Data-model alliance network for the online multi-step thermal warning of energy storage system based on surface...

Area of Science:

• Energy storage systems
• Thermal management
• Artificial intelligence in engineering

Background:

• Battery energy storage systems are crucial but prone to high-temperature explosions.
• Existing temperature prediction methods are either model-driven or data-driven.
• Accurate thermal prediction is vital for preventing battery failures.

Purpose of the Study:

• To develop an advanced online multi-step-ahead thermal warning structure for battery systems.
• To integrate model-driven and data-driven approaches for improved temperature prediction.
• To provide early warning of critical temperature thresholds to prevent explosions.

Main Methods:

• Established a two-node electrothermal model and a multi-scale long short-term memory network.
• Formulated a Data-Model Alliance Network (DMAN) for surface temperature diffusion.
• Employed an improved adaptive boosting algorithm to enhance model integration.
• Integrated a Data-Model Alliance Module (DMAM) and Multi-Step-Ahead Thermal Warning Network (MATWN).

Main Results:

• The DMAN effectively predicts surface temperature diffusion.
• The integrated approach enhances the accuracy of thermal prediction.
• The system provides reliable multi-step-ahead thermal warnings.
• Experimental results demonstrate the technique's effectiveness and progressiveness.

Conclusions:

• The proposed DMAN offers an advanced solution for battery thermal management.
• This integrated approach significantly improves early warning capabilities for temperature excursions.
• The technique enhances the safety and reliability of battery energy storage systems.