基于小波分析和神经网络的老旧小区沉降监测信息化研究

Research on Settlement Monitoring Informatization of Old Community Based on Wavelet Analysis and Neural Network

临近工程引起的老旧小区沉降是一种危害较大的环境灾害,为此该文提出运用Matlab实现精度较高的神经网络预测模型,应用于B/S架构的老旧小区安全管理系统进行信息化监测。为了减少监测数据噪音的干扰,采用小波分析进行信号降噪,运用BP神经网络和改进RBF神经网络进行对比分析。通过对南京某地铁工程周边小区的沉降监测数据实例分析,得出给定阈值消噪后的信号处理效果最好,改进的RBF神经网络具有更精准的预测能力,Matlab程序打包应用于系统后,得到了较好的实际应用。

The settlement of the old community caused by the adjacent project is a kind of environmental hazard with great harm. So, a neural network prediction model with high accuracy is implemented by using Matlab and applied to the old community safety management system with B/S structure for informational monitoring.In order to reduce the interference of monitoring data noise, the wavelet analysis is used to reduce the noise of the signal. The BP neural network and improved RBF neural network are used for comparative analysis. An example analysis of the settlement monitoring data of the surrounding area of a subway project in Nanjing shows that the signal processing effect is best after denoising with a given threshold. The improved RBF neural network has more accurate prediction ability. The Matlab program is packaged and applied in the system, and it gets better practical application.