Fault diagnosis of rotating machinery based on deep learning

In the current machinery manufacturing industry, rotating machinery occupies a very important position. Rotating machinery mainly refers to the machinery that can complete specific functions with the help of rotary action. The main vibration faults of rotating machinery include rotor unbalance, rotor misalignment, friction between moving and static parts and looseness of support parts. Generally, large rotating machinery is generally equipped with vibration monitoring protection and fault diagnosis system to ensure the safe operation of rotating machinery, but there are some limitations. In order to find the hidden danger in the machinery in time and avoid serious mechanical accidents, this paper combines the deep learning technology in machine learning to increase the accuracy and reliability of fault identification, reduce the risk of failure, and provide more guarantee for the production safety of machinery industry. Aiming at the main faults of rotating machinery, this paper proposes a fault recognition method of rotating machinery based on LSTM. Firstly, the sample data is screened and judged effectively, and then the abnormal data are classified by LSTM and softmax. Finally, the evaluation index and result of the fault category are obtained, which is compared with other classification algorithms to show the good applicability of the algorithm. At present, there are still some problems in the model, which need to be further discussed.