摘要
煤炭废水处理对煤炭行业生态环保和国家“双碳目标”具有重要意义,针对煤炭废水处理过程中絮体识别评价困难导致的固液分离效果差、药耗高、调控滞后的问题,提出了基于深度学习的动态絮体综合评价体系,首先,获取大量不同类别的动态絮体样本图像,对絮体图像进行预处理,以絮体纹理参数为依据,提取有效絮体的视觉特征。然后,采用主成分分析法获取综合评价指标,构建基于机器视觉的絮体特性综合评价体系。最后,对核密度分析法拟合的絮体综合特征概率分布曲线进行动态聚类,以其获取的特征点作为分类器的输入,构建分类器模型,以此评价煤炭废水沉降效果。结果表明,所提出的方法能够获取动态絮体图像特征以实现煤炭废水沉降效果的自动识别与评价,最终不同状态下煤炭废水沉降效果检测的准确度均在90%以上,为智能化选煤厂的建设提供了一种新方法、新思路。
The treatment of coal wastewater is of great significance to the ecological and environmental protection of the coal industry and the national “carbon peaking and carbon neutrality goals”. In order to solve the problems of poor solid-liquid separation effect, high drug consumption and lag in regulation and control caused by the difficulty of flocs identification and evaluation in the process of coal wastewater treatment, a dynamic flocs comprehensive evaluation system based on deep learning was proposed. Firstly, a large number of images of dynamic floc samples with different types were obtained and preprocessed, and the visual features of effective flocs were extracted based on the flocs texture parameters. Then, the principal component analysis method was used to obtain the comprehensive evaluation indexes of flocs, the comprehensive evaluation system of flocs characteristics based on machine vision was constructed. Finally, dynamic clustering was conducted on the comprehensive characteristic probability distribution curve of flocs fitted by nuclear density analysis and the obtained characteristic points were used as the input of the classifier to build a classifier model, so as to evaluate the settling effect of coal wastewater. The results show that the proposed method can obtain the image characteristics of dynamic flocs to realize the automatic identification and evaluation of the settling effect of coal wastewater. Finally, the detection accuracy of the settling effect of coal wastewater is more than 90% under different states. The study can provide a new method and new idea for the construction of intelligent coal preparation plant.
作者
韩宜纯
樊玉萍
马晓敏
董宪姝
陈茹霞
宋书宇
武志伟
HAN Yichun;FAN Yuping;MA Xiaomin;DONG Xianshu;CHEN Ruxia;SONG Shuyu;WU Zhiwei(College of Mining Engineering,Taiyuan University of Technology,Taiyuan,Shanxi 030024,China;State Key Laboratory of Mineral Processing,Beijing 102628,China;Tashan Coal Preparation Plant of Jinneng Holding Group,Datong,Shanxi 037000,China)
出处
《矿业研究与开发》
CAS
北大核心
2022年第9期188-196,共9页
Mining Research and Development
基金
国家自然科学基金(51820105006,52074189,520041185)
山西省留学回国人员择优资助项目(20210036)
矿物加工科学与技术国家重点实验室开放基金(BGRIMM-KJSKL-2022-10)。
关键词
选煤厂
煤炭废水
有效絮体
综合评价体系
动态聚类
Coal preparation plant
Coal wastewater
Effective flocs
Comprehensive evaluation system
Dynamic clustering
