Design Of Intelligent Power Distribution Monitoring System And Research On Fault Early Warning

In recent years,my country’s communication industry has developed very rapidly,and the scale of the operator’s communication network has also grown.In the carrier’s communication network,the data center computer room is an indispensable and important hub that bears huge communication pressure.The power distribution system failure rate and the risk of security accidents in the data center computer room are also increasing,which also leads to the computer room.The difficulty of operation and maintenance and the cost of operation and maintenance are increasing rapidly.There are many reasons for equipment failure in the equipment room,such as: abnormal termination of various equipment,abnormal changes in environmental factors such as internal temperature and humidity in the equipment room,etc.The occurrence of faults is often caused by the accumulation of abnormal conditions.These abnormal conditions will be reflected in the various parameters of the equipment during operation.If these operating parameters can be accurately collected,potential laws can be discovered from them,and preventive work can be made.Can significantly reduce the probability of failure.In addition,in the current operation and maintenance work of many computer rooms,the operation and maintenance personnel are accustomed to using their own experience to set the alarm threshold.However,due to the large number of equipment and the complex operating environment,the use of such experience values will cause high leakage.The alarm rate and false alarm rate,the alarm result cannot provide effective information for the maintenance personnel of the computer room,and it also increases the difficulty and cost of maintenance.In response to the above-mentioned power distribution monitoring problems,this paper has done the following research work:(1)Taking the data center computer room power environment monitoring system as the background,it explained its research significance and development status at home and abroad,consulted a large number of documents and technical manuals,conducted a demand analysis for these problems,introduced STM32,RS-485,ucos-II,Modbus and other related technologies analyze the random forest algorithm and research status and serve as the main theoretical support of this article;(2)Designed and implemented an intelligent power distribution monitoring system with STM32 series chips as the core processor,using modular design ideas to implement embedded hardware circuits and software for the smallest system,data acquisition,data processing,communication and other modules;(3)A fault warning method based on improved random forest algorithm is proposed.Aiming at the characteristics of the unbalanced data collected by the system,an improved ADASYN algorithm is proposed and the sampling process adopts a mixed sampling strategy;for the traditional problems of false alarms,false alarms and inaccurate prediction accuracy caused by traditional experience setting thresholds,K-means is introduced.The algorithm generates a threshold,and based on the data set collected by the intelligent power distribution monitoring system,it verifies the fault early warning capability of the improved random forest algorithm;(4)According to the requirements,the test plan is proposed,and the intelligent power distribution monitoring system designed in this paper is integrated and tested according to the test plan,which verifies that the system has the advantages of good real-time performance,high accuracy,and low cost,which meets expectations and is actually put into application.