Laboratory testing of the method of parameter identification of dynamic object by simulation

E.V. Bolovin, A.S. Glazyrin.

Department of Electric Drives and Equipment, National Research Tomsk Polytechnic University, Tomsk, Russia

Abstract

In this paper a laboratory testing of the method of parameter identification of dynamic object is presented. The error of identification of the parameters is analyzed using the developed method.

Key words: first-order dynamic object, identification of the parameters, difference schemes.

Introduction

Testing consisted of two phases. The first step - getting the transients based on numerical solutions of differential equations by Euler's method. The second step - the identification of parameters based on the current values of state variables of the object. The third step - the calculation of error estimates of the parameters.

The test procedure

Thus the mathematical model of the process parameter identification RL-circuit in a matrix form is derived, giving the possibility of obtaining solutions by the inverse matrix. MathCad to model the processes occurring in the object of the first order - RL-circuit. Then test the presented method of parameter identification.

Written in the MathCad parameters of the circuit. All entries made in the program MathCad performed with the font Courier New. All used units owned by SI system.

The resistance
The inductance
The angular frequency
The step time

The index of the iteration process is defined

Will submit to the input sine wave impact

The function of the right side of equation is written (1)

The zero initial conditions is defined

Calculate the current and current derivative during the numerical solution of differential equations by Euler's method

The index of the iteration process is redefined

This redefinition is necessary for the performance of the inverse of the matrix, otherwise the determinant of the output values equal to zero.

The values of resistance rr_j and inductance ll_j are found by (8)

The average values of estimates of active resistance RR and inductance LL are defined

The error of estimates of mean values of the parameters RL circuit are defined

Conclusion

The difference between actual and estimated values are less than 1 percent, respectively, the identification of parameters by the inverse of the matrix is effective.