EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Gain-Scheduling Offline MPC for Air Path qLPV System Control in Spark Ignition Engine Using Electric Supercharger

Rosana C. B. Rego, Camila O. De Lima, Matheus H. R. Miranda, Jony J. Eckert, Andre V. Bueno, Grace Deaecto, Ludmila C. A. Silva

IEEE Access · 2026 · 14 · 77050-77064

DOI: 10.1109/access.2026.3694810 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Access, 2026. Verfasst von Rosana C. B. Rego und 6 weiteren Autor:innen. Die Arbeit nennt 43 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Research Development Foundation-Federal University of Minas Gerais (FUNDEP) National Green Mobility and Innovation Program (MOVER)/Line V, São Paulo Research Foundation, Brazilian National Council for Scientific and Technological Development.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Electric superchargers eliminate turbo lag in spark ignition (SI) engines by providing instantaneous boost pressure response independent of exhaust gas energy. However, coordinating the throttle valve and eBoost motor to track driver torque demands poses significant control challenges due to highly nonlinear air path dynamics, strict actuator constraints, and real-time execution requirements on resource-constrained automotive Electronic Control Units (ECUs). This paper presents a quasi-Linear Parameter-Varying (qLPV) modeling and gain-scheduled offline Model Predictive Control (MPC) with integral action for air path regulation in SI engines equipped with electric superchargers. A qLPV model is derived through exact factorization of nonlinear throttle flow, volumetric efficiency, and compressor torque terms into three scheduling parameters. The controller gains are computed offline via Linear Matrix Inequality (LMI)-based optimization and scheduled online through trilinear interpolation, enabling real-time implementation. Simulation results on standardized (WLTC, US06) and real-world driving cycles demonstrate superior tracking performance, constraint satisfaction, and computational efficiency compared to PID, Fuzzy, and nonlinear MPC controllers. The proposed strategy achieves fuel savings exceeding 33% relative to baseline PID control while maintaining execution times compatible with automotive sampling rates, demonstrating its viability for real-time embedded implementation on production ECUs.

Förderung

  • Research Development Foundation-Federal University of Minas Gerais (FUNDEP) National Green Mobility and Innovation Program (MOVER)/Line V
  • São Paulo Research Foundation
  • Brazilian National Council for Scientific and Technological Development
  • Incentive Program for New Professors at Universidade Estadual de Campinas (UNICAMP)-New Faculty Incentive Program
  • UNICAMP
  • Federal Rural University of the Semi-Arid Region (UFERSA) for the Institutional Support provided through the Authorization for Postdoctoral Qualification
  • Article Processing Charge (APC) was funded by the Brazilian Federal Agency for Support and Evaluation of Graduate Education–CAPES

Zitieren

APA

Rego, R. C. B., Lima, C. O. D., Miranda, M. H. R., Eckert, J. J., Bueno, A. V., Deaecto, G., & Silva, L. C. A. (2026). Gain-Scheduling Offline MPC for Air Path qLPV System Control in Spark Ignition Engine Using Electric Supercharger. IEEE Access, 14, 77050-77064. https://doi.org/10.1109/access.2026.3694810

Vancouver

Rego RCB, Lima COD, Miranda MHR, Eckert JJ, Bueno AV, Deaecto G, et al. Gain-Scheduling Offline MPC for Air Path qLPV System Control in Spark Ignition Engine Using Electric Supercharger. IEEE Access. 2026;14:77050-77064. doi:10.1109/access.2026.3694810

BibTeX
@article{Rego2026,
  author = {Rego, Rosana C. B. and Lima, Camila O. De and Miranda, Matheus H. R. and Eckert, Jony J. and Bueno, Andre V. and Deaecto, Grace and Silva, Ludmila C. A.},
  title = {Gain-Scheduling Offline MPC for Air Path qLPV System Control in Spark Ignition Engine Using Electric Supercharger},
  journal = {IEEE Access},
  year = {2026},
  volume = {14},
  pages = {77050--77064},
  doi = {10.1109/access.2026.3694810}
}
Kostenlos starten →