锻压技术

基于ELM的H型钢组织预测专家系统

英文标题：H-beam steel structure prediction expert system based on ELM

作者：刘亚玓1 2 马劲红1 2 王小巩1 2

单位：1. 华北理工大学冶金与能源学院 2. 华北理工大学现代冶金技术教育部重点实验室

分类号：TG335.4+2

出版年,卷(期):页码：2024,49(1):241-248

摘要：

采用极限学习机建立了H型钢组织预测专家系统，可根据H型钢的化学成分、工艺参数及冷却速度来预测轧后钢种的微观组织、晶粒等级和平均冲击功。以某钢厂H型钢的生产数据为专家系统的数据库，批量预测了H型钢的微观组织、平均冲击功和晶粒等级，预测结果表明，决定系数R2较高，拟合程度好；平均绝对百分比误差较小，精度高。为验证专家系统的可行性，采用与预测相同钢材、相同轧制工艺参数的H型钢进行金相实验和夏比冲击实验，由实验得到的微观组织占比、平均冲击功及晶粒等级与预测值基本一致。结果表明：基于ELM建立的H型钢组织预测专家系统，其预测结果精度高，对实际生产有较好的参考价值。

H-beam steel structure prediction expert system was established by the extreme learning machine, and based on the chemical composition, process parameters and cooling speed of H-beam steel, the microstructure, grain grade and average impact energy of the steel grade after rolling were predicted. Then, the microstructure, average impact energy and grain grade of H-beam steel were predicted in batches by the expert system with the data base from production data of H-beam steel in a certain steel plant. The prediction results indicate that the determination coefficient R2 is high and the fitting degree is good, and the average absolute percentage error is smaller and the accuracy is high. In order to verify the feasibility of the expert system, H-beam steel with the same steel material and the same rolling process parameters as predicted was used to conduct metallographic experiment and Charpy impact experiment, and the proportion of microstructure, average impact energy and grain grade obtained by experiment are basically consistent with the predicted values. The results indicate that the H-beam steel structure prediction expert system established based on ELM has high prediction accuracy and good reference value for the actual production.

基金项目：

河北省自然科学基金资助项目（E2021209041)

作者简介：刘亚玓（1998-），男，硕士研究生 E-mail：yydd981@163.com 通信作者：马劲红(1973-),女，博士，教授 E-mail：majinhong@126.com

参考文献：

[1] Li W，Chen H. Tensile performance of normal and high-strength structural steels at high strain rates[J]. Thin-Walled Structures，2023，184: 110457.

[2] Liu X C，Wu X T，Wang Y，et al. Seismic performance of bolted connection between H-section beam and SST column welded with inclined braces[J]. Journal of Building Engineering，2022，61: 105270.

[3] Dissanayake M，Nguyen H，Poologanathan K，et al. Prediction of shear capacity of steel channel sections using machine learning algorithms[J]. Thin-Walled Structures，2022，175: 109152.

[4] Abedrabbo A F，Osorio J C，Abolghasem S，et al. Predicting subgrain size and dislocation density in machining-induced surface microstructure of nickel using supervised model-based learning[J]. Materials Today Communications，2022，30: 103162.

[5] Meade E D，Sun F W，Tiernan P，et al. A multiscale experimentally-based finite element model to predict microstructure and damage evolution in martensitic steels[J]. International Journal of Plasticity，2021，139: 102966.

[6] Dong X Y，Shin Y C. Predictive modeling of microstructure evolution within multi-phase steels during rolling processes[J]. International Journal of Mechanical Sciences，2019，150: 576-583.

[7] Ling Y，Ni J Y，Antonissen J，et al. Numerical prediction of microstructure and hardness for low carbon steel wire arc additive manufacturing components[J]. Simulation Modelling Practice and Theory，2023，122: 102664.

[8] Srivastava A，Sinha A N，Verma S K. A mini-review on numerical approach of microstructure prediction in eutectoid steel[J]. Materials Today: Proceedings，2022，50: 2241-2248.

[9] Chegni A M，Ghavami B，Eftekhari M. A GPU-based accelerated ELM and deep-ELM training algorithms for traditional and deep neural networks classifiers[J]. Intelligent Systems with Applications，2022，15: 200098.

[10]Wang S，Li J，Zuo X W，et al. An optimized machine-learning model for mechanical properties prediction and domain knowledge clarification in quenched and tempered steels[J]. Journal of Materials Research and Technology，2023，24: 3352-3362.

[11]Ganguly S，Wang X，Chandrashekhara K，et al. Modeling and simulation of mass flow during hot rolling low carbon steel I-beam[J]. Journal of Manufacturing Processes，2021，64: 285-293.

[12]Zhang T，Gao D Y. Tensile behavior analysis and prediction of steel fiber-reinforced-carbon/glass hybrid composite bars[J]. Journal of Building Engineering，2023，64: 105669.

[13]李东宽，郭岩，杨立新，等. TC4钛合金两相区的热变形行为及微观组织[J]. 铸造技术，2022，43(2): 114-119.

Li D K，Guo Y，Yang L X，et al. Thermal deformation behavior and microstructure of TC4 titanium alloy in two-phase region[J]. Foundry Technology，2022，43(2): 114-119.

[14]Zhang X G，Ren Y J，Zhang J，et al. Effects of prior austenite grain size on reversion kinetics of different crystallographic austenite in a low carbon steel[J]. Materials Characterization，2022，190: 112025.

服务与反馈：

【文章下载】【加入收藏】