Research On Crankshaft Rolling Alignment Based On Artificial Intelligent Expert System

Crankshaft is a important mechanical part of high machining quality. As a result of its complicated structure properties, crankshaft is far more complex than ordinary crankshaft when it produces bending distortion in making process.According to the distortion rule of crankshaft rolling, Contrary distortion of crankshaft can be produced to align itself by rolling some positions of crankshaft. Being satisfied with the requirement of fatigue intention, the whole intensity and reliability of crankshaft can improved by rolling and aligning, and the ability of crankshaft counteracting metabolic stress can be improved too. And the cost can also be reduced if rolling and aligning is used in the process of crankshaft. Because of the considerable economic value at market, the new technique is more and more prevalent in machining crankshaft of automobile engine.In the dissertation the rolling movement of crankshaft is analysed, the mathematic model of crankshaft in rolling and work mechanism in aligning crankshaft is discussed, and some theories for establishing crankshaft rolling and aligning expert system are provided. By using the neural network technique and the programme language of C++ Builder, a expert system aiming at supervising the rolling and aligning technics of crankshaft is founded in the dissertation. Neural network is provided with the ability of self-studying and self-organizing. Neural network can realize nonlinear mapping between system input and output. In the dissertation it is demonstrated that expert system has better approaching ability after inputting validated data, and has better spreading ability after ameliorating neural network. In the last, the neural network- hopfield is discussed and compared it with the neural network-BP in the dissertation.