AbstractA spanish research group has developed a new laser material surface heat treatment process monitoring and control system, in which a hybrid control system is used as basis. This hybrid control system combines a proportional-differential fuzzy controller with an uncoupled integral term of the error, which furthermore incorporates an adaptive neuro-fuzzy inference system, ANFIS, (Adaptive NeuroFuzzy Inference System), which can estimate the process state in real time. Licensing of the patent.
DetailsIt has been developed a laser material surface heat treatment process real time monitoring and control system by a hybrid fuzzy controller, which uses the maximum temperature on the material surface as input variable and the laser source output power as the actuation variable.
Furthermore the process carrying out is supervised in real time, verifying the consistency between the temperature measured by the infrared pyrometer used as sensor and the estimated temperature.
The system calculates in a cyclic way the signal to actuate the laser source power based on the monitoring error signal. The process supervision step verifies the consistency between the maximum temperature signal on the material surface and the temperature value estimated by a neuro-fuzzy model of the process, implemented by a neuro-fuzzy inference system, of ANFIS type, making it possible to detect in real time failures in the monitoring and control system operation and to take corrective actions.
The novelty is based on the control procedure itself, both in the way the problem is tackled, using an hybrid control system, combining a fuzzy PD controller with an integral term of the error, both uncoupled, in such a way that the fuzzy controller compensates for process non-linearities, whereas the integral term corrects the steady state error, as well as in the way it is implemented, using standard equipments in addition to a neuro-fuzzy inference model ANFIS in order to validate and contrast data coming from the sensors.