Fuzzy Control of Mechanical Stress for Fatigue Test in Electrodynamic Shaker

The control of systems with traditional methodologies often faces problems of nonlinearity, difficulty in the mathematical definition of the problem and uncertainty of system parameters as well. In this context, controllers based on fuzzy logic shows up as an alternative to solve this type of problem since these methodologies deal with both linear and nonlinear systems. Besides they may control complex multivariable systems. The main objective of this paper is to develop and implement a control algorithm based on fuzzy logic in order to control the mechanical stress of a fatigue test in a metallic material using an electrodynamic shaker. In order to reach the proposed objectives, it is used an electrodynamic shaker, an amplifier, a signal conditioner and strain-gages attached to a test specimen to measure the deformation and therefore the mechanical stress. The system is developed in object-oriented language (Agilent VEE), responsible to acquire stress signals and to process and send the control signals to the shaker. It is developed a routine which allows the user to set up desired cycling frequency and the prescribed mechanical stress for the test. The Agilent VEE software was used jointly with Matlab software for the fuzzy logic implementation. Comparisons are made with traditional PID control at the end of the paper. Data concerning the controllers performance for changes in stress for a step type variation are presented and the corresponding parameters for the system are obtained. Tests are performed with disturbances on amplifier gain in order to synthetically simulate non-linearites of the system. The results show that the proposed metheodology is robust against such perturbations. Besides, the accuracy and response time of the implemented fuzzy control is compatible with the traditional PID control. Finally it is concluded that the fuzzy controller is easier to implement and has fewer parameters to be adjusted when compared with the PID one.