Real-time MPC Design Based on Machine Learning for a Diesel Engine Air Path System

Moriyasu, Ryuta; Ueda, Matsuei; Ikeda, Tokuhei; Nagaoka, Makoto; Jimbo, Takashi; Matsunaga, Akio; Nakamura, Toshihiro

doi:10.1016/j.ifacol.2018.10.119

Cited by 10 publications

(2 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…NMPC with polynomials models combined with a variety of constraint handling strategies has been shown to be solvable in real time [11]. Approximating NMPC control law with an artificial neural network also offers computationally feasible alternatives [20].…”

Section: Introductionmentioning

confidence: 99%

Simultaneous control of NOx, soot and fuel economy of a diesel engine with dual-loop EGR and VNT using economic MPC

Liu

Dizqah

Herreros

et al. 2021

Control Engineering Practice

View full text Add to dashboard Cite

Section: Introductionmentioning

confidence: 99%

Simultaneous control of NOx, soot and fuel economy of a diesel engine with dual-loop EGR and VNT using economic MPC

Liu

Dizqah

Herreros

et al. 2021

Control Engineering Practice

View full text Add to dashboard Cite

“…Model predictive control (MPC) handles MIMO system with constraints perfectly [9–13]. The MPC control system has a wide range of applications in the diesel engine control [14, 15]. In recent researches, the linear MPC theory has developed very well, however, as the non‐linear characteristic inside the system, the linear MPC is unable to achieve an accurate control result in some specific working condition [16, 17].…”

Section: Introductionmentioning

confidence: 99%

Integrated control of torque and emission of a diesel engine based on LPV‐MPC

Sun

et al. 2020

IET Control Theory & Appl

View full text Add to dashboard Cite

This study presents an integrated control scheme of a diesel engine which aims to ensure quick torque response while limiting the average NOx emission below the certain specification. To start with, the sensitivity analysis is performed to recognize the key factors determined to balance emission, torque and fuel consumption. Considering the strong nonlinearity of the diesel engine, a multi‐input multi‐output (MIMO) linear parameter varying (LPV) model is developed. Unlike the conventional state‐depended LPV modelling process, the proposed LPV model is built based on a kernel function with a novel form of state‐free predictive equation so that the system outputs are directly computed. The control objective of coordinating torque tracking and emission reduction is then accomplished by designing a model predictive controller that regulates the fuel injection quantity, fuel injection angle, exhaust gas recirculation (EGR) rate and the opening degree of variable geometry turbine (VGT) valve. Finally, the developed LPV model and the MPC controller are testified with a high‐fidelity commercial diesel engine model. The results show that the LPV‐MPC controller is able to satisfy the NOx emission specification and the engine torque tracks the desired reference fast in transients, while the control variables are suitably actuated within the actuator constraints.

show abstract

Machine learning-based reduced-order modeling and predictive control of nonlinear processes

Zhao

Zheng

Gong

et al. 2022

Chemical Engineering Research and Design

View full text Add to dashboard Cite

Real-time MPC Design Based on Machine Learning for a Diesel Engine Air Path System

Cited by 10 publications

References 13 publications

Simultaneous control of NOx, soot and fuel economy of a diesel engine with dual-loop EGR and VNT using economic MPC

Simultaneous control of NOx, soot and fuel economy of a diesel engine with dual-loop EGR and VNT using economic MPC

Integrated control of torque and emission of a diesel engine based on LPV‐MPC

Machine learning-based reduced-order modeling and predictive control of nonlinear processes

Contact Info

Product

Resources

About