Steady state determination using bond graphs for systems with singular state matrix

Abstract

A bond graph procedure to get the steady state value for linear time-invariant systems is presented. The general case of a singular state matrix is considered. The procedure is based on a junction structure configuration with derivative causality assignment, and on relationships of the bond graphs with integral and derivative causality assignments. It is shown that the structurally null modes, i.e. the poles at the origin, are cancelled for steady state. The key to cancel the poles at the origin is that the adjugate matrix of sIn − Ap multiplies Bp yielding the zeros at the origin with the same order that the structurally null modes, where ( Ap, Bp, Cp, Dp) is a state space realization of a linear time-invariant system, s is the Laplace operator and In, is an n × n identity matrix. Hence, this unstable part of the system is cancelled and the steady state can be obtained. Thus, the singularity of the state space matrix is avoided, and the steady state is obtained from the bond graph with derivative causality assignment. Since the singular state matrix is considered, it is shown that by using the bond graph with derivative causality assignment, an equivalent system with linearly independent state variables can be obtained. An example of an electrical system with an electrical transformer modelled by an I-field whose state matrix is singular is presented. Also, the proposed methodology for a load driven by two DC motors is applied.