基于机器视觉的水处理絮凝过程中絮体检测与控制模型研究

项目名称： 基于机器视觉的水处理絮凝过程中絮体检测与控制模型研究

项目编号： No.61272197

项目类型： 面上项目

立项/批准年度： 2013

项目学科： 自动化技术、计算机技术

项目作者： 胡锋平

作者单位： 华东交通大学

项目金额： 80万元

中文摘要： 针对水处理絮凝过程复杂，影响因素多，具有非线性、大滞后的特点，絮体检测采样点代表性不强，絮体跟踪算法复杂，控制模型难以建立等研究现状。采用机器视觉方法，絮凝池和沉淀池双池采样，对絮凝池检测絮体粒径和数量，基于计算流体力学Fluent软件建立沉淀池中絮体运动轨迹方程，采用粒子滤波多目标跟踪算法检测絮体沉速；基于絮体检测结果，进一步揭示絮凝机理；采用差分进化算法建立投药量与原水水质、絮凝池絮体粒径与数量、沉淀池絮体沉速、滤前水水质等参数关系的多源数据融合模型；以沉淀池中检测的沉降速度为主控，絮凝池中检测获取的粒径、数量等参数为辅控，构建一个既考虑原水水质变化的实时性又考虑滤前水浊度的滞后性的混凝投药闭环串级控制系统。研究成果可为复杂水处理过程絮体检测、控制模型的建立探索新途径，推动絮凝理论的发展，具有一定的学术价值。成果的应用对提高供水水质、保障饮用水安全、降低水厂运行成本具有重大的现实意义。

中文关键词： 机器视觉；水处理；絮体检测；控制模型；粒子滤波

英文摘要： As for the characteristics of complex flocculation process, different influential factors, nonlinearity and large time-delay, while the current study such as representative of sampling points of floc measurement being weak, tracking algorithm being complex, along with the difficulty of setting up control model in water treatment, the theory and methods of machine vision and sampling both in the flocculation basin and the sedimentation basin will be used to measure size and quantity of floc. Floc movement track equation will be established in the sedimentation basin based on the Fluent and floc sedimentation velocity will be measured based on particle filter multi-target tracking algorithm. The flocculation mechanism will be further discovered based on the measurement results. The fusion model of multi-source data, which reveals the relations of dosage, raw water quality, floc size and quantity in the flocculation basin, floc sedimentation velocity in the sedimentation basin and the quality of precipitation water, will be set up based on dedifferential evolution(DE). Taking sedimentation velocity in the sedimentation basin as master control, and floc size, quantity in the flocculation basin as auxiliary control, a closed loop bunch-rank coagulant dosage control system, which both consider raw water quality change

英文关键词： Machine Vision；Water Treatment；Floc Measurement；Control Model；Particle Filter