Position Control of Pneu1natic Actuator Using Artificial Neural Network

Abstract

The mam ISSUe of this work IS to design a proper position control system of a pneumatic system using artificial neural network. The mathematical dynamic model of the pneumatic system IS derived. This model indicates that the pneumatic system IS highly nonlinear system due to the compressibility of the pressurized air signals, nonlinear flow characteristics through variable area valve orifice and mechanical friction. Also, this model shows that pneumatic systems are of time varymg characteristics. Due to the nonlinearity, a conventional neural network based controller, CNNC, [14] is applied to control the position of the pneumatic system. This controller shows reasonable settling time and less overshooting, but its parameters need to be modified for every input conditions. Therefore a proposed neural network controller, PNNC, IS designed and implemented to overcome the disadvantage of the CNNC. The PNN controller IS a rule based ANNC where both the slope and amplitude of the actuation function of each neuron is modified to meet all values of the input range. Application of PNN controller yields to a considerable improvement of the system response for different input conditions without changing the initial values of its parameters. A comparison between the results of proposed rule

based neural network controller, PNNC, and a conventional PI

11

controller shows that the results of the proposed controtler is more cllicicnt and robust than that of the PI controller in terms of the system settling time and stability