基于改进FOA-GM(1,1)灰色预测模型的箱涵沉降预测

Deformation Prediction of Box Culvert Based on Improved FOA-GM (1,1) Grey Prediction Model

箱涵沉降变形是多因素共同耦合作用的结果,具有不确定性、灰色性、贫信息性等特点,传统GM(1,1)模型对箱涵沉降变形的预测结果精度较低,因此引入迭代权值和遗传变异思想改进果蝇优化(FOA)算法,以提高其搜索能力和速度。对传统GM(1,1)模型赋值背景值系数d和时间系数λ,利用改进FOA算法进行参数寻优,建立改进FOA-GM(1,1)预测模型。将该预测模型应用于南水北调中段天津干线段箱涵沉降变形预测,结果表明改进模型模拟值与实际沉降值相对误差均在6%以下,明显高于传统模型的精度,说明改进模型不仅对了解箱涵未来沉降变形趋势和实时控制箱涵不均匀沉降具有重要意义,亦为解决水利工程行业其他类似问题提供了参考。

Box culvert deformation is usually caused by coupling effects of various factors and it is an uncertain system that contains highly gray characteristic, uncertainty and few information. The traditional GM （1,1） model is not enough accurate in predicting the settlement deformation of box culvert. The FOA algorithm was improved by introducing the iteration weight and genetic mutation strategy, which can greatly improve its search capabilities and speed. An improved FOA GM（1,1）prediction model was established to optimize the background value coefficient d and time factor lambda in the traditional GM （1, 1） model. This prediction model was applied to predict deformation of box culvert in the main trunk line of Tianjian in the middle routing of South to North water diversion project. The results show that the improved FOA GM（1,1）prediction model has a better predictive ability than the traditional one, and the relative error between predicted and measured value is mainly less than 6%. It is of great significance to predict the deformation of box culvert and control its uneven settlement, and can provide a reference for solving similar problems in the field of water conservancy project.