Dilute Combustion Control Using Spiking Neural Networks

Maldonado Puente, Bryan; Kaul, Brian; Schuman, Catherine; Young, Steven; Mitchell, Parker

doi:10.4271/2021-01-0534

Title: Dilute Combustion Control Using Spiking Neural Networks

Full Record
Other Related Research

Abstract

Dilute combustion with exhaust gas recirculation (EGR) in spark-ignition engines presents a cost-effective method for achieving higher levels of engine efficiency. At high levels of EGR, however, cycle-to-cycle variability (CCV) of the combustion process is exacerbated by sporadic occurrences of misfires and partial burns. Previous studies have shown that temporal deterministic patterns emerge at such conditions and certain combustion cycles have a significant influence over future events. Due to the complexity of the combustion process and the nature of CCV, harnessing all the deterministic information for control purposes has remained challenging even with physics based 0-D, 1-D, and high-fidelity computational fluid dynamics (CFD) models. In this study, we present a data-driven approach to optimize the combustion process by controlling CCV adjusting the cycle-to-cycle fuel injection quantity. Readily available data from in-cylinder pressure was used to train a spiking neural network (SNN) which learns the optimal way to manage fuel injection in order to reduce CCV while maintaining acceptable levels of fuel consumption. SNNs are particularly well suited for powertrain control applications due to their ability to be deployed on FPGA-based neuromorphic hardware which are small, inexpensive, and have a low power demand. The high-performance computing (HPC) resources of Oak Ridgemore »« less

Authors:

^[1];

^[1]; Mitchell, Parker ^[1]

Oak Ridge National Lab. (ORNL), Oak Ridge, TN (United States)

Publication Date:: Tue Apr 06 00:00:00 EDT 2021

Research Org.:: Oak Ridge National Lab. (ORNL), Oak Ridge, TN (United States)

Sponsoring Org.:: USDOE Office of Energy Efficiency and Renewable Energy (EERE)

OSTI Identifier:: 1785202

Grant/Contract Number:: AC05-00OR22725

Resource Type:: Accepted Manuscript

Journal Name:: SAE Technical Paper Series

Additional Journal Information:: Journal Volume: 2021; Journal Issue: 01; Journal ID: ISSN 0148-7191

Publisher:: SAE International

Country of Publication:: United States

Language:: English

Subject:: 33 ADVANCED PROPULSION SYSTEMS

Citation Formats


                    Maldonado Puente, Bryan, Kaul, Brian, Schuman, Catherine, Young, Steven, and Mitchell, Parker. Dilute Combustion Control Using Spiking Neural Networks.  United States: N. p., 2021. 
Web.  doi:10.4271/2021-01-0534.

Copy to clipboard


                    Maldonado Puente, Bryan, Kaul, Brian, Schuman, Catherine, Young, Steven, & Mitchell, Parker. Dilute Combustion Control Using Spiking Neural Networks.  United States.  https://doi.org/10.4271/2021-01-0534

Copy to clipboard


                    Maldonado Puente, Bryan, Kaul, Brian, Schuman, Catherine, Young, Steven, and Mitchell, Parker. Tue .  
"Dilute Combustion Control Using Spiking Neural Networks".  United States.  https://doi.org/10.4271/2021-01-0534.  https://www.osti.gov/servlets/purl/1785202.

Copy to clipboard


                    
@article{osti_1785202,

  title        = {Dilute Combustion Control Using Spiking Neural Networks},

  author       = {Maldonado Puente, Bryan and Kaul, Brian and Schuman, Catherine and Young, Steven and Mitchell, Parker},

  abstractNote = {Dilute combustion with exhaust gas recirculation (EGR) in spark-ignition engines presents a cost-effective method for achieving higher levels of engine efficiency. At high levels of EGR, however, cycle-to-cycle variability (CCV) of the combustion process is exacerbated by sporadic occurrences of misfires and partial burns. Previous studies have shown that temporal deterministic patterns emerge at such conditions and certain combustion cycles have a significant influence over future events. Due to the complexity of the combustion process and the nature of CCV, harnessing all the deterministic information for control purposes has remained challenging even with physics based 0-D, 1-D, and high-fidelity computational fluid dynamics (CFD) models. In this study, we present a data-driven approach to optimize the combustion process by controlling CCV adjusting the cycle-to-cycle fuel injection quantity. Readily available data from in-cylinder pressure was used to train a spiking neural network (SNN) which learns the optimal way to manage fuel injection in order to reduce CCV while maintaining acceptable levels of fuel consumption. SNNs are particularly well suited for powertrain control applications due to their ability to be deployed on FPGA-based neuromorphic hardware which are small, inexpensive, and have a low power demand. The high-performance computing (HPC) resources of Oak Ridge National Laboratory were used to run an evolutionary-based training approach for choosing the best SNN configuration that minimizes the size of the network while achieving the desired goal. The neuromorphic hardware with the optimized SNN deployed was connected to the rapid prototyping engine control system for real-time control implementation and tested on a single cylinder version of a GM LNF 4-cylinder engine. The results show a significant reduction of CCV with a small percentage of additional fuel used to stabilize the charge.},

  doi          = {10.4271/2021-01-0534},

  journal      = {SAE Technical Paper Series},

  number       = 01,

  volume       = 2021,

  place        = {United States},

  year         = {Tue Apr 06 00:00:00 EDT 2021},

  month        = {Tue Apr 06 00:00:00 EDT 2021}

}

Copy to clipboard

Journal Article:

Free Publicly Available Full Text

Accepted Manuscript (DOE)

Publisher's Version of Record

https://doi.org/10.4271/2021-01-0534

Other availability

Search WorldCat to find libraries that may hold this journal

Save / Share:

Export Metadata

Save to My Library

Similar Records in DOE PAGES and OSTI.GOV collections:

Evaluation of residual gas fraction estimation methods for cycle-to-cycle combustion variability analysis and modeling

Journal Article Maldonado, Bryan P. ; Kaul, Brian C. - International Journal of Engine Research

Cycle-to-cycle combustion variability in spark-ignition engines during normal operation is mainly caused by random perturbations of the in-cylinder conditions such as the flow velocity field, homogeneity of the air-fuel distribution, spark energy discharge, and turbulence intensity of the flame front. Such perturbations translate into the variability of the energy released observed at the end of the combustion process. During normal operating conditions, the cycle-to-cycle variability (CCV) of the energy release behaves as random uncorrelated noise. However, during diluted combustion, in either the form of exhaust gas recirculation (EGR) or excess air (lean operation), the CCV tends to increase as dilutionmore »« less
https://doi.org/10.1177/1468087420983087

Full Text Available
Next-Cycle Optimal Fuel Control for Cycle-to-Cycle Variability Reduction in EGR-Diluted Combustion

Journal Article Maldonado Puente, Bryan ; Kaul, Brian ; Schuman, Catherine ; ... - IEEE Control Systems Letters

Dilute combustion using exhaust gas recirculation (EGR) is a cost-effective method for increasing engine efficiency. At high EGR levels, however, its efficiency benefits diminish as cycle-to-cycle variability (CCV) intensifies. Here, cycle-to-cycle fuel control was used to reduce CCV by injecting additional fuel in operating conditions with sporadic misfires and partial burns. An optimal control policy was proposed that utilizes 1) a physics-based model that tracks in-cylinder gas composition and 2) a one-step-ahead prediction of the combustion efficiency based on a kernel density estimator. The optimal solution, however, presents a tradeoff between the reduction in combustion CCV and the increase inmore »« less
https://doi.org/10.1109/lcsys.2020.3046433

Full Text Available
CHARACTERISTICS OF CYCLE-TO-CYCLE COMBUSTION VARIABILITY AT PARTIAL-BURN LIMITED AND MISFIRE LIMITED SPARK TIMING UNDER HIGHLY DILUTED CONDITIONS

Conference Maldonado, Bryan ; Stefanopoulou, Anna ; Scarcelli, Riccardo ; ...

Diluted combustion with exhaust gas recirculation (EGR) has been widely employed to improve the fuel economy of spark ignition engines. The combustion kinetics, however, are affected and the flame propagation speed is decreased. In order to compensate for this adverse effect, the spark timing needs to be recalibrated to achieve maximum brake torque (MBT). At high levels of EGR dilution, the spark timing is constrained by two ignition limits: 1) the partial-bum limit where the spark timing is retarded from MBT and 2) the misfire limit where the spark timing is too advanced. This work uses a probabilistic framework tomore »« less
https://doi.org/10.1115/ICEF2019-7256
Near Zero Emissions at 50 Percent Thermal Efficiency

Technical Report

Detroit Diesel Corporation (DDC) has successfully completed a 10 year DOE sponsored heavy-duty truck engine program, hereafter referred to as the NZ-50 program. This program was split into two major phases. The first phase was called Near-Zero Emission at 50 Percent Thermal Efficiency, and was completed in 2007. The second phase was initiated in 2006, and this phase was named Advancements in Engine Combustion Systems to Enable High-Efficiency Clean Combustion for Heavy-Duty Engines. This phase was completed in September, 2010. The key objectives of the NZ-50 program for this first phase were to: Quantify thermal efficiency degradation associated with reductionmore »« less
https://doi.org/10.2172/1054511

Full Text Available
Multi-Objective Optimization for Size and Resilience of Spiking Neural Networks

Conference Dimovska, Mihaela ; Johnston, Travis ; Schuman, Catherine ; ...

Inspired by the connectivity mechanisms in the brain, neuromorphic computing architectures model Spiking Neural Networks (SNNs) in silicon. As such, neuromorphic architectures are designed and developed with the goal of having small, low power chips that can perform control and machine learning tasks. However, the power consumption of the developed hardware can greatly depend on the size of the network that is being evaluated on the chip. Furthermore, the accuracy of a trained SNN that is evaluated on chip can change due to voltage and current variations in the hardware that perturb the learned weights of the network. While effortsmore »« less
https://doi.org/10.1109/UEMCON47517.2019.8992983

Full Text Available

Similar Records