忠武输气管道沿线滑坡预测方法研究

The Study of Zhong-Wu Gas Pipeline Landslide Prediction

基于忠武管道沿线滑坡不具有成面、成片分布的特点及其管道沿线滑坡监测方案,系统地分析了忠武输气管道沿线滑坡预测过程,根据灰色GM(1,1)模型具有对数据量需求少、对时间有关的序列有很好的预测效果等优点,重点研究了灰色GM(1,1)模型及其改进模型,并对各种改进的灰色模型预测结果进行比较分析;同时顾及BP神经网络的各种优点,本文还较深刻地研究了灰色神经网络组合预测模型,在此之上提出了一种改进的串联灰色神经网络模型,通过实例分析对该模型在忠武管道沿线滑坡预测方法进行了探讨。

Based on landslides along the Zhong-Wu gas pipeline have not the characteristics of the distribution into a surface and a pieces,and the landslides monitoring cycle of this pipeline.this paper systematically analyzes forecasting process of the landslide along the Zhong-Wu gas pipeline.As GM（1,1） model is less demand on the amount of data,with a good predictive advantages of the sequence of time,this paper focuses on the gray GM（1,1） model and its improved model,and a variety of improved gray model for comparative analysis.In view of the BP neural network has many kinds of advantages.This paper has thoroughly studied the combination of gray neural network model,and based on the combination forecasting model,putting forward an improved gray neural network model series and probe the model through an example.