Industrial large model-based new energy equipment health state estimation method

Objective To address the limitations of current health state estimation methods for naval vessel energy equipment, such as insufficient cross-device generalization ability and high model adaptation complexity, this study proposes a new energy equipment health state estimation method based on industrial large model. By developing a unified large - model framework and employing parameter-efficient fine-tuning, it can quickly build dedicated estimation models by utilizing the operating data of different equipment, achieving high-precision health state estimation results. Methods A large industrial model is constructed based on the pre-trained Transformer architecture, combined with the attention mechanism and low-rank parameter fine-tuning technology. While retaining the general knowledge of the large model, the intelligent health state estimation suitable for a variety of new energy equipment can be realized by dynamically adjusting key parameters to adapt to multiple types of equipment tasks. Results In the validation experiments of two typical new energy devices, lithium battery and fuel cell, the model showed excellent estimation accuracy and generalization ability. Conclusion The method based on industrial large model can effectively improve the accuracy and reliability of health state estimation of new energy equipment, and provide a new technical means for equipment health management.