塑料自然老化力学性能的人工神经网络预测

PREDICTION OF THE PLASTIC MECHANICAL PERFORMANCE IN THE AGING PROCESS BY THE ARTIFICIAL NEURAL NETWORK

针对塑料老化过程,利用人工神经网络方法建立自然老化力学性能时间序列的预测模型。该模型可很好地解决老化试验中数据少且统计规律不明显的问题。从给出的拉伸强度和断裂伸长率的预测例子和计算结果表明,该方法具有良好的精度,可满足工程实际对塑料老化性能预测精度的要求。同时可大量缩短试验时间,节约试验费用。

A model using the artificial neural network to predict the plastic mechanical performance in the aging process is presented. This method can satisfactorily solve the problem in which the test data are less and their statistic rule is not evident. The calculation results from the given tensile strength and elongation at break show that it can reach a satisfactory precision, satisfy the need in engineering practice and save much more time and cost in test.