Management method for energy storage batteries of railway machine room

The invention discloses a railway machine room energy storage battery management method, and relates to the technical field of energy storage batteries, and the method comprises the steps: collecting the voltage, current and SOC of a battery as battery state parameters; the collected battery state parameters are preprocessed; constructing a battery health state feature vector including voltage, current and state SOC; obtaining a battery health state prediction model by training a multi-layer perceptron neural network according to the constructed battery health state feature vector; modeling the voltage time sequence data by adopting an ARIMA (Autoregressive Integrated Moving Average) time sequence analysis method to obtain a voltage time sequence model; performing hyper-parameter optimization on the battery health state prediction model based on a Bayesian algorithm; and integrating the hyper-parameter optimized battery health state prediction model and the voltage time sequence model to obtain a final batter