Application of a digital adaptive controller to a hydraulic system

Deals with the application of model reference adaptive control based on Lyapunov's stability theory to the control of a hydraulic drive. The aim is to move the piston fast and without overshoot to commanded positions and to keep its position constant against different disturbances due to changes in load forces. The proposed strategy guarantees a stable operation, also if the discrete transfer function of the plant contains zeros outside the unit circle. The paper is organized as follows. Starting from the description of the basic linear control loop, the adaptive version of the algorithm is introduced, where minor changes to the historical Lyapunov approach are pointed out, especially with respect to computation of the adaptive error. Two estimation strategies, namely stochastic approximation, as a fast and simple to implement method, and a recursive least squares square-root-filter approach with variable exponential weighting, which results in fast convergence and numerical stability, are given in connection with the adaptive control law. The next section introduces the plant under consideration and finishes by giving an approximate linear model for the hydraulic drive. Following that, implementation aspects of the adaptive control scheme are discussed. The last section represents results of real-time experiments and discusses the performance of the closed loop system for fixed and adaptive control, for load changes and initial adaptation with different estimation methods. >