基于TX2的污水COD软测量装置研究与实现  

Research and Implementation of Sewage COD Soft Measurement Device Based on TX2

作　　者：廉小亲[1,2] 陈群 王俐伟 魏伟[1,2] LIAN Xiao-qin;CHEN Qun;WANG Li-wei;WEI Wei(School of Computer and Information Engineering,Beijing Technology and Business University,Beijing 100048,China;China Light Industry Key Laboratory of Industrial Internet and Big Data,Beijing Technology and Business University,Beijing 100048,China)

机构地区：[1]北京工商大学计算机与信息工程学廉院,北京100048 [2]北京工商大学中国轻工业工业互联网与大数据重点实验室,北京100048

出　　处：《计算机仿真》2020年第9期194-198,316,共6页Computer Simulation

基　　金：北京市自然科学基金北京市教委联合资助项目(KZ201810011012)。

摘　　要：针对污水处理过程中,传统测量方法对化学需氧量(Chemical Oxygen Demand,COD)测量耗时较长、成本高并且容易造成产生误差,需要经常标定等问题,研究了一种结合自组织特征映射(Self-organizing Map,SOM)和径向基函数(Radial Basis Function,RBF)神经网络的COD软测量算法并进行仿真实现,给出了COD软测量装置的总体设计方案及其组成,并阐述了基于STM32的水质参数采集模块和基于Jetson TX2的软测量模块的硬件、软件设计思路及实现方法。测试结果表明:上述测量装置能够实时在线预测COD,具有较高的预测精度,为污水COD参数检测提供了一个快速、低成本和稳定可靠的解决方案。Previous methods for sewage treatment are time-consuming and costly due to the measurement of Chemical Oxygen Demand(COD),which is prone to error and requires frequent calibration.In this work,we pro⁃posed a COD soft measurement algorithm which combines self-organizing Map(SOM)and Radial Basis Function(RBF)neural network.We first presented the overall scheme and the components of COD soft measurement instru⁃ment,then described the hardware and software design.Specifically,the implementation of STM32-based water qual⁃ity parameter collection module and Jetson TX2-based soft measurement module was discussed.The test results show that the measurement device can predict COD online in real time with high prediction accuracy,which provides a fast,low-cost,stable,and reliable solution for COD parameter detection.

关 键 词：化学需氧量 神经网络 软测量 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]