| Electrical discharge machining (EDM) has extensive application in lots industries, such as automobile, spaceflight, micro-electronics, etc. The appearance of Powder mixed fluid EDM (PEDM) extremely enriches the connotation of EDM technology, meanwhile breaks the restriction of people's imagination to this technologies. But, the basic mechanism of PEDM is unclear, dimensional precision is very difficult to get. So that, To analyze the model of discharging, to find the inbeing of PEDM and to promote technology of PEDM practical become Focus of EDM research.This paper, through a series of PEDM experiments, summarizes the relation of some technologic parameters, finds the rule about discharging position. Based on those conclusions, the paper utilizes MATLAB6.5 software to set up the artificial neural network model, verification experiments is carried out to check the validity of the developed models and the method. Those works has been done as follows:(1) Based on the Experiments, it is concluded that in a pulse cycle, the randomness of the discharging position is confining to certain regional range. As time goes on, the area of discharge transfers from a location to another slowly.(2) The samples, according to Orthogonal Design principle, fully represent the systematic characteristic of PEDM. The powder disperses discharge energy, increases the randomness of discharging, all those make the result more accurate.(3) Based BP neural topological structure of network of "tower type" brought forward in this paper, the model, representing the quantitative relationship between the technological parameter and quality targets > was established after trained by samples formed by orthogonal design. The results show that the model reflects the complex relationship and has high predicting accuracy.The conclusions are helpful to guide application and will be the experimental base for expert system of PEDM, and the work also accelerates research on PEDM mechanism.
|
PDF Full Text Request