大坝安全监控的门限回归预测模型及其应用

Dam Safety Monitoring Threshold Regression Model and Application

大坝监测效应量在工作过程中受内部和外部因素的共同作用,经常出现跳跃、渐变、再跳跃等复杂周期变化,表现出高度的非线性,为建立常规的大坝安全监控预测模型带来了不少麻烦。借鉴门限自回归模型的门限思想,在常规回归模型中引入门限因子和门限值,利用提出的最佳拟合相关分析法确定门限因子和门限值,建立了门限回归模型。计算实例表明该方法用于大坝安全监控预测可以取得满意的效果。

For the combined action of inner and external factors, the high nonlinearity of the effective variables appears in dam safety monitoring such as jump-gradualness-jump phenomena, which brings about much trouble in analysis. Based on the threshold idea of threshold regression model, the threshold factors and threshold values are introduced and determined by the correlation analysis method of best curve fitted in conventional regression model. It is proved by an illustration to he available with satisfactory results in practical application of dam safety monitoring.