European Transactions on Electrical Power · 1991 · 41 citations · 0 references
EngineeringComputer EngineeringComputer ScienceDynamical SystemsProcess ControlSystem DynamicsDiscrete Dynamical SystemsDynamic Systems ModelingState EstimationState-space ModelingState-space RepresentationTime IntegrationDiscrete ModelsDiscrete-time ModelsDiscrete-time State EstimationDiscrete-time State-space ModelsInduction Motor TheoryMotor DynamicsMotor ModelingTime Discretization
Abstract If a state‐space observer of a continuous‐time process is to be realized on a sampling microprocessor system, a discrete time model of the process is needed. Time‐varying continuous‐time parameters require the on‐line computation of the discrete‐time model. For this purpose numerical integration algorithms are used to get an approximate model instead of the exact model in discrete‐time. For the model of an induction motor it is pointed out that the often‐used first‐order algorithm of Euler (RK1) is crucial. Even the choice of the continuous‐time differential equation system, which has to be discretized, influences the quality of the discrete‐time approximation in a decisive manner. On the one hand even stability problems appear while on the other hand, by choosing the equation system in a skilful manner, one obtains a simple and accurate discrete‐time model, which can be used for microprocessor implementations.