Inferential Sensor Calibration via Engine Simulation Model
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The manual or semi-manual calibration of inferential sensing algorithms in diesel engines and other applications is time-consuming, expensive, and often results in substandard accuracy due to the need for extensive engineering and test cell measurements.
Innovation Solution
The use of a medium-fidelity, grey box, control-oriented model (COM) for turbocharged engines and after-treatment systems allows for the automatic calibration of inferential sensing algorithms by leveraging an engine simulation model to predict engine properties and minimize differences between inferred and measured signals, enabling systematic and optimal calibration.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If manual or semi-manual calibration of inferential sensing algorithms is performed, then the calibration can be achieved through direct engineering judgment, but the process becomes time-consuming and expensive requiring extensive test cell measurements
Solution Approach 1:
The patent applies preliminary action by pre-calculating optimal calibration parameters through offline system identification algorithms before actual engine operation. The calibration data is pre-processed and stored in lookup tables, allowing the ECU to quickly retrieve and apply calibration values without performing time-consuming manual adjustments during testing or operation.
Solution Approach 2:
The patent uses copying by creating a virtual model of the engine system that replicates the physical engine's behavior. The system identification algorithm develops a mathematical model that copies the engine's dynamic characteristics, allowing calibration to be performed on the virtual model rather than requiring extensive physical testing on the actual engine.
2Measurement precision
If manual manipulation of calibration parameters is performed to achieve acceptable performance, then the calibration can be adjusted to fit measured data, but the process requires significant engineering and test cell time
Solution Approach 1:
The patent implements feedback through an iterative system identification process that continuously compares the model's predicted outputs with actual measured engine data. The algorithm automatically adjusts calibration parameters based on the difference between predicted and measured values, refining the model's accuracy through multiple cycles of comparison and adjustment without manual intervention.
Solution Approach 2:
The patent applies parameter changes by systematically varying calibration parameters through the system identification algorithm to optimize model accuracy. The algorithm automatically modifies parameter values based on sensitivity analysis and error minimization criteria, replacing manual trial-and-adjustment with automated parameter optimization.
3Measurement precision
If extensive test cell measurements are conducted to calibrate inferential sensing algorithms, then sufficient accuracy can be achieved, but the cost and time requirements increase significantly
Solution Approach 1:
The patent applies partial action by using a selective subset of measured variables for calibration rather than requiring comprehensive measurement of all engine parameters. The system identification algorithm identifies which measurements are most critical for accurate calibration and focuses computational resources on those specific parameters, reducing the overall measurement burden.
Solution Approach 2:
The patent uses an intermediary approach by introducing a mathematical model as a mediator between raw sensor measurements and the final inferential sensing outputs. The system identification algorithm develops this intermediate model that transforms limited measurements into accurate predictions of difficult-to-measure quantities, reducing the need for direct expensive measurements.
Data Source
Figure 1
Figure 2
Figure 3A
AI summary
An inferential sensor module is incorporated into an engine simulation model. One or more parameters for the inferential sensor module are calibrated using one or more of engine measurement data and the engine simulation model. The calibration is performed such that a difference between an inferred signal predicted by the inferential sensor module and a signal measured on an engine is minimized. The inferential sensor module and the one or more calibrated parameters are loaded into an engine control unit in order to predict inferred variables.