摘要
针对目前景观湖泊富营养化严重的问题,提出了一种基于粒子群算法(PSO)优化支持向量机(SVM)的叶绿素a含量的预测方法。利用2017年5~10月广西大学碧云湖的水质监测数据和气象数据进行主成分分析,确定影响水体叶绿素a含量的主要因素为TN、TP、浊度、温度、光照时长和pH值,并将其作为PSO-SVM模型的输入量,以预测景观湖泊水体叶绿素a的含量;将该模型应用于镜湖、鉴湖和月牙湖水体叶绿素a含量的预测以验证模型的适用性。结果表明,基于PSO-SVM模型的碧云湖的叶绿素a含量预测的平均平方误差仅为1.25%,平均相对误差为2.46%;该模型对镜湖、鉴湖和月牙湖水体叶绿素a含量拟合值的平均平方误差分别为3.17%、4.05%、2.42%,平均相对误差分别为3.48%、4.31%、2.80%。PSO-SVM模型可以很好地运用于景观湖泊水体叶绿素a含量的预测,可为湖泊富营养化防治提供参考。
Aiming at the serious problem of eutrophication of landscape lakes,apredicted method based on particle swarm optimization(PSO)optimized support vector machine(SVM)for chlorophyll a content was proposed.The water quality monitoring data and meteorological data of Biyun Lake of Guangxi University from May to October 2017 were used to principal component analysis.The main factors affecting the chlorophyll a content of water bodies were TN,TP,turbidity,temperature,illumination duration and pH.These values were chosen as the input of PSO-SVM model to predict the content of chlorophyll a in landscape lake water.The model was applied to the prediction of chlorophyll a content in Jinghu,Jianhu and Yueya lakes to verify the applicability of the model.The results show that the average squared error of chlorophyll a content in Biyun Lake by the PSO-SVM model is only 1.25%,and the average relative error is 2.46%.The fit average squared errors of the chlorophyll a content of Jinghu,Jianhu and Yueya Lake water bodies are 3.17%,4.05%,and 2.42%,and the fit average relative errors are 3.48%,4.31%,and 2.80%,respectively.The PSO-SVM model can be better applied to the prediction of chlorophyll a content in landscape lake waters,which can provide reference for lake eutrophication control.
作者
李国进
王雪茹
黄鹏
LI Guo-jin;WANG Xue-ru;HUANG Peng(School of Electrical Engineering,Guangxi University,Nanning 530004,China)
出处
《水电能源科学》
北大核心
2019年第11期58-61,共4页
Water Resources and Power
关键词
主成分分析法
粒子群算法
支持向量机
叶绿素A含量
水体富营养化
principal component analysis
particle swarm optimization
support vector machine
chlorophyll a content
water eutrophication
