Eindhoven University of Technology MASTER Design of a dispersion system for metal fuels Hameete, J.

(1)

Eindhoven University of Technology

MASTER

Design of a dispersion system for metal fuels

Hameete, J.

Award date:

2020

Link to publication

Disclaimer

This document contains a student thesis (bachelor's or master's), as authored by a student at Eindhoven University of Technology. Student

(2)

Design of a dispersion system for metal fuels

Master of Science thesis

For obtaining the degree of Master of Science in Mechanical Engineering

at the department of Power & Flow at the Eindhoven University of Technology

Graduation committee:

Prof. Dr. L.P.H. de Goey Dr. Y. Shoshin

Prof. Dr. Ir. H.C. de Lange Advisors:

Ir. M. Hulsbos

Jesse Hameete

November 20, 2020

(3)

Abstract

When the energy demand needs to be fulfilled with an increasing share of sustainable and renewable energy sources, energy storage is a topic that is gaining a lot of attention. The ignition of iron particles and the regeneration of these particles is a proposed method to store large amounts of energy. However, properties of iron fuel combustion are largely unknown.

The adiabatic flame temperature can be studied by using a flux burner that is designed specifically for two-phase flows. For this purpose, an iron-air dispersion system was developed in this study. With the help of a literature study, where different dispersion systems were analyzed, different concepts are created and considered. One of these methods; a vibration feeder with a venturi vacuum ejector for mixing of the flow, was chosen and built. This design can produce a constant and reproducible aerosol for very low to low mass flows that is able to break up large agglomerates. The downside of such a system is that different powders need to be calibrated individually, which is a is time consuming process. Therefore, future improvements to the system can are proposed by designing a system that can constantly monitor the mass flow.

(4)

Acknowledgements

I want to thank everybody that contributed in any way to the work that went into this project. First, I would like to thank Dr. Prof. L.P.H. de Goey for overseeing this project, his expertise has been crucial for the entire Metal2Power project. I would also like to thank Ir. M. Hulsbos for the (almost) daily guidance, his matlab scripts that helped me get started and for creating the SEM-images. He has been extremely valuable for the succes of this project. Lastly, there have been contributions in the form of knowledge, software and advice from professors, lab technicians and fellow students that will not be mentioned individually.

3

(5)

Chapter 1 Introduction

As the global need for energy continues to rise, mankind is looking for alternative ways to generate energy. Currently, almost all users that rely on large amounts of energy to continue operation, such as energy production facilities and the transport industry (shipping, rail, road & aviation), still use fossil fuels as the primary source of energy. Because of this, the primary energy demand in 2017 was produced for 81 % by fossil fuels. In the sustainable energy scenario developed by the International Energy Agency and published in the World Energy Outlook 2018 [1], this share of fossil fuels should be brought down to 60 % by 2040.

This renewable energy revolution requires alternative ways to produce and store power with a similar energy density to that of fossil fuels.

Nowadays, Lithium batteries are used as a means to store energy. These batteries are found in electric cars, mobile phones, and even the laptop that was used to write this report. Al- though the use of Lithium batteries is widespread, the energy density and specific energy compared to fossil fuels is inferior. This can be seen in Figure 1.1a, both the specific energy and the energy density of Lithium batteries are at least an order of magnitude smaller when compared to fossil fuels. Furthermore, Lithium batteries are expensive to manufacture, have a limited lifetime and are toxic [2]. All of this combined proves that new methods are required to store the ever growing supply of sustainable energy.

Lately, studies have shown that metals can be a green alternative for storing energy. Specif- ically iron has been studied as a suitable energy carrier because of the abundance of iron in nature. When iron powder is exposed to a high enough temperature, oxidation occurs. This oxidation releases energy that varies on the type of reaction from 272 to 1118.4 mole of product^kJ . The oxidated iron can be converted back into iron powder by means of reduction. A popular reduction mechanism to regenerate oxidated iron is hydrogen reduction. When hydrogen is produced by means of renewable energy sources (e.g. solar farms), then a renewable energy cycle without any carbon emissions can be created as described by Bergthorson [3].

In order to get this technology to a point where the above-mentioned energy cycle can be created, several parameters of an iron flame need to be explored experimentally. For this purpose, a heat flux burner for iron-methane-air mixtures is being built at the Eindhoven University of Technology.

(8)

CHAPTER 1. INTRODUCTION

(a) (b)

Figure 1.1: (a) Energy density plotted against the specific energy of different energy storage mediums [3]. (b) Schematics of a heat flux burner as developed by Bosschaart & De Goey [4]

.

1.1 Heat flux burner

The heat flux method was created to accurately determine the laminar burning velocity from gas-oxidiser mixtures. The heat flux burner is made to keep the flame as one-dimensional as possible to eliminate problems such as heat losses to the burner and flame stretching when measuring the burning velocity. This is done by cooling the mixing chamber and heating the burner plate, as is shown in Figure 1.1b. This method can now be used to determine the burning velocity of an iron-air mixture. However, introducing an aerosol containing solid particles instead of gasses brings forth a new set of challenges. One of these challenges is the generation of the aerosol. Previously, a gaseous fuel was mixed with an oxidizer with two simple mass flow controllers, making a stable aerosol is not so straightforward.

A dispersion system was previously built at the Eindhoven University of Technology by student team SOLID to create a proof of concept of metal fuel combustion [5]. This dispersion system proved to be incompatible for the heat flux burner, as it is impossible to accurately predict the aerosol density due to leakage and agglomeration. Therefore, a new system has to be built.

1.2 Project aim

The goal of this project is to build and validate a new dispersion system. This is done by means of a literature research on solid particles suspended in a gas. This research includes the analysis of agglomerates in different types of iron powder ¹. Subsequently, concepts are generated based on this literature and the best possible concept is selected. The building of a dispersion system is described and finally, the dispersion system is validated. This validation describes how constant the powder flow is, the repeatability of different feed rates, and whether the system is able to break up these agglomerates.

1Agglomerates are clumps of iron powder that stick together

(9)

Chapter 2 Dispersion & iron particle parameters

This chapter describes the research and experiments that have been conducted in order to get a better understanding in the definition and working of a dispersion system.

2.1 The aerosol

An aerosol is described as liquid or solid particles which are suspended in a gas. Examples of aerosols in nature are mist clouds, which are essentially water droplets suspended in the air. Aerosols have a widespread use. Asthma medicine for example, is suspended in a gas or air and inhaled by users to transport the specific medicine to their lungs. To compare other types of aerosols to iron-air aersols, two relevant parameters will be considered; particle size and dust cloud density.[6]

Figure 2.1 shows the particle size and definitions for aerosols. Metal powders that are considered in this project have a diameter between 3 and 100µm [5] [7] . This puts the size distribution of metal particles in the micrometer range. This range is comparable with flour, dust or cloud droplets.

The other important parameter when studying aerosols is the dust cloud density of the aerosol. This density can be explained as the mass of the matter that is suspended in a volume of the aerosol. A higher dust cloud density means it takes more energy to lift the aerosol. The range of dust cloud density for different aerosols in practice is shown in Figure 2.2. The aerosol density that is studied when creating a metal fuel aerosol is in the range of 50 to 700_m^g3. This is fairly high due to the high density of iron powder, and according to figure 2.2, this puts the metal fuel aerosol somewhere in the range of dust aerosols.

(10)

CHAPTER 2. DISPERSION & IRON PARTICLE PARAMETERS

Figure 2.1: Particle size ranges and definitions for aerosols, reprinted from Aerosol Technol- ogy; Properties, behaviour, and measurement of airborne particles [6]

Figure 2.2: Range of aerosol concentration, reprinted from Aerosol Technology; Properties, behaviour, and measurement of airborne particles [6]

Dispersion of solid particles in the form of powders is often called dry-dispersion. According to Hinds, dry-dispersion aerosol generators have two basic requirements, a means to continuously disperse the powder into a generator and a way of dispersing the powder in the form of an aerosol [6]. Yang & Evans have reviewed several methods of dispersing powder and divided these methods into pneumatic, volumetric, screw/auger, electrostatic and vibratory methods [8]. In Appendix 7.1, dispersion systems used for the dispersion of solid powders found in literature are divided in a similar way, although volumetric dispersion is not mentioned in this study. These systems will be used when creating concepts for dispersion systems in Chapter 3.

(11)

2.2 Rotating disk dispersion

Currently, a dispersion system for metal fuels is present at the Eindhoven University of Technlogy. This system, as designed by team SOLID, can be seen in Figure 2.3. It consists of two rotating disks, a hopper and a flushing system. The bottom rotating disk has a groove where the powder is deposited from the hopper. The bottom disk rotates continuously and, as the powder reaches the opposite side of the dispersion system. A needle flushes air through the groove with a high velocity. This air forces the powder to exit the dispersion system through the pipe, from where it can be transported to the burner.

Figure 2.3: Dispersion system that is currently used for the dispersion of metal fuel aerosols.

This version was built by student team SOLID [5]

This system has proven useful for a number of applications. However, this system shows a number of drawbacks that makes the use of this system in the heat flux burner impractical.

Drawbacks of this system are:

• The mass flow and air velocity are too high. The apparatus is not designed for a very low mass flux, and a certain air velocity must be produced in order to flush the powder out of the groove. Scaling down these parameters results in other problems, such as pulsing.

• The aerosol is not constant enough, especially if the angular velocity of the rotating disk is low, pulsing problems occur. Pulsing behaviour can be defined as, instead of a constant angular velocity, the system starts lagging and moving intermittently.

• The system shows leakage. Since air is blown through the groove, the powder will try to find another way out. Most of the powder will exit through the dispersion pipe, but if some powder creeps between the 2 disks, the way is cleared for other particles. This

(12)

For these reasons that are described above, the decision to make a new dispersion system was made. This new dispersion system should be able to create an aerosol out of metal fuels and a gaseous oxidiser without the drawbacks that the rotating disk dispersion system has. To this end, some research regarding flowability of powder is in order. The flowability of a powder links directly to the amount of agglomerations that a powder shows. These agglomerations are inconvenient in dispersion, but have the potential to be even worse in a flame. If large agglomerations enter the flame front, they may disturb the measurements and alter the results. For the best possible results, the powder should be de-agglomerated before entering the mixing chamber.

2.3 Flowability of powders

The flowability of a powder describes how likely the powder is to agglomerate. Agglomeration, or the ’sticking together of particles’ is a process which is highly disruptive in a dispersion system. When designing a dispersion system, one must therefore take the flowability into account. One easy way to estimate the flowability of a powder is to make a link with the particle diameter, powders with a smaller diameter will generally have a poor flowability compared to powders with a larger particle size. This is demonstrated in Abdullah & Geldart [9]. In their paper, the Hausner ratio is plotted against the Sauter diameter. The Hausner ratio is known to accurately represent the flowability of a powder. It is calculated by:

H = ρT

ρ_B (2.1)

In this equation, ρBis the bulk density of a powder and ρT is the tapped density of a powder.

The bulk density can be obtained by pouring a known mass of powder into a cylinder and measuring the volume. The tapped density can then be obtained by using the same cylinder and tapping it until no volume reduction can be obtained. Several tapping methods may be used in this process. The Sauter diameter is the diameter of a particle independent of the sphericity. Abdullah & Geldart conclude that particles with a higher Sauter diameter showed a decrease in the Hausner ratio, and thus increasing flowability [9]. Liu et al. showed the same result, but also showed that an irregular particle size (lower sphericity) and a greater spread in particle size distribution also led to a decrease in Hausner ratio [10]. From these studies can thus be concluded that the three factors that influence the flowability of a powder are particle size, particle size distribution and mean sphericity of a powder.

Sphericity, in this case, describes the ’roundness’ of a particle. In the next chapter, this sphericity is described as a factor between 0 and 1, where 1 is a perfectly round particle.

This sphericity is taken from a 2D still image of a particle, and is thus only an estimation.

The sphericity S can be described as the perimeter of a perfectly round particle with the same area as the particle, divided by the actual perimeter;

S = 2·√ π· A

P (2.2)

Where A is the area of the 2d image of he particle, and P is the perimeter.

(13)

2.3.1 Quantification of flowability

At the Eindhoven Universtiy of Technology, various iron powders are used to study various characteristics of metal fuel combustion. Three of these powders are studied to investigate the link between agglomeration, powder size and sphericity that is insinuated above. The three powders have been selected for their difference in observed flowability and the wide range in particle diameters. The powders were studied at Sympatec with their particle analysers, which claim to break up all agglomerates following the dispersion concept that can be seen in Appendix 7.1 in Figure 7.4. These powders are then investigated the Eindhoven University of Technnology under the Scanning Electron Microscope (SEM).

Sigma Aldrich (5-9 µm)

Sigma Aldrich has the powder with the smallest mean diameter of the three powders. The product specifications of their powder claim a particle size from 5 to 9 µm. To test if this is the actual distribution, the powder was tested at Sympatec with the HELOS laser diffraction.

The full report on the powder is available in Appendix 7.2. The report showed a particle size distribution where x10 ≈ 3 µm and x90 ≈ 13 µm. While this is not in agreement with the numbers that are given by the supplier, the powder is still the finest of the 3 that were tested.

Since this powder was tested with the HELOS, no sphericity data was recorded for this powder. However, there is some sphericity information in the Sauter Mean Diameter (SMD).

This parameter is defined as:

SM D= d³_v

d²_s = 6V_p

A_p (2.3)

In this equation, Vp and Ap are defined as the particle volume and the particle surface area respectively. Practically, the SMD can be described as the diameter of a sphere with the same volume/surface area ratio as the particle and thus, an SMD that is smaller than the average diameter of the particle can be an indicator of a non-spherical particle. In the experiments performed with the HELOS system, the mean diameter was found to be 6.57 and 6.41 µm respectively and the SMD was recorded at 5.43 and 5.37 µm respectively. This is a decrease of about 16 %, which seems to indicates that the powder is fairly spherical. For further analysis, Figure 2.4a and 2.4b show the SEM images of the powder. These images show that the particles are indeed fairly spherical. These images also show both small agglomerations and large clumps of powder that have been agglomerated, the latter can be seen in Figure 2.4a.

Gongyi City Meiqi industry (10 µm Carbonyl Iron Powder)

Gongyi City Meiqi industry supply Carbonyl Iron powder with a diameter of approximately 10 µm. The powder was tested using QICPIC dynamic image analysis, which showed a particle size distribution where x10 ≈ 6 µm and x90 ≈ 23 µm, with an average particle size of x₅₀= 10.83 µm. The full report on the powder is available in Appendix 7.3. While there may be some overlap with the particles from Sigma Aldrich, these particles are clearly larger on

(14)

TLS Technik (20-50 µm)

The powder from TLS Technik is claimed to be 20-50 µm. This powder is also tested with the QICPIC dynamic image analysis device of which the report is shown in Appendix 7.3.

The report shows a particle size distribution of x10 ≈ 26 µm and x⁹⁰ ≈ 59 µm. These are therefore the largest particles that will be tested. Using the same method as with the carbonyl powder above, the sphericity of this powder is also tested. The powder was tested twice to show an average sphericity of 0.865, which is similar to what was found in the previous test. This is confirmed when the SEM images of Figure 2.4d are reviewed. The particles seem roughly as spherical as the other powders. However, the TLS powder shows no agglomeration whatsoever.

2.3.2 Discussion of flowability results

Flowability of a certain powder is inversely proportional to agglomeration. A powder with excellent flowability properties will most likely have very few or no agglomeration. It was hypothesised in studies [9][10] that the particle size, sphericity, and particle size distribution are parameters which influence the flowability of a powder. It was shown that at an approximately constant sphericity, smaller particles are more likely to agglomerate under the same conditions, and thus have a lower flowability. A clear indication that a spread in particle size distribution also increases the agglomeration rate was not found. It should be noted that the current study made use of the iron powders available at the Eindhoven University of Technology. A more thorough investigation of the individual parameters with more powders is neccesary if a better understanding of flowability is to be obtained.

As can be seen in some of the results, especially those which use powder with a smaller diameter such as the Sigma Aldrich powder in Figure 2.4b, the agglomerations are severe and can grow up to 100 micrometer. If these clumps of powder get through the dispersion system, they could give some problems in the heat flux burner. It would therefore be beneficial if the dispersion system could break these large agglomerates. Furthermore, this study has only taken the physical properties particle size, sphericity and particle size distribution into account. Studies have shown that the moisture content is also an indication for flowability, and decreasing the moisture content in a powder by employing drying methods might be beneficial for the flowability of the powder [11].

(15)

(a) (b)

(c) (d)

Figure 2.4: Images from different types of powder, captured by the scanning electron microscope (SEM). Image (a) and (b) show the Sigma Aldrich powder, which has a reported mean diameter of 5-9 micrometer. Image (c) shows the Gongyi City Meiqi Industry powder, which has an average diameter of 10 micrometer. Image (d) shows TLS Technik powder that has a mean diameter of 20-50 micrometer. From these images it is clear that the relatively large TLS Technik powder is very spherical when compared to the other powders. This powder also shows no sign of agglomeration.

(16)

Chapter 3 Generating dispersion concepts

This chapter describes and rates concepts for the dispersion of iron powder based on the literature. One system is chosen to be built.

3.1 Morphological chart

Concepts are generated using a morphological chart. This chart enables the generation of new ideas by using the information of current dispersion systems to create a list with the four basic functions of a dispersion system; storing powder, metering powder, dispersing powder and deagglomeration and/or distribution of the powder. For more information on the references to the different authors in this chapter, the reader is referred to Appendix 7.1.

The chart with all the possible options for the functions of the dispersion system can be seen in Table 3.1. These options will be elaborated in following sections.

Storing Metering Dispersing Distributing Hopper Rotating disk Flushing Nozzle

Bed Linear actuator Fluidized bed Dispersion pipe Valve Venturi effect

Auger Mixing in flow Electrostatic

Vibration

Table 3.1: Morphological chart with all the options for a dispersion system that have been introduced in Chapter 2.

Storing powder

Storing of powder indicates the means to store the powder that is about to be dispersed.

This can be done in one of two ways. The first method is using a hopper, a hopper has a large storage capacity, but it has a problem of being guided by gravity. Therefore, a hopper can not be applied in all systems. For systems that cannot contain a hopper, a fluidized bed is usually the only option. The fluidized bed is a small storage where the powder on the bed will be dispersed continuously. A disadvantage of the bed in contrast with the hopper is the relatively small powder storage capacity, an advantage is the minimal space required compared to a hopper.

15

(17)

CHAPTER 3. GENERATING DISPERSION CONCEPTS

Metering powder

Metering of powder is needed to ensure a constant volume flux. In an ideal scenario, the metering of the powder is as accurate and as flexible as possible. Metering of powder is one of the most important parts in a dispersion system. There are a lot of possible methods for accurate metering of powder. The current demonstrator that is described in Section 2.2, and the TSI dispersion system as tested by Chen et al. [12], use a rotating disk to meter the powder. This disk system relies on gravity to fill the disk from a hopper, and the rotation can be varied to vary the volume flux of the powder. Another way to vary the flux is with the use of a linear actuator. this concept has been demonstrated by Risha et al. [13], the actuator acts as a packed bed where the powder is continuously flushed into the outlet nozzle.

If a powder is free flowing, a valve can be used to meter the powder. This concept is shown by Prenni et al. [14]. In their work, a solenoid valve is used to control the volume flux of powder. An Auger was also used in the literature [15] [16]. An auger works really well with a hopper, but has the disadvantage that metering accurately is rather difficult. Some examples of electrostatic metering have been shown by Stiphout [17], Chen et al. [18] and Olansen et al. [19]. While most of these methods rely on a packed bed instead of a hopper, the results show that metering of iron powder can be done accurately. However, since this method is relatively new, testing of accurate metering is required. Matsusaka et al. [20]

showed a method of metering using vibration, and showed that this is a reliable and accurate way of metering.

Dispersion of powder

There are several methods of dispersion. Dispersion of a powder is a way to entrain the metered powder in a stream of gas, after which it can be ejected. One way that is often used is flushing the powder. This method is employed by Chen et al. [12], student team SOLID [5] with a rotating disk design and by Risha et al. [13] with a linear actuator design.

This method is relatively simple and ensures that all powder will be displaced. One downside however, is that it is uncertain whether all the powder ends up in the nozzle, since the inserted gas will be flushing throughout the system. A good design is needed to prevent leakage in the system. Another method of dispersion is the fluidized bed which is demonstrated by Prenni et al. [14]. A packed bed of powder is fluidized and the aerosol of nitrogen and particles is gradually fed out through the outlet. While this is a good way to achieve a constant aerosol, the packed bed does not enable a sustained aerosol for an extended period of time.

The venturi effect, or the sucking in of powder into an air flow, has been demonstrated by Sympatec in their powder dispersion system [21]. The suction of particles does have the advantage of gaining more control over the trajectory of the particles, however, the inflow of air has to be accurately controlled to maintain a sufficient vacuum flow. The last option for dispersing powder is the mixing in a flow of air. This principle is demonstrated by Atkins [15], who used gravity to drop the powder from an auger into an air stream. This is the simplest solution, but it can be a problem with powders that are likely to agglomerate, since they might stick to the walls.

(18)

Deagglomeration / distribution of the aerosol

As suggested in the previous sections, it would be preferable to deagglomerate powder clumps so that the powder is distributed in a constant aerosol. Distribution of the aerosol is every- thing that happens from the dispersion to the point where the aerosol exits the system. One of these methods is using a nozzle. A nozzle is a small orifice which turns the aerosol into a constant spray, it is often used in paint spray and coatings. The other method is a simple dispersion pipe. The length of the pipe helps in dispersing the powder since agglomerations are believed to be reduced as clumps of particles collide with other clumps or the wall.

3.2 Concepts

The concepts will be chosen with help of the morphological chart. There will be four different concepts, each one of these concepts will be based on a different principle. These four principles are:

• No moving parts.

• Reliability.

• Flexibility.

• Simplicity.

In the sections below, each of these principles will be explained, and a concept will be described based on the choices within the morphological chart. The advantages and disadvantages of each concept will then be summarized. These advantages and disadvantages are mostly qualitative, because accurate quantitative differences cannot be measured at this stage. The advantages and disadvantages are based on observation in current methods and comments in the reference papers as described in Chapter 2 and Appendix 7.1.

(19)

3.2.1 No moving parts

Moving parts are generally not preferred when working with fine powders because the powder can jam the moving parts, or leakage can occur on the entrance of a moving part. Therefore, it is important that a dispersion system has as few moving parts as possible.

Both a hopper and a bed can be used in this concept, since both these parts cannot move. For the metering of the powder, electrostatic metering has a preference over the other options, since there are no moving parts involved in the examples that are shown in Appendix 7.1.

Mixing in the flow has a clear advantage since the powder does not come into contact with any part other than the walls. For the distribution, a dispersion pipe can be used for the deagglomeration of the powder. This option is used because mixing in the flow can cause some agglomerations. All these choices lead to the following concept:

Hopper/Bed−→ Electrostatic metering −→ Mixing in flow −→ Dispersion pipe The concept can be interpreted in multiple ways, and since there are several ways to imple-

(a) (b)

Figure 3.1: (a) Concept 1a: No moving parts with hopper & flow retardation. (b) Concept 1b: No moving parts with a bed & electrodes.

ment electrostatic metering, the concept is divided into two concepts. Both of these concepts can be seen in Figure 3.1.

Concept 1a uses electrostatic retardation of the flow, which is a concept that still needs to be proven. The concept is fairly simple, since it only requires a variable current for the

(20)

Positive

• Simple to operate.

• Hopper allows for aerosol that can be sustained for a long time.

• Large range of aerosol densities by varying current and mass flux.

• No moving parts, so no leakage of powder.

Negative

• Electrostatic retardation of iron powder is an unproven concept

• Dumping powder in air flow can cause particle build up and unsteady aerosol.

Concept 1b is based on the apparatus from Stiphout [17] and the apparatus from Shoshin &

Dreizin [22]. The concept, which can be seen in Figure 3.1b, uses a bed which can be used to fluidize the particles and allow them to bounce up and down between the electrodes. The air comes in from underneath the electrodes and the stream of air takes iron particles with them as it is forced through the outlet. Although this is quite an effective concept, the downsides are that the density of the aerosol cannot be easily controlled, and that this density varies over time. In addition, Shoshin reported that, since the particles that exit the outlet are generally positively charged, the particles will repel each other. This problem can possibly be fixed by making a dispersion tube that is grounded, as is sketched in Figure 3.1b. This, as with the previous concept, will need to be tested in order to validate the effectiveness. The positive and negative factors of this concept are listed here:

Positive

• Proven concept of dispersion.

• No agglomeration since the charged particles will not stick together.

• No moving parts, so no leakage of powder.

Negative

• Very difficult to control aerosol density.

• Particles repelling each other at the outlet.

• Aerosol is not consistent.

(21)

3.2.2 Reliability

Reliability considers the consistency of an aerosol, and aims to decrease the possibility of a concept failing. A reliable concept has parts that are proven to work, and is able to create a constant aerosol that can be maintained for a long time.

A hopper is a great way to get a consistent aerosol that can be maintained for several minutes. A hopper can also be as large as is required. A stirring device can be added in the hopper to prevent blockage through agglomeration [5]. Venturi vacuum ejectors are already implemented in the current generation of powder dispersion systems. Sympatec uses venturi devices in the latest version of their powder dispersion system used in dynamic image analysis [16]. To create a constant reliable aerosol, a dispersion pipe is the last factor to ensure that the iron particles exit deagglomerated and constantly entrained in the flow. These choices have led to the following concept:

Hopper−→ Vibration −→ Venturi effect −→ Dispersion pipe

(a) (b)

Figure 3.2: (a) Concept 2: Reliability. The main components of this concept are the vibration feeder and the venturi vacuum ejector. (b) Concept 3: Flexibility. The linear actuator enables the air stream to draw a constant mass from the pile.

A sketch of the concept can be seen in Figure 3.2a. The concept has the advantage that all of

(22)

and disadvantages for such a system are:

Advantages

• Good control of aerosol density.

• Hopper allows for aerosol that can be sustained for a long time.

• Both feeder and outlet minimize the chance of agglomeration.

• Modular system.

Disadvantages

• More complex with more input parameters.

• Open system.

3.2.3 Flexibility

A dispersion system that can produce a wide range of aerosol densities can be used in many different set ups, which is a clear advantage. For this reason, the most flexible concept is assembled here.

A linear actuator is one of the most flexible devices for metering the powder. Due to the accuracy of the feed speed and the flexibility in bed size of linear actuators that are widely available, a actuator can be chosen based on the needs of the project. The downside to this is that the actuator acts as a bed, and therefore a hopper cannot be an option in this concept.

For the dispersing, flushing the powder is very flexible since the density of the aerosol can be varied with the mass flux of the inflow. A nozzle can be implemented in the design to reduce agglomeration. All these choices lead to the following concept:

Bed−→ Linear actuator −→ Flushing −→ Nozzle

This concept is sketched in Figure 3.2. with only two input parameters, the concept is simple, and the density of the aerosol has a wide range. The accuracy of this density however, is directly dependent of the accuracy of the linear actuator. Another downside is that due to the use of the linear actuator, small particles may penetrate the sealing of the actuator and cause leakage. This can reduce the lifetime of the system. The advantages and disadvantages of this system are summarized below:

• A lot of flexibility in mass flow and aerosol density.

• Simple to operate .

• Compact design.

Disadvantages

• Linear actuator may cause leakage.

• Accuracy is dependent on actuator manufacturer.

(23)

3.2.4 Simplicity

A concept that is simple has the benefit that the construction is relatively cheap and the apparatus is easy to operate. In this concept, the simplest concept will be chosen.

The use of a hopper is simpler than the use of a bed, since a hopper is widely available and can be filled much easier than a bed in most cases. For the metering, an auger is available for purchase and easy to operate. When it is combined with mixing in flow, there are only 2 parameters that determine the aerosol density, namely the Auger and the air flux. A nozzle is the last part of this concept since it is a simple way to control the flow speed and width of the aerosol. These choices lead to the following concept:

Hopper−→ Auger −→ Flushing −→ Nozzle

A sketch of this concept is shown in Figure 3.3. It is flexible and because the powder is kept away from the motor of the auger, it is not likely to cause leakage. The concept is simple to operate and the auger is reliable. It does, however, have a few drawbacks. The auger is generally not suited for very low powder feed rates, so lower densities would be hard to achieve with this concept. On top of that, flushing powder off the auger is likely to provide an aerosol that does not have a constant density. The advantages and disadvantages of this concept are:

Figure 3.3: Concept 4: Simplicity. A simple auger/hopper combination is the simplest form of a dispersion system.

(24)

• Simple to operate.

• Flexible range of aerosol densities.

• Robust, low chance of leakage.

• Auger is reliable.

Disadvantages

• Not suited for low aerosol densities.

• Not suited for low feed rates.

• Inconsistent aerosol density.

3.3 Concept choice

Now that all concepts have been defined, a choice has to be made. The concepts are rated on a couple of criteria. The criteria should be noted such that they can be rated on a scale of 1-5 and have been generated from the strengths of the different systems.

• Reliability

• Flexibility

• Simplicity

• Modularity

• Does the system have any moving parts?

For the first four criteria, a rating of 1-5 is given to the system, where 1 means that the system rates poorly, and a 5 means that the system has an excellent rating to the criteria.

For the last criteria, 2 bonus points are given to a system that doesn’t have any moving parts.

Addition of all the ratings gives a total score with a maximum of 22 points. The concept with the highest score is chosen and will be built. The ranking is displayed in Table 3.2, and the motivations for each concepts will be explained in the next chapter.

3.3.1 Motivation rating concepts

As stated before, a machine should be reliable. The reliability of the concept is rated in the first column in Table 3.2. Concept 1a has the lowest ranking in reliability since the principle of metering relies on a concept that has not been proved yet. The reliability of this concept is very low. Concept 1b is proven, but only on particle scale. There is a chance that the apparatus can’t be scaled up. This needs to be investigated further, and therefore this concept does not score good on reliability. Concept 3 has a metering concept that has been proven with aluminium particles, but in another environment and with other aerosol densities, therefore it also scores low on reliability. Concept 4 has been proven on a large scale, it does have some problems with aerosol density, therefore it scores slightly inferior to concept 2, which has been designed for reliability.

(25)

Concept Reliable Flexible Simple Modular Moving parts Total

1a: No moving parts 1 4 3 1 2 11

1b: No moving parts 2 2 2 1 2 9

2: Reliability 4 3 3 5 2 16

3: Flexibility 2 4 4 3 0 13

4: Simplicity 3 3 4 4 0 14

Table 3.2: Ranking of all the concepts to the criteria. A ranking of 1-5 is given in the first four criteria, and an additional 2 bonus points can be given for a system without any moving parts.

The second column rates the flexibility, the ability of a concept to create an aerosol with a wide variety in density. Concept 1b can produce arosols with a really low density, but high density-aerosols are most likely difficult. Therefore, it scores the worst of all concepts.

Concept 2 and 4 both score three points on flexibility, Concept 2 is highly flexible, but the limitations of the venturi will most likely result in limitations for high density aerosols. Con- cept 4 has the opposite problem, since large densities are no problem, but smaller densities may be a problem. Concept 1a and 3 both have a high degree of flexibility, and will therefore get 4 points.

The simplicity of the design is also taken into account. This applies to both the effort it will take to make the machine, and the effort it will take to operate the machine. For this requirement, Concept 1b scores the worst, since it is a quite complicated design, and it will most likely prove very difficult to operate the machine to maintain a constant aerosol density.

concept 1a and 2 are still quite complex, but are rated a bit better since the machines will most likely be simple to operate. Concept 3 rates equally good to concept 4, because both designs will be reasonably simple to make and simple to operate. The reason that these concepts do not receive the maximum score is because both the concepts still consist of different parts, which means that it will still take some effort to make the machine.

A modular system is desirable, as modularity gives the ability to swap out components.

This increases the usability of this machine in other projects, or it can increase the flexibility of a certain system. The first two concepts 1a and 1b are not modular, which is why these have been given a score of 1. Concept 3 has some modularity, since the actuator can be swapped out, which is why this concept has a modularity score of 3. Concept four is already a bit more modular, since the hopper and the auger can be seen as seperate from the nozzle and tube. Concept two is also highly modular with a seperate feed system and metering system that can be swapped out. These concepts are scored with a 4 and 5 respectively.

In conclusion, adding up these scores results in a score of 16 for concept 2: reliability. This will be the concept which will be investigated further both theoretically and experimentally, in order to determine if this concept can be made into a dispersion system.

(26)

Chapter 4 Building a dispersion system

This chapter elaborates on the design of the dispersion system, and the choices that have been made in the design process. Both of the major component of the dispersion system, the vibrating feeder system and the venturi vacuum ejector, are discussed in this chapter, and the setup of the system is shown.

4.1 General design considerations

In order to design an apparatus, some parameters need to be considered. These parameters have come forth from the experience with the previous dispersion systems, along with some calculations in Matlab. A parameter, β, is introduced. This parameter shows the molar fraction of iron in the fuel. Thus, when β is 0, the fuel contains no iron and is pure methane.

When β is 1, the fuel only contains iron. The assumptions that have been made in these calculations are:

• Diameter of the burner is 3 cm.

• The flame speed of the methane / air flame (β = 0) is assumed to be 0.37^m_s .

• The flame speed of the iron / air flame (β = 1) is assumed to be 0.05^m_s. This assumption is based on the limited simulation results that are available in literature (eg.[23] ).

• The equivalence ratio φ = 1.

• All the iron powder in is converted to Hematite according to the reaction: 4F e+3O² = 2F e2O₃.

• The flame speed is assumed to decline linearly with an increasing amount of iron powder.

Using these assumptions, the reaction for the iron-air-methane flame can be written as a function of β;

(1− β)CH4+ βF e + (2− 5

4· β)O2 = (1− β)CO2 + 2(1− β)H2O+β

2F e₂O₃ (4.1)

25

(27)

CHAPTER 4. BUILDING A DISPERSION SYSTEM

Using these reaction rates, the stoichiometry of CH4 : F e : O2 can be described as (1− β) : β : (2− ⁵₄ · β). These factors are, of course, also dependent on the equivalence ratio. Since the assumption that the equivalence ratio is 1 has been made, the equation simplifies. The other reaction rates describe the molar fractions Nspecies, and using the molar volume of these species and the diameter of the burner, the mass flux of iron powder can be calculated.

Q_tot = π

4 · D² (4.2)

V_molaravg = (X

N_species· V^molarspecies)/Ntotal (4.3) Q_{F e}= VmolarF e· N^{F e}· Q_tot

Vmolaravg

(4.4)

˙

m_{F e}= ρF e· QF e (4.5)

Similarly, the mass and volume fluxes for the oxidisers can also be calculated. A dust cloud density can then be calculated by:

ρ_aerosol= m˙_total

Q_total (4.6)

All these parameters are plotted in Figure 4.1. The iron mass flow shows a spike at β≈ 0.8.

This shows that, as the mixture changes from a 100% iron powder to a mixture with a little bit of methane, the velocity increase of the mixture is larger than the iron powder decrease of the mixture, causing an increase in iron powder flow up to a maximum at β≈ 0.8.

Figure 4.1: Modelling of the first assumptions. These figures give a first estimation of the mass flow and air flow that is required for the dispersion system.

(28)

As can be seen in Figure 4.1, the dispersion system will need to be designed for an iron mass flow of approximately 5− 40^mg_s , and an oxidizer flow of approximately 2− 15_min^l . As stated before, the hopper system should be designed such that it can sustain an aerosol for at least 10 minutes, and the system should be designed such that is both modular, and the outflow is as constant as possible.

4.2 Venturi vacuum ejector

A venturi vacuum ejector is a static generator that uses compressed air to generate a vacuum.

The ejector principle is schematically drawn in Figure 4.2 and is a textbook example of Bernoulli’s principle. In the figure, the colors indicate the pressure, where red indicates a relatively high pressure and yellow indicates a relatively low pressure. Bernoulli’s law states that:

P₁+ 1

2· ρ · v1² = P2+1

2 · ρ · v2² (when ∆h = 0) (4.7)

∆P = 1

2 · ρ · (v2²− v1²) (4.8)

Since the conservation off mass holds:

v₁· A¹ = v2· A² (4.9)

Because A2 < A1 (4.10)

v₂ > v₁ (4.11)

Therefore, P2 < P₁ (4.12)

In this example, P2 is the vacuum pressure. Air leaves the nozzle with a velocity roughly

Figure 4.2: Schematic drawing of a venturi vacuum ejector.

equal to: v2 = v1(^D_D¹

2)² = 3600v1 if the nozzle has a diameter of 0.5 mm. In the conditions described in the previous section, this would mean that the velocity would be exceeding mach velocities. The vacuum pressure P2 is lower than the atmospheric pressure P3, and can thus be used to suck the particles into the venturi.

For the current setup, an air flow of 2-15 normal litre per minute (NLPM) is required.

For this purpose, the Vacuum Ejector ZH05DSA-06-06 by SMC is chosen. This ejector works according to the venturi principle that is explained above. It has an inflow, outflow and vacuum port with a diameter of 6 millimeter and an internal nozzle with a diameter of 0.5 millimeter. The ejector can handle flow rates from 0-19 NLPM with up to a 95 % vacuum.

With this air flow, the mass flow that was predicted in the previous section should not be a problem.

(29)

4.3 Vibration feeder & Hopper

The hopper and vibration feeder are integrated into the same part. This has the advantage that deagglomeration is started to some degree in the hopper, and thus agglomeration is believed to be reduced even further. The hopper & vibration feeder will be called the vibration feeder for the remainder of this report. The vibration feeder consists of three parts;

a vibrator which will be responsible for the feeding of the particles through vibration, a thyristor regulator which regulates the current that is flowing to the vibrator, and finally a gutter and hopper assembly that are integrated and designed to allow a minimal mass flow of powder to flow through the system. The vibrator and thyristor regulator are parts that have been purchased, while the gutter and hopper are designed. The reason for this is that the commercially available gutters are designed for a mass flow that exceeds the required mass flow for the system. The end result of this system can be seen in Figure 4.3.

Figure 4.3: Image of the complete vibration feeder. The three distinctive parts that can be recognized are the vibrator, the thyristor regulator and the hopper & gutter unit.

4.3.1 Vibrator

The vibratory feeder consists of a base, a tray and a set of springs. The base unit contains an electromagnetic drive unit that generates vibrations from an AC signal that is varied with a regulator. The amplitude of the signal that is provided to the drive can be increased or decreased to control the power of the drive [24]. This type of vibrator is designed for small- scale applications and is therefore ideal for controlling the mass flow of the iron powder.

(30)

4.3.2 Thyristor regulator

Although the name does suggest it, the regulator contains more than just a thyristor. A thyristor is a device quite similar to a diode. Where a diode only contains a P-type and an N-type semiconductor, the thyristor contains four layers with an P-N-P-N type semiconductor respectively. This causes three possible depletion regions to be formed, and a gate terminal is then necessary in order to allow electrons to pass. In short, a thyristor is a device that acts as a conductor when a gate is triggered, and will continue to conduct (even when the trigger is taken away) until the voltage reaches a threshold value and shuts off. Using this feature, the positive side of the sine wave of an AC signal can be ’cut off’ with a system that is schematically pictured in Figure 4.4. The potentiometer can be adjusted such that it changes the rate at which the capacitor charges. When the capacitor is charged, the diac gives the gate signal to the thyristor, which will then power the vibrator. This cycle repeats itself every time the sine wave becomes positive, and the charging rate of the capacitor impacts the moment that the thyristor starts to conduct. Using this principle, the amplitude that is sent to the vibrator can be varied.

Figure 4.4: Simplified electric scheme of the thyristor regulator

4.3.3 Hopper & gutter

The hopper & gutter part is designed to feed very low to medium mass fluxes. An exploded view of the system can be seen in Figure 4.5. The powder stopper (the small plate in front of the hopper) can be mounted in the most downward position to accommodate a mass flow that is as low as possible, and can be mounted further upward to accommodate a higher mass flow. Since the mass flow is expected to rely heavily on the flowability of the powder, and the flowability of the powder can be influenced by a number of parameters mentioned in Section 2.3, the mass flow will need to be calibrated for every powder that will be used.

Information on how to calibrate the powder can be found in Appendix 7.4.

(31)

Figure 4.5: Exploded view of the vibration system

(32)

4.4 Experimental setup

The previous section describes the parts that have been designed or ordered in order to make a working dispersion system. A schematic overview of the dispersion system that has been constructed can be found in Figure 4.6. The Venturi vacuum ejector and the vibration feeder are locked in an airtight container, to ensure that the volume flux of the air flow can be controlled. The air supply to the supply vacuum ejector of the venturi is controlled by a mass flow controller (MFC) with a maximum air flow of 18 NLPM, and the box is connected to another MFC with a maximum air flow of 30 NLPM that indirectly controls the the airflow to the vacuum side of the venturi.

Figure 4.6: Schematics of the final setup

The Vibration feeder delivers the mass flow of powder to the venturi, and is controlled through the thyristor regulator. A voltage meter is connected between the thyristor and the vibration feeder to monitor the voltage and increase the repeatability of experiments. The mass flow of powder is entrained in the air in the venturi, and the resulting aerosol exits the exhaust side of the venturi and is then transported to the heat flux burner.

A picture of the setup as it is built can be seen in Figure 4.7. In first experiments, the airtight container seems to be showing some signs of leakage. Air is leaking in through the handles on the container. If this gives any problems, the container should be revised.

(33)

Figure 4.7: Final setup of the dispersion system. The top of the airtight container has been removed for a clear overview.

(34)

Chapter 5 Validation of the dispersion system

This chapter describes the experiments that have been performed to test the capabilities of the dispersion system, and to test if the system can operate in the ranges that have been described in the previous chapter. The mass outflow, consistency of the aerosol, and the ability to break agglomerates are tested in this chapter.

5.1 Mass outflow validation

One of the advantages of this dispersion system is that it can produce an aerosol that is both continuous and reproducible. To test this, a setup has been made where the output of the dispersion system is weighed continuously.

5.1.1 Experimental setup

The vibration system is described in chapter 4.3. This system is designed specifically for low mass flows. The mass flows that can be obtained with this device is tested with a simple weighing setup that can be seen in Figure 5.1. It consists of a scale with an accuracy of 0.01 gram that outputs a value to the computer every second. For an accurate determination of the continuity of the system, the measurements are continued for at least 180 seconds. The output of this scale is then post processed with a Matlab program that makes a linear fit of the data of each measurement. The coefficient of determination is then checked for the goodness of fit of the linear fit. The coefficient of determination (R²) is calculated as:

R² = 1−SS_res

SS_tot (5.1)

SS_res=X

i

(yi− fi)² (5.2)

SS_tot =X

i

(yi− y)² (5.3)

In this equation, SStot gives the variance of the data, and SSres the sum of squares of the residuals, where yi is the mass value of the i-th second, y is the average mass value, and fi

is the mass value that is predicted using the equation of the linear fit. For a perfect fit, the sum of squares of the residuals would be zero, and the value for R² would be one. For the measurements in this section, the coefficient of determination should be higher than 0.99 in order for a measurement to be considered ’linear’.

33

(35)

CHAPTER 5. VALIDATION OF THE DISPERSION SYSTEM

Figure 5.1: Setup of the mass measurement experiment. The basic components of the setup are the dispersion system, a weighing system and a computer for the processing of data.

5.1.2 Results

In this project, the 2 most used powders are tested. Both the Rio Tinto and the TLS Technik powder are tested. The results of these tests can be seen in Figure 5.2. Figure 5.2a shows the calibration data for the Rio Tinto powder, this powder has a diameter x90= 75µm, and is therefore the largest powder. Figure 5.2b shows the calibration data for the TLS powder.

This powder is smaller but, as is shown in Section 2.3, has great flowability properties.

(a) (b)

Figure 5.2: Calibration data for the 2 powders that were considered in this experiment, the Rio Tinto powder and the TLS technik powder.

(36)

During the measurements that can be seen in figure 5.2, the matlab script was programmed to give an error for measurements that had an R² value that was less than 0.99. These values were checked manually after the measurements, and show values that all exceed 0.99. There- fore, it can be confirmed that the mass flow is constant. The figures show that the error for the lower voltages is a lot smaller than the error for the higher voltages. This has to do with the specific settings of the vibration device. This device is calibrated for low mass flows and as a consequence, it is less accurate for higher mass flows. More information on how to calibrate a powder and how to change this sensitivity can be found in the user manual in Appendix 7.4.

The dispersion system should be able to work with a range of powders that is as wide as possible. In Section 2.2, it was shown that the rotating disk dispersion system was not able to disperse powders with a low flowability. To test the capabilities of the new dispersion system. Sigma Aldrich powder was tested. This is the powder with the lowest flowability of the samples that are available at the Einhoven University of Technology, as is shown in Section 2.3. It was found that this dispersion system shows the same drawbacks that the rotating disk dispersion system has in that it is not able to make a constant and reproducible aerosol. Due to these limitations, the Sigma Aldrich powder will not be tested at this time.

However, some improvements to the system can be made. Such as changing the design of the hopper to allow the powder with agglomerations to pass through with ease, or decreasing the moisture content of the powder. These improvements will be added to the recommendations.

(37)

5.2 Breaking up agglomerates

The principle of the dispersion system is believed to be able to break up most agglomerates.

An analysis with a Scanning Electron Microscope (SEM) was conducted in order to study the effect of vibrations on the large agglomerates. These agglomerates are mostly found in powders with a small average diameter and powders with a low sphericity, as was confirmed in Section 2.3.

5.2.1 Experimental setup

For this experiment, a sample of powder was collected before and after the dispersion system.

The sample after dispersion was collected by turning on the dispersion system and setting the potentiometer of the vibration system to roughly half the intensity ( 165V). These samples are studied under a SEM to look for agglomerations. As could be seen in Section 2.3, there are different types of agglomerations. Small agglomerations consist of a few particles and generally have a maximum size of roughly five times the particle diameter. Larger agglomerations consist of larger clumps of particles that can grow up to one hundred times the particle diameter. For this experiment, Sigma Aldrich powder with a particle size of 5 - 9 µm was used. This powder has shown the most agglomerations in previous tests.

5.2.2 Results

Figure 5.3 shows the SEM image of the powder before dispersion and after dispersion. In Figure 5.3a, the powder before dispersion clearly shows a lot of agglomeration. The powder clusters show both small and large agglomerations to be present in large quantities. Figure 5.3b shows the powder after dispersion and as is visible in the figure, a lot of agglomeration is still present. Especially the small agglomerations are still visible. This leads to the conclusion that only the very large agglomerates are broken up by the dispersion system. According to the literature that is available in Appendix 7.1, possible ways to further reduce the agglomerations include using a nozzle or a dispersion pipe. These options could be considered when further reduction of the smaller agglomerations is necessary. It should also be noted that higher voltage vibrations may be able to break agglomerations better, this was not studied in the current experiment.

(38)

(a) (b)

Figure 5.3: SEM images from Sigma Aldrich powder with a nominal particle size between 5 and 9 micrometer. (a) shows the powder before dispersion. The large agglomerations can clearly be seen as the big clumps of powder. (b) shows the powder after dispersion. In this figure, the large clumps of powder have disappeared. However, smaller agglomerations are still present.

5.3 Consistency validation

Consistency of the aerosol is important to assure a flame of a constant quality. A setup is being built by a bachelor student to asses the consistency of the aerosol from the dispersion system and, if possible, to monitor the mass flow of iron powder in situ. The schematics of this system can be found in Figure 5.4. A laser sheet is formed by a point source laser using sheet optics. The laser sheet passes the air/powder mixture just above the tube of the dispersion system, and is then focused on a photodetector. The output of the sensor is continuously recorded (The sensor gives a signal 10 times per second) and monitored on the computer. The metal particles that pass through the laser sheet will absorb photons, and this will influence the signal. Preliminary results of this project are not yet available at this time.

Figure 5.4: Setup of the consistency experiment. The basic components of the setup are the laser, optics, dispersion system and detector, which is connected to the computer.

(39)

Chapter 6 Conclusion

Society is continuously searching for new and better ways to store energy without the emissions that are usually found in burning hydrocarbons. Metal fuels are a promising alternative, presenting a similar energy density to most hydrocarbons and, in the case of iron, a better specific energy than the next best alternative, Lithium batteries. In this research, a dispersion system for the heat flux burner is developed. This dispersion system needs to be able to sustain an aerosol with a volume flux between 2 and 15 _min^L with a dust cloud density between 5 and 40 ^mg_sec.

Looking at dispersion systems that produce dry aerosols. A literature study showed different promising alternatives. There have historically been multiple ways to disperse dry powder in a gas, and all relevant options have been considered for the design of an iron-air dispersion system. The system that is chosen is the most reliable to produce a steady aerosol, while also being modular, flexible and having no moving parts that may cause leakages. This system is built and tested in the lab. The setup consists of a vibration device with a thyristor regulator, a hopper / powder gutter and a venturi vacuum ejector which are all sealed in an airtight box.

Due to particle size, particle shape and particle size distribution, every type of powder has different flowability properties. Therfore, every powder that is going to be used in the heat flux burner should be calibrated according to the user manual which can be found in Ap- pendix 7.4. Rio Tinto and TLS Technik powders were tested, and the results can be found in this Section 5.1. It is found that the vibration device can produce a mass flow that is constant, but the repeatability of the mass flow becomes lower for higher mass flows, and thus the error in the calibration data becomes larger. This is a known problem due to the settings of the vibration device, and the repeatability is excellent for the mass flows that are required for the experiments with the heat flux burner. Depending on the type of powder, the dispersion system has a mass flow that can go as low as 10 ^mg_sec.

It is proposed in literature that vibrations may reduce the amount of agglomerations in the powder. This is tested by looking at a powder that is known to be likely to agglomerate, a sample of this powder before and after dispersion was analysed under the scanning electron microscope (SEM). It is found that large agglomerations are broken down by the dispersion system, while smaller agglomerations were still present in the powder after dispersion.

(40)

CHAPTER 6. CONCLUSION

The building of the dispersion system revealed some challenges in creating an aerosol using air and metal powder. Some parts of the system can still be improved. A list of the points that can still be improved is given below:

• The dispersion system has to be calibrated for each type of powder, and any changes to the settings of the vibration system would invalidate the calibration data. A system that could continuously monitor the iron mass flow would eliminate this problem, meaning that setting up the system would require a lot less time. The laser system proposed in Section 5.3 could be used for this purpose, but the concept would need to be proven first.

• It is shown in Section 5.2 that the system is able to break up large agglomerations, but a lot of smaller agglomerations still remain in the powder after agglomeration.

Additionally, the system is not able to handle powders with severe agglomerations.

Adaptions to the system should be made if the system has to disperse these powders.

Additionally, literature shows that drying of the powders may enhance the flowability and make them usable by the dispersion system.

• The airflow that is going to the vacuum side of the venturi vacuum ejector and the vibration system add 2 new degrees of freedom to the heat flux burner. It would be benificial for the control of the system if all these degrees of freedom would be controlled from a single file in, for example, LabVIEW. A start to this goal has been made in this project, and can be found in the user manual.

To conclude, very little is known about burning iron particles in a controlled manner. Re- search towards the understanding of iron fuel as a sustainable energy source is a great way to build towards applications that use iron as a sustainable cyclic fuel.

(41)

Bibliography

[1] International Energy Agency. World Energy Outlook 2018. 2018.

[2] M. Stanley Whittingham. Lithium batteries and cathode materials. Chemical Reviews, 104(10):4271–4301, 2004.

[3] J. M. Bergthorson, S. Goroshin, M. J. Soo, P. Julien, J. Palecka, D. L. Frost, and D. J.

Jarvis. Direct combustion of recyclable metal fuels for zero-carbon heat and power.

Applied Energy, 160:368–382, 2015.

[4] K. J. Bosschaart and L. P.H. De Goey. Detailed analysis of the heat flux method for measuring burning velocities. Combustion and Flame, 132(1-2):170–180, 2003.

[5] SOLID. Realising the world ’ s first Circular Energy Station using Metal Fuels. Technical report, 2018.

[6] W.C. Hinds. Aerosol Technology; Oroperties, behaviour, and measurement of airborne particles. John Wiley & Sons, Inc., 2nd edition, 1999.

[7] N. van Rooij, V.J.T. Seijger, and T. Spee. IJzer als brandstof voor schepen. Technical report, Eindhoven University of Technology, Eindhoven, 2019.

[8] S. Yang and J. R.G. Evans. Metering and dispensing of powder; the quest for new solid freeforming techniques. Powder Technology, 178(1):56–72, 2007.

[9] E. C. Abdullah and D. Geldart. The use of bulk density measurements as flowability indicators. Powder Technology, 102(2):151–165, 1999.

[10] L. X. Liu, I. Marziano, A. C. Bentham, J. D. Litster, E.T.White, and T. Howes. Effect of particle properties on the flowability of ibuprofen powders. International Journal of Pharmaceutics, 362(1-2):109–117, 2008.

[11] Dipanshu Chinwan and M. Elena Castell-Perez. Effect of conditioner and moisture content on flowability of yellow cornmeal. Food Science and Nutrition, 7(10):3261–3272, 2019.

[12] B. T. Chen, H. C. Yeh, and B. J. Fan. Evaluation of the TSI small-scale powder disperser.

Journal of Aerosol Science, 26(8):1303–1313, 1995.

Eindhoven University of Technology MASTER Design of a dispersion system for metal fuels Hameete, J.

Design of a dispersion system for metal fuels

Master of Science thesis

For obtaining the degree of Master of Science in Mechanical Engineering

at the department of Power & Flow at the Eindhoven University of Technology

Graduation committee:

Prof. Dr. L.P.H. de Goey Dr. Y. Shoshin

Prof. Dr. Ir. H.C. de Lange Advisors:

Ir. M. Hulsbos

Jesse Hameete

November 20, 2020

Abstract

Acknowledgements

Contents

Chapter 1 Introduction

1.1 Heat flux burner

1.2 Project aim

Chapter 2

Dispersion & iron particle parameters

2.1 The aerosol

2.2 Rotating disk dispersion

2.3 Flowability of powders

2.3.1 Quantification of flowability

2.3.2 Discussion of flowability results

Chapter 3

Generating dispersion concepts

3.1 Morphological chart

3.2 Concepts

3.2.1 No moving parts

3.2.2 Reliability

3.2.3 Flexibility

3.2.4 Simplicity

3.3 Concept choice

3.3.1 Motivation rating concepts

Chapter 4

Building a dispersion system

4.1 General design considerations

4.2 Venturi vacuum ejector

4.3 Vibration feeder & Hopper

4.3.1 Vibrator

4.3.2 Thyristor regulator

4.3.3 Hopper & gutter

4.4 Experimental setup

Chapter 5

Validation of the dispersion system

5.1 Mass outflow validation

5.1.1 Experimental setup

5.1.2 Results

5.2 Breaking up agglomerates

5.2.1 Experimental setup

5.2.2 Results

5.3 Consistency validation

Chapter 6 Conclusion

Bibliography