Yang Datao1, Yang Weiqin2, Dang Xuliang2. Numerical Analysis of High Mn Aluminum Bronze in Centrifugal Casting Based on BP Artificial Neural Network[J]. Special Casting & Nonferrous Alloys, 2010,30(7):621-624.
Yang Datao1, Yang Weiqin2, Dang Xuliang2. Numerical Analysis of High Mn Aluminum Bronze in Centrifugal Casting Based on BP Artificial Neural Network[J]. Special Casting & Nonferrous Alloys, 2010,30(7):621-624.DOI:
基于BP人工神经网络模型的离心铸造高锰铝青铜数值分析
摘要
高锰铝青铜ZCuAl8Mn13Fe3Ni2
由于组成该合金的元素较多
且对力学性能的影响各不相同
通过数学建模方法对化学成分进行优选以得到优异的力学性能。基于在离心铸造条件下的大量实际生产数据
采用MATLAB构建BP人工神经网络模型
建立抗拉强度、屈服强度、伸长率、硬度分别与Al、Mn、Fe、Ni成分之间的关系
并通过回代、预测手段检测其有效性。再利用所建模型得到的500组预测值进行大范围筛选
经与真值筛选结果对比分析并结合实际数据
给出了合金成分的优选范围
并对元素间匹配关系进行了讨论。
Abstract
High Mn aluminum bronze ZCuAl18Mn13Fe3Ni2 is composed of multi-elements
in which the elements affect the mechanical properties of the bronze each other.It is important to optimize chemical composition to realize desirable mechanical properties of the bronze by mathematical modelization.Relationship among tensile strength
yield strength
elongation
hardness with Al
Mn
Fe
Ni elements was established by constructing BP artificial neural network model on the Matlab plateau based on lots of data from practical centrifugal casting
and availability of the model was verified by regression and prediction.500 sets predicted values in extensive scope were obtained by the model
which was comparatively analyzed with screening true value.Furthermore
combining practical data
the optimized scope of the alloy composition was presented.In addition
ingredients
impurity content
control of pouring temperature were approached.
关键词
高锰铝青铜离心铸造BP人工神经网络数值分析成分优选
Keywords
High Mn Aluminum BronzeCentrifugal CastingBP Artificial Neural NetworkNumerical AnalysisComposition Optimization