CCTEG Hangzhou Research Institute Co., Ltd.
Wuhan Water Group Co., Ltd.
School of Environmental Scienceand Engineering, Huazhong University of Science and Technology

目前国内水厂大多采用经验法进行混凝剂投加控制,为实现水厂混凝剂投加智能化,本研究搭建了基于矾花图像识别的智能投药系统。 该系统结合了 YOLOv5 矾花识别算法和 LinearRegression 加药决策算法,并在此基础上添加了一个 7 维的全连接 BP 神经网络,通过(563,7)的样本集(563 条包含矾花数量、矾花平均等效直径、进水流量等 7 项参数的样本的集合) 进行训练,计算确定每一层的最佳权重,得到最低损失值为 0.018 的线性回归模型。 生产试验表明,矾花目标检测准确率为 83.5%,预测投药量相比原有经验值降低 11.0%。 与传统控制方法相比,该系统时延性更低,可靠性更强,药耗更低,有效降低了水厂加药生产和管理成本。

At present, most domestic water plants use empirical methods for coagulant dosing control.In order to realize the intelligent dosing of coagulant in water plants, this research has built an intelli⁃gent dosing system based on alum floc image recognition. The system combines the YOLOv5 alum flocrecognition algorithm and the Linear Regression dosing decision algorithm. And on this basis, a 7-di⁃mensional fully connected BP neural network was added for training through a sample set of (563, 7)(563 samples containing 7 parameters such as the number of alum flocs, the average equivalent diame⁃ter of alum flocs, and the inflow flow rate). The optimal weights for each layer were calculated and de⁃termined, resulting in a linear regression model with a minimum loss value of 0.018. The productiontest showed that the detection accuracy of alum floc target was 83.5%, and the predicted dosage was11.0% lower than the original empirical value. Compared with the traditional control method, thesystem has lower time ductility, stronger reliability and lower coagulant consumption, which have effec⁃tively reduced the production and management costs of dosing in water plants.

付元, 雷智丰, 崔东锋, 等. 基于矾花图像识别的混凝剂智能投加系统研究[J]. 能源环境保护, 2023, 37(4): 83-90.

FU Yuan, LEI Zhifeng, CUI Dongfeng, et al. Research on an intelligent coagulant dosing system based on alumfloc image recognition[J]. Energy Environmental Protection, 2023, 37(4): 83-90.