PVC异型材性能参数的神经网络预测

Neural network forecast of function parameters of PVC profile

神经网络是当前主要的智能控制技术之一。依据厂家提供的大量PVC异型材的实际工艺参数作为样本,建立了PVC异型材性能参数的预测网络。采用BP算法对网络进行了结构参数调节及训练,并对性能参数热收缩率进行了预测,相对误差小于0.13,取得了良好的效果。

Nowadays NN(neural network) is one of the chief intelligence control technologies.Based on a number of real technological parameters provided by a factory, a network model of forecasting PVC function parameters is built and trained by adjusting the network construction parameters. At last, a function parameter called heat contract rate is forecasted accurately with relative errors less than 0.13.