Large eddy simulation of dual-fuel combustion under ICE conditions

The present thesis aims at studying n-heptane/methanol dual-fuel combustion under internal combustion engine conditions and strives to improve the understanding of its ignition, combustion, and pollutant emission mechanisms. Large-eddy simulation (LES) coupled with Eulerian stochastic fields (ESF) approach is employed to simulate single/dual-fuel combustion in a constant-volume vessel to mimic the single/dual-fuel combustion in conventional/dual-fuel premixed engines. The experimental configuration from Engine Combustion Network (ECN) is considered as the baseline case in the simulations. The main works are summarized in two parts: model development and studies of the fundamental physics involved in dual-fuel combustion.
First, the ESF approach with a novel modified method is proposed, implemented, and evaluated. Results show that the modified ESF method removes the numerical error in the element mass conservation and shows capability in predicting both premixed and non-premixed flames relevant to dual-fuel combustion. Second, LES of n-heptane single-fuel and n-heptane/methanol dual-fuel combustion is carried out and validated against ECN Spray-H experiments. A good agreement is obtained in terms of flow, combustion, and emissions characteristics.
Finally, a parameter study is performed to investigate the effects of the dual-fuel strategies, including the primary fuel concentration, the ambient temperature, and the pilot fuel injection timing. It is concluded that: 1) The ambient methanol is found to have an effect of suppressing the two-stage ignition and heat release of n-heptane, this is more significant under high ambient methanol concentration conditions. 2) The effects of methanol on the n-heptane ignition and NOx formation are strongly dependent on the ambient temperatures. The retardation of the n-heptane high temperature ignition is more remarkable under low ambient temperatures. The NOx and soot in the dual-fuel case is lower than that of the single-fuel case in moderately high initial temperatures, while an opposite trend is observed in higher temperatures. 3) A late injection may lead to an overlap of the ambient methanol auto-ignition and the delivery of n-heptane. This overlap results in high soot and NOx formation.

The internal combustion engine (ICE) is an important energy conversion machine, which has been widely used in cars, ships, and aircraft. In the ICE the fuel is combusted with air to produce heat, which gives rise to hot gas expansion that drives the piston to move reciprocatively and thereby provides power to the vehicles. Currently, gasoline and diesel are the common fuels for ICE, which are known as fossil fuels. On the one hand, fossil fuel consumption is not sustainable due to limited resources, and on the other hand, fossil fuel combustion produces pollutants such as NOx, soot particles, CO and unburned hydrocarbon that are harmful to the environment. Furthermore, fossil fuel combustion generates greenhouse gas CO$_2$ emissions that lead to global warming. In 2020, as a consequence of the confinement due to the COVID-19 pandemic, the slump in road transport activity accounted for 50% of the decline in global oil demand, and the drop in aviation for around 35%. In addition, the global carbon dioxide emissions from surface transport fell by 36% by April 2020 and made the largest contribution to the total emissions change. This is interesting data showing that the transportation sector has great potential to reduce fossil fuels usage and emissions. Therefore, it is of great importance to seek alternative fuels and clean combustion technologies for combustion in ICEs.

Methanol is one of the ideal alternative fuels for its rich resources and clean combustion features. It has a promising renewable feature, e.g., it can be produced from the reductive conversion of carbon dioxide with hydrogen. The raw material carbon dioxide could be captured from industrial effluents or the atmosphere, i.e., carbon dioxide is recycled into useful fuel in this process. Since the compression-ignition (CI) engines have a better fuel economy, there is a strong interest in applying methanol for CI engines. The engines in most of the trucks and some of the private cars are CI engines, or often known as diesel engine. In CI engines, liquid fuel is injected into the compressed air and auto-ignited as the compressed air has a high temperature and high pressure. However, it is difficult for pure methanol to ignite in conventional CI engines. The pure methanol-fueled CI engines suffer from incomplete combustion and misfire. As a result, novel combustion strategies, e.g., dual-fuel combustion, are developed to improve the ignition process in methanol-fueled CI engines. Dual-fuel combustion is a promising concept for combustion engines also because of its potential to reduce engine noise and emissions. The engines using dual-fuel combustion strategies are known as dual-fuel engines. In methanol/diesel dual-fuel engines, methanol is premixed with air during intake and compression strokes to form a methanol-air mixture, and then diesel is injected to ignite the methanol-air mixture. Although dual-fuel combustion has many advantages, it has been found that inappropriate use of dual-fuel strategies may not be beneficial and may even deteriorate engine performance.

To enhance its benefits and avoid its shortcomings, detailed knowledge on dual-fuel combustion is needed including the fuel/air mixing, chemical kinetics and its interaction with the flow and mixing in the engine, the mechanisms of fuel ignition and combustion wave propagation, and reasons behind pollutant emissions. Such detailed knowledge can be gained by performing engine experiments and high-fidelity numerical simulations. In this thesis, numerical simulations are carried out to understand the physical and chemical processes in methanol/diesel dual-fuel combustion. Numerical simulation is a powerful method as it provides detailed insights, such as flow velocity, species and temperature distributions in three-dimensional space and time, which are needed to understand the fundamental processes in dual-fuel engines. Although some of these information can be obtained from engine experiments, the data acquisition in engine experiments is limited by the difficulty in accessing the engine through its metal walls, and diagnostic methods for detailed species distribution are available for only limited number of species. Numerical simulations have also limitations. The simulations are performed by solving highly nonlinear transport equations, which gives rise to the generates of a large span of flow scales in the flow. Such flow is known as turbulence. Simulation of all scales of engine turbulence is not possible with current supercomputers, and turbulence models are needed to simplify the problem. In this thesis, a high-fidelity method for the simulation of turbulence is used, which is known as large eddy simulation (LES). In LES, only large-scale turbulence eddies are solved while the small-scale eddies are modelled. In dual-fuel combustion, chemical reactions are interacting with turbulence. The combustion and emission behaviours of dual-fuel combustion are highly sensitive to the turbulence/chemistry interaction (TCI). A novel method for TCI is developed in this thesis, and validation of the method is carried out by comparing it with experiments. The new model is based on the transported probability density function and the Eulerian stochastics fields (ESF) approach.

The results of the simulation are first applied to study spray combustion that had been studied in experiments to make sure that the numerical methods can predict the real behaviour in such complex spray process. Then, the method is adopted to simulate various dual-fuel combustion. For example, altering the methanol concentration to check whether one can use more methanol and less diesel, or what the consequence will be if the methanol and air mixture is compressed to a higher temperature, or how the combustion and emission will be if the diesel is injected in different timing. Those are the questions to be answered in order to design high-performance dual-fuel engines.

The results show that a good agreement with the measurements is obtained in terms of the flow, combustion, and emissions. There are three ignition stages in methanol/n-heptane dual-fuel combustion, the first- and second-stage n-heptane ignition and the ambient methanol auto-ignition. The ambient methanol has an effect of suppressing the ignition. The more methanol is delivered, the later ignition of the n-heptane spray will be. A late diesel injection may lead to an overlap of the ambient methanol auto-ignition and liquid fuel injection, which results in a high pollutant emission. These observations and conclusions from the simulations will help engineers to design advanced engines for cars, trucks, and ships to burn methanol in a cleaner and more efficient way.

The main contributions of this thesis are two-fold: first, the proposed new ESF model can give rise to higher simulation accuracy, which can be used in engine design with high fidelity; second, the more in-depth knowledge on dual-fuel combustion which help engineering to better understand the performance of their engines and in turn help improve the design of the engines.