Research on Health Status of Mobile Energy Storage Equipment Based on Random Forest and Long Term Memory Network algorithm

Random Forest (RF) and Long Term Memory Network (LSTM) are used to assess the health status of mobile energy storage devices in this paper. Firstly, the RF model is used to select the characteristic and the initial classification of the operation data. Then, the key characteristics of the equipment are selected, so that the data can be compressed and extracted effectively. Then, the LSTM model can be used to forecast the condition of the equipment dynamically. In particular, LSTM is able to capture long term dependencies in the process of operation of equipment, and provide more precise prediction results for RUL. The results show that the RF + LSTM model is superior to the conventional one in many aspects. The accuracy of RF + LSTM is improved by 10%, and the real time is reduced by 15%. Furthermore, this method has great adaptability, and it can be used to analyze the data of all types of devices. What's more, the proposed algorithm is scalable and flexible, which can satisfy the demands of the modern intelligent monitoring system in terms of real-time, precision and robustness.