C/SiC复合材料数据库及机器学习性能预测平台的设计与开发
Design and Development of Database and Machine Learning Property Prediction Platform of C/SiC Composites
曹文轩 1史振宇 1邹斌 1王继来 1张成鹏 1袭建人2
作者信息
- 1. 山东大学机械工程学院,高校洁净机械制造教育部重点实验室,济南 250061
- 2. 山东大学材料科学与工程学院,济南 250061
- 折叠
摘要
以MySQL Workbench 8.0为数据库平台,采用三层浏览器/服务器架构,利用PyCharm 2022开发了一套基于Web界面和机器学习技术的C/SiC复合材料数据库系统,该系统包括材料数据库和机器学习模型2部分,具有对材料组分结构、制备工艺、试验性能等材料研发各阶段的信息查询、添加、修改、删除、搜索以及支持信息管理、设计分析视图、案例推理辅助设计、用户与系统管理等功能.基于神经网络回归算法训练机器学习模型,利用材料微观结构参数预测材料力学性能,对不同微观结构参数的影响权重进行评价,并部署在系统平台中;通过Web用户界面调用机器学习模型,对预留的验证集数据进行拉伸模量和弯曲模量预测.结果表明:机器学习模型预测材料力学性能的精度达到95％左右,训练出来的预测模型具有良好的精度与泛化能力.
Abstract
With MySQL Workbench 8.0 as the database platform and adopting a three-layer browser/server architecture,a C/SiC composite database system based on Web interface and machine learning technology was developed by PyCharm 2022.The system included material database and machine learning model,and had the functions of information query,addition,modification,deletion and search,as well as support information management,design analysis view,case-based reasoning aided design,user and system management for material component structure,preparation process,test performance and other stages of material research and development.The machine learning model was trained based on the neural network regression algorithm,the mechanical properties of materials were predicted by the microstructure parameters of materials,and the influence weights of different microstructure parameters were evaluated and deployed in the system platform.The machine learning model was invoked through the Web user interface to predict the tensile modulus and the flexural modulus of the reserved validation set data.The results show that the accuracy of machine learning models in predicting material mechanical properties reached about 95％,and the trained predictive model exhibited good accuracy and generalization capability.
关键词
C/SiC复合材料/材料数据库/Web系统/神经网络/性能预测Key words
C/SiC composite/material database/Web system/neural network/performance prediction引用本文复制引用
基金项目
国家重点研发计划(2022YFB3401902)
国家自然科学基金(U21A20134)
山东省自然科学基金优秀青年基金(ZR2022YQ48)
出版年
2024
NETL
NSTL
万方数据