The main goal of research on reciprocating internal combustion engines (ICEs) consists in increasing the power output while reducing pollutant emission and fuel consumption. Cycle-to-cycle variability (CCV) is closely coupled with the intrinsic turbulent nature of in-cylinder flow and is detrimental in terms of combustion efficiency, fuel consumption, and tailpipe emissions. Due to fluctuations in flame propagation, heat release, and burnt product formation, CCV is now seen as one of the major limiting factors for higher power output and lower fuel consumption in ICEs. Therefore, it is essential to understand and control CCV to improve the overall engine efficiency and performance. Experimental techniques like particle image velocimetry (PIV) provide a powerful technical support for the analysis of the spatial and temporal evolution of the flow field in ICEs. Proper orthogonal decomposition (POD) has been largely used in conjunction with PIV to analyze flow field characteristics. Several methods involving POD have been proposed in the recent years to analyze engine CCV. In this work, phase invariant POD analysis, conditional averaging, and triple and quadruple POD decomposition methods are introduced and applied to a large database of PIV data from the optically accessible TCC-III research engine. Results are discussed with particular emphasis on the capability of the methods to perform both quantitative and qualitative evaluations on CCV. A new quadruple POD decomposition methodology is proposed and compared to those available in the literature. Besides experimental techniques, Computational Fluid Dynamics (CFD) has become a fundamental tool for understanding the complex aero-thermochemical processes that take place in the cylinder and for driving the development of new technological solutions. Large-eddy simulation (LES) is the most practical simulation tool to understand the nature of CCV. This work investigates the CFD capabilities to simulate CCV. Several methods of analysis were assessed on a 50 LES cycles dataset on the TCC-III engine under motored conditions. The accuracy and the reliability of CFD simulations stands in the models used for the discretization of the fluid domain and for the numerical computation of the governing equations. The meshing strategy plays a central role in the computational efficiency, in the management of the moving components of the engine and in the accuracy of results. The overset mesh approach, usually referred to as Chimera grid or Composite grid, was rarely applied to the simulation of ICEs, mainly because of the difficulty in adapting the technique to the specific complexities of ICE flows. This work demonstrates the feasibility and the effectiveness of the overset mesh technique application to ICEs thanks to a purposely designed meshing approach. 50 LES cycles were performed on the TCC-III engine under motored conditions. The proposed POD quadruple decomposition methodology was extensively applied to assess both the accuracy of the simulated results and the potential of the method itself for understanding CCV.

Il principale obbiettivo della ricerca sui motori a combustione interna (internal combustion engines, ICEs) consiste nell’incremento della potenza erogata a fronte di una contemporanea riduzione di consumi ed emissioni inquinanti. La variabilità ciclica (Cycle-to-cycle variability, CCV) è fortemente legata alla natura intrinsecamente turbolenta della fluidodinamica dei motori a combustione interna, ed i suoi effetti sono nocivi sull’efficienza di combustione, sul consumo di carburante e sulle emissioni inquinanti. A causa delle fluttuazioni nella propagazione del fronte di fiamma, nel rilascio di calore e nella formazione dei prodotti di combustione, la CCV è individuata tra i fattori più limitanti per il raggiungimento di sempre più alte potenze specifiche a fronte di minori consumi. È dunque fondamentale capire e controllare la CCV per migliorare l’efficienza e le prestazioni dei motori. Tecniche sperimentali come la Particle Image Velocimetry (PIV) forniscono un consistente supporto tecnico per l’analisi dell’evoluzione spaziale e temporale dei flussi nei motori. La Proper Orthogonal Decomposition (POD) è stata largamente usata insieme alla PIV per analizzare le caratteristiche del campo di moto dei motori. Negli ultimi anni, diversi metodi basati sulla POD sono stati proposti per analizzare la CCV sui motori. In questo lavoro vengono illustrati la phase invariant POD, la media condizionale, e la decomposizione tripla e quadrupla della POD. Questi metodi sono applicati ad un vasto database di dati PIV sul motore di ricerca ad accesso ottico TCC-III. I risultati sono discussi dando particolare risalto alle capacità di ciascun metodo di dare una stima sia qualitativa che quantitativa della CCV. Un nuovo metodo di decomposizione quadrupla della POD viene proposto e comparato ai metodi presenti in letteratura. Oltre alle tecniche sperimentali, la fluidodinamica computazionale (Computational Fluid Dynamics, CFD) è ormai diventata uno strumento imprescindibile per la comprensione dei complessi fenomeni aero-termochimici che hanno luogo nel cilindro e per guidare lo sviluppo di nuove soluzioni tecniche. Le simulazioni LES (large-eddy simulations) sono lo strumento più indicato per simulare la CCV. In questo lavoro è stato valutato il potenziale della CFD nella simulazione della CCV. Diverse tecniche di analisi della CCV sono state valutate su un dataset di 50 cicli LES sul motore TCC-III in trascinato. L’accuratezza e l’affidabilità delle simulazioni CFD sono insite nei modelli usati per discretizzare il dominio fluidodinamico e per risolvere le equazioni che governano i fenomeni fisici. La strategia di discretizzazione (meshing) assume un ruolo centrale nell’efficienza computazionale, nella gestione dei componenti in movimento nel motore e nell’accuratezza dei risultati. L’approccio overset mesh, chiamato anche Chimera o Composite grid, è stato raramente applicato ai motori, soprattutto a causa delle difficoltà nell’adattamento di questa tecnica alle complessità specifiche della fluidodinamica dei motori. In questo lavoro viene dimostrata l’applicabilità della tecnica overset mesh ai motori a combustione interna, attraverso un approccio di discretizzazione sviluppato appositamente, e ne viene mostrata l’efficacia. 50 cicli LES sono stati calcolati sul motore TCC-III in trascinato. La tecnica di decomposizione quadrupla della POD sviluppata è stata ampiamente applicata per valutare sia l’accuratezza dei risultati simulati, sia l’efficacia del metodo stesso nella comprensione della CCV.

UN’INDAGINE SULLA VARIABILITA’ CICLICA NEI MOTORI A COMBUSTIONE INTERNA UTILIZZANDO PROPER ORTHOGONAL DECOMPOSITION E LARGE-EDDY SIMULATIONS / Federico Rulli , 2020 Mar 10. 32. ciclo, Anno Accademico 2018/2019.

UN’INDAGINE SULLA VARIABILITA’ CICLICA NEI MOTORI A COMBUSTIONE INTERNA UTILIZZANDO PROPER ORTHOGONAL DECOMPOSITION E LARGE-EDDY SIMULATIONS

RULLI, FEDERICO
2020

Abstract

The main goal of research on reciprocating internal combustion engines (ICEs) consists in increasing the power output while reducing pollutant emission and fuel consumption. Cycle-to-cycle variability (CCV) is closely coupled with the intrinsic turbulent nature of in-cylinder flow and is detrimental in terms of combustion efficiency, fuel consumption, and tailpipe emissions. Due to fluctuations in flame propagation, heat release, and burnt product formation, CCV is now seen as one of the major limiting factors for higher power output and lower fuel consumption in ICEs. Therefore, it is essential to understand and control CCV to improve the overall engine efficiency and performance. Experimental techniques like particle image velocimetry (PIV) provide a powerful technical support for the analysis of the spatial and temporal evolution of the flow field in ICEs. Proper orthogonal decomposition (POD) has been largely used in conjunction with PIV to analyze flow field characteristics. Several methods involving POD have been proposed in the recent years to analyze engine CCV. In this work, phase invariant POD analysis, conditional averaging, and triple and quadruple POD decomposition methods are introduced and applied to a large database of PIV data from the optically accessible TCC-III research engine. Results are discussed with particular emphasis on the capability of the methods to perform both quantitative and qualitative evaluations on CCV. A new quadruple POD decomposition methodology is proposed and compared to those available in the literature. Besides experimental techniques, Computational Fluid Dynamics (CFD) has become a fundamental tool for understanding the complex aero-thermochemical processes that take place in the cylinder and for driving the development of new technological solutions. Large-eddy simulation (LES) is the most practical simulation tool to understand the nature of CCV. This work investigates the CFD capabilities to simulate CCV. Several methods of analysis were assessed on a 50 LES cycles dataset on the TCC-III engine under motored conditions. The accuracy and the reliability of CFD simulations stands in the models used for the discretization of the fluid domain and for the numerical computation of the governing equations. The meshing strategy plays a central role in the computational efficiency, in the management of the moving components of the engine and in the accuracy of results. The overset mesh approach, usually referred to as Chimera grid or Composite grid, was rarely applied to the simulation of ICEs, mainly because of the difficulty in adapting the technique to the specific complexities of ICE flows. This work demonstrates the feasibility and the effectiveness of the overset mesh technique application to ICEs thanks to a purposely designed meshing approach. 50 LES cycles were performed on the TCC-III engine under motored conditions. The proposed POD quadruple decomposition methodology was extensively applied to assess both the accuracy of the simulated results and the potential of the method itself for understanding CCV.
AN INVESTIGATION ON CYCLE-TO-CYCLE VARIABILITY IN INTERNAL COMBUSTION ENGINES USING PROPER ORTHOGONAL DECOMPOSITION AND LARGE-EDDY SIMULATIONS
10-mar-2020
FONTANESI, Stefano
File in questo prodotto:
File Dimensione Formato  
PhD_thesis_Federico_Rulli.pdf

Open access

Descrizione: tesi di dottorato
Dimensione 12.45 MB
Formato Adobe PDF
12.45 MB Adobe PDF Visualizza/Apri
Pubblicazioni consigliate

Licenza Creative Commons
I metadati presenti in IRIS UNIMORE sono rilasciati con licenza Creative Commons CC0 1.0 Universal, mentre i file delle pubblicazioni sono rilasciati con licenza Attribuzione 4.0 Internazionale (CC BY 4.0), salvo diversa indicazione.
In caso di violazione di copyright, contattare Supporto Iris

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11380/1200665
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
social impact