A Comparison of 1D-3D Co-Simulation and Transient 3D Simulation for EGR Distribution Studies

Pavlos Dimitriou, Calogero Avola, Richard Burke, Colin Copeland, Niall Turner

Research output: Contribution to conferencePaper

3 Citations (Scopus)

Abstract

Computational modeling, an important task for design, research and development stages, is evolving fast with the increase of computational capabilities over the last decades. One-dimensional (1D) CFD simulation is commonly used to analyze the flow rates and pressures of an entire fluid system of interconnected parts such as pipes, junctions, valves, and pumps. In contrast, three-dimensional (3D) CFD simulation allows detailed modeling of components such as manifolds, heat exchangers, and combustion cylinders where the flow contains significant 3D effects. Coupling a 1D model with a 3D domain potentially offers the benefits of both simulation strategies in one co-simulation approach. The present study provides a deep understanding of the co-simulation approach by listing all necessary steps need to be followed before and during the coupling of the 1D and 3D simulation software. It analyses the simulation and convergence time requirements based on the 3D model mesh quality and compares this approach with the current 1D-3D uncoupled approach followed in the industry. The outputs of both simulation approaches are then compared with experimental results. The co-simulation time mainly depends on the mesh quality of the 3D domain and the number of inner iterations per time-step which is entirely determined by the nature and complexity of the simulation. The co-simulation time per engine cycle is almost identical to the uncoupled approach. However, it was found that the number of cycles required for convergence in the coupled approach is nearly double than the uncoupled approach. The comparison between the two simulation approaches and the experimental results demonstrated the very 3D nature of the flows, the sensitivity of the uncoupled approach to input conditions and the sensitivity of co-simulation to the averaged boundary conditions transferred from the 1D model back to the 3D domain.
Original languageEnglish
Publication statusPublished - Oct 2016
EventASME ICEF 2016 - Greenville, USA United States
Duration: 9 Oct 201612 Oct 2016

Conference

ConferenceASME ICEF 2016
CountryUSA United States
CityGreenville
Period9/10/1612/10/16

Fingerprint

Computational fluid dynamics
Engine cylinders
Heat exchangers
Pipe
Flow rate
Boundary conditions
Pumps
Engines
Fluids
Industry

Cite this

Dimitriou, P., Avola, C., Burke, R., Copeland, C., & Turner, N. (2016). A Comparison of 1D-3D Co-Simulation and Transient 3D Simulation for EGR Distribution Studies. Paper presented at ASME ICEF 2016, Greenville, USA United States.

A Comparison of 1D-3D Co-Simulation and Transient 3D Simulation for EGR Distribution Studies. / Dimitriou, Pavlos; Avola, Calogero; Burke, Richard; Copeland, Colin; Turner, Niall.

2016. Paper presented at ASME ICEF 2016, Greenville, USA United States.

Research output: Contribution to conferencePaper

Dimitriou, P, Avola, C, Burke, R, Copeland, C & Turner, N 2016, 'A Comparison of 1D-3D Co-Simulation and Transient 3D Simulation for EGR Distribution Studies' Paper presented at ASME ICEF 2016, Greenville, USA United States, 9/10/16 - 12/10/16, .
Dimitriou P, Avola C, Burke R, Copeland C, Turner N. A Comparison of 1D-3D Co-Simulation and Transient 3D Simulation for EGR Distribution Studies. 2016. Paper presented at ASME ICEF 2016, Greenville, USA United States.
Dimitriou, Pavlos ; Avola, Calogero ; Burke, Richard ; Copeland, Colin ; Turner, Niall. / A Comparison of 1D-3D Co-Simulation and Transient 3D Simulation for EGR Distribution Studies. Paper presented at ASME ICEF 2016, Greenville, USA United States.
@conference{ea907ce9d2564dc9ba7fbc0f66456701,
title = "A Comparison of 1D-3D Co-Simulation and Transient 3D Simulation for EGR Distribution Studies",
abstract = "Computational modeling, an important task for design, research and development stages, is evolving fast with the increase of computational capabilities over the last decades. One-dimensional (1D) CFD simulation is commonly used to analyze the flow rates and pressures of an entire fluid system of interconnected parts such as pipes, junctions, valves, and pumps. In contrast, three-dimensional (3D) CFD simulation allows detailed modeling of components such as manifolds, heat exchangers, and combustion cylinders where the flow contains significant 3D effects. Coupling a 1D model with a 3D domain potentially offers the benefits of both simulation strategies in one co-simulation approach. The present study provides a deep understanding of the co-simulation approach by listing all necessary steps need to be followed before and during the coupling of the 1D and 3D simulation software. It analyses the simulation and convergence time requirements based on the 3D model mesh quality and compares this approach with the current 1D-3D uncoupled approach followed in the industry. The outputs of both simulation approaches are then compared with experimental results. The co-simulation time mainly depends on the mesh quality of the 3D domain and the number of inner iterations per time-step which is entirely determined by the nature and complexity of the simulation. The co-simulation time per engine cycle is almost identical to the uncoupled approach. However, it was found that the number of cycles required for convergence in the coupled approach is nearly double than the uncoupled approach. The comparison between the two simulation approaches and the experimental results demonstrated the very 3D nature of the flows, the sensitivity of the uncoupled approach to input conditions and the sensitivity of co-simulation to the averaged boundary conditions transferred from the 1D model back to the 3D domain.",
author = "Pavlos Dimitriou and Calogero Avola and Richard Burke and Colin Copeland and Niall Turner",
year = "2016",
month = "10",
language = "English",
note = "ASME ICEF 2016 ; Conference date: 09-10-2016 Through 12-10-2016",

}

TY - CONF

T1 - A Comparison of 1D-3D Co-Simulation and Transient 3D Simulation for EGR Distribution Studies

AU - Dimitriou, Pavlos

AU - Avola, Calogero

AU - Burke, Richard

AU - Copeland, Colin

AU - Turner, Niall

PY - 2016/10

Y1 - 2016/10

N2 - Computational modeling, an important task for design, research and development stages, is evolving fast with the increase of computational capabilities over the last decades. One-dimensional (1D) CFD simulation is commonly used to analyze the flow rates and pressures of an entire fluid system of interconnected parts such as pipes, junctions, valves, and pumps. In contrast, three-dimensional (3D) CFD simulation allows detailed modeling of components such as manifolds, heat exchangers, and combustion cylinders where the flow contains significant 3D effects. Coupling a 1D model with a 3D domain potentially offers the benefits of both simulation strategies in one co-simulation approach. The present study provides a deep understanding of the co-simulation approach by listing all necessary steps need to be followed before and during the coupling of the 1D and 3D simulation software. It analyses the simulation and convergence time requirements based on the 3D model mesh quality and compares this approach with the current 1D-3D uncoupled approach followed in the industry. The outputs of both simulation approaches are then compared with experimental results. The co-simulation time mainly depends on the mesh quality of the 3D domain and the number of inner iterations per time-step which is entirely determined by the nature and complexity of the simulation. The co-simulation time per engine cycle is almost identical to the uncoupled approach. However, it was found that the number of cycles required for convergence in the coupled approach is nearly double than the uncoupled approach. The comparison between the two simulation approaches and the experimental results demonstrated the very 3D nature of the flows, the sensitivity of the uncoupled approach to input conditions and the sensitivity of co-simulation to the averaged boundary conditions transferred from the 1D model back to the 3D domain.

AB - Computational modeling, an important task for design, research and development stages, is evolving fast with the increase of computational capabilities over the last decades. One-dimensional (1D) CFD simulation is commonly used to analyze the flow rates and pressures of an entire fluid system of interconnected parts such as pipes, junctions, valves, and pumps. In contrast, three-dimensional (3D) CFD simulation allows detailed modeling of components such as manifolds, heat exchangers, and combustion cylinders where the flow contains significant 3D effects. Coupling a 1D model with a 3D domain potentially offers the benefits of both simulation strategies in one co-simulation approach. The present study provides a deep understanding of the co-simulation approach by listing all necessary steps need to be followed before and during the coupling of the 1D and 3D simulation software. It analyses the simulation and convergence time requirements based on the 3D model mesh quality and compares this approach with the current 1D-3D uncoupled approach followed in the industry. The outputs of both simulation approaches are then compared with experimental results. The co-simulation time mainly depends on the mesh quality of the 3D domain and the number of inner iterations per time-step which is entirely determined by the nature and complexity of the simulation. The co-simulation time per engine cycle is almost identical to the uncoupled approach. However, it was found that the number of cycles required for convergence in the coupled approach is nearly double than the uncoupled approach. The comparison between the two simulation approaches and the experimental results demonstrated the very 3D nature of the flows, the sensitivity of the uncoupled approach to input conditions and the sensitivity of co-simulation to the averaged boundary conditions transferred from the 1D model back to the 3D domain.

M3 - Paper

ER -