DIPLOMA OF MECHANICAL ENGINEERING SESSION 02/2009 DMCC 3733: CONTROL ENGINEERING ASSIGNMENT 3 NAME: MATRIC NO: ISBOLLAH BIN HARUN D040!00!" MHD FUAD BIN ABDUL HALIM D040!0039 MOHAMAD FIKRI BIN MOHAMAD FAU#I D040!00!7 KALAI GANES A/L KARUNAKARAN D040!0003 SUBMITTED TO: EN #AIRULA#HAR BIN SUBMISSION DATE: 27/03/2009 MATHEMATICAL MODELING OF DYNAMICS SYSTEM M$%&'($%)*$+ M,-'+ A mathematical model uses mathematical language to describe a system. Mathematical models are used not only in the natural sciences and engineering disciplines (such as physics, biology, earth science, meteorology, and electrical engineering) but also in the social sciences (such as economics, psychology, sociology and political science); physicists, engineers, computer scientists, and economists use mathematical models most extensively. Eykhoff (!"#) defined a mathematical model as $a representation of the essential aspects of an existing system (or a system to be constructed) %hich presents kno%ledge of that system in usable form$.&' Mathematical models can take many forms, including but not limited to dynamical systems, statistical models, differential e(uations, or game theoretic models. )hese and other types of models can overlap, %ith a given model involving a variety of abstract structures. *opulation +ro%th. A simple (though approximate) model of population gro%th is the Malthusian gro%th model. )he preferred population gro%th model is the logistic function. Model of a particle in a potential,field. -n this model %e consider a particle as being a point of mass m %hich describes a tra.ectory %hich is modeled by a function x / 0 1 02 given its coordinates in space as a function of time. )he potential field is given by a function 3/02 1 0 and the tra.ectory is a solution of the differential e(uation 4ote this model assumes the particle is a point mass, %hich is certainly kno%n to be false in many cases %e use this model, for example, as a model of planetary motion. 5 Model of rational behavior for a consumer. -n this model %e assume a consumer faces a choice of n commodities labeled ,6,...,n each %ith a market price p, p6,..., pn. )he consumer is assumed to have a cardinal utility function 7 (cardinal in the sense that it assigns numerical values to utilities), depending on the amounts of commodities x, x6,..., xn consumed. )he model further assumes that the consumer has a budget M %hich she uses to purchase a vector x, x6,..., xn in such a %ay as to maximi8e 7(x, x6,..., xn). )he problem of rational behavior in this model then becomes an optimi8ation problem, that is/ 9r his model has been used in general e(uilibrium theory, particularly to sho% existence and *areto optimality of economic e(uilibria. :o%ever, the fact that this particular formulation assigns numerical values to levels of satisfaction is the source of criticism (and even ridicule). :o%ever, it is not an essential ingredient of the theory and again this is an ideali8ation C+$..)/0)12 ($%&'($%)*$+ (,-'+. Many mathematical models can be classified in some of the follo%ing %ays/ . ;inear vs. nonlinear/ Mathematical models are usually composed by variables, %hich are abstractions of (uantities of interest in the described systems, and operators that act on these variables, %hich can be algebraic operators, functions, differential operators, etc. -f all the operators in a mathematical model present linearity, the resulting mathematical model is defined as linear. A model is considered to be nonlinear other%ise. 6. <eterministic vs. probabilistic (stochastic)/ A deterministic model is one in %hich every set of variable states is uni(uely determined by parameters in the model and by sets of previous states of these variables. )herefore, deterministic models perform the same %ay for a given set of initial conditions. =onversely, in a stochastic model, randomness is present, and variable states are not described by uni(ue values, but rather by probability distributions. 2. >tatic vs. dynamic/ A static model does not account for the element of time, %hile a dynamic model does. <ynamic models typically are represented %ith difference e(uations or differential e(uations. #. ;umped vs. distributed parameters/ -f the model is homogeneous (consistent state throughout the entire system) the parameters parameters are distributed. -f the model is heterogeneous (varying state %ithin the system), then the parameters are lumped. <istributed parameters are typically represented %ith partial differential e(uations C,13,+4%),1 I1%'25$+ A convolution is an integral that expresses the amount of overlap of one function as it is shifted over another function . -t therefore ?blends? one function %ith another. @or example, in synthesis imaging, the measured dirty map is a convolution of the ?true? =;EA4 map %ith the dirty beam (the @ourier transform of the sampling distribution). )he convolution is sometimes also kno%n by its +erman name, faltung (?folding?). A )he time domain output of an ;)- system is e(ual to the convolution of the impulse response of the system %ith the input signal A Much simpler relationship bet%een fre(uency domain input and output A @irst look at graphical interpretation of convolution integral =onvolution is implemented in Mathematical as =onvolve&f, g, x, y' and <iscrete =onvolve &f, g, n, m'. Abstractly, a convolution is defined as a product of functions and that are ob.ects in the algebra of >ch%art8 functions in . =onvolution of t%o functions and over a finite range is given by () %here the symbol denotes convolution of and . =onvolution is more often taken over an infinite range, (6) =onvolution -ntegral as sum of impulse responses @or continuity %ith the page deriving the convolution integral %e can approximate the input by a series of impulses...
plot the response of the system to each of these impulses... and plot the response as the sum of the individual responses (also sho%n is the exact solution). L)1'$5 S0.%'( A linear system is a system %here input or output relationships may be represented by a differential e(uation. As a mathematical abstraction or ideali8ation, linear systems find important applications in automatic control theory, signal processing, and telecommunications. @or example, the propagation medium for %ireless communication systems can often be modeled by linear systems. A general deterministic system can be described by operator, H, that maps an input, x(t), as a function of t to an output, y(t), a type of black box description. ;inear systems satisfy the properties of superposition and scaling. +iven t%o valid inputs B(t) B6(t) as %ell as their respective outputs y (t) C : DB(t)E y6 (t) C : DB6(t)E then a linear system must satisfy (t) F y6 (t) C : D (t) F 6 (t)E for any scalar values and . L)1'$5 T)(' I13$5)$1% S0.%'( A time invariant system is a system described by a differential e(uation %ith constant coefficient. )hus, the plant is time invariant if the parameters do not change as a function of time. A time,invariant system is one %hose output does not depend explicitly on time. )hat is, treating time as the independent variable, it is an autonomous system. -f the input signal x produces an output y then any time shifted input, t, x (t F ) results in a time,shifted output t, y (t F ) @ormally, if S is the shifting operator (SGx (t) C x (t H G)), then the operator T is called time, invariant, if T (SGx) C SG (Tx). )his property can be satisfied if the transfer function of the system is not a function of time except expressed by the input and output. )his property can also be stated in another %ay in terms of a schematic. -f a system is time,invariant then the system block is commutative %ith an arbitrary delay S)(6+' '7$(6+' )o demonstrate ho% to determine if a system is time,invariant then consider the t%o systems/ >ystem A/ y(t) C t . x(t) >ystem I/ y(t) C J . x(t) >ince system A explicitly depends on t outside of x (t) and y (t) it is time,variant. >ystem I, ho%ever, does not depend explicitly on t, so it is time,invariant. F,5($+ '7$(6+' A more formal proof of the previous example is no% presented. @or this proof, the second definition %ill be used. >ystem A/ >tart %ith a delay of the input xd(t) C x( t F ) K(t) C txd(t) C tx( t F ) 4o% delay the output by G K(t) C tx(t) K6(t) C y(t F ) C (t F )x(t F ) =learly y(t) y6(t), therefore the system is not time,invariant. >ystem I/ >tart %ith a delay of the input xd(t) C Jx(t F ) =learly y(t) y6(t), therefore the system is time,invariant. Although there are many other proofs, this is the easiest. L)1'$5 T)(' V$50)12 S0.%'( )ime,varying distributed systems have not been examined in any detail in the scattering simulation literature, though time,varying L<@s and <L4s have both been proposed, %ith a focus on vocal tract modeling. )hough it is true that time,variations in material parameters generally render a system non,passive, %e %ill sho% here ho% passive net%ork representations may be developed for an important class of systems. =onsider a system of the form (*(x, t) %) F F I(x, t) % F f(x, t) C J )ime,varying impulse response )he time,varying impulse response h(t6,t) of a linear system is defined as the response of the system at time t C t6 to a single impulse applied at time t C t. -n other %ords, if the input x(t) to a linear system is x(t) C (t M t) %here G(t) represents the <irac delta function, and the corresponding response y(t) of the system is y(t) t C t6 C h(t6, t) then the function h(t6,t) is the time,varying impulse response of the system. L)1'$5)8$%),1 ,1 1,19+)1'$5 ($%&'($%)*$+ (,-'+ -n mathematics, a nonlinear system is a system %hich is not linear, that is, a system %hich does not satisfy the superposition principle, or %hose output is not proportional to its input. ;ess technically, a nonlinear system is any problem %here the variable(s) to be solved for cannot be %ritten as a linear combination of independent components. A nonhomogeneous system, %hich is linear apart from the presence of a function of the independent variables, is nonlinear according to a strict definition, but such systems are usually studied alongside linear systems because they can be transformed to a linear system of multiple variables. 4onlinear problems are of interest to physicists and mathematicians because most physical systems are inherently nonlinear in nature. 4onlinear e(uations are difficult to solve and give rise to interesting phenomena such as chaos. )he %eather is famously nonlinear, %here simple changes in one part of the system produce complex effects throughout -n control engineering a normal operation of the system may be around an e(uilibrium point and the signal may be considered small signal around the e(uilibrium. :o%ever, if the system operates an e(uilibrium point and if the signals involved are small signals, then it is possible approximate the nonlinear system by a linear system. >uch a linear system e(uivalent to the nonlinear system considered %ithin a limited operating range. >uch a lineari8ed model or also kno%n as linear, time,invariant model is very important in control engineering. L)1'$5 A665,7)($%),1 ,/ N,1+)1'$5 M$%&'($%)*$+ M,-'+: )o obtain a linear mathematical model for a nonlinear system, %e assume that the variables deviate only slightly from some operating condition. =onsider a system %hose input is ( t ) and output is y( t ). )he relationship bet%een ( t ) and y( t ) is given by/ y = f ( x ) (e(uation ) -f the normal operating condition corresponds to x, y, then E(uation maybe expanded into a )aylor series about this point as follo%s/ y = f ( x ) ; f(x)+ (x-x)+ (x-x)N +(e(uation 6) Lhere df O dx, df / dxare evaluate x = x. -f the variation x - x is small, %e may neglect the higher,order terms in x x. )hen e(uation 6 may be %ritten as/ y= y +P(x x) (e(uation 2) Lhere, y C f ( x ) P C x x E(uation 2 may be %ritten as/ y y = P ( x x ) Q.. (E(uation #) %hich indicates that y y is proportional to x x. E(uation give a linear mathematical model for the non linear system give by e(uation near the operating point x x, y y. 4ext, consider that a nonlinear system %hose output y is a function of t%o output x and x6, so that/ y = f ( x, x6 ) (e(uation R) )o obtain a linear approximation to this nonlinear system, %e may expand E(uation R into )aylor series about the normal operating point x, x6. )hen E(uation R becomes/ y = f (x, x6 ) F& ( x1 x1) F ( x2 x2 ) ' F & ( x1 x1 )N F 6 ( x1 x1 ) ( x2 x2 ) F ( x1 x2 )N FQQQ Lhere the partial derivative are evaluate at x1 = x1, x2 = x2. 4ear the normal operating point, the higher order terms may be neglected. )he linear mathematical model of this nonlinear system in the neighborhood of the normal operating condition is then given by y y =P ( x1 x1 ) F P6 ( x2 x2 ) %here, y = f ( x1, x2 ) P C x1 = x1, x2 = x2. P6 C x1 = x1, x2 = x2. )he lineari8ation techni(ue presented here is valid in the vicinity of the operating condition. -f the operating condition vary %idely, ho%ever, such lineari8ed e(uations are not ade(uate, nonlinear e(uation must be dealt %ith. -t is important to remember that a particular mathematical model used in analysis and design may accurately represent the dynamic of and actual system for certain operating condition, but may not be accurate for the operating condition. TRANSFER FUNCTION -n control theory, functions called transfer function are commonly used to characteri8e the input, output relationships of components or systems that can be described by linear, time,invariant, differential e(uations. )he transfer function of a linear, time,invariant, differential e(uation system is defined as the ratio of the ;aplace transform of the output (response function) to the ;aplace transform of the input (driving function) under the assumption that all initial conditions are 8ero. )ransfer function C +(s) C Syste m -nput 9utput ()ransfer @unction) 9utput C Seroes -nput C *oles 0(>) C C K(>) Example; F 6y(t) C r(t) ;aplace )ransform >K(s) M y(J) F 6K(s) C 0(s) K(s) &(s F6)' C 0(s) C C A )ransfer @unction is the ratio of the output of a system to the input of a system, in the ;aplace domain considering its initial conditions to be 8ero. -f %e have an input function of B(s), and an output function K(s), %e define the transfer function :(s) to be/ &)ransfer @unction' 0eaders %ho have read the =ircuit )heory book %ill recogni8e the transfer function as being the ;aplace transforms of a circuit$s impulse response. Iy using the concept transfer function, it is possible to represent system dynamics by algebraic e(uations in s. -f the highest po%er of s in the denominator of the transfer function is e(ual to n, the system is called an nth-order system. )he applicability of concept of the transfer function is limited to linear, time,invariant, differential e(uation system. )he transfer function approach ho%ever is extensively used in the analysis and design of such system. -n %hat follo%, it shall list important comments concerning the transfer function. )he transfer function of a system is a mathematical model in that it is an operational method of expressing the differential e(uation that relates the output variable to the input variable. )he transfer function is a property of a system itself, independent of the magnitude and nature of the input or driving function. )he transfer function includes the units necessary to relate the input to the output, ho%ever it does not provide any information concerning the physical structure of the system. -f the transfer function of a system is kno%n, the output and response can studied for various form of inputs %ith a vie% to%ard understanding the nature of the system. -f the transfer function of a system is unkno%n, it may be established experimentally by introducing kno%n inputs and studying the output of the system. 9nce established, a transfer function gives a full description of the dynamic characteristics of the system, as distinct from its physical description. P,+'. $1- #'5,. *oints in the s plane at %hich the function +(s) is analytic are called ordinary oints, %hile points in the s plane at %hich the function +(s) is no analytic are called sin!"#ar oints. >ingular points at %hich the function +(s) or its derivatives approach infinity are called o#es. >ingular points at %hich the function +(s) e(uals 8ero are called $eros. -f +(s) approaches infinity as s approaches and if the function has a finite, non8ero value at s = -, then s C ,p is called pole of order a. -f a= , the pole is called a simple pole, and so on. +(s) (s F p)T, for a C ,6,2,Q. )o illustrate, consider the complex function +(s) has 8eros at s C ,6, s C ,J, simple poles at s C J, s C ,, s C ,R, and a double pole at s C ,R. 4ote that +(s) became 8ero at s C U. +(s) C >ince for large values of s, +(s) possesses a triple 8ero at s C U. -f points at infinity are included, +(s) has the same number of poles as 8eros. )o summari8e, +(s) has five 8eros (s C ,6, s C ,J, s C U, s C U, s CU) and five poles ( s C J, s C ,, s C ,R, s C ,R, s C ,R). +(s) C I(64+.' R'.6,1.' F41*%),1 =onsider the output (response) of a system to a unit,impulse input %hen the initial conditions are 8ero. >ince the ;aplace transform of the unit,impulse function is unity, the ;aplace transform of the output of the system is K(s) C +(s) C )he inverse ;aplace )ransform of the output gives the impulse response of the system. )he inverse ;aplace transforms of +(s), or is called the impulse,response function. )his function g(t) is also called the %eighting function of the system. -nverse ;aplace &+(s)' C g(t) )he impulse response functions g(t) is thus the response of a linear system to a unit impulse input %hen the initial conditions are 8ero. )he ;aplace transform of this function gives the transfer function. )herefore, the transfer function and impulse response function of a linear, time, invariant system contain the same information about the system dynamics. -t is hence possible to obtain complete information about the dynamic characteristics of the system by exciting it %ith an impulse input and measuring the response. @rom a ;aplace transform table, %e kno% that the ;aplace transform of the impulse function, G(t) is/ V&G(t)' C >o, %hen %e plug this result into our relationship bet%een the input, output, and transfer function, %e get/ K(s) C 0(s):(s) K(s) C ():(s) K(s) C :(s) -n other %ords, the ?impulse response? is the output of the system %hen %e input an impulse function. S%'6 R'.6,1.' F41*%),1 >imilarly to the impulse response, the step response of a system is the output of the system %hen a unit step function is used as the input. )he step response is a common analysis tool used to determine certain metrics about a system. )ypically, %hen a ne% system is designed, the step response of the system is the first characteristic of the system to be analy8ed. @rom the ;aplace )ransform table, %e can also see that the transform of the unit step function, u(t) is given by/ V&u(t)' C *lugging that result into our relation for the transfer function gives us/ K(s) C 0(s):(s) K(s) C ( ):(s) K(s) C And %e can see that the step response is simply the impulse response divided by s. r(t) (*osition) t Example/ F 6y(t) C r(t) ;aplace )ransform >K(s) M y(J) F 6K(s) C 0(s) K(s) &(s F6)' C 0(s) C C 0(s) C K(s) C & ' x ( ) K(s) C R$(6 R'.6,1.' F41*%),1 r(t) (3elocity) ) Example / F 6y(t) C r(t) ;aplace )ransform >K(s) M y(J) F 6K(s) C 0(s) K(s) &(s F6)' C 0(s) C C 0(s) C K(s) C & ' x ( ) K(s) C P$5$<,+)* R'.6,1.' F41*%),1 r(t) (Acceleration) ) Example / F 6y(t) C r(t) ;aplace )ransform >K(s) M y(J) F 6K(s) C 0(s) K(s) &(s F6)' C 0(s) C C 0(s) C K(s) C & ' x ( ) K(s) C