.In this thesis, the use of partially recurrent neural networks with the conjugate gradienttraining algorithm to model a particular hydraulic load sensing pump was investigated. Asimulation study was first conducted using "noise-free" data to examine the modeling errors inorder to provide a clear insight into the mechanism of the modeling error accumulation over thetransient state with a recurrent type of model structure. The established concepts and approachwere then applied to experimentally modeling a load sensing pump. An experimental system wasdesigned and constructed with particular attention paid to the design and generation of sufficientlyrich input signals, and to the selection of an appropriate sampling rate. The data obtained on thetesting of the load sensing pump dynamics are used in the training and testing of the neural models.The analysis and discussion showed that the training accuracy and the error accumulation were thetwo most critical factors in examining and interpreting the overall modeling accuracy.
--------------------------------------------------------------------------------
library2.usask.ca
