State of Charge Prediction for Electric Loader Battery Based on Extreme Learning Machine

In order to accurately predict the state of charge (SOC) of electric loader battery, a multi-strategy improved dung beetle optimization algorithm (MIDBO) is proposed to optimize the SOC prediction method of the extreme learning machine (ELM). Firstly, principal component analysis (PCA) is used to screen the input features and reduce dimensionality. Secondly, logistic chaos mapping is introduced in the population initialization stage to improve the diversity and quality of the population. The dynamic spiral search strategy is utilized in the breeding stage of the dung beetle algorithm, and the Lévy flight mechanism is incorporated in the foraging stage to escape local optima. Finally, adaptive t-distribution variation and a dynamic selection strategy are employed to update the position of the dung beetle, enhancing the convergence speed of the algorithm. This approach enables accurate prediction of the electric loader battery SOC. Experiments using real vehicle data from electric loaders show that the accuracy of the MIDBO-ELM model reaches 99.985%, with an error of less than 5‰. The robustness of the model is verified by adding different types of noise to simulate measurement errors. Additionally, the accuracy of the MIDBO algorithm is validated through test functions, and the generalization capability of the algorithm for SOC prediction is confirmed using a public data set.