摘要
海水混凝过程的建模对海水淡化的水质控制和加药控制有重要意义。针对海水混凝大滞后的特点,提出一种基于机器学习序列模型的建模方法。采用门控循环单元(Gate Recurrent Unit,GRU)编码器和线性网络解码器等结构化单元,建立出水浊度与入水浊度、流量、絮凝剂及助凝剂用量等因素的关系模型。基于实际运行数据的数值实验验证了所建模型的有效性,也为絮凝剂及助凝剂的减量控制提供了可靠的依据。
The model of seawater coagulation process is of great significance for water quality control and dosing control of seawater desalination.A modeling method based on machine learning and sequence modeling is proposed because of the characteristics of large time delay and time-varying of seawater coagulation.Using structured units such as GRU and linear network decoder,a model between turbidity of inflow,flow rate,flocculant and dosage of coagulant aid was established.The numerical experiments based on the actual operation data verify the validity of the model,and provide a reliable basis for the reduction of flocculant and coagulant aids.
作者
钟骅
朱力
方昱杰
金伟剑
许力
ZHONG Hua;ZHU Li;FANG Yujie;JIN Weijian;XU Li(College of Electrical Engineering,Zhejiang University,310013;Hangzhou Water Treatment Technology Development Center,310012,Hangzhou,China)
出处
《水处理技术》
CAS
CSCD
北大核心
2021年第4期101-105,共5页
Technology of Water Treatment
基金
国家重点研发计划项目(2017YFC0403701)。
关键词
海水淡化
预处理
混凝沉淀
序列模型
机器学习
seawater desalination
preprocessing
coagulation
sequence model
machine learning
