AI Battery SOH Prediction Using Driving Data and BMS Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional technologies for predicting the state of health (SOH) of batteries have low prediction accuracy as they rely solely on battery information.
Innovation Solution
An apparatus and method that utilize an artificial intelligence (AI) model trained with a dataset including battery information, driving information, and SOH of a probe vehicle to predict the SOH of a battery in a target vehicle, with the option to correct the prediction using data from a battery management system (BMS).
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional battery SOH prediction methods are used, then the prediction process is simple, but the prediction accuracy is low
Solution Approach 1:
The patent merges multiple data sources (battery information, driving information, and SOH data from probe vehicles) into a unified training dataset for the AI model. This combination of diverse data types enables more accurate SOH prediction by capturing both battery degradation patterns and their relationship with actual driving conditions, resolving the contradiction between simple prediction methods and accurate results.
Solution Approach 2:
The patent introduces an AI model as an intermediary between raw battery data and SOH prediction results. This intermediary component processes complex relationships between battery parameters and degradation, enabling high accuracy prediction while managing system complexity through a dedicated predictive layer that can be trained and optimized independently.
2Measurement precision
If only battery information is used for prediction, then the data collection is simple, but the prediction accuracy is low
Solution Approach 1:
The patent combines battery information with driving information and probe vehicle SOH data to create a comprehensive training dataset. This merging of multiple data types increases the information content available for prediction, enabling the AI model to learn more accurate degradation patterns that account for real-world operating conditions.
Solution Approach 2:
The patent adds new dimensions to the prediction problem by incorporating driving information (speed, acceleration, route) and probe vehicle SOH data alongside traditional battery parameters. This dimensional expansion transforms the prediction from a single-dimension battery parameter analysis to a multi-dimensional assessment that captures the complex relationship between usage patterns and battery degradation.
3Measurement precision
If AI model training is implemented, then prediction accuracy improves, but the processing time increases
Solution Approach 1:
The patent performs AI model training in advance using historical battery information, driving information, and probe vehicle SOH data. This preliminary action creates a pre-trained model that can then make predictions quickly in real-time applications, separating the time-consuming training phase from the time-sensitive prediction phase to resolve the contradiction between accuracy and processing time.
Solution Approach 2:
The patent uses data from probe vehicles (other vehicles with similar batteries) to create training examples that replicate real-world degradation patterns. This copying of actual operational data from multiple sources provides the AI model with comprehensive training material without requiring extensive real-time data collection and processing for each prediction.
Data Source
AI summary
In an apparatus for predicting a state of health (SOH) of a battery and a method thereof, the apparatus includes a training dataset including battery information, driving information, and an SOH of a probe vehicle, trains an artificial intelligence (AI) model by use of the training dataset, and predicts the SOH of the battery corresponding to the battery information and driving information of the target vehicle based on the AI model, predicting the SOH of the battery provided in the target vehicle with predetermined accuracy.