摘要
为提高电缆微量气体浓度预警精度与速度,该文设计基于光声光谱的电缆微量气体浓度标定和自动化预警系统。系统通过采集装置检测光声信号二次谐波信息,并利用标定仪标定气体浓度。预警端基于RBF神经网络启动异常诊断模型,提取统计特征并进行分类诊断。若发现异常,系统启动声光报警装置,以语音报警、灯光警示的方式,进入自动化预警状态。经实验测试后,所设计系统利用光声光谱技术,远程检测电缆隧道微量气体后,对微量一氧化碳气体的异常诊断结果精准,且对异常浓度气体的自动化预警时延在0.4 s之内。
To improve the accuracy and speed of cable trace gas concentration warning,a cable trace gas concentration calibration and automated warning system based on photoacoustic spectroscopy is designed.The system detects the second harmonic information of the photoacoustic signal through the acquisition device and scales the gas concentration with the calibration meter.The early warning terminal starts the abnormal diagnosis model based on the RBF neural network,extracts statistical features and makes classified diagnosis.If there is an abnormality,the system will activate the sound and light alarm device to enter the automatic warning state through voice alarm and light warning.After experimental testing,the designed system utilizes photoacoustic spectroscopy technology to remotely detect trace gases in cable tunnels.The abnormal diagnosis results of trace carbon monoxide gases are accurate,and the automatic warning time delay for abnormal concentration gases is within 0.4 seconds.
作者
高建国
GAO Jianguo(State Grid Shuozhou Electri Powre Suplly Company,Shuozhou 036002,China)
出处
《自动化与仪表》
2024年第3期84-87,92,共5页
Automation & Instrumentation
基金
国网山西省电力公司朔州供电公司科技项目(5205F0220007)。
关键词
光声光谱
电缆微量气体浓度标定
自动化预警
异常诊断
RBF神经网络
photoacoustic spectroscopy
cable trace gas concentration calibration
automated early warning
abnormal diagnosis
RBF neural network
作者简介
高建国(1970-),男,本科,高级工程师,研究方向为输变电设备运维检修。
