模糊辨识法预测马氏体时效不锈钢的力学性能

Prediction to Mechanical Properties of Maraging Stainless Steel by Fuzzy Identification Method

应用模糊辨识方法对 (% ) :0 0 1～ 0 0 5C ,0～ 0 89Si,0～ 1 2 5Mn ,10 2～ 16 0Cr,0～ 9 0Ni,0～1 8Cu ,0～ 0 8Al,0～ 0 9Ti马氏体时效不锈钢的力学性能进行预测。该方法是基于T S模糊模型 ,采用三角形隶属函数计算给定样本的隶属度 ,利用稳态卡尔曼滤波器辨识模糊模型的结论参数。根据收集整理国内外文献的实验数据 ,使用模糊模型建立化学成分和热处理主要工艺参数与产品力学性能之间的关系。离线仿真表明 ,产品力学性能的预测值与实际值拟合良好 ,抗拉强度 (10 0 0～ 2 0 0 0MPa)相对误差 91%在± 2 %以内 ,断面收缩率 (4 5 %～ 77% )的相对误差 85 %在± 2 %以内 ,延伸率 (6 %～ 2 2 % )的相对误差 84 %在± 4 %以内。

Prediction to mechanical properties of maraging stainless steel - 0.01～0.05C, 0～0.89Si, 0～1.25Mn, 10.2～16.0Cr, 0～9.0Ni, 0～1.8Cu, 0～0.8Al and 0～0.9Ti has been carried out by fuzzy identification method.It is based on T-S fuzzy model using triangle-shaped membership functions to calculate the grade of membership for each given pattern and using Kalman filtering to identify the consequent parameters of fuzzy model. On the basis of data collected from domestic and foreign literature, the relation between chemical compositions/main heat treatment parameters and mechanical properties of maraging stainless steel products has been established by using fuzzy model. Off line simulation indicated that predicted and measured values were in good agreement. Relative error for 91% predicted data of tensile strength ( 1 000～ 2 000 MPa) was less than ±2%, relative error for 85% predicted values of reduction of area (45%～77%) was less than ±2% and relative error for 84% predicted values of elongation (6%～22%) was less than ±4%.