Ressearch On Optimization Strategy For Wireless Temperature Measurement System Of Substation

Substation safety issue has been the focus of attention in the industry, and there exists high average failure rate problem of power grids in China. More safe and reliable substation monitoring system becomes an integral part of the substation operation system. In particular, as an important index parameters of safe operation of power network equipment-temperature monitoring is becoming more and more important. However, the existing design model of temperature monitoring system has some issues in temperature measurement, display and data storage, such as not real time, high redundancy, and difficult to extend. The studies of the performance optimization strategy of temperature online monitoring system in transformer substation are as follows.(1)Multithreaded collection algorithm of wireless temperature measuring system in substation is studied. Existing multithreaded sampling schemes are mainly implemented by lock-based producer-consumer model with queue, but the model have its own problem because of its implementation are based on lock and cannot fully utilize the advantage of multi-core CPU. Based on a comparative study of block and non-block data structure, this article presented a lock-free data structure and a new algorithm based on double CAS operations. When the structure is used as a data buffer of producer-consumer model, the system has better performance than traditional models.(2) The substation temperature update algorithm with tree data structure is studied. Traditional temperature updating algorithm used the strategy of concentrating measuring points temperature data, one times refresh, updating all temperature of measuring points strategy, is effective in smaller data scale. But for large scale substations which have a lot of measuring points, due to the temperature sensor information are all updating at once, there will be large amounts of data to update which will lead to more system resources and network bandwidth consuming and cause system performance degradation, even feign death, crashes and other problems. This article studied the bottleneck problems updating algorithm:effects of dirty points coverage are found on these algorithms, new temperature update policy is proposed based on write-back of dirty data, then using Bitmap methods, Hash list, and B-tree method implements strategy. We find that coverage rate of dirty measuring points even around 90% still have good efficiency.(3) The issues of nonlinear temperature data access and store in database is studied. Relational databases store the nonlinear relationship between data generally using foreign keys, multiple table joins etc. These methods not only can’t exert advantage of relational database query and wasting large amounts of storage space and lead to data redundancy. Based on the study of several conventional like path notation, adjacency list representation, representation of closures and nested set representation algorithm, this article present an improved representation of nested collections. The method has greatly improved the efficiency of accessing data, such as add, delete, modify, search nodes, and has an interface which is easy to use.This article design and deployment the HCES-8011 temperature online monitoring system of transformer substation which has features of measurement point temperature real-time on-line monitoring and discovery of failure points, then prove the effectiveness of the proposed algorithm for data collection, processing and storage based on this system.Research results show that schemes proposed in this paper reduce the bandwidth of the system, improve the timeliness and responsiveness of the system, reduce data redundancy and save a lot of storage space.