Design And Implement Of Control System For Crystal Grinding Machine Based On Neural Network

With people's living standards rising,the demand for crystal becomes bigger,and the modern crystal processing industry also faces new demands.Existing crystal grinding machine NC system has more or less some problems,these problems led to the efficiency of the machine can not be further improved,and the quality of crystal products exist some unstable situation.At the same time with the development of some existing technologies,the traditional crystal grinding machine numerical control system also needs further improvement.In this paper,the status of crystal processing technology at home and abroad and the future development trend of crystal processing technology is analyzed.The design idea of crystal grinding machine control system and the principle of crystal grinding are introduced in detail,which uses STM32F103ZET6 microprocessor as control core to control each function module.The hardware design and software design of control system are described in detail.Aiming at the instability of crystal during the grinding process,the neural network was used to establish the model to control the crystal grinding.This paper analyzes the force in the process of crystal grinding and discusses several factors that affect the quality of crystal products.In order to solve the problem that grinding discs are not stable because of abrasive wear,this paper propose to use neural network to optimize the grinding time to improve the grinding condition.Reducing the requirements of operators.Then introduces the development of BP neural network and the application of BP neural network,and discusses the improvement of BP algorithm.After the analysis of several key parameters in the grinding process,the BP neural network model was established.The torque,contact pressure and grinding times of the grinding machine were taken as the input of the neural network and the grinding time was taken as the output to establish the single hidden layer BP neural network.The network is trained after error elimination and normalization of the data.The trained model can fit the target value well.The various modules of NC system have a lot of bus connection,and the scene of the work environment is very complex,the traditional digital keyboard input is not very convenient.This paper uses the Wi-Fi network to set parameters and transfer data through the host computer.which increases the convenience of the system,and each modules of the equipment were tested.This paper uses the neural network to optimize the grinding process and control the grinding time of the crystal.The optimized system is stable in practical work.The results show that the system has achieved some results,laying the foundation of further improvement for the system.