Intelligent gas relay for transformer and method

The invention relates to the field of intelligent detection, and particularly discloses an intelligent gas relay for a transformer and a method. A pressure value, a temperature value and an oil level value which are acquired are monitored in real time through a sensor group; and time sequence correlation collaborative analysis of the pressure value, the temperature value and the oil level value is carried out by using a data processing and analysis algorithm based on a deep learning neural network, so that whether the transformer has a fault or not is automatically judged according to the real-time operation state of each parameter in the transformer. And an early warning signal prompt is generated when the transformer has a fault. Therefore, fault problems can be found in time and early warning can be carried out, so that corresponding countermeasures can be taken, the timeliness of fault diagnosis is improved, meanwhile, the problem of false alarm caused by gas accumulation in a traditional scheme can be av