基于分布式优化与机器学习的电气火灾预警系统设计  

Design of Electrical Fire Warning System Based on Distributed Optimization and Machine Learning

作　　者：袁绪锋 YUAN Xufeng(Gansu Daxu Engineering Consulting and Design Co.,Ltd.,Lanzhou 730000,China)

机构地区：[1]甘肃大旭工程咨询设计有限公司,甘肃兰州730000

出　　处：《通信电源技术》2025年第1期31-33,共3页Telecom Power Technology

摘　　要：设计一种基于分布式优化与机器学习的电气火灾预警系统,旨在构建具备高实时性和高扩展性的智能预警体系。系统采用分布式计算、边缘计算与机器学习相结合的架构,依托多种传感器节点实时监测电气火灾早期征兆。数据通过高效传输协议传送至分布式计算平台进行处理,并利用Spark和Kafka实现低延迟数据处理。仿真验证表明,系统在正常与异常工况下的火灾预警准确率和召回率均超过95%,验证了系统的有效性和可靠性。An electric fire early warning system based on distributed optimization and machine learning is designed to build an intelligent early warning system with high real-time and high expansibility.The system adopts the architecture of distributed computing,edge computing and machine learning,and relies on various sensor nodes to monitor the early signs of electrical fire in real time.The data is transmitted to the distributed computing platform for processing by efficient transmission protocol,and low-latency data processing is realized by Spark and Kafka.The simulation results show that the fire warning accuracy and recall of the system under normal and abnormal working conditions are over 95%,which verifies the effectiveness and reliability of the system.

关 键 词：分布式优化 电气火灾预警系统 机器学习 

分 类 号：TP277[自动化与计算机技术—检测技术与自动化装置] TP18[自动化与计算机技术—控制科学与工程]