基于人工神经网络的铝合金力学性能预测方法

Aluminum alloy mechanical properties prediction method based on artificial neural network

传统方法对铝合金力学性能测量精度低、误差率高、计算复杂。本文提出一种基于BP神经网络的铝合金力学性能预测方法研究。首先对铝合金拉伸试验获取的数据进行整理和分类,基于BP神经网络强大的映射和分析功能,对铝合金在多样的热冲击温度下的力学性能进行分析和研究。预测结果表明,BP神经网络算法具有较高的预测精度,误差率能够控制在5%以内。对比常温条件下铝合金力学性能,在高温短时热冲击的情况下,铝合金力学性能大幅度下降,提出的神经网络算法能够为改善铝合金力学性能提供数据上的支撑。

The traditional method for aluminum alloy mechanical properties measurement precision is low,the error rate is high,the complicated calculation.In this paper,a kind of aluminum alloy mechanical properties prediction method based on BP neural network research.First on the tensile test of aluminium alloy to obtain the data arrangement and classification,based on the BP neural network powerful mapping and analysis,in a variety of thermal shock temperature of aluminum alloy mechanical properties is analyzed and studied.Predicted results show that the BP neural network algorithm has higher prediction precision,error rate can be controlled within 5%.Contrast under the condition of normal temperature aluminum alloy mechanical properties in high temperature short time under the condition of thermal shock,aluminum alloy mechanical properties,a big drop in neural network algorithm to improve the aluminum alloy mechanical properties to provide data support.