浓缩与压滤过程药剂协同控制系统研究与应用

发布时间：2018-04-15 09:14

本文选题：煤泥水处理 + 药剂优化　；参考：《太原理工大学》2017年硕士论文

【摘要】：煤泥水处理作为选煤厂的一项重要工艺流程,直接关系到厂区的闭水循环指标,同时也对选煤厂整体效率与生产指标造成影响。浓缩与压滤是煤泥水处理中的两个关键环节,其目的就是为了实现煤泥水中细微颗粒与清水的分离,回收煤泥,清水循环利用。絮凝剂与助滤剂的添加主要是为了改变煤泥水微粒的表面电性,加速煤泥微粒絮团的形成,加速沉淀,并提高脱水性能。本文针对成庄矿选煤厂原有絮凝剂与助滤剂药剂添加装置系统中添加量由人工设置或针对单个环节进行添加量控制的问题,没有兼顾浓缩与压滤两个相关环节的协同作用,造成药剂添加不合理,造成药剂的浪费,为了解决上述问题,提出了针对浓缩与压滤过程药剂协同控制的研究。通过对原有絮凝剂与助滤剂药剂添加装置的分析,现具有完善的药剂溶液制备装置与自动添加装置,药剂添加量控制策略有待完善。系统利用原有的制备与添加装置,药剂添加量由协同系统求解所得。煤泥水处理过程作为一个典型的物化反应过程,在浓缩与压滤过程中药剂添加量主要影响变量为入料浓度、入料流量、底流浓度、溢流浓度、压滤周期与煤泥饼水分,该过程具有强耦合、非线性、大滞后等特点,很难通过数学推导建立其模型,本文提出了通过BP神经网络对药剂添加量进行模型的建立,并通过APSO算法进行最优量求解的策略。经过对BP神经网络原理和推理算法的分析,基于煤泥水处理中变量影响关系,分别建立4×5×1结构的絮凝剂添加模型与3×5×1结构的助滤剂神经网络模型,并利用现场50组数据对网络训练。依据要实现目标与现场工况建立药剂添加优化量最优化模型,确定优化约束条件,选定PSO算法对药剂最优化模型进行求解计算,并利用惯性权重值与粒子飞翔速度线性递减的自适应策略对PSO算法进行改进。在Matlab平台对APSO算法进行程序的设计与运行。为了实现算法对优化量的在线求解计算,通过Simulink仿真平台利用S函数调用APSO程序,并通过OPC技术实现与PLC控制器的联合运行。本系统选用AB 1756-Control Logix PLC为协同控制器,研华科技ACP4000作为上位机,系统进行硬件结构的搭建,使用RSLogix5000进行控制器程序的编写,选用Matlab/Simulink作为APSO算法在线计算平台。为了实现数据的交互,使用MSG功能模块与原系统控制器进行通讯,利用OPC接口技术实现控制器与Matlab、FT VIEW之间的数据通讯。系统通过现场传感器采集工况数据,通过OPC技术反馈到Matlab的APSO算法相应变量中,由算法在线对该工况下最优药剂添加量求解计算,将优化药剂量经协同控制器返回给原药剂添加系统,之后原药剂添加系统按优化药剂量执行动作,达到在线协同优化的目的。系统在成庄矿选煤厂运行稳定可靠,且通过对系统运行前后三个月的数据分析,煤泥生产总量也略有提高,同时吨煤泥PAC药耗由2.453Kg/T降低到2.341Kg/T,吨煤泥PAM药耗由0.182Kg/T降低到0.172Kg/T。药剂消耗的经济指标,由4.119降低到3.914。说明本系统不仅保证了煤泥水处理系统生产速率,同时降低了药剂的消耗,提高了药剂间的协同作用与选煤厂经济效益。
[Abstract]:Slime water treatment is an important process of coal preparation plant, is directly related to the closed water circulation index of the plant, but also the impact on the overall efficiency of coal preparation plant and production index. Concentration and filter press are two key links of slime water treatment, its purpose is to realize the separation of fine particles and water in coal slurry recycling water slime, recycling. Adding flocculant and filter aid is mainly to change the surface electric properties of coal slurry particles, the formation of slime particles, accelerate floc accelerated precipitation, and improve the dewatering performance. This paper focused on the effects of Chengzhuang Coal Preparation Plant original flocculant and filter aid agent adding device system by artificial set or for a single link to add volume control problems, did not take into account the synergistic effect of the two related aspects of concentration and filter press, by adding medicament is not reasonable, resulting in chemical waste, for To solve the above problems, put forward in the research process of the drug concentration and filter press cooperative control. Through the analysis of the original flocculant adding device and filter aid agent, reagent solution preparation device has perfect device and automatic adding reagent addition, control strategy needs to be improved. The system uses the original preparation and adding device the reagent addition by solving the cooperative system. The slime water treatment process as a typical chemical reaction process, adding amount of main influence variable feeding concentration in concentration and filtration process of traditional Chinese medicine agent, feeding rate, bottom flow concentration, overflow concentration, filtration cycle and slime cake moisture, the process with strong coupling. The characteristics of nonlinear, large delay, it is difficult to establish the model through mathematical derivation, this paper proposes to establish the BP neural network of the reagent addition model, and the optimal APSO algorithm The amount of solving strategy. Through the analysis of the BP neural network principle and inference algorithm, variable slime water treatment effect based on the established 4 * 5 * 1 flocculant structure filter aid addition neural network model with 3 x 5 x 1 structure, and using the data of the 50 groups on the basis of network training. To achieve the goal of establishing pharmacy and site conditions add optimization optimization model to determine the optimal conditions, selected PSO algorithm to solve this optimization model of agents, and the use of inertia weight value to improve the PSO algorithm and the particle velocity decreases linearly flying adaptive strategy. In the design and operation of the Matlab platform program of APSO algorithm. In order to on line optimization algorithm of computing quantity, by using the S function call APSO program Simulink simulation platform, and through the implementation of joint operation with the PLC controller of OPC technology. This system The AB 1756-Control Logix PLC for collaborative controller, Advantech ACP4000 as the host computer, the hardware structure of the system was set up, prepared using RSLogix5000 controller program, Matlab/Simulink is selected as the APSO algorithm of online computing platform. In order to realize data exchange, MSG module is used to communicate with the original system controller, controller and Matlab using OPC interface technology, data communication between FT VIEW. This system uses the sensor to collect performance data by OPC technology, feedback to the Matlab APSO algorithm the corresponding variables, by the method of online optimal agent under the condition, adding amount of calculation, the optimization dosage by synergetic controller is returned to the technical agent system, after the original agent according to the optimized dosage action, achieve online collaborative optimization. System in Chengzhuang coal mine coal preparation plant stable operation Reliable, and through the analysis of the data of three months before and after the operation system, the total coal production also increased slightly, while PAC tons of slime drug consumption decreased from 2.453Kg/T to 2.341Kg/T, the economic index decreased from 0.182Kg/T to 0.172Kg/T. pharmaceutical consumption consumption tons of slime PAM drugs, reduced from 4.119 to 3.914. that the system not only ensured the production rate coal slurry treatment, and reduce drug consumption, improve the pesticides and the synergy between coal preparation plant and economic benefits.

【学位授予单位】：太原理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TD94

【参考文献】