Engine Model State Reuse for Multi-Solver Vehicle Calibration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current engine modeling techniques require sequential running of solvers, which can be time-intensive and computationally costly, especially for iterative processes, making it difficult to determine multiple engine parameters simultaneously, and often necessitate multiple controllers, increasing hardware and maintenance costs.
Innovation Solution
Implementing a system where multiple solvers can be run nearly simultaneously using a single engine model by incorporating an input selector unit to manage solver priority and previous engine states, allowing solvers to resume calculations from previous states, thus reducing the need for separate controllers and improving computational efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple solvers are run sequentially through the engine model, then each solver can determine desired parameters accurately, but the computational time and complexity increase significantly, taking 20 minutes or longer
Solution Approach 1:
The patent applies preliminary action by maintaining previous engine states between solver executions. Instead of starting each solver from scratch, the system preserves the engine model state after each solver completes, allowing subsequent solvers to resume from where they left off. This dramatically reduces computational time while maintaining accuracy, as solvers don't need to recompute already-determined parameters.
Solution Approach 2:
The patent implements continuity of useful action by keeping the engine model in a continuous state across multiple solver operations. The engine model state is maintained and carried forward between solvers, ensuring that the computational work done by each solver contributes to the overall process without redundant recalculations. This creates a seamless workflow where solvers operate continuously on the same model instance.
2Measurement precision
If multiple solvers are run sequentially, then comprehensive engine parameters can be determined, but multiple controllers are required, increasing hardware costs and maintenance complexity
Solution Approach 1:
The patent applies universality by designing a single engine model that can serve multiple solvers simultaneously. The engine model is configured to handle various solver types (steady-state, transient, optimization solvers) and can determine different engine parameters through its flexible state management. This multi-functional approach eliminates the need for separate controllers for each solver, reducing hardware complexity while maintaining comprehensive parameter determination capabilities.
3Measurement precision
If iterative processes are used by solvers to determine parameters, then accurate results are achieved, but the computational intensity and time requirements increase dramatically
Solution Approach 1:
The patent reduces computational energy by performing preliminary action through state preservation. Previous engine states are maintained and reused by subsequent solvers, preventing redundant iterative computations. This approach allows iterative processes to converge faster or be skipped entirely when states are already available, significantly reducing the computational energy required while maintaining accurate parameter determination.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Systems and methods for modelling an engine in a vehicle are provided. A method can include determining a first solver for modelling the engine. The first solver can include a first routine configured for determining one or more parameters of the engine. The method can further include obtaining a first set of inputs associated with the first solver. The first set of inputs can include one or more inputs obtained from the first solver and one or more previous engine states. The method can further include modelling the engine based at least in part on the first set of inputs to generate a first set of outputs. The first set of outputs can include one or more parameters of the engine and one or more engine states. The method can further include providing the one or more parameters to the solver.