Friction force estimation in pneumatic cylinders by full automation of experimental procedures

Abstract

The effective utilization of pneumatic cylinders in precise control applications requires the correct determination of unpredictable, variable, and non-linear friction resistance characteristics. Even though the friction characteristics depend on many structural and operating variables, the existing expensive experimental methodologies are prone to error, time-consuming, and have excessive computational burden. In this study, the full automation of friction force estimation processes has been designed and implemented with MATLAB/Stateflow including the execution of all experimental steps, processing of signals, extraction and classification of data, and estimation of parameters. An operational graphical user interface has been developed for entering the required input values of the experimental procedures, connecting to testing equipment, carrying out experiments, and presenting obtained model parameters. The friction force parameters of three pneumatic cylinders, two of which were identical, have been determined both with the automated method and the manual method. The automated method has produced better results compared to the manual method while reducing the total estimation duration drastically. The pressure tracking performance tests conducted to verify the friction model parameters have exhibited better tracking performances with the automated method results.