Training - Courses

SLMC: Model-Based Calibration

This hands-on, two-day course focuses on tools and techniques for using Design of Experiments, statistical modeling, and optimization to calibrate modern powertrain systems in MATLAB^® and Simulink^®. The course is designed for engineers who intend to calibrate and test engines, develop control algorithms, and simulate powertrain behavior. By the end of the course, attendees will be able to produce a set of optimal base calibration tables for a typical modern gasoline or diesel engine. Two subject variants, based on gasoline or diesel materials, are available. Topics include:

Working with Design of Experiments
- Setting up test plans
- Designing for constraints
- Applying classical, space-filling, and optimal methods
Data loading and handling

Loading, visualizing, filtering, and augmenting measured data

Response surface modeling

Creating one-stage and two-stage response-surface models using radial basis functions, polynomials, splines, and neural nets

Calibration generation

Setting up optimization problems
Reviewing, executing, filling, and exporting tables
Applying application-specific, multiobjective, and drive-cycle optimization

VIEW SCHEDULE and Register SHARE with Manager/Colleague

	Detailed course outline

Day 1 of 2
Problem Definition	Objective: Get an overview of the general effects of the engine actuators used in the course and review a block-diagram problem statement for the calibration problem to be solved over the two-day course. Definition of engine type and actuator technologies Definition of dynamometer test-setup Definition of actuator control ranges Qualitative review of engine actuator effects on emissions, performance, and fuel economy Preview of optimal calibration table solutions to be achieved by the end of the course
Design of Experiments	Objective: Develop an efficient Design of Experiments (DoE) test plan for the calibration problem definition using the minimal amount of expensive dynamometer test points. Set up test plan in MBCMODEL interface Enter engine actuator variable ranges Design DoE Apply design constraints Export DoE to CSV or MATLAB for dynamometer testing
Test Data Import, Filtering, and Meta Variable Setup	Objective: Import data into Model-Based Calibration Toolbox™ and filter it for bad data. Using the directly measured data, calculate meta variables that may not be directly measured (e.g., mechanical power, brake specific fuel consumption). Putting measured data into simple MBC-compatible Microsoft^® Excel^® format Importing data to the MBC Data Editor Filtering bad data by test and record Adding meta variables
Statistical Modeling of Measured Engine Data	Objective: Develop accurate statistical engine models required for calibration development. Export those models to MATLAB and Simulink so that others can reuse them in HIL and powertrain simulation models. Associating the engine data set with the engine test plan Generating statistical response models Removing outlier data Refining response models Visually reviewing response models Exporting response models to MATLAB and Simulink
Day 2 of 2
Set Up and Run Calibration Optimization Problem in CAGE Tool	Objective: Import models into the Calibration Generator tool (CAGE) in Model-Based Calibration Toolbox, set up empty lookup table structures, and enter the optimization problem definition. Import engine models into the CAGE tool Set up calibration lookup tables Set up calibration optimizer objective function and constraints for operating-point optimization and operating mode optimization Add table-gradient and sum constraints to operating-point optimization for drive-cycle-based optimization Run optimizations for point and sum
Fill Calibration Tables, Judge Results, and Export for ECU Implementation	Objective: Fill the optimal calibration tables, judge the sensitivity of the final calibrations and their validity relative to a validation data set, and export calibration tables for ECU implementation. Use the Fill Tables button to fill the empty calibration tables with optimal results Use the Tradeoff tool to judge the sensitivity of the final calibration tables and determine if additional constraints are needed in the optimization process Use the Dataset tool to import test points measured at optimal settings and compare to predicted engine emissions, performance, and fuel economy to validate the calibration tables Export calibration tables to ATI Vision, INCA .DCM, and MATLAB formats for ECU implementation
Student Walk-In Example Application Support	Objective: Work with the instructor to solve specific calibration problems using the skills learned in the formal course.

Prerequisites

Background in powertrain calibration required; basic knowledge of MATLAB recommended.

Course Length - 2 days