Research Article

A Physics-Constrained Neural Network Framework for Surrogate Modeling and Multi-Objective Optimization of an RCCI Engine

Volume: 10 Number: 3 August 19, 2026

A Physics-Constrained Neural Network Framework for Surrogate Modeling and Multi-Objective Optimization of an RCCI Engine

Abstract

Reactivity Controlled Compression Ignition (RCCI) engines offer a promising pathway toward simultaneous reduction of NOx, soot, and fuel consumption; however, their inherently nonlinear input-output behavior makes systematic calibration challenging. This study proposes a physics-constrained neural network (PCNN) surrogate model for a single-cylinder, naturally aspirated RCCI engine fueled with isopropanol and n-heptane blends (compression ratio 17.5, bore × stroke 85 × 90 mm, displacement 510 cm³). The model maps six operator-controllable inputs, namely premixed fuel ratio, excess air ratio, engine speed, intake air temperature, injection advance, and fuel mass flow rate, to ten simultaneous engine outputs covering torque, brake-specific fuel consumption, CO, HC, CO₂, NO, soot, IMEP, peak pressure rise rate, and indicated thermal efficiency. A thermodynamic consistency constraint derived from the BSFC identity is embedded directly into the composite training loss, coupling the predicted torque and fuel consumption outputs through known engine speed and fuel flow inputs. The PCNN is benchmarked against Random Forest, Support Vector Regression, and a classical feedforward ANN under an identical 80/20 train–test protocol. The proposed PCNN achieved the highest average test R² of 0.9584 across all ten outputs, outperforming Random Forest, SVR, and ANN baselines with corresponding average R² values of 0.8664, 0.9332, and 0.8657, respectively. The PCNN delivered the best individual R² in nine of the ten target variables, while SVR slightly outperformed PCNN for CO prediction. The trained surrogate is subsequently employed within a multi-start Sequential Quadratic Programming framework to identify optimal RCCI operating conditions under four engineering scenarios: maximum performance, minimum emissions, balanced multi-objective operation, and maximum fuel economy. Results confirm that no single operating point simultaneously satisfies all objectives, and the proposed framework provides an efficient, physics-informed tool for navigating these trade-offs.

Keywords

References

  1. Gowthama Krishnan M, Rajkumar S, Devarajan Y, Rajiv A. A comprehensive review on advancement and challenges of renewable biofueled reactivity controlled compression ignition (RCCI) engine. J Energy Inst. 2024;113:101540. https://doi.org/10.1016/j.joei.2024.101540
  2. Karczewski M, Chojnowski J, Szamrej G. A review of low-CO2 emission fuels for a dual-fuel RCCI engine. Energies. 2021;14(16):5067. https://doi.org/10.3390/en14165067
  3. Savaş A, Uslu S. Optimization of diethyl ether-modified sesame biodiesel–diesel fuel blends for enhanced engine performance and emission mitigation. Int J Automot Sci Technol. 2025;9(3):294–304. https://doi.org/10.30939/ijastech..1724235
  4. Halis S, Kocakulak T. RSM based optimization of lambda and mixed fuel concentration parameters of an LTC mode engine. Energy. 2024;306:132550. https://doi.org/10.1016/j.energy.2024.132550
  5. Uyumaz A, Solmaz H. Emme havası giriş sıcaklığı ve ön karışımlı yakıt oranının RCCI yanma karakteristiklerine ve motor performansına etkileri. J Polytech. 2017;20(3):689–698. https://doi.org/10.2339/politeknik.339401
  6. Poorghasemi K, Khoshbakhti Saray R, Solmaz H, Mousavi SM, Zehni A. Multi-dimensional 3D-CFD/chemical-kinetics modeling of natural gas–diesel RCCI engines: combustion phasing, efficiency and emissions. Engineering Perspective. 2026;6(1):69–82. https://doi.org/10.64808/engineeringperspective.1802746
  7. Moradi J, Gharehghani A, Aghahasani M. Application of machine learning to optimize the combustion characteristics of RCCI engine over wide load range. Fuel. 2022;324:124494. https://doi.org/10.1016/j.fuel.2022.124494
  8. Bogar E, Arabaci E, Halis S, Uyumaz A. Modeling and optimization of performance and emissions in a gasoline-isopropanol SI engine: multi-model prediction and a PID-based search algorithm. Sci Rep. 2026;16:13568. https://doi.org/10.1038/s41598-026-44323-9

Details

Primary Language

English

Subjects

Automotive Mechatronics and Autonomous Systems, Automotive Combustion and Fuel Engineering, Automotive Engineering (Other)

Journal Section

Research Article

Publication Date

August 19, 2026

Submission Date

April 21, 2026

Acceptance Date

June 25, 2026

Published in Issue

Year 2026 Volume: 10 Number: 3

APA
Sanci, M. E. (2026). A Physics-Constrained Neural Network Framework for Surrogate Modeling and Multi-Objective Optimization of an RCCI Engine. International Journal of Automotive Science And Technology, 10(3), 572-585. https://doi.org/10.30939/ijastech..1935208
AMA
1.Sanci ME. A Physics-Constrained Neural Network Framework for Surrogate Modeling and Multi-Objective Optimization of an RCCI Engine. IJASTECH. 2026;10(3):572-585. doi:10.30939/ijastech.1935208
Chicago
Sanci, Muhammet Emre. 2026. “A Physics-Constrained Neural Network Framework for Surrogate Modeling and Multi-Objective Optimization of an RCCI Engine”. International Journal of Automotive Science And Technology 10 (3): 572-85. https://doi.org/10.30939/ijastech. 1935208.
EndNote
Sanci ME (August 1, 2026) A Physics-Constrained Neural Network Framework for Surrogate Modeling and Multi-Objective Optimization of an RCCI Engine. International Journal of Automotive Science And Technology 10 3 572–585.
IEEE
[1]M. E. Sanci, “A Physics-Constrained Neural Network Framework for Surrogate Modeling and Multi-Objective Optimization of an RCCI Engine”, IJASTECH, vol. 10, no. 3, pp. 572–585, Aug. 2026, doi: 10.30939/ijastech..1935208.
ISNAD
Sanci, Muhammet Emre. “A Physics-Constrained Neural Network Framework for Surrogate Modeling and Multi-Objective Optimization of an RCCI Engine”. International Journal of Automotive Science And Technology 10/3 (August 1, 2026): 572-585. https://doi.org/10.30939/ijastech. 1935208.
JAMA
1.Sanci ME. A Physics-Constrained Neural Network Framework for Surrogate Modeling and Multi-Objective Optimization of an RCCI Engine. IJASTECH. 2026;10:572–585.
MLA
Sanci, Muhammet Emre. “A Physics-Constrained Neural Network Framework for Surrogate Modeling and Multi-Objective Optimization of an RCCI Engine”. International Journal of Automotive Science And Technology, vol. 10, no. 3, Aug. 2026, pp. 572-85, doi:10.30939/ijastech. 1935208.
Vancouver
1.Muhammet Emre Sanci. A Physics-Constrained Neural Network Framework for Surrogate Modeling and Multi-Objective Optimization of an RCCI Engine. IJASTECH. 2026 Aug. 1;10(3):572-85. doi:10.30939/ijastech. 1935208


International Journal of Automotive Science and Technology (IJASTECH) is published by Society of Automotive Engineers Turkey

by.png