The invention belongs to the technical field of battery detection, and particularly relates to a battery module internal
short circuit detection method and
system based on
magnetic field detection. Comprising the steps of constructing a three-dimensional
magnetic field sensing array; screening an effective observation window based on quasi-static logic, and triggering a magnetic sensor in the effective observation window to carry out magnetic induction intensity
data acquisition; decoupling of background
magnetic field and topology current-carrying contribution
magnetism is carried out on the original magnetic induction intensity data, and
residual magnetic field features are obtained; inputting the
residual magnetic field features into the constructed physical constraint neural
network model, and inverting the reconstructed
current density based on physical driving; and dividing the battery states into different safety levels by adopting a multi-level fault
response strategy based on the reconstruction
current density, and completing the internal
short circuit detection of the battery module. According to the invention,
short circuit detection in the battery module is realized based on magnetic field three-dimensional array
perception in combination with a physical constraint
deep learning model, and an optional implementation path is provided for early accurate diagnosis of short circuit faults in the battery module.