The Simulation Of Stone Sawing Process

The circular sawblades with diamond impregnated segments have been extensively applied in the sawing of granites. At present, the mechanism of the cutting process of granite can't be well explained. Therefore, the cutting parameters are usually chosen according to practical experience.In this paper, the technique of simulation is used to solve this problem. The BP Artificial Neural Net (ANN) model of the sawing process is established to simulate the sawing power and sawing force, and the simulated result is validated by experiment. By optimizing the sawblades life, some optimal parameters of sawing of granites are provided.Chapter one is a preface about the background of the research project. Chapter two describes the basic knowledge of ANN and BP (Error Back Propagation), and how to use in MATLAB. Chapter three introduces the method of sawing experiment and how to sample and transform the data of sawing experiment. The sawing power and sawing force is simulated, and the BP model is compared with the least square method in chapter four. The wear resistance of the sawblades is simulated and optimized with three different object functions in chapter five.Through the simulation, the following conclusions can be drawn. Sawing power and sawing force could be simulated within 20% error by the BP model which was well-trained. By some tips, the BP model can simulate the continuous input and gain continuous results; moreover, some parameters beyond the boundary could be simulated, too. BP model could identify the relationship between sawing power and tangential force, the simulated curve is relatively stable, given the initial weight and threshold value at random, which proves a single hidden layer with the structure of 2-5-3 can preferably reflect the rules involved in the sawing process and filter the noise produced in the sawing process. Compared with the classical BP model, the learning algorithm of Levenberg-Marquardt trains more quickly and simulates more precisely. Another new conception to find optimal sawing parameters is presented. Through the simulation, the optimal sawing parameters to obtain the longest sawblade life can be determined under the conditions of fixed removal rate or depth of cut.