Differential Equations and Linear Superposition

Similar documents
Chapter 7. Response of First-Order RL and RC Circuits

9. Capacitor and Resistor Circuits

Inductance and Transient Circuits

Module 4. Single-phase AC circuits. Version 2 EE IIT, Kharagpur

Capacitors and inductors

RC (Resistor-Capacitor) Circuits. AP Physics C

The Transport Equation

Signal Processing and Linear Systems I

Differential Equations. Solving for Impulse Response. Linear systems are often described using differential equations.

CHARGE AND DISCHARGE OF A CAPACITOR

Mathematics in Pharmacokinetics What and Why (A second attempt to make it clearer)

Transient Analysis of First Order RC and RL circuits

cooking trajectory boiling water B (t) microwave time t (mins)

1. y 5y + 6y = 2e t Solution: Characteristic equation is r 2 5r +6 = 0, therefore r 1 = 2, r 2 = 3, and y 1 (t) = e 2t,

AP Calculus AB 2013 Scoring Guidelines

Stochastic Optimal Control Problem for Life Insurance

Second Order Linear Differential Equations

Full-wave rectification, bulk capacitor calculations Chris Basso January 2009

2.5 Life tables, force of mortality and standard life insurance products

Economics Honors Exam 2008 Solutions Question 5

4 Convolution. Recommended Problems. x2[n] 1 2[n]

17 Laplace transform. Solving linear ODE with piecewise continuous right hand sides

Appendix A: Area. 1 Find the radius of a circle that has circumference 12 inches.

AP Calculus BC 2010 Scoring Guidelines

MTH6121 Introduction to Mathematical Finance Lesson 5

Equation for a line. Synthetic Impulse Response Time (sec) x(t) m

Random Walk in 1-D. 3 possible paths x vs n. -5 For our random walk, we assume the probabilities p,q do not depend on time (n) - stationary

Making Use of Gate Charge Information in MOSFET and IGBT Data Sheets

Lectures # 5 and 6: The Prime Number Theorem.

Signal Rectification

Technical Appendix to Risk, Return, and Dividends

THE PRESSURE DERIVATIVE

PROFIT TEST MODELLING IN LIFE ASSURANCE USING SPREADSHEETS PART ONE

Steps for D.C Analysis of MOSFET Circuits

µ r of the ferrite amounts to It should be noted that the magnetic length of the + δ

Part II Converter Dynamics and Control

Voltage level shifting

The Torsion of Thin, Open Sections

AP Calculus AB 2007 Scoring Guidelines

AP Calculus AB 2010 Scoring Guidelines

Chapter 2 Kinematics in One Dimension

CAPACITANCE AND INDUCTANCE

11/6/2013. Chapter 14: Dynamic AD-AS. Introduction. Introduction. Keeping track of time. The model s elements

Fourier Series and Fourier Transform

Chapter 2 Problems. 3600s = 25m / s d = s t = 25m / s 0.5s = 12.5m. Δx = x(4) x(0) =12m 0m =12m

= r t dt + σ S,t db S t (19.1) with interest rates given by a mean reverting Ornstein-Uhlenbeck or Vasicek process,

Cointegration: The Engle and Granger approach

Name: Algebra II Review for Quiz #13 Exponential and Logarithmic Functions including Modeling

Chapter 4: Exponential and Logarithmic Functions

Module 3. R-L & R-C Transients. Version 2 EE IIT, Kharagpur

4. International Parity Conditions

Newton s Laws of Motion

Present Value Methodology

A Curriculum Module for AP Calculus BC Curriculum Module

A Note on Using the Svensson procedure to estimate the risk free rate in corporate valuation

Permutations and Combinations

Answer, Key Homework 2 David McIntyre Mar 25,

Morningstar Investor Return

Suggested Reading. Signals and Systems 4-2

Communication Networks II Contents

Switched Mode Converters (1 Quadrant)

Analogue and Digital Signal Processing. First Term Third Year CS Engineering By Dr Mukhtiar Ali Unar

Switching Regulator IC series Capacitor Calculation for Buck converter IC

DC-DC Boost Converter with Constant Output Voltage for Grid Connected Photovoltaic Application System

Acceleration Lab Teacher s Guide

Individual Health Insurance April 30, 2008 Pages

LLC Resonant Converter Reference Design using the dspic DSC

A Probability Density Function for Google s stocks

Imagine a Source (S) of sound waves that emits waves having frequency f and therefore

Option Pricing Under Stochastic Interest Rates

Chapter 8: Regression with Lagged Explanatory Variables

Kinematics in 1-D From Problems and Solutions in Introductory Mechanics (Draft version, August 2014) David Morin,

Pulse-Width Modulation Inverters

Astable multivibrator using the 555 IC.(10)

Duration and Convexity ( ) 20 = Bond B has a maturity of 5 years and also has a required rate of return of 10%. Its price is $613.

Lecture 2: Telegrapher Equations For Transmission Lines. Power Flow.

Vector Autoregressions (VARs): Operational Perspectives

Full-wave Bridge Rectifier Analysis

Task is a schedulable entity, i.e., a thread

Chapter 8 Student Lecture Notes 8-1

Fourier Series & The Fourier Transform

ECEN4618: Experiment #1 Timing circuits with the 555 timer

Term Structure of Prices of Asian Options

ON THE PRICING OF EQUITY-LINKED LIFE INSURANCE CONTRACTS IN GAUSSIAN FINANCIAL ENVIRONMENT

ANALYSIS AND COMPARISONS OF SOME SOLUTION CONCEPTS FOR STOCHASTIC PROGRAMMING PROBLEMS

1 HALF-LIFE EQUATIONS

The Fourier Transform

WHAT ARE OPTION CONTRACTS?

A Re-examination of the Joint Mortality Functions

Stability. Coefficients may change over time. Evolution of the economy Policy changes

Making a Faster Cryptanalytic Time-Memory Trade-Off

Mortality Variance of the Present Value (PV) of Future Annuity Payments


NOTES ON OSCILLOSCOPES

Return Calculation of U.S. Treasury Constant Maturity Indices

I. Basic Concepts (Ch. 1-4)

Mechanical Fasteners Tensile and Shear Stress Areas

Life insurance cash flows with policyholder behaviour

Chapter 13. Network Flow III Applications Edge disjoint paths Edge-disjoint paths in a directed graphs

Transcription:

Differenial Equaions and Linear Superposiion Basic Idea: Provide soluion in closed form Like Inegraion, no general soluions in closed form Order of equaion: highes derivaive in equaion e.g. dy d dy 2 y 0 d is a rd order, non-linear equaion. Firs Order Equaions: (separable, eac, linear, ricks) A separable equaion can be wrien: e.g. dy A( ) B( y) A( )d d A ( )d B ( y)dy C dy y 2 d 2 0 B( y)dy dy y 2 d y 2 y 2 C 0 2

More Generally, an equaion can be an eac differenial: A(, y)d B (, y)dy 0 if he lef side above is he eac differenial of some u(,y) hen he soluion is: u(, y) C A necessary and sufficien condiion for his is: A(, y) y B(, y) e.g. ( 2 2y) ( 2 y 2 )y is eac. We can hen parially inegrae A(,y) in and B(,y) in y o guess a from for u(,y): ( 2 2y)d ( 2 y 2 )dy u(, y) 2y 2y y y C ( y) C( ) 2y C For firs order equaions, i is always possible o find a facor λ(,y) which makes he equaion eac. Unforunaely, here is no general echnique o find λ(,y)...

Consider he general linear firs-order equaion: dy d f( )y g( ) Les ry o find an inegraing facor for his equaion... λ ( y, ) We wan λ ( )dy λ ( ) ( f( )y g( ) )d 0 o be eac. Equaing he parial derivaives: A(, y) y B(, y) d d λ ( ) λ( )f( ) This new equaion for lambda is separable wih soluion: λ ( ) e f( ) d So we can always find an inegraing facor for a firs order linear differenial equaion, provided ha we can inegrae f(). Then we can solve he original equaion since i is now eac. e.g. The inegraing facor is: y ( )y e y y ( ) e λ ( ) e f( ) d e So we ge he eac equaion: e y e y( ) e 2 Inegraing boh sides: e y e 2 C y 2 e 2 Ce

Why do we call he equaion below linear? Consider a se of soluions: dy d f( )y 0 y { y ( ), y 2 ( ),... } Any linear combinaion of he above soluions is anoher valid soluion of he original equaion. If we se: y ay ( ) by 2 ( ) we ge: dy a d dy 2 b af ( )y d ( ) bf ( )y 2 ( ) 0 which is clearly saisfied if y and y 2 are indeed soluions. Linear equaions are saisfied by any linear superposiion of soluions. We divide he se of soluions ino a se of linearly independen soluions saisfying he linear operaor, and a paricular soluion saisfying he forcing funcion g(). In general, i can be shown ha over a coninuous inerval, an equaion of order k will have k linearly independen soluions o he homogenous equaion (he linear operaor), and one or more paricular soluions saisfying he general (inhomogeneous) equaion. A general soluion is he superposiion of a linear combinaion of homogenous soluions and a paricular soluion. A wide variey of equaions of ineres o an elecrical engineer are in fac linear...

Consider he circui below: I R V C Assuming ha C is iniially uncharged, when he swich is closed, a curren flows across he resisor, charging he capacior. We know ha he oal charge Q across a capacior is jus C imes he volage across i: Q() CV C () Also, he charge on he capacior mus be he inegral of he curren I hrough he resisor: Q() I C () d The curren hrough he resisor is deermined by he volage across i: I R () V R () R From Kirchoff s laws, we have ha he oal volage drop around a loop is zero and ha he oal curren a each node is zero. So we can find V C from he consan baery volage V and he volage across he resisor: V V R () V C (), I R () I C () So we ge: V C () d d V C () C ( V V C () ) d R ( V V C () )

Given ha V C (0) 0, find he volage across he capacior as a funcion of ime. We have: The homogenous soluion is separable: d d V C () And he paricular soluion is rivial since d/d (V) 0: So he general soluion is: d d V C () Since V C (0) 0, we have: V C () V V C () 0 V C () e V C () V V C () V ae V C () V Ve V e I is cusomary o se o ime scale of such a circui by he quaniy which has he unis of ime. A, V C V(-e - ) 0.62V. 0.8 0.6 0.4 0.2 2 4

Now consider he case in which he driving volage is a sine wave such as power grid disribuion curren. We hen have: d d V C () This paricular soluion is a bi harder han he las, bu he equaion sill falls ino he firs-order linear mold: f() / and g() A/*Sin(w). The inegraing facor is: so he equaion below should be eac: e d d V C () V C () Asin ( w) λ () e Carrying hrough he inegraion as before: e V C () Asin ( w)e V C sin ( w) wcos ( w) A R 2 C 2 w 2 A ( sin ( w) wcos ( w) ) w2 is he paricular soluion. Since we can superpose soluions, if we have a consan volage and a sinewave drive, he general soluion is: V C () ae A ( sin ( w) wcos ( w) ) w2 An eample of such a soluion is he circui wih a consan volage and a small amoun of AC noise. Alhough he soluion above looks dauning, remember ha any linear combinaion of sin and cos funcions of he same frequency can always be wrien in he form: Asin(wθ). Thus he capacior is changing he phase of he impressed volage. The morass of algebra above is why we ofen choose o wrie such funcions as he real par of e i(wθ). (Phasor noaion).

Eample If he Baery was in fac a source of alernaing curren, wih uni frequency and ampliude, and if he inial charge on he capacior was zero, we ge he following plo for he volage across he capacior: 0.2 0. 2 4-0. Noe ha here is a ransien adjusmen o he phase and poenial. A wice he frequency, he phase and he ampliued are boh changed: 0. 0.05 2 4-0.05 However, noe ha he ime scale of he ransien is unchanged.

We can eend his noion of superposiion of soluions o he case of muliple sources and muliple simulaneous equaions. To solve, we simply consruc soluions for each of he sources acing independenly, and hen superpose he (add) hem o find he ne behavior. For eample, if he inpu o he previous circui was wo waves, one a uni frequency and zero phase and anoher a wice he frequency and zero phase, we ge he following superposed soluion: V C () 2π 4π 2 e sin ( 2π) 2πcos ( 2π) 4π 2 4π 6π 2 e sin ( 2π) 2πcos ( 2π) 4π 2 Noe, ha he area of he wave ends o balance across he capacior: 0. 0.2 0. -0. 2 4-0.2

Phasors Represen an alernaive curren by a single comple number Trick -- use o represen boh he ampliude and phase Ae iϕ The idea here is ha i is much easier o manipulae polar form comple numbers han rigonomeric funcions. We shall represen: Acos ( ω ϕ) Re Ae iϕ e iw ( ) So far -- i doesn look much beer. However, in some applicaions all pars of he circui have he same frequency of operaion. In his case, we only need o carry he comple phasor Ae iϕ o keep rack of boh. A phasor of he sum of wo alernaing signals is he same as he sum of he phasors So if we have wo volage sources A and A2, represened by A e iϕ and bya 2 e iϕ 2 heir sum is jus A e iϕ A 2 e iϕ 2. (Noe ha his doe no hold for he produc of wo phasors. I is herefore safe o represen soluions o linear differenial equaions as phasors and do calculaions in shorand, as long as one realizes ha he real-par is he desired soluion. A second applicaion will be he calculaion of branch volages and currens in alernaing curren (A.C.) circuis. In a general nework of resisors, inducors and capaciors, he general behavior can be shown o be linear. Thus, he volage beween any wo nodes or he curren along any branch will have he same frequency as he source. Thus only wo numbers, (a phasor) sufices o deermine he circui behavior--- he ampliude and he phase. (We shall come back o his opic in a laer lecure).

Oscillaing (Resonan) Circuis Consider he circui below: I C L R If we ake posiive curren as shown, we have: dq di I Q CV V L RI d d To solve his, we will eliminae boh Q and I -- o ge a differenial equaion in V: di d I d d ( CV) dv C d 2 2 dv dv C V LC 2 d 2 d 2 dv R d 2 dv Ld LC V 0 dv d This is a linear differenial equaion of second order (noe ha solve for I would also have made a second order equaion!). Such equaions have wo indepeden soluions, and a general soluion is jus a superposiion of he wo soluions. We shall assume ha inially, he volage sored in C is A.

Suppose we ry: V Âe iw for a comple phasor Â. We have: ω 2 Âe iw iω R L Âeiw LC Âeiw 0 This equaion holds for all, w, A only if he following quadraic equaion is idenically zero: We ge: ω 2 i R L ω 0 LC ω i R 2L ± LC R 2 4L 2 So w is in general comple. This equaion has several special cases: R 0 (lossless case) hen w is pure real and he soluion is an alernaing volage of frequency:. V LC Re( Âe iw ) A cos ( LC ϕ) Here he phase is zero for he iniial condiions. R posiive, bu w sill has a real par: LC R 2 > 4L 2 R < 2 L C V R 2L Ae cos LC R 2 4L 2 ϕ In his case, he real par of w defines he frequency of oscillaion, while he

imaginary par deermines he energy loss. Noe ha he ampliude decreases eponenially for increasing ime. R large -- w is pure imaginary: V R 2L A e R 2 2L 2 LC R 2L e R 2 2L 2 LC In his case, he volage is no oscillaory -- in fac, i simply falls eponenially o zero. Finally, here is a special criical value: R 2 L C In his case, he equaion above is singluar -- or equivalenly, he quadraic equaion for w has a repeaed roo. ω i R 2L V ( A B)e i LC LC In each of he above cases, we have consisanly used he noion ha he value was represened by he real par of he comple polar form. This rick has saved an enormous amoun of work -- compared o using rigonomeric funcions.

More Techniques for Differenial Equaions A paricularly powerful echnique is ha of making an appropriae change of variable in an equaion. For eample, y' f( a by c) is easily solved if we make he subsiuion: u a by c A variable change is performed by wriing he form o be subsiued, find each (oal) derivaive and hen subsiuing. e.g. u a by c, u' a by' So he equaion above becomes: u' a b f( u) du a bf( u) d which is separable. Noe ha afer he inegraion, we mus replace u by abbyc o ge he final soluion. In general, a appropriae subsiuion is a very powerful echnique o simplify or solve a ypical differenial equaion. Recognizing ha a paricular subsiuion is useful is ofen a ricky proposiion. Some special cases are lised below: y' f( )y g( )y n Bernoulli Equaion can be made linear by seing u y n. Noe ha: u' ( n)y n y'

and: y' y n f( ) y n g( ) So subsiuing: u' n uf ( ) g( ) which is linear. Homogenous firs-order equaions: denoes a homogenous funcion of degree r. A general firs order equaion is homogenous if boh A and B are homogenous funcions. Then can be solved by subsiuing: A( a, ay) a r A(, y) A(, y)d B (, y)dy 0 y( ) u ( ), dy u( )d du E.g. yd ( 2 y )dy 0 Here r. If we subsiue: u ( )d ( 2 2 u ) ( u( )d du) 2 2u d ( 2 u )du 0 which is separable. 2d ( 2 u ) u 2 du

If he dependen variable y is missing, subsiue y u, lowering he order by one. e.g. y'' f( )y' g( ) 0 u' f( )u g( ) 0 Then solve and inegrae when done... If he independen variable is missing, le y be he new independen variable, and le u y, u y ec. This also lowers he order by one. e.g. e.e.g. y'' f( y' ) g( y) 0 u' f( u) g( y) 0 y'' y' y 2 0 u' u y 2 0 The homogenous soluion: u' u 0 u ae y The paricular soluion: u y 2 2y 2 In general: u y' ae y y 2 2y 2 So finally: ae y y 2 dy 2y 2

Always wach for an eac equaion or for an appropriae inegraing facor: e.g. y'' f( y) can be inegraed immediaely if boh sides are muliplied by y... y''y' f( y)y' 2 ( y' )2 f ( y)dy e.e.g. y'' y 2 2 ( y' )2 y y 6 2 y' y 2 ± 6 ( C) 2 A paricularly useful subsiuion is one of he simples: for a linear second order equaion. Given he equaion: Subsiuing, we obain: y u( )v( ) y f( )y g( )y 0 2u u fu gu v ( f( ) )v ( )v 0 u u Now, i is clear ha special chioces for u will poenially give us simplified equaions for v... For eample, if u is a soluion o he original equaion, he hrid erm drops ou, leaving an equaion han can be reduced o firs order, linear. If we ry: u hen u saisfies he second erm and i drops ou, leaving a simpler second order equaion. ( ) f( )d 2 e

e.g. If our equaion is: y 2 y 4 y 0 and we happen o guess ha one soluion is: y e we can use he previous subsiuion o find he oher soluion... So we subsiue: y( ) v( )e v( ), y ( ) ( v ( ) )e 2 y ( ) v 4 v v v e 4 And we ge: v v v 0 2 This equaion can be simplified by leing w v : Which is separable-- we ge: w w ( ) 0 2 log ( w) C 2 log ( ) 0 2 w ae 2 v So inegraing, one las ime: v ( c)e 2 y is he oher soluion... ( c)e 2 e ce

We can use he preceeding rick o help find paricular soluions as well. Consider he following 2nd order linear equaion: If we know a soluion of he homogenous equaion, u(), we can subsiue as before o ge a new linear s order equaion in v (). Solving his equaion and hen inegraing again will allow he paricular soluion o be found. This migh be an easier procedure han variaion of parameers... e.g. y f( )y g( )y h( ) y ( 2 4)y 4 2 y 2e Noe ha a soluion o he homogenous equaion is simply: y ae 4 So we subsiue: y v( )e 4, y v e 4 4ve 4, y 6e 4 v 8e 4 v e 4 v To obain: v ( 2 4)v 2e This equaion is firs order, linear in v -- we ge: v ( ) 4 8e 4 ae