Identification of a pneumatic actuator using non-linear black-box model

A pneumatic actuator is a device that is capable of converting energy from a pressurized gas into motion. While motion can be created through other means, such as a hydraulic or electric motor, pneumatic actuators are safer, cheaper, and cleaner. Therefore, pneumatic actuators have been used widely in the field of industry automation. However, the compressibility of air and the inherent non-linearity of pneumatic actuators cause challenges in controlling accurately position of pneumatic actuators. This paper presents an accurate non-linear back-box model (NBBM) for identifying the dynamic behavior of pneumatic actuators. Once the optimized NBBM of the pneumatic actuator is obtained, it can give a generation of an effective solution for designing a position controller of that. Here, the NBBM is a multi-player perceptron neural network (MLPNN), whose parameters are optimized by using the Lervenberg-Marquardt Back Propagation (LMBP) algorithm. For the model verification, a pneumatic actuator was set up to investigate the dynamics of it as well as to generate the training data. Next, the advanced NBBM for the pneumatic actuator is performed with suitable inputs to estimate the cylinder piston displacement. Finally, the NBBM ability is evaluated by a comparison of the estimated and real pneumatic actuator performance.