根据神经网络的非线性和良好的函数逼近特性,提出了基于人工神经网络的灰色模型、多项式回归模型组合的输气管道腐蚀速率预测模型.此组合模型将最佳组合权重隐含在网络的连接权中,兼具灰色预测、回归预测和神经网络预测的优点,克服了原始数据少,数据波动大对预测精度的影响,也增强了预测的自适应性,在客观地反映输气管道腐蚀速率变化趋势方面具有一定的优势．通过实例分析,表明预测值与实际结果有很好的一致性.
Based on the nonlinear character and excellent character of function approximation of BP neural networks,we proposed a grey regression combination forecasting model for gas pipeline corrosion.In this model the optimum combination weight was included in the weight of networks,the model has combined the advantages of grey estimation,regress and BP neural networks estimation,whith then can overcome the influence of little raw data and high data fluctuation on the estimation precision,meanwhile it has enhanced the self adaptability of the estimation and can reflect the objective trend of gas pipeline's corrosion rate system.Case study proved the effectiveness of the proposed model in this paper.
参考文献
| [1] | |
| [2] | |
| [3] | |
| [4] | |
| [5] | 龙学渊,袁宗明. 不等时距GM(1,1)模型在预测输气管道腐蚀中的应用[J]. 管道技术与设备, 14(1):35
上一张
下一张
上一张
下一张
计量
文章评分
|