Heat flux measurement device

(1)

Heat flux measurement device

Designing an experimental system for determining the effectiveness of thermal barrier coating inside a

combustion chamber

Pisasale Salvatore

Master Thesis in

Mechanical Engineering

Södertälje, Sweden 2015

(2)

2

(3)

3 Master of Science Thesis MMK 2015:102 MFM 163

Heat flux measurement device.

Designing an experimental system for determining the effectiveness of thermal barrier coatings inside a

combustion chamber

Salvatore Pisasale

Approved Examiner

Andreas Cronhjort

Supervisor

Christian Binder

Commissioner

SCANIA CV AB

Contact person

Anders Thibblin

Abstract

This Masters Thesis has been carried out in collaboration with SCANIA CV AB and it concerns the development of an experimental measurement set up to analyze the heat losses in a Diesel engine. This measurement equipment will be used to test a type of coating, called TBC (Thermal Barrier Coating). Scania has been studying this kind of coating for some years and it has been noticed as a way to improve the efficiency of the engine. It is then important for the company to understand the behavior of this coating considering all the combustion’s features of an internal combustion engine.

The target of the project has been the replacement of one of the valves of a Diesel engine with a stationary sample holder equipped with a measurement set up in order to measure the heat losses from the combustion chamber. The design has been dimensioned considering the size and the working conditions of a single cylinder test engine at Scania.

The concept of the project is the placement of some thermocouples in the holder so that a difference of temperature can be detected and the relative heat flux can be computed. The TBC will be attached to one of the surfaces of the holder in order to test a decrease in the heat loss through the holder itself.

The conclusion of the project shows the good operation of the design and a substancial decrease in the heat loss when using the TBC. Scania should continue investigating the behavior of TBC with the use of the same design or a different one which fits different operating conditions of the engine.

(4)

4

(5)

5 Examensarbete MMK 2015:102 MFM 163

Värmeflödesmätanordning

Design av ett experimentellt system för utvärdering av termiska barriärskikt i ett förbränningsrum

Salvatore Pisasale

Godkänt Examinator

Andreas Cronhjort

Handledare

Christian Binder

Uppdragsgivare

SCANIA CV AB

Kontaktperson

Anders Thibblin

Sammanfattning

Detta examensarbete, som har utförts i samarbete med Scania CV AB, handlar om att utveckla en provmetod för att analysera värmeförluster i en dieselmotor. Provmetoden kommer att användas för att undersöka effekten av termiska barriärskikt – TBC (Thermal Barrier Coatings). Scania har under en tid studerat dessa beläggningar, då de har identifierats som ett möjligt sätt att öka motorns verkningsgrad. Det är då viktigt för företaget att förstå hur dessa beläggningar beter sig under de förhållanden som råder i en förbränningsmotor.

Målet med detta projekt har varit att ersätta en av ventilerna i en dieselmotor med en stationär provhållare med mätutrustning för att kunna mäta värmeförluster från förbränningsrummet.

Provhållaren och omgivande komponenter har dimensionerats utifrån mått och driftpunkter för en encylindermotor på Scania.

Termoelement placeras i provhållaren så att temperaturskillnader kan detekteras och värmeflöden beräknas. En av provhållarens ytor kan beläggas med TBC för att kunna mäta förändringen i värmeflöde genom själva provhållaren.

Slutsatserna i detta examensarbete är att provhållarens konstruktion fungerar bra i motorn och att det är en väsentlig minskning av värmeflödet genom provhållaren då TBC används.

Scania bör fortsätta undersöka TBC med denna konstruktion, eller med en modifierad variant som passar olika driftpunkter.

(6)

6

(7)

Acknowledgement

The possibility to carry out my master thesis at Scania has been a great opportunity for me and this is why first of all I would like to thank my supervisor Anders Thibblin. His

knowledge on the material science field together with a really close support helped me a lot in the realization of the project. The discussions held with him have been beneficial for my personal and academic improvements.

Tack så mycket!

I want also to thank my academic supervisor Christian Binder. His advices and his critical judgements increased my understanding of the topic and changed also my way of working, making it more constructive.

A special thanks goes to Daniel Norling for his outstanding knowledge in internal combustion engines and for all the support he gave me during the realization of the system and also during the tests on the engine. His smart interpretation of the results helped me in the understanding of the combustion process and pushed me to gain a deeper explanation of the related

phenomena.

I would also like to thank the Mechanical Workshop of the UTPW for being so patient with me every time I showed up with some changes in the parts of the system or every time I needed their help.

My gratitude goes also to James Davy and to the guys of the Kanalprovrummet working for the NMGP. They helped me in the definiton of some details of the project and above all they fixed most of the problems that came out with the cylinder head.

Un sentito ringraziamento va ai miei genitori e a tutti i miei amici, vecchi e nuovi, che non hanno mai smesso di mostrarmi il loro sostegno nonostante la lunga distanza.

A mia sorella Lucia. Per te, per sempre.

(8)

8

(9)

9

1. Introduction

1.1 Background

The major part of the power–generating systems that nowadays are utilized require the combustion of fuels. Diesel engines, gasoline or turbojet engines are just few examples of the huge number of devices that have a combustion chamber and that require then the combustion of fossil fuels (coal, natural gas, etc…)

The combustion leads to the emission of exhaust gases and some of them are really dangerous for life on Earth. Carbon dioxide (CO2) above all is what mainly causes the Greenhouse Effect that causes the global warming. This is why the emission of CO2 has been discussed for many years and nowadays it is still a problematic issue. Other emissions are dangerous (like NOx [0] and particulates [1]) but are not relevant for the content of this project.

Moreover world’s resources of fossil fuels will not last forever and it has been estimated that the production of fossil fuels will stop in the near future. For these reasons it is necessary to design new devices which consume fuel in order to have less emissions of harmful gases; a way to reach this purpose is having more efficient combustion engines.

For a Scania Diesel engine (which is going to be the subject of this project) the efficiency is normally around 45-46%. The losses in a Diesel engine in general can be considered evenly distributed between the heat going to the cooling system and the heat to exhaust gases [2]. Since the high values of pressure and temperature of the exhaust gases can be partly converted to work at the outlet of the engine, for example in a turbocharged engine or using a Waste Heat Recovery (WHR systems), it can be easily noticed that a lot of importance is given to the heat losses to the cooling system. From the end of the 20^thcentury this has been one of the main topics of the studies that tried to see the insulation of the hot parts of the engine as a mean of saving energy. That is why there have been many works regarding the design of low heat rejection engines (LHR) that use a certain type of coating, called TBC (Thermal Barrier Coating) ([3],[4]) The efficiency of this coating is influenced by many factors, like temperature, pressure, soot present in the combustion chamber, infrared radiations from the combustion and of course service time. All these factors have been tested by others in real working

conditions of an engine [5], except for the radiative heat transfer.

(12)

12

1.2 Objectives of the project

Measuring the heat losses from the combustion chamber becomes extremely important to understand how it is related with the insulation of the combustion chamber and try to get an optimal solution to reduce it.

The target of this project is then to develop an experimental measurement device. This set up is then used to evaluate the effectiveness of different TBC materials inside a combustion chamber of a Diesel engine. The behavior and the aging of some coatings inside the combustion chamber will be evaluated in order to analyze an effective decrease of the heat loss through the TBC itself.

1.3 Contribution of the project

The evaluation of how TBC influences the heat flux from the combustion chamber has been already studied by others using a coated probe [5]. This project is based on the same concept but it has been designed on a Scania engine and moreover it shows more technical means for a better understanding of the combustion process.

In particular, the contribution of infrared radiation inside the combustion chamber in real working conditions of the engine can be studied.

Furthermore, the temperature transient inside the coating can be studied and a more precise study of the heat flow across the measurement device can be done.

1.4 Limitation

The restrictions of this project are based on the available single cylinder engine where the measurement equipment will be designed on (Diesel engine at Scania) and on the number of coatings that are going to be tested. There will not be any kind of ranking among different types of TBC coatings in this project. Only two different types of coating are procured for testing to see the effective reduction of the heat losses compared to the not coated system and to analyze other aspects related to the heat transfer over the coating. Moreover the heat loss that will be analyzed is not the entire amount of heat lost to the head of the engine; only the quantity transferred to one of the valves will be taken into consideration.

(13)

13

2. Theory

2.1 Heat transfer

The heat transfer is the exchange of thermal energy between two systems which have different temperatures. This exchange of energy can happen by three ways: conduction, convection and radiation.

2.1.1Conduction

The conductive heat transfer is related to molecular or atomic level activity and it occurs when a gradient of temperature between two systems is present. The systems don’t have to move relative one to another and this requires then a physical contact between them in order to have conductive heat transfer. In solid materials it occurs with the vibration of the lattice; if the material is a conductor, the exchange of energy occurs with the movement of the free electrons. The equation that describes the conductive heat transfer is the Fourier’s law; if a mono-dimensional case of a surface A, which has two

dimensions much bigger than the thickness L, is analyzed, then the Fourier’s law is:

𝑞̇

_𝑥

=

^{𝑘∙𝐴∙(𝑇}_𝐿²^−𝑇¹⁾

(1)

where q̇_x is the conductive heat transfer rate along the x-direction (perpendicular to A), k is the thermal conductivity, T₂ is the temperature of the hotter side of the wall and T₁ is the colder one [6].

2 .1.2 Convection

The convective heat transfer is present when two systems are in motion relative each other. It is more complex to describe than the conductive heat transfer but basically two different mechanisms can be recognized which operate simultaneously. One is related to the random motion of the molecules (that is called diffusion) and it is then related to the conductive heat transfer. The other one refers to the macroscopic motion of the

molecules.

Two different kinds of convection can be recognized and they are defined as “natural”

and “forced”. Convection is defined natural when a mass force like buoyancy or centrifugal forces is present. It is instead defined forced when the motion is due to an

(14)

14

external force like a pump, a fan or the wind. In both cases the resulting motion can be defined external if it is on an external surface (on a plate, on an aerodynamic profile,…) or internal if it occurs in an area bounded by surfaces(pipes, canals, cavities,…).

The expression that describes the convective heat transfer is the Newton’s cooling law:

q̇ = h ∙ A ∙ (T₂− T₁) (2)

where q̇ is the heat transfer rate, h is the convective heat transfer coefficient, A is the surface of the heat transfer, T2 is the temperature of the hotter system and T1 is the temperature of the colder one [6].

The convective heat transfer coefficient depends on some physical properties of the fluid like the dynamic viscosity (μ), density (ρ), specific heat capacity (cp), thermal conductivity (k) and the flow characteristics like the velocity (u).

2.1.3 Radiation

The mechanism of the radiative heat transfer is very different compared to the ones discussed before and it is related to the condition of the material of the system. The energy emission is in fact linked to the configuration of the electrons of the atoms and it is conveyed by electromagnetic waves. For this reason the energy will be a function not only of the temperature but also of the wavelength and of the direction of the emission.

Since the energy is conveyed with waves, radiation does not need a material mean, it is even more efficient in a vacuum. The emissive power E_b is the rate at which the energy is released from a surface, per unit of surface area [W/m²] and the highest value it can reach is given by the Stefan-Boltzmann equation [6]:

E_b= σ ∙ T_s⁴ (3)

where Ts is the temperature of the surface in K and σ is the Stefan-Boltzmann constant (𝜎 = 5,67 ∙ 10⁻⁸ 𝑊/𝑚²∙ 𝐾⁴)

This is valid only in the case of an ideal body, which is called black body, that absorbs all the radiation but does not reflect any of it. A real body instead does not absorb all the radiation since some of it is reflected. The ratio between the emissive power of a real body E and the one of the corresponding black body E_b is called emissivity ε (𝜀 = 𝐸/𝐸_𝑏) [6].

(15)

15

2.1.4 Heat transfer in internal combustion engine

In an internal combustion engine the combustion chamber is perhaps the most critical part of the engine when heat transfer is discussed. At the end of the compression of the charged fresh air, liquid high pressure fuel is injected into the combustion chamber.

Since the temperature in the combustion chamber is high, the fuel starts evaporating and reaches the self-ignition temperature and the combustion begins. Forced convection and radiation are present at this point: the convective heat transfer depends on the

characteristics of the flow field and includes the major part of the heat transferred while radiation contribution has been estimated to be between 20-40% [7]. During the

combustion a very high temperature can be reached, also over 2000 K [8]. Even if the air has then high density, the fuel can pass through the air and evolve into jet-like flames that imping the piston bowl [9] (Fig.1). This has the resultant effect to increase the heat transfer since the difference in temperature between the flame and the walls is high. The flames consume oxygen so that the combustion slows down and some intermediate combustion products can appear and can then evolve into soot.

Fig.1 Impinging jet flow ([9])

While running the engine a layer of soot will accumulate on the walls of the combustion chamber. Even if it has some good properties as an insulant material, it has negative aspects for the radiative heat transfer. The soot in fact behaves like a black body on the walls of the combustion chamber, absorbing all the radiation and subsequently it reradiates it through the TBC. Wahiduzzaman and Morel ([10],[11]) proved that thin ceramics are partially transparent to the thermal radiation typical of a diesel engine. This

(16)

16

layer of soot can modify the temperature profile and influence the heat transfer over the TBC materials. Borman [12] says that this effect doesn’t raise the temperature but will cause deep heating of the transparent ceramics while Siegel [13] affirms that for soot coated thermal barrier the radiative effect increases the temperature of the metal face.

The radiative heat transfer is composed of two main components. A part of the radiation is generated by the incandescent soot particles with temperature of 2500-2800 K. Since this radiation generates short wavelengths near the visible and in the infrared-spectrum (λ = 0.77-1.5 µm), this radiative heat transfer contribution should be taken into

consideration for semi-transparent coatings [14]. The other part of radiation is generated by the hot gases (at temperature T= 1200-1900 K) and by the heated up walls of the combustion chamber. Experiments have anyway shown that soot emission is often considerably stronger than the emission from combustion gases in a Diesel engine [15].

After the convective heat transfer in the combustion chamber the conductive heat transfer is present through all the parts beside the walls; forced convective heat transfer with oil and coolant fluid is present in the cooling system’s ducts in order to remove the heat and so to cool the engine. A simplified description of the heat transfer can be seen in Fig. 2.

Fig.2 Heat transfer in an internal combustion engine [16]

(17)

17

2.2 Thermal Barrier Coating (TBC)

As mentioned before, a special insulating system will be tested in this project. It is called TBC (Thermal Barrier Coating) and can be used to get a decrease in the heat loss from the combustion chamber

TBC is an advanced materials system usually attached to a metallic surface which operates under high values of temperature. It has a lot of application in gas turbine and aeromechanical field with some application in the automotive industry. Thanks to its physical properties (like low thermal conductivity and low heat capacity), TBC is not only used to insulate the underlying material in order to reduce the heat losses but also to protect it from thermal stresses.

The TBC is mainly composed of 3 layers (Fig.3) and each of them has its function. The outer layer is the ceramic top coat and it has to withstand the high temperature of the working condition and to decrease the thermal conductivity of the system [17]. The most used ceramic top coat today is the yttria-stabilized zirconia 8YSZ because it has shown a low thermal conductivity also for high temperature [18]. The inner layer is the bond coat and it has to protect the metallic layer from high temperature oxidation, to promote the adherence between the top coat and the metallic surface and to reduce the thermal expansion differences. As a good oxidation resistant, CoNiCrAlY is often used as bond coat [19]. During the service of the TBC a thermally grown oxide layer (TGO) is formed between the top coat and the bond coat and it is basically the result of the slow oxidation of the bond coat.

Fig.3 Structure of the TBC

(18)

18

The application of the TBC inside a combustion chamber has been a very discussed topic in many works. One of the most critical points to take into consideration is the heat transfer from the combustion chamber. Woschni ([20], [21], [22]) stated that, since the temperature of the ceramic layer of combustion chamber’s wall increased, there would be a drastic increase of the convective heat transfer coefficient. The studies of Furuhama and Enomoto [23] lead to the same conclusion of Woschni. On the other side there have been other studies that have shown opposite results: Morel ([24],[25]) stated that the increase of the temperature in the wall leads to a reduction of the heat transfer and defined “clearly not realistic” the results from Woschni. Jackson ([26],[27]) also showed that the mean and the peak value of the heat transfer in the coated combustion chamber decreased compared to the not coated one.

Another aspect to take into consideration is the volumetric efficiency of the combustion chamber, because the hot walls from the combustion can influence negatively the refilling of the chamber with fresh air. This problem has been discussed in SAE publication by the Toyota Motor Company [28] and will be now described. The aspect that has been analyzed is the so called “Swing Temperature” which is the fluctuation of the temperature of the combustion chamber walls. In the publication the metallic surface of the combustion chamber is compared with two different coatings: one is the

traditional TBC with low thermal conductivity and heat capacity comparable with the metallic surface, the other one is a theoretical material with an extremely low thermal conductivity and low heat capacity. The comparison is shown in Fig.4: the x axis shows the Crank Angle (CA) of the camshaft while the y axis presents the temperature profiles of the gas and of the combustion chamber during a Diesel cycle.

(19)

19

Fig.4 Gas and combustion chamber walls temperature profiles [28]

When no coating is used, the hot gas release heat to the walls of the combustion chamber. However, since the metallic surface has high thermal conductivity and high heat capacity, its temperature remains low and a relevant temperature difference is retained. There is then a big heat transfer from the combustion chamber which is one of the main causes of energy loss from the combustion chamber itself.

In the case of traditional insulation (TBC), there is a smaller difference of temperature between the gas and the walls because the material of the coating has a low thermal conductivity but, since its heat capacity is still high, the temperature average of the walls of the combustion chamber is higher than the metallic surface case. Therefore the fresh charged air is heated up by the walls and expands its volume in the chamber so that there is less air available for the combustion.

If the “Swing Temperature” insulation is instead used, the coating can follow the behavior of the gas so that for a small amount of heat transferred to the gas there is a rapid increase in temperature of the coating surface thanks to its properties of low thermal conductivity and low heat capacity. The heat present in the wall can be released during the exhaust stroke of the engine so that the volumetric efficiency should not be deteriorated. During all the cycle the difference in temperature between gas and walls is low then and this leads to a small heat flux from gas to walls and vice versa.

(20)

20

2.3 Thermocouples

A thermocouple is a sensor which is used to measure the temperature. It consists of two wires of different metallic materials welded together at one end, creating then a

junction. This junction is called “hot junction” and it is where the temperature is measured. The other ends of the wires are welded as well and it is called “cold junction”. This junction is called also “reference junction” and it is important that the temperature measured by this junction is constant over the time (usually 0°C) so that the temperature measurement depends only on the temperature at the hot junction.

Electronic methods are used to keep the reference temperature constant and the

operation is called “compensation of the reference temperature”. It is very important in order to get precision and accuracy from the measurement. Nowadays thermocouples are widely used because they are inexpensive, they can be replaced easily, they are standardized and they can cover a wide range of temperatures. The main limitation of the use of the thermocouples is the accuracy of the measurements: it is very difficult to get a systematic error smaller than 1°C. The principle of operation of a thermocouple is based on a thermoelectric effect, which is also called the Seebeck effect.

For the Seebeck effect, if a circuit made of two different electric conductors is subjected to a thermal gradient, then a voltage is created in the circuit (Fig.5). This value of voltage is then given as an input to a device and the value of temperature at the hot junction is computed.

Fig 5 Scheme of the electric circuit in a thermocouple

(21)

21 In real conditions also two other thermoelectric effects should be analyzed: Peltier and Thomson effects. The first one regards an electric current present between two different conductors. The current produces a heat transfer between the conductors and in a thermocouple this is prevalent in the junctions. The Thomson effect regards instead a current-carrying conductor which has two different temperatures at its ends. Then a heat transfer is present and the heat will be released by the conductor if the electric current and the thermal flux have the same direction. If they are in opposite directions the heat will be absorbed. Fortunately for the materials usually used these two contributes are relatively small and they can then be neglected.

The relation between the voltage and the temperature is not linear and can be approximated by the following equation [29]:

𝛥𝑇 = ∑^𝑁_𝑛=0𝑎_𝑛∙ 𝑉^𝑛 (4)

where 𝛥𝑇 is the difference of temperature between the hot junction and the reference junction, 𝑎_𝑛 depends by the materials used in the thermocouple, V is the voltage measured in the circuit and N depends by the desired precision.

Different types of thermocouples are used today. The difference between each other relies on the range of temperature to measure and in which environment the

thermocouples are used. The most common ones are type K and N thermocouples.

(22)

22

(23)

23

3. Description of the system

3.1 Design of the parts

The main purpose of the project is to replace one of the valves of a Diesel engine (Fig.6) with a TBC sample holder. This holder will be stationary and equipped with a

measurement equipment in order to be able to measure the heat loss through the holder itself.

One of the inlet valves was chosen to be removed instead of an exhaust valve and the reason is given by the design of the engine head. Since the two outlet ducts converge to only one duct, there would have been some problems with the exhaust gas. If one of the exhaust valves was replaced, the exhaust gas coming from the other exhaust valve could easily reach the backside holder leading to a heat transfer on the surface of the holder and this would affect the measurements. Moreover since the values of pressure and temperature at the outlet duct are very high, there would be bigger problems regarding the leakages through the holder to the rocker cover.

Fig 6 Scheme of an inlet valve [30]

(24)

24

Fig. 7 shows the ordinary design of an intake valve of a Diesel engine and beside the pictures of the main part of the new design.

Fig. 7 Traditional (on the left) [31] and suggested (on the right) design of an intake valve

The design of the system has been realized trying to get a one dimensional heat flux through the holder in order to be able to use the Fourier’s law (1). This is an assumption whose validity will be discussed after the measurement from the holder because no heat flow simulation was done before the realization of the design.

The material used to manufacture the parts was a steel, denominated as Steel SS 2541 (Rp0.2=800 MPa, Rm=1000-1200 MPa, A%=11, HB=300-355).

The main part of the design of the system is the sample holder (Fig.8). The only constraint for this part is given by the hole in the head which is used to host the valve guide. This hole will host the holder but it is not possible to make it larger since the risk to reach the water-cavity is very high so the nominal diameter for the inlet valve has been used (16 mm). The hole in the holder (Fig.8 - A) has been created for two reasons.

The first one is to be able to put a pipe for a compressed air flow; this will adjust the temperature profile through the holder in case that the measured heat flux is not one- dimensional. The other reason is given by the most common use of TBC inside the combustion chamber. Generally TBC is used to insulate the piston and the distance between the piston and the oil cooling system is known in a Scania engine as well as the oil temperature inside the piston. The compressed air flow can then be used to change the temperature profile in the holder in order to get the same temperature profile as of the piston. The limitation in the manufacturing of this hole was given by the drill used;

given the diameter of the tool in the drill, it was not possible to manufacture a deeper

(25)

25 hole. The holder will be equipped with four thermocouples so four holes (Fig.8 - B) have been created on the surface and they reach the center of the valve at different heights. In order to let the wires of the thermocouples reach the upper side of the engine head four grooves have been made on the surface of the holder (Fig.8 - C). The

positioning of the grooves is such that a cylindrical symmetry is achieved in the holder in order not to have inhomogeneities in the heat transfer. A pocket (Fig.8 - D) has also been made to insert a thin metallic plate. This is used to hold the thermocouples tight to the holder and also to ensure that thermocouples reach the tip of the holes.

Fig.8 CAD model (on the left) and photo of the holder (on the right)

One of the holes for the thermocouples reaches a certain point in the holder such that the distance between the bottom of the holder and the thermocouple is the same as that mentioned before between the piston and the oil cooling system (Fig.9). The number of the thermocouples has been chosen so that it is possible to compute the behavior of the heat flux in the holder and so notice any possible deviation from the one-dimensional flux. Two thermocouples are in fact necessary to estimate a difference of temperature and compute then a heat flux. With three thermocouples it is possible to check whether the heat flux is constant between two pair of thermocouples. Four thermocouples are

(26)

26

even better because in case of malfunctioning of one thermocouple, this will not cause problems during the measurements.

Fig 9 Distance between the piston and the oil cooling system [32] (on the left) and distance between the combustion chamber and one of the thermocouple (on the right)

The second part of the system is a ring (Fig.10) which is placed at the bottom of the engine head. Its main purpose is not to let the holder fall down to the piston. The central conical hole (Fig.10 - A) is used to host the holder while the eight holes (Fig.10 - B) are used to connect the ring to the engine head with the use of screws. Another hole has been created in the ring (Fig.10 - C) and it will be used to place a thermocouple. This thermocouple and one of the thermocouples placed in the holder will have the same distance from the bottom of the system so that a possible horizontal heat flux can be measured (Fig.11). The ring has also been designed to host a copper washer. The

washer will act as a proper sealing system and ensure a good heat transfer to the head of the engine. Since the heat flux coming from the combustion chamber has mainly two ways to go, either to the holder or to the cylinder head, it is very important to ensure a good heat transfer through the washer in order to get a one-dimensional heat flux

through the holder. The geometrical constraints in the design of the ring are mainly two.

The first one is its external conical shape to avoid any interference with the inner side of the engine head. The second one is the outer diameter of the bottom side of the ring and in this case the constraint is given by the other inlet valve. A too large hole for the ring in the engine head would reach the other intake seat affecting the working of the engine.

(27)

27

Fig 10 CAD model (on the left) and photo (on the right) of the ring

Fig. 11 One of the thermocouples in the sample holder and the thermocouple placed in the ring will have the same distance from the bottom of the system.

The third part of the system is a cover (Fig. 12) and its purpose is to keep the valve in its stationary position while the engine is running. To face the force coming from the combustion gas pressure (considering the maximum value of the pressure and the dimension of the holder it has been estimated to be almost 1200 N), this cover has three pins. Two of them (Fig. 12 - A) show two threaded holes and they are used to connect the cover to the engine head thanks to two screws. One of these two pins is in direct contact with the head of the engine. The height of the other one has been chosen considering also the dimensions of the lower part of the rocker cover. In this way the cover can be easily removed and it is possible to have then a faster set up time. The connection has been assessed considering the axial force the screws had to withstand and the inner diameter of the screws themselves. It was not possible to get further information about the screws so a reasonable class of resistance 8,8 has been supposed because it is typically used in mechanical connections. In this way a safety factor over

(28)

28

ten has been obtained. The connection is then considered secure even if a lower class of resistance is used. Since the positions of the valve and of these two holes are not aligned due to geometrical constraints in the upper side of the engine head, a bending torque is generated. In order to avoid any problem related to the bending of the part, a third pin (Fig. 12 - B) has been designed. This pin is in contact with a border of the lower side of the rocker cover, counteracting the bending force. The design of the cover needed to be adjusted since the threads in the holes do not ensure the tightening of the cover to the head. Since the thickness of the pins was thin and there was then the risk to break the holes, the threads in the upper part the screws have been removed. The cover has a central hole (Fig. 12 - C) to host the holder and at its tip the pattern of the holes (Fig. 12 - D) in the holder has been reproduced in order to let the pipe and the thermocouples come out from the engine head. The constraints in this case were given by the presence of other parts in the cylinder head (like valves, injector,…). Another aspect to take into consideration was the depth of the central hole for the holder. This feature has been designed such that there is no joke between the tip of the holder and the surface of the hole in order to press the valve down and not to have any leakage at the interface between the ring and the holder.

Fig 12 CAD models (on the left and centre) and photo of the cover

(29)

29 The fourth part of the system is a disk with four bended arms (Fig.13) that is used as a shield during the combustion. This part will be placed under the ring (thanks to four screws) and it will be used to protect the bottom of the ring and the valve from the infrared radiation coming from the combustion. The purpose is to evaluate the

contribution of the radiative heat transfer during the combustion. Unlike the other parts, this shield is made of Inconel 600, a special super alloy that can be easily manufactured and can withstand high temperatures. This is an important aspect for this component since the shields is protruding into the combustion chamber and also because the only way to transfer heat is the contact with the ring so a lot of heat will be store in this part.

It is linked to the head of the engine by using the same holes of the ring. The

geometrical constraint is the height of the shield because when the piston is at the top dead center the distance between the piston and the shield is critical. The diameter of the disk has to be large enough to stop most of the infrared radiations and it does not have to be too close to the sample holder not to interfere too much with the heat transfer between the combustion gases and the holder.

Fig 13 CAD model (on the left) and photo (on the right) of the shield

(30)

30

Fig.14 shows a sectional view of the CAD model of the complete design of the system.

Fig. 14 Sectional view of the design, CAD mode

(31)

31

3.2 Technical means on the cylinder head

The head of the engine was also changed in order to fit the design of the holder and the other components (Fig.15-16). The intake valve seat has been manufactured to fit the ring and not to have any interference with it. Eight threaded holes (M4) have also been made in order to hold the ring. The constraints here were given by the presence of the other intake seat of the inlet valve (as mentioned before), the cavity for the water and the standard design of the intake seat. The first one sets the upper limit of the outer diameter of the ring, the second one is a limit condition for the depth of the threaded holes and the last one limits the choice of the holes’ position. All these constraints set also a rigid limitation on the size of the head of the screws to use in the ring. It was not possible to find a type of screw with a so small head so it was necessary to adjust the head of the screws by manufacturing them with a lathe.

Fig. 15 Bottom side of the engine head, CAD model

Fig. 16 Bottom side of the engine head

(32)

32

On the upper side of the engine head a threaded hole (M8) has been made in order to be able to connect the cover to the head (Fig.17). For the other pin in the cover the

threaded hole already present in the head is used. The position of the hole has been chosen considering the sizes of the adjacent parts of the head.

Fig. 17 Threaded hole in the upper side of the engine head; CAD model (on the left) and photo (on the right)

It has been decided to cover the inlet duct related to the holder with a metal plate (Fig.

18). This has been done in order to reduce the leakage from the combustion chamber and not to have any mixing between the leakage itself and the fresh charged air.

Moreover, this inlet duct is used to get out from the head the thermocouple of the ring and for this reason a groove between the inlet duct and the inlet manifold has been made (Fig. 18).

Figure 18 Inlet manifold with the cover plates (on the left) and inlet ducts on the head (on the right)

(33)

33 The pipe and the four thermocouples in the holder need the get out from the head of the engine so a groove has been made on the upper surface of the rocker cover (Fig. 19). To avoid any leakage of oil from this hole, it will be filled with some silicon.

Fig. 19 Modified rocker cover

Since the holder has to be stationary over the time, the mechanism that controls the movements of the valves had to be changed. For this reason the valve bridge of the rocker arm has been cut on the extremity related to the holder (Fig. 20). In this way the rocker arm does not have any control over the holder which can stay in its position while the engine is running.

Fig. 20 Cut valve bridge of the rocker arm

Another aspect to take into consideration is the Swirl number of the engine. The swirl number is a measure of how fast the air rotates in the cylinder. This number is

dimensionless and it is obtained by dividing the rotational speed of the air with the rotational speed of the engine. It is really important to get the right value of the Swirl number because if it is too small the air and the fuel would burn slowly and this would

(34)

34

result in high fuel consumption and high soot emission. If it is too high instead, the fuel sprays is bended too much and it cannot reach the edges of the combustion chamber.

The air close to this area will not take part in the combustion and this would result in soot emission and probably to a slower combustion. In a Scania engine the regular value of the Swirl number is about 1,5 but in this case, since for the project one of the inlet valves has been blocked, it would get to 3 more or less. To decrease the swirl number the shape of the port of the working inlet valve has been modified so that the air enters to the cylinder with another angle. In this way the same Swirl number as of an engine with two working inlet valves has been obtained

3.3 Coating of the parts

Three sets of systems have been designed in order to test different coatings. The first one (hardware A) has no coating on any surface and it has been used to test the design of the system. It will be possible also to take measurements from the traditional case when no coating is used. The holders and the rings of the second (hardware B) and the third (hardware C) set show instead some technical means in order to be able to attach the coating. Both of the holders have been manufactured with a reduced length to compensate for the thickness of the coating at the bottom of the holder. The rings instead have a pocket at the bottom so that a disk of coating can be placed. The coating put on a single system (for example holder B and ring B) is of the same type and of the same thickness in order to try to get the same temperature gradient and not to have then any heat flux from the ring to the holder or vice versa.

The hardware B (Fig. 21) has been equipped with a layer of 8YSZ TBC which has been attached by an external company. The method used to attach the TBC is APS (Air Plasma Spray) and it has been chosen instead of the also commonly also used EB-PVD (Electron Beam Physical Vapor Deposition) method. EB-PVD requires in fact high temperature for the coating process and this could damage the steel decreasing its mechanical properties The thickness of the coating is specified to 0.5 mm (Bond coat 100 µm Amdry 9700, Top Coat 400 µm Metco204 NS-1). Moreover the expected tolerance on the coating is about ±100 µm. The choice of this coating is due to recent tests carried out at Scania with this type and thickness of TBC. Since these tests did not report any failure, it has been decided to test this TBC in the project.

(35)

35

Fig. 21 Holder and ring coated with TBC

The hardware C was designed for a different kind of TBC. It can be defined as “Coating C” and it can be applied by brushing or spraying it onto the surface. The purpose of this hardware was (together with the test of the painted coating) to investigate the

temperature profile near the combustion chamber (aforementioned as “Swing

Temperature”). This can be done using two thinner thermocouples at the bottom of the holder. One can be attached to the metallic surface, then a thin layer of “Coating C” can be painted on the metallic surface followed by a heat treatment. Then the other

thermocouple is attached to the TBC and the process finishes with another thin layer of

“Coating C” painted at the bottom of the holder followed by heat treatment once again.

For this reason the design of the third set is slightly different; it shows two smaller grooves on the conical side of the holder that are supposed to host the wires of the two thinner thermocouples.(Fig.22).

Fig. 22 Detail of the hardware C

(36)

36

Since it has not been possible to receive the painted TBC nor the thin thermocouples on time, a different material will be tested on the third set. It is a metallic glass, defined as

“Fe SP529”. It has shown [33] high resistance to corrosion and wear and thanks to its high content of hard phases (66.2% of carbides and borides) it is assumed to have relatively low thermal conductivity. Regarding the thermal conductivity then, in a simple classification among the different coatings, Fe SP529 can be placed between the traditional steel and the more advanced TBC.

The holder and the ring of every hardware have been designed such that their surfaces facing the combustion chamber reach the same height. This is an important factor in order to have a good heat transfer through the system. After the manufacturing of the parts however, there was a certain difference of height between the ring and the holder of each hardware. These differences have been detected by using an confocal

microscope and for all the steps of the manufacturing they have been listed in Table 1:

Table 1 Difference of height between holder and ring

HARDWARE Difference of height

[µm]

A B C

After the first manufacturing

45 200 130

After the second manufacturing

45 55 45

After the coating process

/ 220 Not measured

After the first manufacturing of the parts, the holders of all the hardware were protruding a bit more toward the combustion chamber. Holders B and C in particular showed a relevant difference of height compare to hardware A. An estimate of the comparison in the heat transfer between the case with and without the height difference has been computed. Assuming that the hole in the ring is a bit larger than the designed one and assuming also the same convective heat transfer coefficient h and the same temperature difference ΔT, according to (2) the ratio between the convective heat transfers depends on the ratio between the surfaces of heat transfer. For the hardware A has been computed an increase of 4.2%, for hardware B of 9.8% and for hardware C of

(37)

37 6.3%. Since it was difficult to improve the difference on the hardware A, in order to reduce this difference in height it has been decided to adjust the length of the holder B and C. The result of the second manufacturing shows that the differences in the holders B and C reached the same order of magnitude of holder A, which is the finest tolerance achievable. It has been decided then not to manufacture manually the parts since it was not possible to use more precise tools and also because there was the risk to remove too much material. After the placement of the TBC, ring B showed this time a bigger height than the holder. The difference in the heat transfer in this case is more difficult to determine. With the same assumptions done before it is not possible to estimate a difference because the TBC attached on the surface has the shape of a disk so no change in the heat transfer area can be noticed. There is a change in the heat transfer but this will require a study of the flow close to surface of the holder in order to analyze how the convective heat transfer coefficient h and ΔT change. It was not possible moreover to adjust the thickness of the coating since there was the risk to damage the TBC. The coating for hardware C was delivered later than the TBC for hardware B and moreover there has not been enough time to test it so the difference in height after the coating process has not been done.

3.4 Thermocouples

As mentioned before, the holders and the rings will be equipped with thermocouples in order to get temperature data while running the engine. The thermocouples used are K type (Fig. 23) and they have an accuracy of ±0.3 °𝐶. This type of thermocouples can be used (from supplier information) in a range between -200 to 1260°C and it is good for oxidizing environments but not for reductive ones. Moreover, to fit the design of the systems, 1 mm wire thermocouples have been chosen.

Fig. 23 Type K thermocouple

(38)

38

Fig. 24 shows the placement of the thermocouples in the holder and in the ring. In the holder, the thermocouples pass through the holes on the top of the cover, then they come along the grooves on the holder and after that they are place inside the holes.

Fig. 24 Placement of thermocouples in the holder and in the ring

(39)

39

3.5 Accuracy of measurements

The heat fluxes through the sample holder are measured from experimental measured variables which are therefore affected by a certain variability. It is necessary then to compute the effect of these errors on the final value of the heat flux and so determine the range of its variability.

For the heat flux through the holder, since as mentioned before it is assumed to be one- dimensional, the Fourier’s law (1) is used:

𝑞̇_𝑥 = 𝑘 ∙ 𝐴 ∙ (𝑇₁− 𝑇₂) 𝐿

In this case x is the vertical direction through the holder, 𝑇₂ and 𝑇₁ are the temperatures measured by the thermocouples, L is the distance between the thermocouples (defined as x) and A is the heat transfer surface. Since the flux is assumed to be 1D, a perfect contact is assumed between the sample holder and the ring and the same material has been used for the ring and the holder, A can be assumed constant and equal to the circular surface of the sample holder (𝐴 = 𝜋𝑅²) (Fig. 25).

Fig.25 Vertical heat transfer in the holder

The error propagation formula can now be taken into consideration:

𝜎_𝑞 = √∑ (_𝜕𝑥^𝜕𝑞

𝑖𝜎_𝑥_𝑖)²

𝑁𝑖=1

(5)

(40)

40

where σq is the error in the computed heat flux q, σxi is the error on each variable xi. In this case the variables are the two temperatures measured by the thermocouples and the distance between them.

The thermal conductivity k is a parameter of the material which depends on temperature. However no data have been found regarding the dependence on temperature of the steel used in the project. It has been then assumed a linear

relationship between temperature and k and the same dependence on temperature (Table 2) of another steel (40CrMnMo7) similar to the one used in the project:

Table 2. Dependence on temperature for the thermal conductivity of 40CrMnMo7

If a linear relationship is assumed also for 40CrMnMo7 between 20 °C and 250 °C:

𝑘 = 𝑚 ∙ 𝑇 + 𝐶 (6)

the slope m of the straight line can be computed and it is m = -0.0026 W/mK². This value will be therefore used also for the steel SS 2541 and since the k at 20 °C is known (k = 38 W/mK), also the other parameter C can be computed and it is equal to C

= 38,052 W/mK. In equation (6) k depends on a temperature T which changes between the two thermocouples so it has been decided to consider the mean value of the two temperatures. The new equation for the conductive heat transfer is:

𝑞̇_𝑥 =^{𝑚∙𝐴∙(𝑇}_2∙𝑥¹²^−𝑇²²⁾+^{𝐶∙𝐴∙(𝑇}_𝑥¹^−𝑇²⁾ (7)

The error from the measurement of thermocouples, as mentioned before, is 𝜎_𝑇 =

±0.3 °C while the errors in the distance x has been computed with a X-ray analysis.

Three sets of two measurements (six in total) have been done per each couple of thermocouples and every two measurements a recalibration of the measurement equipment has been done. This has led to an error of 𝜎_𝑥 = ±0.115 mm.

For the horizontal heat flux the system has been assumed as a thick pipe (Fig. 26).

T[°C] k[W/mK]

20 34

250 33.4

500 33

(41)

41

Fig.26 Horizontal heat transfer in the system

The expression that describes this heat transfer is similar to the one used before but this time the radial direction has to be considered:

𝑞̇_𝑟 = −𝑘 ∙ 𝐴 ∙^𝑑𝑇_𝑑𝑟 (8)

The only thing that changes is the value of A which in this case depends on the radius (𝐴 = 2𝜋𝑟𝐿; where L is the height of the ring).

Solving the differential equation, the horizontal heat flux is given:

𝑞̇_𝑟 = 2𝜋𝑘𝐿 ∙ ^𝑇¹^−𝑇²

ln(^𝑟2

𝑟1) (9)

where r₁ is the radius of the holes for the thermocouples and r₂ is the distance between the thermocouples.

The error from the temperature measurements is the same but in this case there was no possibility to measure on the X-ray the distance between the thermocouples. The UNI EN ISO 22768-m has been used to determine the tolerance for the manufacturing of the holes and an error of 𝜎_𝑥= ±0.2 mm has been computed.

Since data from the engine tests show that the lower side of the system seems to be affected in a relevant way by secondary heat fluxes, it has been decided to analyze the heat flux between the thermocouples 3 and 4 (see figure 27), which is less affected by secondary heat fluxes. The accuracy of the measurements changes a lot among different percentages of fuel since for low injected fuel (and so low heat flux) the error affects the measurements in a significant way. One can state then that the accuracy (as relative error) of the measurements is of 18% for fuel load ≤25% and 7% for fuel load ≥50%

(42)

42

(43)

43

4. Engine test

Tests have been carried out in a single cylinder engine at Scania (Table 3):

Engine technical information

Displacement [cm³] 2.123

Bore/Stroke [mm] 130/160

Compression ratio (nominal value) 22

Number of valves 4

Swirl number 1.5

Length of connecting rod [mm] 255

Type of injector NGS High Power 10 holes 300 pph

Table 3 Technical information of the engine

The test plan is mainly composed of three parts, which are used to get information about different aspects of the combustion in the engine.

The first test has been done to analyze the heat flux when changing the fuel injected in the cylinder. It has been decided to have a nominal load of fuel equal to 240 mg/inj since this is a normal full load condition for the production of the engines at Scania. At full load the sweep in the pressure of the compressed air flow has been used in order to try to see any better scenario for the one-dimensionality of the heat flux.

The second test has been carried out by sweeping the rail pressure while retaining the fuel load at 50% of the nominal quantity. The rail pressure is the fuel pressure in the pipe of the injector. A high rail pressure, given a certain quantity of fuel, results in a shorter duration of fuel injection and in a higher velocity of the fuel entering in the combustion chamber. This leads to a faster combustion but also to a higher emission of NOx because of the higher temperatures. A low rail pressure results instead in a lower injection velocity of the fuel and in a slower combustion; it leads so to a poor mixing of the fuel with the air and gives soot emission.

The last test has been carried out by sweeping lambda while retaining the fuel load at 75% of the nominal quantity. Lambda is defined as:

𝜆 = ⁽

𝑚𝑓𝑢𝑒𝑙𝑚𝑎𝑖𝑟)

(^{𝑚𝑎𝑖𝑟} 𝑚𝑓𝑢𝑒𝑙)

𝑠𝑡

(10)

where (_𝑚^𝑚^𝑎𝑖𝑟

𝑓𝑢𝑒𝑙)

𝑠𝑡

refers to a combustion in which the exact amount of air is used to burn all the fuel. If the amount of fuel does not change, when lambda is high there is more

(44)

44

gas in the combustion chamber but the same quantity of energy is released from the combustion. This leads to lower values of temperature and so less heat transfer. High lambda also leads to a faster combustion and higher emission of NOx in terms of mass.

The purpose is to perform all the test points for the three different hardware in

succession (A, then B and then C) and then test again the A or the B hardware in order to notice a certain repeatability of the measurements. After that the shield can be

mounted in at least one of the coated hardware to study the contribution of the radiative heat transfer.

The test plan is summed up in Table 4.

Table 4 Test plan

First of all hardware A has been tested to verify that the design could withstand the working conditions of the engine and that there were no leakages affecting the design.

After the first tests it has been noticed that the parts of the design did not brake but there was a relevant leakage of oil from the upper side of the head of the engine. This

probably happened because the O-ring placed around the holder was not tight enough to ensure a good sealing system. To seal the system properly some silicon has been added around the O-ring and between the cover and the head. No leakage were detected in the following tests.

No. Hardware

load

rail pressure lambda load

rail pressure lambda sweep in

sweep in sweep in sweep in

Test 1

2

3

4

A

B

C

B + shield

(45)

45

5. Results

5.1 One-dimensional heat flux validity

Fig.27 Numeration of thermocouples

To check the validity of the one-dimensional flux assumption, the fluxes between thermocouples 3-4 and 5-3 (Fig. 27) have been compared. Fig. 28 shows the ratio between the vertical flux (3-4) and the horizontal one (5-3) when the hardware B has been tested.

Fig.28 Ratio between the vertical and the horizontal heat flux with sweep in the air flux for hardware B

(46)

46

On the y axis the ratio between the vertical heat flux and horizontal one is shown and on the x axis the distance from the bottom of the holder is considered. The different points in the plot show the ratio between the heat fluxes when the compressed air flow is used with different values of pressure. The graph shows that the ratio is over two for all the working conditions. All the points are also really close to each other and considering the accuracy of the measurements any difference could be noticed at all so this shows how the compressed air flux was not so effective on the system. This can also be seen in Fig.

29 which shows the heat flux profile between thermocouples no. 3 and 4 for the same hardware B.

Fig.29 Heat flux profile between thermocouples no. 3 and 4 for hardware B

The x axis shows the values of pressure of the air flow while the y axis presents the heat fluxes computed between thermocouples no.3 and 4. The errors in the plot refer to the error propagation mentioned before. For every couple of thermocouples (and so for every heat flux measurement) the relative value of error has been measured and plotted;

all the similar plots below will show the same error. One can notice that there seems to be an increase in the heat flux when increasing the pressure of the compressed air flow even if the errors are so big that nothing certain can be said about the measurement.

Another aspect can be noticed in Fig. 30 which shows the temperature profile through the holder. The y axis in this case shows the temperatures measured by the

(47)

47 thermocouples of the holder. The different lines show again the different temperature profiles with different values of pressure for the compressed air flow. Three

measurements have been taken for every value of temperature and the mean value of them has been considered in the plot. In this case the errors in the plot refer to the precision of the device used in the test cell in terms of standard deviation of the three measurements; all the similar plots below will show the same type of error.

Fig.30 Temperature profile with sweep in air for hardware B

The overall result is a decrease of the temperature in the holder when using the compressed air flow; moreover, since there is no relevant difference between the

temperature profiles, it gets difficult to simulate the temperature profile of the piston, as mentioned in the description of the holder, with this cooling system.

The real values could also differ from the ones shown before. To estimate the heat transfer it has been assumed a perfect contact between the ring and the holder. This is a conservative way to estimate it because there has been no possibility to check any possible errors in the manufacturing of the interface between the ring and the holder and also because there is always a certain contact resistance between two solids in contact.

This would result in an increase of the thermal resistance and so in a decrease of the horizontal heat flux.

(48)

48

5.2 Test of the hardware

After testing the working of the new design of the system, the thermocouples have been placed inside the holder and the ring in order to get some measurements with different working conditions. Due to some technical problems in the test cell and to the oil leakage during the first tests there has not been enough time to test the hardware C. It was not possible to have an ABA test in order to test the repeatability of the

measurements neither. Moreover after the test on the hardware A, high and unrealistic values of the volumetric efficiency have been noticed, probably due to some air leakage between the inlet manifold and the cylinder head. For this reason it will not be possible to compared strictly the results from the two hardware but at least an estimate of the difference in the use of the TBC can be given.

Fig. 31 and Fig. 32 refer to the first test and show respectively the temperature profiles of the uncoated and coated holder. The x axis presents the distance from the bottom of the holder while the y axis shows the temperature measured by the thermocouples;

different lines in the plot show different percentages of injected fuel.

Fig.31 Temperature profile with sweep in fuel load for hardware A

(49)

49

Fig.32 Temperature profile with sweep in fuel load for hardware B

From both of the plots one can notice that temperatures get higher as fuel injected increases because the energy released during the combustion increases. A decrease of temperature in the coated parts for high percentages of fuel is also visible.

The plot in Fig. 33 shows the behaviors of the two hardware A and B at full load condition. It can be easily noticed that the two heat flux profiles do not have the same behavior. This could be due to some differences in the manufacturing of the two sets of hardware leading to different contact points at the interface between the ring and the holder. Another alternative could be the fact that the force of the pressure coming from the combustion gas deforms the two hardware differently; hardware B has in fact a pocket at the bottom and so it can be bended a bit more easily.

(50)

50

Fig. 33 Comparison of heat flux profile at full load between hardware A and B

Also for this reason only the flux between the thermocouples number 3 and 4 is considered and it is visible in Fig. 34. The plots show how the heat flux changes when the percentage of injected fuel is changed. One can notice that not only the temperature increases (as shown in Fig. 31 and Fig. 32), but also the heat flux increases as the quantity of fuel rises. In both cases a certain linearity can be detected and a smaller heat flux is present when coated parts are used.

Fig. 34 Heat flux between thermocouples no. 3 and 4 for hardware A (on the left) and B (on the right)

(51)

51 Fig. 35 refers instead to the second part of the test plan and shows the temperature profile of the design when the coated parts are used. This time the different lines in the plot refer to different values of the rail pressure.

Fig. 35 Temperature profile with sweep in rail pressure for hardware B

As expected, as the rail pressure increases, the temperatures increase as well and also some kind of linearity between temperature and rail pressure can be detected as it is shown in Fig. 36 (Data from thermocouple n.4)

Fig.36 Behavior of temperature as a function of the rail pressure for hardware B

(52)

52

The test with sweep in rail pressure for the hardware A shows the same trend of the hardware B but with higher temperatures. The study of the heat flux shows the same problem as the test with the sweep in load regarding the possible errors in

manufacturing and the different deformation of the parts (Fig. 37).

Fig.37 Comparison of heat flux profile at rail pressure 2000 bar between hardware A and B

For this reason also in this case the heat flux between the thermocouples 3 and 4 is taken into consideration. Fig. 38 shows the behavior of that heat flux respect with the rail pressure. In this plot it can be seen that the errors in the measurements are so big that nothing can be said about how the heat transfer changes when sweeping the rail pressure even if there seems to be a trend of an increase of heat flux when increasing the rail pressure.

(53)

53

Fig. 38 Heat flux profile between thermocouples no. 3 and 4 for hardware B

The plot in Fig. 39 refers instead to the last part of the test plan and shows different temperature profiles for different values of lambda for the system with coated parts. On the x axis the distance from the bottom of the holder is given while on the y axis the temperature measured by the thermocouples is shown. Different lines in this case refer to different values of lambda.

Fig. 39 Temperature profile with sweep in lambda for hardware B

As expected when lambda increases the overall result is a decrease of the temperature.

Heat flux measurement device