Research And Applications Of Ant Colony Algorithm In Transformer Fault Diagnosis

The rapid development of China's electricity industry resulted into significant surging of the retainment of transformers, and the transformer-related operating data has ever been increasing. Thus there is an urgent need to provide a means for transformer fault diagnosis based on intelligent technologies. Referencing to the ant colony algorithm, in-depth research work is conducted on the diagnosing methods for transformer faults, which include:1.A transformer fault diagnosing system is designed by adopting expert systems. the system is composed of four modules: information collection module; data classification storage module;integrated management module; analysis and diagnose module. Analysis work is then performed on the transformer fault diagnosis system, showing that the system guarantees the high efficiency, reliability, and easy to maintenance.2.ACO has been successfully introduced to various optimization applications. Parpinelli and his colleagues proposed a new algorithm, named as Ant Miner, and used it into data mining. A refinement on the Ant-Miner algorithm is performed in this paper, avoiding the algorithm to fall into local optimal solution while significantly reducing the searching time. Thereby the overall convolution performance is enhanced. Meanwhile, based on digital simulation, the optimization of the main parameters of the algorithm is studied, which result is a provision of optimal set of parameters. Comparing result of CN2 with Ant-Miner proved that the refined algorithm provides better classification rule and more powerful predictability, bearing a reduced rule set and much more simple rules.3.The refined algorithm is applied on transformer fault diagnosing systems. And the analysis on the real application evaluated that the refined Ant-Miner algorithm is a feasible means for the transformer diagnosis.