摘要
为提高水质监测系统覆盖范围并提升系统鲁棒性,设计一种以LoRa技术为通信方式,结合BP神经网络的水质监测系统。利用多节点采集水质的温度、pH值、总溶解固体(TDS)、氧化还原电位(ORP)等参数,通过无线传输技术将数据传输至汇聚节点,之后上传至云端物联网平台并实时下载到本地数据库,以支持网络模型处理和数据可视化分析,实现了多区域信息采集。再结合粒子群优化(PSO)算法优化BP神经网络的水质参数预测模型,实现对水质参数的预测补充,以提高系统的鲁棒性。通过实验验证系统水质信息采集的准确性以及参数预测模型的可靠性,结果表明,粒子群优化算法优化的BP神经网络模型对于pH值、温度、TDS和ORP四个参数的预测平均绝对百分比误差分别降低0.8269%、1.9475%、1.1039%和0.3125%,能够满足监测系统的需求。
In order to expand the coverage of the water quality monitoring system and enhance its robustness,a water quality monitoring system using LoRa technology as the communication method and combining with BP neural networks is designed.The multiple nodes are used to collect parameters such as temperature,pH,total dissolved solids(TDS),and oxidation-reduction potential(ORP)of the water quality.The wireless transmission technology is used to transmit the data to convergence node,then upload to the cloud-based IoT platform and download to the local database in real time,so as to support network model processing and data visualization analysis,and achieve multi-regional information collection.The water quality parameter prediction model of BP neural network is optimized by combining with the particle swarm optimization(PSO)algorithm to realize the prediction supplement of water parameter,so as to improve the robustness of the system.The accuracy of water quality information collection of the system and the reliability of the parameter prediction model are verified by the experiments.The results show that the BP neural network model optimized by the PSO algorithm can reduce the mean absolute percentage errors in predicting the four parameters of pH value,temperature,TDS,and ORP by 0.8269%,1.9475%,1.1039%,and 0.3125%,respectively,which can meet the requirements of the monitoring system.
作者
张凌飞
赵明玉
赵展文
陈博行
陈洋洋
ZHANG Lingfei;ZHAO Mingyu;ZHAO Zhanwen;CHEN Bohang;CHEN Yangyang(School of Intelligence Science and Engineering,Qinghai Minzu University,Xining 810007,China;EDA Technology Research Institute,Qinghai Minzu University,Xining 810007,China)
出处
《现代电子技术》
北大核心
2026年第4期33-41,共9页
Modern Electronics Technique
基金
青海省“昆仑英才·高端创新创业人才”项目。
关键词
水质监测
无线传输
LoRa技术
粒子群优化算法
BP神经网络
参数预测
water quality monitoring
wireless transmission
LoRa technology
particle swarm optimization algorithm
BP neural network
parameter prediction
