Research On Remote Online Monitoring System Of Dry Type Transformer

Dry-type transformers have good fire-proof and explosion-proof performance.They can be applied to public places such as commercial centers and residential quarters,and can meet the power supply requirements of the industry.Therefore,their power supply reliability requirements are high,but the current transformer industry monitors dry-type transformers.Most of them use thermostats to collect information and only meet the monitoring needs of power customers.Based on the above background,this thesis proposes a dry-type transformer monitoring system,in which the cloud server is applied to realize the real-time monitoring of multiple dry-type transformers remotely.The main research contents are as follows:Firstly,the requirements of the system lower position machine are analyzed.The selected monitoring data is the winding temperature and the three-phase electrical parameters.The hardware circuit design is carried out with the PIC single chip as the main control chip,and the real-time data acquisition,display and control of the fan are realized by the lower computer software program,sound and light alarm,protection trip and other functions;wireless communication module selected by the company developed a combination of GPRS and GPS technology USR-GPRS232-7S3 module,through the AT command and the company’s cloud server to establish a connection to achieve remote data transmission.Secondly,the improved SVR and RBF algorithm is used to predict the dry-type transformer winding temperature.The algorithm dynamically updates the particle’s velocity and position through a weighted adaptive particle swarm optimization algorithm(DAPSO)and introduces backward prediction.The factor optimizes the SVR,and then the improved SVR algorithm is used as the initial structure of the RBF model for temperature prediction.The project was built with VS2012 C++.The winding load voltage,current,active loss,ambient temperature and winding temperature value at the previous moment were taken as the influencing factors.The improved RBF algorithm model of PSO_RBF and DAPSO_SVR was established for prediction and comparison.The results show that the improved RBF algorithm is more accurate for the dry-temperature prediction.Finally,C#is selected as the remote monitoring interface development software,and the B/S architecture is established to complete the development of the dynamic Web page.The program is remotely transmitted,displayed,stored,and queried by calling the DLL(Dynamic Link Library)command provided by the company.The design realizes the remote monitoring function;completes the prototype debugging of the monitoring terminal,performs the no-load experiment on the transformer,and proves that the data acquisition accuracy of the monitoring terminal reaches the national standard through the analysis of the experimental data.The monitoring system combines the monitoring terminal of the dry-type transformer with the cloud server,so that the staff can understand the operating state of the transformer without entering the site,facilitating the large-scale unified management of the device,and facilitating the early prediction of the fault.And take measures to minimize losses,which is of great significance for enterprises to achieve unattended and industrial automation.