Research On Power Optical Fiber Fault Monitoring And Early Warning System

In recent decades,the demand for power supply has been increased and the orderly operation of power supply system has played an important role in supporting the development of various industries.And power optical fiber communication is an important guarantee for the orderly dispatch of all links in the power system.As the increasing number of power optical fibers and the extension of service hours,the failure rate of electric fiber is increasing,and the aging of electric fiber is getting more and more serious.As the number of power fiber is increasing,the use time is longer and longer,the failure rate of power fiber is higher and higher,the aging phenomenon of the line is becoming more and more serious,which makes the power fiber fault repair and line maintenance more and more difficult and cost.In order to ensure the smooth operation of power dispatching system,it is an urgent problem for the power industry to study the power fiber fault monitoring and early warning system.In this context,the research content of this paper mainly includes the following two points:(1)Design of power fiber fault monitoring equipment.We proposed a scheme and designed an optical fiber fault monitoring equipment,which combine the optical power meter and the optical time-domain reflectometer(OTDR).And we described the proposed scheme from two aspects,the method of hardware circuit designing and software programming.For the first part we described the working principle and design process of STM32 controller,optical power meter,OTDR and other modules in power optical fiber fault monitoring equipment.And then we described the designing of STM32 controller and upper computer software.(2)Research on the algorithm of power fiber optic line state warning.Because of many factors affecting the change of optical power,the 250 day optical power data of a power supply company in Jilin Province is divided into two time-domain data,namely,low frequency and high frequency.Then we adopted the Gated Recurrent Unit(GRU)and the Autoregressive Integrated Moving Average Model(ARIMA)algorithm to make a prediction respectively and formed a GRU-ARIMA optical power combined prediction model finally.We also used particle swarm algorithm to optimize the parameters of GRU-ARIMA combination model,and the prediction accuracy was improved by 65% compared with the previous algorithm model.After various tests,the equipment has good functions.When the fiber fails,the power fiber fault monitoring equipment can detect the fault point type and distance within 16 s,and the distance error is within 10 m.In the research of power fiber line state warning algorithm,a combined prediction model of GRU-ARIMA based on particle swarm optimization is formed.The value of root mean square error(RMSE)value is only 0.024,the proposed method can successfully predict the trend of power optical fiber optical power,and it is beneficial to the maintenance of power optical fiber lines.Hardware equipment has been installed in Baojia substation and Jinzhu substation in Jilin Province,and the server has been installed in the communication room of new building of Jilin national grid.So far,the system has been working stably for one and a half years.