LM-BP神经网络水质评价模型应用研究——以新乡市卫河为例

Applied Research on LM-BP Neural Networks for Water Quality Assessment:Taking the Weihe River of Xinxiang City for Example

介绍了LM-BP神经网络模型的原理及算法和模型的优点。针对实际水质评价问题,利用随机内插方法在地表水环境质量分级标准阈值间生成训练样本和检验样本,建立了新乡市卫河地面水环境质量综合评价的LM-BP神经网络模型,将模型应用于卫河2011年3月份、9月份的水质评价,并与单因子评价法、模糊综合评价法进行了比较分析。实验结果表明该模型设计合理,泛化能力强,收敛速度快,算法稳定,推导严谨,有较充分的理论依据,应用于水质评价具有其合理性、实用性和有效性,适用于作深入的水环境质量分析。

The working principle ,algorithm and advantages of LM-BP Neural Network is introduced .According to the actual prob‐lem of water quality assessment ,the random interpolation method is used to generate training and testing samples at the surface wa‐ter environmental quality grading standard threshold and the surface water environmental quality comprehensive assessment of the BP neural network model is established for the Weihe River of Xinxiang City .The model was applied to water quality assessment in March 2011 and in September 2011 of Weihe River .Then the results are compared with single factor assessment method ,fuzzy math‐ematical assessment method .The experimental results show that the model design is reasonable ,the generalization ability is strong , the convergence is fast ,derivation is rigorous ,algorithm is stable ,and theoretical basis is sufficient .The model applied in water quality evaluation has its rationality ,practicability and validity ,which is available to a thorough analysis of the water environmental quality .