Application Of Neural Networks PID Algorithm In Rotation And Location Control For The Crucible

A new method of controlling the rotation and location of the single crystal furnace crucible is present in the dissertation, which bases on the technology of single crystal silicon growth, and designs the hardware circuit according and program to the require. It respectively controls the rotation and location of the single crystal furnace by a piece of single-chip micyoco 80C196KC. The drive motor of the single crystal furnace crucible is the permanent magnetism direct-current servo motor, and the drive of the motor adopts MOSFET which is as the master element in the control. There are many merits of the permanent magnet direct-current servo motor, such as the volume is small, the weight is light, the output power is big, the efficacy is high, the structure is simple, the initiating performance is good, operation is smooth, and it can control the speed in the extensive range. In the growth of the single crystal silicon, it adopts the signal of diameter and weight that are detected by the sensors of diameter and weight as the controlling data of diameter. The way can improve the accuracy of the diameter.Because the controlling object has the characters of nonlinearity, time-varying, time delay, the general PID controller is difficult to come to the exact controlling. It adopts the calculation of PWM space ratio by the combination of ANN and traditional PID algorithm, and then adjusting the output of the PWM of the single-chip computer, to realize the adjustment of armature voltage of the motor, and to control the rotation and location of the single crystal furnace crucible. The system can ensure the stability and robustness in the process of the crystal growth, and make the error of the diameter much smaller than the present standard +/-1 mm, it also can improve the quality of the crystal growth and the rate of finished products.