Nowadays, one of the hottest topics in the automotive engineering community is the reduction of the dependency on fossil fuels. This is motivated by the need to reduce GHG emissions, which in turn explains the increased interest towards Carbon-Neutral ones. Carbon-Neutral per se is the balance between emitting and absorbing carbon emissions, this means that any CO2 released into the atmosphere is balanced by an equivalent amount being removed. In such a category we could list fuels like synthetic fuels, biofuels (such as biodiesel, or bioethanol), Hydrogen, and Ammonia. These last two, not producing CO2 at all because of their carbon-free nature. Additionally, for the carbon-neutral cycle, producing hydrogen from biomass is one promising sustainable route and it could be realized by thermochemical and biological processes. In the present work focus will be made particularly on H2-fuelled internal combustion engines (H2ICEs) which could be a readily applicable solution to reduce GHG emissions in the transportation sector, with much faster carbon-footprint reduction than BEVs. The well-established technology of modern ICEs could be combined with the advantages related to hydrogen combustion, with minor modifications needed for conventional liquid-fuelled engines to run on hydrogen. The direct gaseous injection of H2 into the combustion chamber is a hard challenge for both the ICE designers and injection system suppliers. To reduce uncertainties, time, and development costs, CFD tools appear extremely useful, since they can accurately predict mixture formation and combustion before the expensive prototyping/testing phase. The gaseous injection, involved in the Direct-Injected H2-ICEs operation, promotes a super-sonic flow which leads to very high gradients in the mixing-zone, i.e. between the bulk of the injected gas and the flow, already, inside the combustion chamber. To develop a methodology for an accurate simulation of these phenomena, a preliminary focus is made on turbulent flow validation in an optically accessible engine, i.e. the Darmstadt engine of TU Darmstadt, which is representative of current four-valve GDI production engines. The so-called operating point A is investigated hereafter, with a focus being made on statistical convergence and relevance of the datasets, by made RANS, as well as LES, 3D-CFD simulation. A new index is introduced, based on the statistical relevance of the k-th experimental ensemble average velocity vector; it is applied to evaluate the influence of the population of cycles on the representation of flow patterns. Simulated fields are then compared to the experimental counterparts using the currently best-in-class comparative indices and Proper Orthogonal Decomposition (POD). POD is employed to characterize the Cycle-to-Cycle Variability (CCV) of the analyzed operation and its possible causes. Additionally, an in-house developed quadruple decomposition method based on POD is being elapsed to decompose the compared flow fields. The comparison between experimental flow fields and simulated ones provides an insight into both strengths and weaknesses of the adopted modeling approach, as well as of the recently proposed grid methodology. Tested and validated the adopted framework, attention is shifted to actual H2-ICE operation by working on the SOpHy Engine which is one of the ECN group experimental rigs. This engine is fed through a single nozzle H2-injector and PLIF data are available for comparison with the CFD outcomes. A detailed comparison is carried out, alongside different sensitivities to the main key parameters of a 3D-CFD simulation. This aims to produce new best practices for such simulations with scientific evidence being provided by the comparison against experimental data.
Al giorno d'oggi uno dei temi più discussi nella comunità dell'ingegneria automobilistica è la riduzione dei combustibili fossili. Vi è, infatti, la necessità di ridurre le emissioni di GHG, il che spiega l’incremento di interesse verso i combustibili Carbon-Neutral. Per Carbon Neutral si intende l'equilibrio tra l'emissione e l'assorbimento di CO2, ossia qualsiasi particella rilasciata nell'atmosfera è bilanciata da un’equivalente quantità rimossa. Una tale categoria annovera combustibili sintetici, i biocarburanti (come biodiesel o bioetanolo), l’idrogeno e l’ammoniaca. Questi ultimi due non producono CO2 per la loro struttura molecolare. Inoltre, per il ciclo a emissioni zero, la produzione di H2 dalla biomassa, mediante processi termochimici e biologici, è una strada promettente e sostenibile. Il presente lavoro è focalizzato in particolare sui motori a combustione interna ad H2, i quali potrebbero essere una rapida soluzione per ridurre le emissioni di GHG nel settore dei trasporti, con un abbattimento più rapido delle emissioni rispetto ai veicoli a batteria. La tecnologia dei motori a combustione interna può essere combinata con i vantaggi della combustione dell'H2, sebbene siano necessarie delle modifiche per il corretto funzionamento. L'iniezione gassosa diretta di H2 nella camera di combustione è una sfida difficile sia per i motoristi che per i fornitori di sistemi di iniezione. Per ridurre le incertezze, i tempi e i costi di sviluppo, le simulazioni mediante fluidodinamica computazionale (CFD) sono estremamente utili, potendo prevedere con precisione la formazione di miscele e la combustione, ottimizzando la costosa fase di produzione/test. L'iniezione gassosa, derivante dall’iniezione diretta di H2, favorisce un flusso supersonico che porta a gradienti elevati nella zona di miscelazione, ovvero tra il gas iniettato e il fluido già presente all'interno della camera di combustione. Per sviluppare una metodologia che garantisca una simulazione accurata di questi fenomeni, in primo luogo considereremo un motore ad accesso ottico, ovvero il motore Darmstadt della TU Darmstadt, rappresentativo dei motori attualmente in produzione. Verrà simulato il punto operativo A mediante approccio RANS e LES, con particolare attenzione alla convergenza statistica e alla rilevanza dei dataset utilizzati. È stato introdotto un nuovo indice, basato sulla rilevanza statistica del vettore k-esimo di velocità del campo sperimentale medio; questo verrà usato per valutare l'influenza della popolazione di cicli sulle strutture dei campi di moto. I campi simulati verranno confrontati con i rispettivi sperimentali utilizzando gli attuali indici comparativi più diffusi, nonché la POD, usata per caratterizzare la variabilità ciclica (CCV) del punto operativo analizzato e le sue possibili cause. Inoltre, un metodo, sviluppato internamente, di decomposizione quadrupla basato sulla POD verrà usato per decomporre i campi di moto analizzati. Il confronto tra i campi di moto sperimentali e simulati fornisce una panoramica dei punti di forza e di debolezza dell'approccio modellistico adottato, nonché della metodologia di griglia computazionale proposta. Testata e validata la metodologia adottata, si passerà alla simulazione vera e propria dei motori ad H2 lavorando sul SOpHy Engine che è uno degli apparati sperimentali del gruppo ECN. Questo motore è alimentato da un iniettore ad H2 a singolo ugello e dati PLIF sono disponibili per il confronto con i risultati CFD. Verrà quindi effettuato un confronto dettagliato, insieme con diverse sensibilità ai parametri chiave di una simulazione CFD-3D. L’obbiettivo finale è la produzione di nuove linee guida per la simulazione dei motori ad H2, con prove scientifiche fornite dal confronto con dati sperimentali.
Una Metodologia per la Simulazione CFD-3D di Motori a Combustione Interna Alimentati con Carburanti Carbon-Neutral: Sviluppo e Validazione rispetto ai Dati Sperimentali
BARBATO, ALESSIO
2023
Abstract
Nowadays, one of the hottest topics in the automotive engineering community is the reduction of the dependency on fossil fuels. This is motivated by the need to reduce GHG emissions, which in turn explains the increased interest towards Carbon-Neutral ones. Carbon-Neutral per se is the balance between emitting and absorbing carbon emissions, this means that any CO2 released into the atmosphere is balanced by an equivalent amount being removed. In such a category we could list fuels like synthetic fuels, biofuels (such as biodiesel, or bioethanol), Hydrogen, and Ammonia. These last two, not producing CO2 at all because of their carbon-free nature. Additionally, for the carbon-neutral cycle, producing hydrogen from biomass is one promising sustainable route and it could be realized by thermochemical and biological processes. In the present work focus will be made particularly on H2-fuelled internal combustion engines (H2ICEs) which could be a readily applicable solution to reduce GHG emissions in the transportation sector, with much faster carbon-footprint reduction than BEVs. The well-established technology of modern ICEs could be combined with the advantages related to hydrogen combustion, with minor modifications needed for conventional liquid-fuelled engines to run on hydrogen. The direct gaseous injection of H2 into the combustion chamber is a hard challenge for both the ICE designers and injection system suppliers. To reduce uncertainties, time, and development costs, CFD tools appear extremely useful, since they can accurately predict mixture formation and combustion before the expensive prototyping/testing phase. The gaseous injection, involved in the Direct-Injected H2-ICEs operation, promotes a super-sonic flow which leads to very high gradients in the mixing-zone, i.e. between the bulk of the injected gas and the flow, already, inside the combustion chamber. To develop a methodology for an accurate simulation of these phenomena, a preliminary focus is made on turbulent flow validation in an optically accessible engine, i.e. the Darmstadt engine of TU Darmstadt, which is representative of current four-valve GDI production engines. The so-called operating point A is investigated hereafter, with a focus being made on statistical convergence and relevance of the datasets, by made RANS, as well as LES, 3D-CFD simulation. A new index is introduced, based on the statistical relevance of the k-th experimental ensemble average velocity vector; it is applied to evaluate the influence of the population of cycles on the representation of flow patterns. Simulated fields are then compared to the experimental counterparts using the currently best-in-class comparative indices and Proper Orthogonal Decomposition (POD). POD is employed to characterize the Cycle-to-Cycle Variability (CCV) of the analyzed operation and its possible causes. Additionally, an in-house developed quadruple decomposition method based on POD is being elapsed to decompose the compared flow fields. The comparison between experimental flow fields and simulated ones provides an insight into both strengths and weaknesses of the adopted modeling approach, as well as of the recently proposed grid methodology. Tested and validated the adopted framework, attention is shifted to actual H2-ICE operation by working on the SOpHy Engine which is one of the ECN group experimental rigs. This engine is fed through a single nozzle H2-injector and PLIF data are available for comparison with the CFD outcomes. A detailed comparison is carried out, alongside different sensitivities to the main key parameters of a 3D-CFD simulation. This aims to produce new best practices for such simulations with scientific evidence being provided by the comparison against experimental data.File | Dimensione | Formato | |
---|---|---|---|
Alessio_Barbato_PhD_thesis.pdf
accesso aperto
Dimensione
16.59 MB
Formato
Adobe PDF
|
16.59 MB | Adobe PDF | Visualizza/Apri |
I documenti in UNITESI sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/20.500.14242/80133
URN:NBN:IT:UNIMORE-80133