Tuning PI speed controllers for electric drives using simulated annealing

This paper introduces the application of a stochastic search technique, known as simulated annealing, to the problem of tuning electric drive speed controllers. Tuning by conventional analysis methods is complicated by the presence of a significant nonlinearity caused by saturation of the current controller. The nonlinearity ensures that optimum controller settings depend on the form of the input stimulus, for example the magnitude of a step change in speed demand. For the purposes of illustration, the paper concentrates on finding the settings for the parameters of a proportional-integral (PI) speed controller which minimise the integral of absolute speed error following a step change in speed demand. After describing the basic principles of simulated annealing, the paper presents results showing the development of optimum controller settings. Simulated annealing avoids the requirement for mathematical modelling of the drive and its load and is able to deal with the inherent nonlinearities of the system. The approach described is applicable to both off-line and on-line tuning of a wide variety of controller configurations.