Fault Diagnosis of Dry-type Air-Core Reactor Based on Temperature Field with Multi-Component Normal Distributions

In the power system, dry-type air-core reactor is an important equipments, and its fault can bring great hidden danger to whole system. Most of the existing Eault diagnosis methods of the reactor can be divided into three categories: current-based methods, current-based additional effects methods, and temperature-field-based methods. Generally, for constructing the temperature field, the sensor needs to meet the position constraint, that is, all sensor remains the same location on each equipment. However, it is difficult to satisfy this constraint in real work scenarios. To solve this problem. We propose a new method which uses a multi-component normal distribution to fit the temperature field, and then compares the measured temperature with the ideal temperature to infer the running status of the dry-type air- core reactor. Because the RFID chip have stable performance and is wireless passive. We use this chips to collect the temperature of the reactor. The verification experiments show that our proposed method is feasible and effective.