基于SSA-BP算法的超高温陶瓷裂纹长度预测  

Crack Length Prediction of Ultra-High Temperature Ceramics Based on SSA-BP Algorithm

作　　者：王一宁 刘宝良 刘洋[2] 李长青[3] WANG Yining;LIU Baoliang;LIU Yang;LI Changqing(School of Architectural and Civil Engineering,Guangdong University of Petrochemical Technology,Maoming 525000,China;Intelligent Technology Institute,HE Power Equipment National Engineering Research Center Co.Ltd,Harbin 150001,China;Science School,Heilongjiang University of Science and Technology,Harbin 150022,China)

机构地区：[1]广东石油化工学院建筑工程学院,广东茂名525000 [2]哈电发电设备国家工程研究中心有限公司智能技术研究所,黑龙江哈尔滨150001 [3]黑龙江科技大学材料学院,黑龙江哈尔滨150022

出　　处：《广东石油化工学院学报》2024年第4期104-107,共4页Journal of Guangdong University of Petrochemical Technology

基　　金：国家自然科学基金面上项目(11572113);广东省自然科学基金项目(2023A1515012366);黑龙江省自然科学基金项目(LH2021E107)。

摘　　要：超高温陶瓷构件在航天航空中的运用往往会出现检测方面的困难,在构件产生裂纹后会在一定范围内失效。针对使用传统的BP神经网络预测超高温陶瓷构件的裂纹长度存在的对连接权值和阈值具有较强依赖性导致收敛速度较慢、易陷入局部最优和稳定性差等问题,提出一种基于麻雀搜索算法SSA优化的BP神经网络关于裂纹长度的预测方法。以ABAQUS有限元分析软件得出的超高温陶瓷裂纹长度相关参数构成的基础数据集作为模型的输入。利用SSA优化BP神经网络的初始权值与阈值得到了更好的拟合结果。结果表明利用SSA-BP神经网络进行预测的可行性。The use of ultra-high temperature ceramic components in aerospace applications often presents difficulties in detection and failure within a certain range after cracks are produced in the component.This paper proposes a method based on the Sparrow Search Algorithm(SSA)to predict the crack length of ultra-high temperature ceramic components using traditional BP(back propagation)neural networks,which has a strong dependence on the connection weights and thresholds,resulting in slow convergence,easy to fall into the local optimum and poor stability.This paper presents a method for predicting crack length based on the SSA optimization of BP neural networks.This paper uses the parameters related to the crack length of ultra-high temperature ceramics derived from ABAQUS finite element analysis software as the base data set as the input to the model.The initial weights and thresholds of the BP neural network are optimized by using SSA to obtain better fitting results.The results show the feasibility of using SSA-BP neural for prediction.

关 键 词：超高温陶瓷 裂纹长度预测 SSA-BP 数值模拟 

分 类 号：TH164[机械工程—机械制造及自动化]