Transformer fault dynamic early warning method based on hidden Markov model

Abstract

The invention discloses a transformer fault dynamic pre-warning method based on a hidden Markov model, which includes: (1) using the data of dissolved gas concentration in oil during the process of faulty transformer equipment from normal state to fault state and the oil concentration data of normal transformer equipment; Dissolved gas concentration data is used to train the hidden Markov model to obtain a transformer fault model M that fits each fault type m and transformer normal model M; (2) using transformer fault model M m With the normal model M of the transformer, find the model M′ that matches the dissolved gas concentration data of the transformer to be tested after linear interpolation processing, and predict the condition of the transformer to be tested according to the model M′, the current health status of the transformer to be tested and the data. Momentary state of health. This method can use the extracted dynamic features to predict the future operation status of the transformer, realize the dynamic early warning function of the transformer equipment, and has a wide application prospect in the maintenance and maintenance of the equipment.

Description

technical field [0001] The invention relates to the field of statistical analysis, in particular to a method for dynamic early warning of transformer faults based on a hidden Markov model. Background technique [0002] Dissolved gas in oil is an important indicator of transformer faults. Studies have shown that the electricity and heat released when transformer equipment fails will cause transformer oil to decompose and produce different dissolved gases in oil, including hydrogen, methane, ethane, ethylene, etc. By calculating and comparing the concentration and composition of the gas generated during the operation of the transformer, the potential risks of the equipment can be analyzed, and a preliminary judgment can be made on the health status of the transformer equipment. Combined with practical experience, relevant researchers have summarized some laws and judgment methods for different fault types, such as IEC 60599 (three ratios) method, Rogers (four ratios) method an...

AI Technical Summary

Problems solved by technology

However, this method still belongs to the ratio method. This kind of method has certain limitations in fault description, which is manifested in two aspects. First, the actual situation of transformer equipment faults is very complicated, and only a simple ratio is used to analyze it. There is a certain deficiency in the description, and the ratio operation cannot reflect the complex characteristics of gas indicators

Second, the essence of the ratio method is a threshold method, and these thresholds mostly come from the practical experience of researchers, which lack certain rigor. The fault discrimination near the threshold is often vague, and the ratio method has further improvement in the accuracy of discrimination. room for improvement

However, overall, the above methods still have certain limitations.

First, the existing technology only performs fault judgment on the current running state of the equipment, without considering the historical operation of the equipment

In fact, the historical operation data of equipment contains a lot of information, including the overall level of dissolved gas content in oil, the growth trend information of gas concentration, etc., and the evaluation of the current state only ignores the above information

Second, the existing technology cannot realize the early warning function of the equipment operation status

The existing technology is a static method that does not consider the time factor, and can only distinguish the normal state and fault state of the equipment, and cannot estimate the potential risks of the equipment

Images