Speed control of dc engines using labview-based experimental setup with artificial neural networks

DC (Direct Current) Engines have been used in widespread applications not only in industry but also in everyday life. Un-stationary situations observed at the loads complicate the control of DC engines. In this thesis an experimental setup is designed to realize the control of DC engines by different methods and to compare these methods as a first step. The experimental setup is able to be connected with personal computers (PC). It includes a control card and a data acquisition card that are designed within this study, besides. The data measured and collected are transferred to the PC, evaluated by controller structure and necessary control signals are transferred to the DC engine by the help of this experimental setup. Controller design and data evaluation operations are carried out using LabWIEV (grahical programming language) and Matlab software?s. At the second step of the study, An Artificial Neural Network (NN) based controller and a PID controller is designed for speed control of a DC engine. NN structure is formed not only at Matlab but also at LabVIEW environments. Measured and collected data are transferred to PC via experimental setup and presented to NN structure formed at Matlab environment.The ANN structure is trained and final values of neuron weights are obtained. Following, these weights are entered to the NN structure formed at LabVIEW environment. Finally, real time control is applied to DC engine. To compare the performance of NN controller a PID based controller is designed, besides. All results are compared, discussed and presented.