Neural Network And Simulation In Machining Applications

This article has recommended some basic theories of high-speed cutting and cutting force at first, it is that the work obstrution of material resists the cutter to cutting force. It is important factor which influences systematic intensity of the craft, rigidity , and the quality of processed workpiece .It is main basis of designing of lathe , cutter Jig and calculation of power consume. At present during the course of automatic produce and accurate processing , it is used to examine processing surface quality , machining accuracy and degree of cutter wearing. The artificial neural network is a kind of simulation of the structure of the human nervous system, it announces non-line relations contained in the data sample, A large amount of dealing unit make up the non-linear self-adaptation dynamic system, It has good adaptivity , organizes by oneself and very strong study , association , fault-tolerant and anti-interference ability, It can imitate information processing mechanism of brain at various level and in various degree and carries on modeling to the multifactor complicated unknown coefficient system very flexibly and conveniently . Especially BP network applies field of the pattern- recognition and prediction-assessing etc extensively in recent years and make the good result. BP network adopt error against propagate algorithm to leam and train the neural network at present, this algorithm is on the basis of the error function gradient of the network. To overcome slow convergence speed and putting some little deficiency extremely of BP network in the course of studying and training, this paper has used theLevenberg-Marquardt back-propagation algorithm. This article utilize neural network theory mainly and set up cutting force BP neutal network prediction system on the basis on MTLAB neural network toolbox.and It confirmed connection between the exporting amount the inputting amount. And examines the result of the model through the cutting force experimental data. The experimental result has proved feasibility of this model emulation for cutting force, more accurate than the experience formula,It has very good using value.