基于BP神经网络的黄河水质预测研究

Prediction of water quality in Yellow River based on BP neural network model

为了进一步研究高度非线性和非确定性水环境系统的变化规律,预测复杂、模糊、高度非线性河流水质。以黄河内蒙古河段为例,构建以上游断面监测数据预测下游水质变化的BP神经网络模型。模型选用L-M数值优化算法,采用3-6-1型网络结构,对黄河内蒙古河段2013—2014年各监测断面COD监测值进行训练。结果表明:BP神经网络具有很强的非线性映射能力和柔性的网络结构,可以很好地应用于黄河内蒙古河段COD预测,平均相对误差为5.66%,预测精度较高,为河段水环境治理、水质监测和控制污染提供借鉴意义。

In order to study the variation of highly nonlinear and non - deterministic water environ-ment systems, and predict the complexity, fuzzy and highly nonlinear river water quality, taking the Inner Mongolia reach of the Yellow River as an example, the B P neural network model for predicting the change of downstream water quality was established. The model uses the L - M numerical opti-mization algorithm, and uses the 3 - 6 - 1 type network structure to train the monitoring COD moni-toring value of the monitoring section of the Yellow River Inner Mongolia during 2013 and 2014. The results show that the B P neural network has a strong nonlinear mapping ability and flexible network structure, which can be applied to the C O D prediction of the Yellow River in Inner Mongolia, with an average relative error of 5. 6 6% and high prediction accuracy. The results can provide reference for environmental management, water quality monitoring and pollution control of the Yellow River.