基于ARIMA-RTA组合模型的海堤工程沉降预测

海堤工程沉降预测对工程维护和灾害防治具有重要意义,但因影响沉降的因素较多,目前尚无一种普适的分析模型. 以海堤工程的沉降数据为研究对象,充分利用ARIMA 模型较高的拟合能力和RTA 的等维递补预测思想提高海堤工程监测数据的预测长度,建立了ARIMA鄄RTA 组合预测模型,并以浙江某海堤工程为例进行了实例计算,对沉降实测值和模型预测值进行比较. 结果表明,ARIMA鄄RTA 组合模型能够充分挖掘监测数据的动态信息,具有物理机制明确,预测精度高的优点,为海堤工程监测资料分析提供了新的研究思路.

Prediction of seawall settlement based on ARIMA-RTA model

Analysis and prediction of seawall foundation settlement are very important for engineering maintenanceand disaster prevention, but there is no unified analytical model because of too many factors influencing thesettlement. Based on the seawall monitoring data, the paper makes full use of the good fitness of AutoregressiveIntegrated Moving Average Model ( ARIMA) and dimension-fixed and cursion-compensated idea of RealtimeTracing Algorithm (RTA), and a new model named ARIMA-RTA is developed. As a test, the prediction model isapplied to calculating a case of the seawall monitoring data, and the results show that the model has carefully-defined physical conception and thus opens up broad prospects for application to engineering data analysis for highprecision and noise immunity.