摘要
钢筋锈蚀是混凝土结构破坏的主要原因之一,通过研究钢筋锈蚀情况,进而分析混凝土构件的耐久性,对提高建筑物的使用寿命来说,具有非常重要的现实意义。钢筋锈蚀受多种因素影响,但传统的钢筋锈蚀检测法容易受到人为和条件因素的干扰,进而产生较大的误差,因此构建一种钢筋锈蚀预测模型,可以快速地进行钢筋锈蚀程度的预测。通过钢筋锈蚀实验获得了23组实测数据,建立基于钢筋中间距、保护层厚度、裂缝宽度、锈蚀电流和锈蚀直径5个因素的预测指标体系,通过主成分分析(PCA)对数据进行降维处理,建立最小二乘支持向量机(LSSVM)模型对钢筋锈蚀率进行预测,选用粒子群算法(PSO)寻找出LSSVM中正则化参数和核函数宽度系数的最优参数组合。结果表明,PSO-LSSVM模型的平均绝对误差和均方根误差为1.88%和1.94%,并同BP神经网络、未优化的LSSVM模型的预测结果做对比分析,验证了该模型的预测精度更高,为复杂环境下的钢筋锈蚀情况监测提供了一种新的途径。
The corrosion of steel bars is one of the main reasons for the destruction of concrete structures. It is of great practical significance to improve the service life of buildings by studying the corrosion of steel bars and then analyzing the durability of concrete components. The reinforcement corrosion is affected by many factors, but the traditional reinforcement corrosion detection method is easy to be interfered by human and conditional factors, resulting in large errors. Therefore, building a reinforcement corrosion prediction model can quickly predict the reinforcement corrosion degree. In this paper, 23 groups of measured data were obtained through the reinforcement corrosion experiment, and a prediction index system was established based on the five factors of reinforcement spacing, cover thickness, crack width, corrosion current and corrosion diameter. The data were reduced by principal component analysis(PCA), and a least squares support vector machine(LSSVM) model was established to predict the reinforcement corrosion rate. The particle swarm optimization(PSO) algorithm is used to find the optimal combination of regularization parameters and kernel width coefficients in LSSVM. The results show that the mean absolute error and root mean square error of the PSO-LSSVM model are 1.88% and 1.94%, and compared with the prediction results of the BP neural network and the unoptimized LSSVM model, it is verified that the prediction accuracy of the model is higher, which provides a new way to monitor the corrosion of steel bars in complex environments.
作者
刘爽
樊文辉
成思维
顾力冬
郭顺军
LIU Shuang;FAN Wen-hui;CHENG Si-wei;GU Li-dong;GUO Shun-jun(School of Architecture and Civil Engineering,Qiqihar University,Heilongjiang Qiqihar 161006,China)
出处
《齐齐哈尔大学学报（自然科学版）》
2023年第2期37-43,共7页
Journal of Qiqihar University(Natural Science Edition)
基金
2022年度黑龙江省大学生创新创业训练计划项目(x202210232134)
2022年度黑龙江省省属本科高校基本科研业务费青年创新人才项目(145209208)。
关键词
钢筋锈蚀
最小二乘支持向量机
粒子群算法
主成分分析
reinforcement corrosion
least squares support vector machine
particle swarm optimization
principal component analysis
