The Research In Value Forecasting Methods Of Body Panels Stamping Die Based On Neural Networks

The value of mould and die shows the technology, economy, efficiency and management of mould manufacturers, it is also an important factor that determining whether mould and die manufacturers could play a role in the keen market competition. With the fast development of the mould and die manufacturing industry, quotation relying on traditional method obviously can not meet the need of market's competition and variety. At the same time the development of CE and CSCW require enterprises to offer mould and die's forecasting cost and price at the different stages of mould's design and manufacture in time too. In order to overcome these practical situations, a more quick reflection and accurate mould and die cost calculation and quotation system should be introduced.In this paper, after analyzing the traditional cost forecasting and mould quotation methods, a forecasting method of stamping die body panels's material and machining cost based on BP neural networks was put forward. Based on the result of tests and manufacture experience a predication model of cost forecasting was built up. The feature of body panels stamping die and parts effect on the cost of die's material and machining, which were forecasted and simulated. The results showed that a three-layer BP neural networks model in one hidden layer could accurately forecast the value of cost. To make a detail analysis of the geometry characteristics and quotation method of body panels die, selecting and disposal of body panels die and parts's feature, foundation method and learning process and learning parameter's selection of BP neural networks. The trained BP neural networks was applied to forecast production cost of body panels die and to make a analysis of the reason of existed forecast error and how to improve forecast's precision. We can make a mould quotation based on it, and at the same time it provides a reliable support of the value data management of die.