Local Coolant Flow Modeling for Heat Exchanger Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing thermal management systems in automobiles inaccurately predict coolant flow at individual heat exchange components due to the complexity of these components, leading to insufficient accuracy when using overall flow models.
Innovation Solution
Construct a local flow model by obtaining a target physical model from thermal management system data, calculating second flow data at specific components, and training a preset model with characteristic parameters to determine accurate local coolant flow.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If an overall flow model of the thermal management system is used to predict coolant flow, then the prediction covers the entire system, but the accuracy at individual heat exchange components is insufficient
Solution Approach 1:
The patent divides the thermal management system into multiple independent local flow models, each corresponding to a specific heat exchange component. Instead of using a single overall flow model, the system creates segmented models that can independently predict coolant flow at each component, thereby improving local prediction accuracy while maintaining manageable complexity through modular design.
Solution Approach 2:
The patent implements local quality by developing component-specific flow models tailored to the characteristics of each heat exchange component. Each local flow model is trained on data specific to its corresponding component, enabling predictions that reflect local flow conditions rather than averaging across the entire system, thus achieving higher precision at individual components.
2Productivity
If the thermal management system uses a simplified approach for flow prediction, then the calculation is faster, but the accuracy of local coolant flow determination deteriorates
Solution Approach 1:
The patent applies preliminary action by pre-training multiple local flow models offline using historical data and component characteristics. Once trained, these models can rapidly predict coolant flow at each heat exchange component without requiring complex real-time calculations, thus achieving both high calculation efficiency and accurate local flow predictions during system operation.
Data Source
AI summary
A method for constructing a local flow model includes: obtaining a target physical model corresponding to a thermal management system based on acquired first flow data of the thermal management system; calculating second flow data of coolant at a target heat exchange component in the thermal management system based on the target physical model; and obtaining a local flow model for the coolant at the target heat exchange component by training a preset model according to the second flow data and a target characteristic parameter for controlling operation of the thermal management system corresponding to the second flow data.

