System Simulation for Robust Calibration and Diagnostics

Cummins Inc. is the world’s largest independent diesel engine manufacturer, celebrating their 100th year of operation in 2019. Cummins manufactures and supplies engines ranging in displacement from 2.8L to 95L using a variety of fuels to a diverse customer and applications base across the globe. The combination of tightening global emissions standards, increased customer expectations with regards to reliability, and demands for improved fuel economy from production line to Emissions-Useful-Life (EUL) presents a complicated set of technical challenges across all their product lines.

To help meet these technical challenges, Cummins is utilizing MathWorks tools to perform simulations both in a hardware-in-loop (HIL) environment as well as using a fully virtual model-in-loop (MIL)/software-in-loop (SIL) environment with engine and aftertreatment plant models. Using these models, Cummins can perform both sub-system and system-level performance characterization and robustness testing, which is typically expensive, difficult, or impossible to accomplish in the real world. This simulation data is used to provide guidance to programs on controller tuning and to inform decisions on system performance and diagnostic robustness.

System simulation (engine, aftertreatment, and controls) use cases for performance and diagnostic validation using MathWorks tools are demonstrated in this presentation. Evaluation of model fidelity, identification of customer clusters, noise factor robustness testing, system simulation, and evaluation of performance and diagnostic capability use a number of tools to deliver efficient virtual product validation. Using examples from a recent program, Cummins shares how MathWorks tools are being applied in support of continuous product improvement efforts.