Industrial Restructuring And Layout Optimization Of Quipment Manufacturing Industrial In Hebei

Monocrystalline is a type of single crystal formed by silicon atoms regularlyarranged together, which is the basic material of microelectronics technology and alsoimportant for manufacturing solar panels. With the development of China’s science andtechnology, the demand for production and the diameter accuracy of monocrystallinesilicon increase rapidly. The domestic techniques of single crystal furnace diametermeasurement and pulling process are lagging behind, which can neither accuratelymeasure the true diameter nor precisely control pulling speed of the single crystal. So thatthe domestic monocrystalline crystal production with the features of poor accuracy,surface roughness, and low production efficiency.The edge detection image processing technology with Sobel level difference iswidely applied in diameter detection system. The edge gray-scale range detection of thismethod is not precise enough, but improving the measurement precision by hardwareresolution method is very expensive. In order to improve the detection accuracy of singlecrystal furnace system and reduce the hardware cost, this paper presents the concept of subpixel to achieve improved resolution by algorithm.In the pulling process, silicon single crystal diameter is an important parameter,whose variation is influenced by pulling velocity. For the motor system withcharacteristics of multivariable parameters and nonlinear, the traditional PID controltheory in Czochralski single crystal furnace for motor speed control system has significantlimitations. In order to solve the problem that the PID parameters cannot be adaptivelyadjusted by the traditional control theory with the motor system parameters changimg, anovel control method of PID neural network is applied to the crystal pulling speed controlsystem and the PID neural network system has been established in this paper. The optimalPID parameters can be adjusted at any time to achieve the adaptive control and theeffectiveness of this method has been demonstrated by using MATLAB software.This paper has presented that the PID neural network control hardwareimplementation process of single crystal furnace speed control system.The communication configuration of MATLAB and PLC is built by MCGS. TheMATLAB program is used to implement the neural network control algorithm. The PIDcontrol algorithm is completed by STEP7ladder control procedures, which enables the entire neural network PID control.