BME Renewable Fuel Research: The Atomization Hurdle for Energy Technologies

At the Budapest University of Technology and Economics (BME), a dedicated team of scientists and engineers is leading the way in sustainable development through their research in energy industry technologies. Under the guidance of Associate Professor Viktor Józsa, the Combustion Research Group at BME is committed to reducing the carbon footprint of the energy sector. This is associated with the “100 EJ problem” – the challenge of transitioning the global energy system into one that is sustainable and low carbon, whose total energy consumption encompasses 100 Exajoules (EJ) per year. While progress has been made in various sectors, the transportation industry poses a significant hurdle when it comes to decarbonization.

Decarbonizing the transportation sector is a complex task that requires innovative approaches and comprehensive solutions. While electric vehicles (EVs) have gained popularity, their widespread adoption is hindered by factors such as limited infrastructure, long charging times, and high costs.

One promising solution here is renewable liquid fuels that can be used with modern injection systems and combustion chambers. These fuels have the potential to significantly reduce carbon emissions while maintaining the existing infrastructure and providing an energy-dense solution for long-distance transportation. However, atomizing renewable liquid fuels is more complicated than for conventional fuels.

To tackle the complexities of atomizing renewable liquid fuels on the software side, Combustion Research Group is using advanced tools like MATLAB and Simulink provided to BME through Campus-Wide Licenses.

Challenge

Engineers at the BME Combustion Research Group have performed experiments with different types of atomizers in order to test their newly developed evaluation methods on various sets of measurement data.

There are many great model platforms that help to gain a better understanding of the automatization phenomena, including the Plain-jet airblast atomization model platform. In twin-fluid and pressure atomizers, the process of droplet formation arises from the intense momentum transfer between a liquid jet or film and the surrounding continuous phase, typically air, in practical scenarios. High-velocity air blows over the surface of a low-velocity liquid jet which leads to rapid dispersal of the liquid jet into ligaments and then into tiny droplets.

The interaction between the surrounding gas and the spray establishes a two-way coupling, meaning that the velocity field of the air stream is affected by the spray, and vice versa, particularly in turbulent flows. When validating numerical simulations, it is essential to measure both the velocity field and the droplet size distribution at characteristic points, while monitoring turbulent fluctuations can also be critical.

Comparing various renewable fuels and technologies is crucial for selecting the best candidate and utilizing it most efficiently. When it comes to combustion applications, the droplets have to evaporate prior to the flame front. Therefore, the volume-to-surface diameter or Sauter Mean Diameter (SMD) is determined instead of other average diameter types. SMD is defined as follows:

Where Di is the diameter of a single droplet at a single measurement point.

The Phase Doppler technique is commonly employed for size characterization purposes, while high-speed imaging is typically utilized for qualitative analysis. In this blog, we will focus on the results of Phase Doppler Anemometry (PDA) measurements.

First, the atmospheric temperature-dependent spray characteristics of selected model liquids need to be thoroughly examined. The selected liquids encompass a diverse viscosity range, including distilled water, standard diesel oil (EN 590:2014), light heating oil, and crude rapeseed oil. The investigation is conducted within a liquid preheating temperature range of 25–100 °C, which also partially covers, e.g., the thermal stability range of pyrolysis oils. By considering these temperature limits, a broad range of physical parameters can be explored, with the liquid viscosity alone spanning more than two orders of magnitude.

Additionally, it is necessary to find a robust and efficient method which makes the characterization of the flow field possible based on typical droplet size and velocity measurement techniques.

Solution

The global SMD of the spray is approximated by six different empirical and semi-empirical formulae. The parameters of each SMD equation were determined using the least squares method. The parameter fitting was implemented in the MATLAB software environment, where the parameters of each equation were defined as symbolic variables. To ensure the best fit, a Globalsearch solver engine was applied; it uses a Scatter Search (ScS) mechanism to generate the starting points. ScS is a population-based metaheuristic algorithm devised to perform a smart search for the problem domain. The solver analyzes a set of starting points and rejects the ones that are unlikely to improve the best local minimum found thus far. The population elements are then updated. ScS differs from other population-based evolutionary heuristics, such as genetic algorithms (GA), mainly in how it generates new population candidates. It uses deterministic combinations of previous members of the population as opposed to a more extensively used randomization in GAs. Furthermore, it embodies principles and strategies that are still not emulated by other methods and which prove to be advantageous for solving a variety of complex optimization problems.

As a next step, the revision of several atomizers’ data sets by a novel algorithm was carried out in MATLAB. Thus, the critical droplet size can be determined, up to which the mean droplet velocity represents the gas-phase velocity. This makes possible an easy and straightforward characterization of the flow, based on size and velocity data. The tricky part here is taking into consideration the presence of those droplets which are either too tiny or too large: filtering droplets that are too small may lead to biased data for estimating turbulence characteristics, while keeping droplets that are too large is also risky, due to the effect of “overshooting.” Excessively large, unrepresentative droplets were filtered out by the “isoutlier” MATLAB function that uses the triple of the median absolute deviations.

Results

Increasing the temperature of a liquid slightly reduces SMDs, while increased atomizing pressure smoothens the droplet size distribution and makes it more monodisperse, as illustrated in the figure below.

It can be concluded from the two plots that atomizing pressure has a greater effect on spray size distribution than liquid preheating, as inertial forces govern spray formation over viscous forces in the conditions under study.

The high-velocity atmospheric atomization of four liquids was analyzed: water, standard diesel oil, light heating oil, and crude rapeseed oil at various liquid preheating temperatures. Since there is no known analytic way to estimate the SMD of an airblast atomizer’s spray in the conditions in question, the given empirical formulae were analyzed to understand the background physics better. Based on the results, the following conclusions were derived:

There is a limiting viscosity which significantly affects spray characteristics. Below this value, further preheating has no additional effect on the spray quality. This finding is in line with the fact that diesel oil can be atomized efficiently without preheating, while crude rapeseed oil has to be introduced well above 100 °C into the combustion chamber to have a sufficiently fine spray for liquid fuel combustion.

However, the liquid type and preheating temperature have a negligible effect on critical droplet size, which is ultimately governed by the air pressure needed for atomization.

Summary

Challenge

Characterizing practical spray patterns is a challenging task in today’s renewable fuel research. The actual goal here is twofold:

• Understand the effect of the liquid temperature on the atomized droplets, and
• determine the gas-phase velocity in particle-laden gas flow based on measurements of the droplet size and velocity.

Solution

1. The droplets’ volume-to-surface diameter has been estimated by fitting non-linear functions of various parameters, including viscosity and density – which are also influenced by temperature.
2. Determination of critical droplet size, up to which the mean droplet velocity can represent the gas-phase velocity well.

Results

1. The temperature, i.e. preheating of the liquid, has a minor effect on the atomization process. However, there is a viscosity limitation that affects spray characteristics: below this value, any further preheating has no effect on the spray’s quality.
2. An algorithm has been developed and published that is capable of preprocessing droplet size- and velocity measurement data. It works well on liquids in a wide range of conditions, including more than a two-magnitude spread in liquid viscosity. Thus, this algorithm serves as a robust and effective tool for spray flow characterization.

Featured products

MathWorks® products:

• MATLAB
• Curve Fitting Toolbox
• Global Optimization Toolbox
• Statistics and Machine Learning Toolbox
• Database Toolbox

Further reading materials:

Read the full research paper by Erika Racz, Milan Maly, Jan Jedelsky, and Viktor Jozsa titled “Gas-phase velocity estimation in practical sprays by Phase-Doppler technique“. The paper published in the International Journal Of Multiphase Flow is open-access and can be downloaded here.

Read the full research paper by Andras Urban, Milan Maly, Viktor Jozsa, and Jan Jedelsky titled “Effect of liquid preheating on high-velocity airblast atomization: From water to crude rapeseed oil“. The paper published in the International Journal Of Multiphase Flow is open-access and can be downloaded here.

The referenced MATLAB code can be found at:

• • The homepage of the Combustion Research Group – Gas-phase velocity estimation (Matlab code with sample file)
  • MathWorks File Exchange – Gas-phase velocity estimation (Matlab code with sample file)

Learn more about Budapest University of Technology and Economics (BME)