STUDY ON MODEL PREDICTIVE CONTROL TO MINIMIZE TEMPERATURE CHANGE OF VERTICAL PLATE WITH VARYING HEAT GENERATION

摘要

Precise process temperature control of 0.001°C under circumstances of noise-temperature change of 0.1 °C is required in semiconductor manufacturing process. We studied optimum control method to minimize temperature change at an object position in a 2-dimensional vertical plate with a varying noise-heat-generation and a control-heater. We numerically calculated 2-dimensional unsteady thermal conduction in the plate with feedback control, feedforward control, and model predictive control of the control-heat-generation. The temperature change at the object position can be decreased 1/80 times smaller than that without control-heat-generation using the feedback control with two monitoring temperatures. The temperature change at the object position can be decreased 1/1000 times (0.002°C) using the model predictive control of 5 s interval with step response pattern as a dynamic predictive model. We found that the accuracy of the dynamic predictive model is very important for precise temperature control. Experiment was performed for the model predictive control with a network model as the dynamic predictive model, and the experimental result agreed with the calculation result.