基于人工神经网络算法的水厂混凝投药控制系统研究与开发

发布时间：2018-03-04 08:17

本文选题：Elman神经网络　切入点：自来水厂　出处：《长江科学院院报》2017年05期 　论文类型：期刊论文

【摘要】：针对自来水生产投药工艺长滞后、非线性、多输入因子、不确定性、时变性、模糊性等特点,采用人工神经网络算法对周围环境自适应和自学习,研究和开发了一套用于水厂混凝投药的自动控制系统。系统以武汉市第一大水厂——宗关水厂为例,研究了Elman神经网络算法对控制系统混凝投药效果的影响,并基于OLE-DB开放性数据访问标准实现对WinCC工控系统样本数据读取和存储的预处理。系统主要包括投药工艺、数据查询、曲线生成、配药查询、报警日志、报警统计、药耗统计、波动评价、报警设置等功能模块,在宗关水厂的成功运行实现了混凝投药工艺生产运行参数的在线监视和全自动化运行。为水厂的安全生产提供了保障,达到了节约药耗、减少人工、降低操作人员劳动强度的目的。
[Abstract]:According to the production process of tap water dosing long lag, nonlinear, multi input factor, uncertainty, time-varying, fuzzy characteristics, artificial neural network algorithm for adaptive and self-learning environment, the research and development of a set of automatic control system for water coagulation system in Wuhan city. The first large - Zongguan waterworks plant as an example, studies the Elman neural network algorithm of control system of coagulant dosage effect, and pretreatment of OLE-DB open data access standard read on the WinCC control system based on sample data and storage. The system mainly includes the administration process, data query, curve generation, dispensing query, alarm log, alarm statistics. The fluctuation of drug consumption statistics, evaluation, alarm settings and other functions, in the successful operation of Zongguan water to realize online monitoring of mixed operation parameters and automatic coagulant dosing process. It provides a guarantee for the safe production of the water plant, which has achieved the purpose of saving the drug consumption, reducing the labor force and reducing the labor intensity of the operators.

【作者单位】：长江科学院仪器及自动化研究所;
【基金】：长江科学院技术开发和成果转化基金项目(CKZS2014004/YQ)
【分类号】：TP183;TU991.6

【相似文献】