将粒子群算法引入大坝安全监控领域,并结合多元回归统计模型,建立基于粒子群算法的混凝土坝变形预报模型。利用粒子群算法的全局寻优能力以及该算法具有正反馈信息的仿生特点。通过优化迭代计算,确定坝体变形统计模型中各回归系数。工程实际应用表明,基于该模型的预报结果与传统的最小二乘法相比,可显著提高混凝土坝变形的预报精度,所以该模型的预报应用是可行的。
A concrete dam deformation forecasting model is established based on the particle swarm optimization (PSO) algorithm and the traditional multi-statistical regression model. Making use of global searching optimization and the PSO algorithm's ability to adopt positive feedback information, the regression coefficients of a multi-statistic regression model are determined by iterative calculation. Application in practical engineering shows that the forecasting results of this model arc better than those of the traditional least-square regression method. Hence, this model is feasible in concrete dam deformation forecasting with high precision.