基于马尔科夫链的曲线拟合法在尾矿坝沉降预测中的应用

Application of Curve Fitting Method Based on Markov Chain in Tailings Dam Settlement Prediction

为了对尾矿坝进行高效准确的沉降预测,笔者提出一种基于马尔科夫链的曲线拟合法。将曲线拟合法中的指数曲线法和Logistic曲线法分别与马尔科夫链构建组合预测模型,在保留曲线拟合法建模简单、易于操作的优势的基础上,提高模型的预测精度。结合尾矿坝沉降监测数据,并与传统的曲线拟合法、后向传播神经网络、基于遗传算法的后向传播神经网络进行比较。实验结果表明,提出的方法预测误差小,预测精度明显优于传统预测方法。

In order to predict the settlement of tailings dams efficiently and accurately,a curve fitting method based on Markov chain was proposed.On the basis of retaining the advantages of simple and easy operation of curve fitting method,the exponential curve and Logistic curve of curve fitting method and Markov chain were used separately to build combined prediction model for improving prediction accuracy.Experimental evaluation with settlement data of the tailings dam has been carried out.The results were compared with the traditional curve fitting method,back propagation(BP)neural network,and genetic algorithm-based back propagation(GA-BP)neural network.The experimental results show that the smaller prediction error can be obtained using the proposed method and the prediction accuracy is significantly better than the traditional prediction methods.