基于卡尔曼滤波、分形和LSTM的大坝变形趋势分析方法

Analysis method of dam deformation trend based on Kalman filter,fractal and LSTM

为了实现对大坝变形趋势的合理分析,提出了一种融合卡尔曼滤波、分形理论和长短时记忆神经网络(LSTM)的大坝变形趋势综合分析方法。该方法利用卡尔曼滤波对原始观测数据进行降噪处理,采用分形理论中的R/S方法对大坝未来变形趋势进行定性判断和解析,通过对滤波后数据建立LSTM定量预测模型,结合定性和定量的分析结果,实现大坝变形趋势的综合评判。实例分析结果表明,该方法能够较好地分析大坝变形趋势,对大坝监测数据的随机性和非平稳性具有较好的适用性,趋势分析和预测符合工程实际情况,为大坝的变形综合分析提供了一种新方法。

In order to realize reasonable analysis of dam deformation trend,a comprehensive analysis method integrated with Kalman filter,fractal and long short-term memory neural network(LSTM)was proposed.In the method,Kalman filter was used to denoise the original observation data,the R/S method of fractal was used to qualitatively judge and analyze the future deformation trend of the dam,and a LSTM quantitative prediction model was established for the filtered data.Combining with the qualitative and quantitative analysis results,comprehensive evaluation of dam deformation trend was realized.The example analysis results show that the proposed method can analyze the dam deformation trend well,and has good applicability to the randomness and nonstationary of dam monitoring data.The trend analysis and prediction are in line with the actual situation of the project,which provides a new method for the comprehensive analysis of dam deformation.