چكيده به لاتين
In most structural components in mechanical engineering, the surface is particularly exposed to stress. Therefore, hardening of these surfaces is doubly important in industrial applications. Since in many applications only the hardening of part of the surface is required, laser hardening is a very good choice for this. Since the quality of hardening is strongly related to the surface temperature, so the most important and in fact the most challenging control parameter in this process is the surface temperature control in the presence of disturbances such as changing the laser scanning speed or changing the thickness or shape of the work piece.Another important point to consider in the design of the controller is that due to the high sensitivity of the laser diodes, the control signal and its changes can not have any desired value, in other words, they must be bound. Also, regarding the temperature measured from the surface, care must be taken that the temperature does not exceed a certain value. A value higher than the allowable limit will melt the workpiece and lower than the allowable limit will disrupt the hardening performance. Therefore, the system output must also be constrained. As will be shown below, open-loop control will not work by applying a constant current for this purpose. So we have to go to the closed loop design.
By reviewing the articles and works done so far by researchers and engineers in this field, it was found that the main effort of researchers in controlling surface temperature by changing parameters such as laser power or laser scan speed and a very important and vital factor is compliance with restrictions. The issue mentioned earlier has not received much attention. Also, the results of the controllers designed in these articles are mainly simulations and have not been implemented in practice. Considering the constraints governing the system and the considerations raised, the proposed method of this dissertation is to use the controller of the linear prediction model.Also, for better comparison, in parallel with the design of this controller, the famous PID controller was designed with two methods. The results of simulation and practical implementation showed that the controller of the constrained predictor model has a much better and more acceptable performance in comparison with the PID controller in the two areas of reference signal tracking and also the elimination of perturbation.
