Abstract
Neural network predicting model for axial piston pump is prepared. Experimental set up was prepared, the model uses results obtained from an experimental work. The neural network model has feed forward structure based on Levenberg Marquardt optimization technique for training process. Also governing equations of an axial piston pump and swash plate torque equations are derived. Two Training algorithms are adopted, one for experimental data and the second for theoretical results. The ability of the trained network to reproduce and to predict is demonstrated by its excellent approximation of experimental and theoretical data. Analysis of flow rate, discharge pressure, leakage, pump efficiency and swash plate torque in a variety conditions is investigated using proposed neural network model. The model was able to predict the behavior of the pump and leads to stable solution.
Key wards
Axial piston pump; Swash plate; Neural network; Levenberg Marquardt algorithm.
|
