System identification model and predictive functional control of an electro-hydraulic actuator system

• Main Author: Mat Lazim, Noor Hanis Izzuddin
• Format: Thesis
• Language: English
• Published: 2015
• Subjects: TK Electrical engineering. Electronics Nuclear engineering
• Online Access: http://eprints.utm.my/id/eprint/54591/1/NoorHanisIzzuddinMFKE2015.pdf; http://eprints.utm.my/id/eprint/54591/; http://dms.library.utm.my:8080/vital/access/manager/Repository/vital:86029

The nonlinearities, uncertainties, and time varying characteristics of electrohydraulic actuator (EHA) have made the research challenging for precise and accurate control. In order to design a good and precise controller for the system, a model which can accurately represent the real system has to be obtained first. In this project, system identification (SI) approach was used to obtain the transfer function that can represent the EHA system. Parametric system identification method was utilized in this research as it emphasizes more on mathematical than graphical approach to obtain the model of the system. Multi-sine and continuous step signals were used as the input for the identification process. The models obtained were validated using statistical and graphical approach in simulation and experimental works to decide which model can represent the EHA system more precisely. Predictive functional control (PFC) was proposed and implemented for position control of the EHA. Besides, an optimal proportional-integral-derivative (PID) controller tuned by particle swarm optimization (PSO) was implemented in simulation and experimental work as comparison with the proposed controller. A comprehensive performance evaluation for the position control of the EHA is presented. As expected from the PFC main objective which is to realize closed-loop behaviour close to first order system with time delay, the experimental work conducted shows the controller capability to reduce the overshoot value by 87% as compared to the PID-PSO. The findings also demonstrated that the steady-state error was reduced by 37% and smaller integral absolute error (IAE).