c + 2ςωn c + ω c = ω u 2 n 2 n ωn : Natural frequency ς : Damping ratio Transfer function (will be studied later): C (s) ωn2 = 2 2 U ( s ) s + 2ςωn s + ωn Roots: poles ADVANCED CONTROL SYSTEM DESIGN Dr. Radhakant Padhi, AE Dept., IISc-Bangalore 26 Unit Step Response of a Second-order System ξ=0 0<ξ<1 ξ=1 ξ>1 ADVANCED CONTROL SYSTEM DESIGN Dr. Radhakant Padhi, AE Dept., IISc-Bangalore 27 Under-damped system response specifications Ref: N. S. Nise: Control Systems Engineering, 4th Ed., Wiley, 2004 Tr : Rise time Tp : Pick time Ts : Settling time ADVANCED CONTROL SYSTEM DESIGN Dr. Radhakant Padhi, AE Dept., IISc-Bangalore 28 Transient Response Specifications π −β R is e tim e : T r = , ωd π P e a k tim e : T p = ωd D a m p e d n a tu ra l fre q u e n c y: ω d = ω n 1 − ζ M a x im u m o v e r s h o o t: M S e ttlin g tim e : T s = = 4 ξω n 3 ξω n p = e ⎛ − πξ ⎜ ⎜ ⎝ 1− ζ 2 ⎞ ⎟ ⎟ ⎠ (2% c r ite r io n ) (5 % c r ite r io n ) ADVANCED CONTROL SYSTEM DESIGN Dr. Radhakant Padhi, AE Dept., IISc-Bangalore 2 29 Percentage over-shoot as a function of damping ratio Ref: N. S. Nise: Control Systems Engineering, 4th Ed., Wiley, 2004 Good to increase damping ratio. ADVANCED CONTROL SYSTEM DESIGN Dr. Radhakant Padhi, AE Dept., IISc-Bangalore 30 Rise time as a function of damping ratio Ref: N. S. Nise: Control Systems Engineering, 4th Ed., Wiley, 2004 Increasing damping ratio too much is Bad..! ADVANCED CONTROL SYSTEM DESIGN Dr. Radhakant Padhi, AE Dept., IISc-Bangalore 31 Second-order System Response as a Function of Damping Ratio ADVANCED CONTROL SYSTEM DESIGN Dr. Radhakant Padhi, AE Dept., IISc-Bangalore 32 ADVANCED CONTROL SYSTEM DESIGN Dr. Radhakant Padhi, AE Dept., IISc-Bangalore 33 ADVANCED CONTROL SYSTEM DESIGN Dr. Radhakant Padhi, AE Dept., IISc-Bangalore 34