基于多尺度特征融合的光伏电站火灾信号智能感知系统  

Intelligent Perception System for Fire Signals in Photovoltaic Power Plants Based on Multi-scale Feature Fusion

作　　者：陈学云 CHEN Xueyun

机构地区：[1]大唐青海能源开发有限公司,青海海南州813000

出　　处：《电力系统装备》2024年第8期8-9,21,共3页Electric Power System Equipment

摘　　要：火灾信号感知是光伏电站安全监测中必不可少的一个环节,感知结果是光伏电站火灾管控的重要依据。现行系统在实际应用中错误感知比例比较高,因此,文章提出基于多尺度特征融合的光伏电站火灾信号智能感知系统。在系统硬件方面对烟雾传感器、温度传感器、报警器进行选型与设计;在系统软件方面采用最大最小值法对信号进行标准化处理,通过对火灾信号进行多尺度特征融合感知,实现系统设计。经实验证明,所设计系统对火灾信号错误感知比例不超过1%,在光伏电站安全监测领域具有良好的应用前景。Fire signal perception is an essential part of safety monitoring in photovoltaic power plants,and the perception results are an important basis for fire control in photovoltaic power plants.The current system has a higher proportion of false perception in practical applications.Therefore,a multi-scale feature fusion based intelligent perception system for photovoltaic power plant fire signals is proposed.Select and design smoke sensors,temperature sensors,and alarms in terms of system hardware;In terms of system software,the maximum minimum method is used to standardize the signal,and multi-scale feature fusion perception is applied to fire signals to achieve system design.Experimental results have shown that the proportion of false perception of fire signals in the design system does not exceed 1%,and it has good application prospects in the field of safety monitoring of photovoltaic power plants.

关 键 词：多尺度特征融合 光伏电站 火灾信号 智能感知 最大最小值法 标准化 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]