A Neural Network Based Fuzzy Controller For Pneumatic Circuit

Abstract

Pneumatic circuits are widely used in industrial automation, such as drilling,sawing, squeezing, gripping, and spraying. Furthermore, they are used in motioncontrol of materials and parts handling, packing machines, machine tools, foodprocessingindustry and in robotics.In this paper, a Neural Network based Fuzzy PI controller is designed andsimulated to increase the position accuracy in a pneumatic servo circuit where thepneumatic circuit consists of a proportional directional control valve connected with apneumatic rodless cylinder. In this design, a well-trained Neural Network with asimplest structure provides the Fuzzy PI controller with suitable input gains dependingon feedback representing changes in position error and changes in external load force.These gains should keep the positional response within minimum overshoot,minimum steady state error and compensate the effect of applying external load force.A comparison between this type of controller with a conventional PID type shows thatthe PID controller failed to keep the cylinder position with minimum steady state errorand failed to compensate the effect of applying external load force as compared withthe results when using a Neural Network based Fuzzy PI type controller. This isbecause of nonlinearities that exist in the pneumatic circuit. Thus, the positionresponse using Neural Network based Fuzzy PI controller is better with an average ofimprovement in position accuracy of (11 %).