STOCKHOLM SWEDEN 2020,
Water blow out phenomena
inside a heavy truck silencer
ROHIT SURAM VENKATA SUBRAMANIYAM
KTH ROYAL INSTITUTE OF TECHNOLOGY
SCHOOL OF INDUSTRIAL ENGINEERING AND MANAGEMENT
A Master Thesis Report on
Water blow out phenomena inside
a heavy truck silencer
Rohit Suram Venkata Subramaniyam
Performed in Engine after treatment control group
(NCFF) at
Scania CV AB, Södertalje, Sweden.
Master of Science Thesis TRITA-ITM-EX 2020:533 KTH Industrial Engineering and Management
Machine Design SE-100 44 STOCKHOLM
Examensarbete TRITA-ITM-EX 2020:533
Vatten blåser ut fenomen i en tung ljuddämpare
Rohit Suram Venkata Subramaniyam
Godkänt
2020-09-22
Examinator
Dr. Andreas Cronhjort
Handledare
Dan Edlund, Dr. Andreas Cronhjort
Uppdragsgivare
Scania CV AB
Kontaktperson
Dan Edlund
Sammanfattning
NOx sensorer har blivit viktiga komponenter i utvecklingen av ett effektivt avgassystem för tunga fordon under de senaste åren. När det ackumulerade vattnet i ljuddämparen stänker på NOx-sensorn kan det orsaka permanenta sprickor i sensorn. För att skydda sensorn från detta misslyckande utvecklas en daggpunktsstrategi på Scania. Detta är viktigt att förutsäga när det är säkert att slå på NOx-sensorn utan att skada den. Men strategin innehåller för närvarande endast fasöverföringsfenomenen och försummar effekten av att vatten blåser ut fenomen inuti ljuddämparen.
För att undersöka effekten av utblåsning av vatten utformas en experimentell testmetod och experimenten utförs på olika platser i ljuddämparen. Resultaten från experimenten visar att effekten av vattenblåsning verkligen är en viktig faktor för att utveckla en bättre daggpunktsstrategimodell. För en vald plats samlas mängden vatten kvar efter utblåsning och den tid det tar för utblåsningsfasen som data från experimenten.
En matematisk modell för fenomen för vattenblåsning utvecklas i MATLAB. Modellen uppskattar den maximala mängden vatten som kan finnas i ljuddämparens undervolymer med tanke på effekten av vatten som blåser ut. Modellen verifieras med experimentdata för en Scania CAS1 ljuddämpare. Kalibreringsriktlinjer för den utvecklade utblåsningsmodellen dokumenteras också i denna rapport.
Vatten blåser in efter behandlingsanordningar, Daggpunktsstrategi, Avgas efter behandlingssystem, NOx-sensorer, Vattenbildning i avgaserna.
Water blow out phenomena inside a heavy truck silencer
Rohit Suram Venkata Subramaniyam
Approved
2020-09-22
Examiner
Dr. Andreas Cronhjort
Supervisor
Dan Edlund, Dr. Andreas Cronhjort
Commissioner
Scania CV AB
Contact person
Dan Edlund
Abstract
NOx sensors have become salient components in the development of efficient exhaust after treatment system for heavy duty vehicles in the past few years. When the accumulated water inside the silencer splashes on to the NOx sensor, it can cause permanent cracks in the sensor. To protect the sensor from this mode of failure, a dew point strategy is developed at Scania. This is important to predict when it is safe to switch on the NOx sensor without causing any harm to it. But the strategy currently includes only the phase transfer phenomena and neglects the effect of the water blow out phenomena inside the silencer.
To investigate the effect of water blow out, an experimental test method is designed and the experiments are conducted at different locations in the silencer. The results from the experiments shows that the effect of water blow out is certainly an important factor to develop a better dew point strategy model. For a selected location, the quantity of water remaining after blow out and the time taken for the blow out phase are collected as data from the experiments.
A mathematical model for the water blow out phenomena is developed in MATLAB. The model estimates the maximum amount of water which could be present in all the sub- volumes of the silencer considering the effect of water blow out. The model is verified with the experimental data for a Scania CAS1 silencer. Calibration guidelines for the developed blow out model are also documented in this report.
Water blow out in after treatment devices, Dew point strategy, Exhaust after treatment system, NOx sensors, Water formation in the exhaust.
Firstly, I would like to thank my supervisors Mr.Dan Edlund at Scania CV AB and Dr.Andreas Cronhjort from Internal Combustion Engines department at KTH Royal Institute of Technology. They consistently guided me in the right path whenever I faced problems in my research work.
I would like to thank my manager Dr.Robin Nyström at Scania CV AB for his continued support throughout the project and for giving me a wonderful opportunity to complete my master thesis at Scania CV AB. I would also like to thank all my group members at Engine after treatment control group (NCFF) for their valuable co-operation.
I would like to thank the technical experts who assisted me with the design of the test method, the blow out experiment and the CAD model of the silencer for the project:
• Daniel Engström - Senior Engineer at Scania CV AB
• Fredrik Thelin - Development Engineer at Scania CV AB
• Jan Karlsson - Senior Engineer at Scania CV AB
In addition to this, I would like to thank Scania CV AB for providing all the support, resources and infrastructure essential to finish the project.
I am also very thankful to my parents, friends and family members for their constant motivation and unswerving support. This achievement would not have been possible without them.
Thank you.
Notations
w Weight of a compound kg
M Molecular weight of a compound kg/mol
V Volume L
p Partial pressure of a compound bar
ρcell Cell density kg/m3
x Total height of cells mm
I Initial amount of water L
λ Air-fuel ratio −
F Fraction of amount of water −
C Filter parameter −
y Mole fraction of a compound −
Abbreviations
C16H34 Hexadecane C8H18 Iso-octane
CAD Computer-Aided Design CAN Controller Area Network CO Carbon monoxide
CO2 Carbon-dioxide
CP SI Cells Per Square Inch DOC Diesel Oxidation Catalyst DP F Diesel Particulate Filter ECU Engine Control Unit H2O Water
HC Hydrocarbons N2 Nitrogen
N Ox Nitrogen oxides O2 Oxygen
P M Particulate matter
SCR Selective Catalytic Reduction Y SZ Yttrium-stabilized zirconium
List of Tables
xList of Figures
xi1 Introduction
11.1 Objectives . . . 2
2 Literature Study
3 2.1 Diesel emissions . . . 42.2 Basics of NOx sensor . . . 5
2.3 Estimation of dew point temperature . . . 6
2.3.1 Gasoline engines . . . 7
2.3.2 Diesel engines . . . 8
2.4 Cold start of the engine . . . 9
2.5 Condensation and evaporation of water . . . 9
2.5.1 Condensation . . . 9
2.5.2 Evaporation . . . 10
2.6 Design of the silencer . . . 11
2.7 Dew point strategies for a sensor . . . 12
2.7.1 Dew point strategy at Scania . . . 13
2.8 Different experimental setups . . . 13
2.9 First order recursive filter . . . 15
2.10 Shape of cells in substrates . . . 16
3 Water blow out Experiment
20 3.1 Setup . . . 203.2 Procedure. . . 25
3.3 Data . . . 27
4 Model Development
334.1 Maximum amount of water (No mass flow) . . . 34
4.2 Fraction of maximum amount of water (For a certain mass flow) . . . . 39
4.3 Equilibrium maximum amount of water . . . 42
4.4 Filter parameter . . . 48
4.5 Model-based maximum amount of water . . . 53
4.6 Verification of the model . . . 56
5 Results
586 Conclusions
637 Future work
65Bibliography
68Appendices
71A Code for Equilibrium maximum amount of water
71B MATLAB code for the filter
76C Code for filter parameter
79D Code for Model-based maximum amount of water
823.1.1 List of various places where water can get trapped in the silencer . . . . 20
3.2.1 Amount of water poured into the silencer at different locations . . . 26
3.3.1 Data obtained from blow out experiment at inlet water pocket . . . 28
3.3.2 Data for 90% blow out phase . . . 29
3.3.3 Data obtained from blow out experiment at water pocket before evaporation tube . . . 29
3.3.4Data obtained from blow out experiment at water pocket after evaporation tube . . . 30
3.3.5 Data for 90% blow out phase . . . 30
3.3.6Data obtained from blow out experiment at outlet . . . 31
3.3.7 Data for 90% blow out phase . . . 31
3.3.8Data for blow out experiment in DPF . . . 31
3.3.9Data for blow out experiment in catalyst 2 . . . 32
4.0.1 Different sub-volumes inside the silencer . . . 33
4.1.1 Maximum amount of water (No mass flow) . . . 39
4.2.1 Fraction of maximum amount of water for sub-volume 2 . . . 40
4.2.2Fraction of maximum amount of water for sub-volume 4 . . . 41
4.2.3Fraction of maximum amount of water for sub-volume 7 . . . 41
4.4.1 Filter parameters for different mass flow rates in sub-volume 2 . . . 50
4.4.2Filter parameters for different mass flow rates in sub-volume 4 . . . . 51
4.4.3Filter parameters for different mass flow rates in sub-volume 7 . . . 52
5.0.1 Details for the equilibrium maximum amount of water code . . . 59
5.0.2Details to tune the filter . . . 60
5.0.3Details for the filter parameter code . . . 61
5.0.4Details for the model-based maximum amount of water code . . . 62
2.1.1 Emissions from diesel combustion [3] . . . 4
2.2.1 Schematic of a NOx sensor [2] . . . 5
2.2.2NOx sensor [15] . . . 6
2.3.1 Vapour pressure curve of water [5] . . . 8
2.5.1 Condensation process [14] . . . 10
2.5.2 Evaporation process [14] . . . 10
2.6.1 Schematic of the silencer . . . 11
2.9.1 First order recursive filter . . . 16
2.10.1Square shaped cell in a substrate [10] . . . 17
2.10.2Sine wave . . . 18
2.10.3One half of a sine wave . . . 18
2.10.4Wave shaped cells in the substrate . . . 19
3.1.1 Inlet water pocket . . . 21
3.1.2 Water pocket after evaporation tube . . . 21
3.1.3 Space before catalyst 2 . . . 22
3.1.4 Outlet . . . 22
3.1.5 IPLEX G Lite portbale videoscope . . . 23
3.1.6 Funnel, Hose pipes and Syringe used in the experiment . . . 23
3.1.7 Weighing machine used in the experiment . . . 24
3.1.8 Control room of AR1 engine test rig . . . 24
3.1.9 Complete setup of the silencer in AR1 engine test rig . . . 25
4.0.1 Schematic of water blow out model . . . 34
4.3.1 Mass flow rate vs Time . . . 43 4.3.2Equilibrium maximum amount of water vs flow rate for sub-volume 2 . 44 4.3.3Equilibrium maximum amount of water vs flow rate for sub-volume 3 . 45
4.3.4Equilibrium maximum amount of water vs flow rate for sub-volume 4 . 46 4.3.5 Equilibrium maximum amount of water vs flow rate for sub-volume 6 . 47 4.3.6Equilibrium maximum amount of water vs flow rate for sub-volume 7 . 48
4.4.1 Sample filter for 10 kg/min in sub-volume 4 . . . 50
4.4.2Filter parameter vs mass flow rate for sub-volume 2 . . . 51
4.4.3Filter parameter vs mass flow rate for sub-volume 4 . . . 52
4.4.4Filter parameter vs mass flow rate for sub-volume 7 . . . 53
4.5.1 Model-based and equilibrium maximum amount of water vs flow rate . 54 4.5.2 Model-based and equilibrium maximum amount of water vs flow rate . 55 4.5.3 Model-based and equilibrium maximum amount of water vs flow rate . 55 4.6.1 Model-based vs equilibrium maximum amount of water . . . 56
4.6.2Model-based vs equilibrium maximum amount of water . . . 57
4.6.3Model-based vs equilibrium maximum amount of water . . . 57
7.0.1 Sprays which can be used for substrates . . . 65
Introduction
Non-ideal combustion of diesel fuel yields many harmful emissions to the environment like carbon monoxide (CO), nitrogen oxides (NOx) and particulate matter (PM) [3].
NOx is a combination of Nitrous oxide (NO) and Nitric oxide (NO2) gases. These gases can cause serious health issues in humans like breathing problems, chronically reduced lung infection, eye irritation [11]. These gases are very hazardous to environment as they can form acid rains and smog [11]. Hence, it is important to reduce the NOx emissions from the diesel exhaust. For this purpose, all the automotive manufacturers are subjected to stringent emission legislation standards [6]. To meet these standards and reduce harmful emissions, an advanced exhaust after treatment system is designed at Scania. NOx sensors are a very important part of this exhaust after treatment system at Scania. They are used to measure NOx levels at different places in the silencer.
NOx sensors are built with small ceramic tiles and they are heated approximately to 800°C to detect the NOx level in the exhaust[10]. There are several failure modes for a NOx sensor and many other ways for the sensor to break. However, one of the most common mode of failure is when the water splash from accumulated water in the after treatment system cracks the heated permeable outer shroud of the sensor [10]. Generally, water can be formed in the silencer either during the idling phase or during the cold start of an engine [11]. The water thus formed can get adsorbed onto the substrates and walls of the silencer. When this water splashes onto the NOx sensor, it causes permanent cracks in the sensor [10]. It is of utmost importance to ensure that the NOx sensor does not get damaged by this mode of failure. Hence, a dew point strategy is essential to evaluate when it is safe to turn on the NOx sensor.
The current dew point strategy at Scania is to wait a set amount of time after a specific temperature is reached inside the silencer [10]. This set amount of time is termed as the startup time for a NOx sensor. It is assumed that all the accumulated water in the silencer has evaporated after reaching this startup time and hence it is safe to turn on the NOx sensor. However, the start-up time for a NOx sensor estimated using this strategy is found to be very high and needs to be reduced substantially.
To reduce the start-up time for a NOx sensor and calculate the accumulated water in the silencer more accurately, a new dew point strategy has already been developed at Scania. This strategy takes into account the phase transfer phenomena (condensation and evaporation of water) during the evaluation of start-up time for a NOx sensor.
However, it neglects the water blow out phenomena from one sub volume to another sub volume inside the silencer. Hence, it is essential to understand the water blow out phenomena inside the silencer and create a model for the same. This will help in developing a better dew point strategy and reduce the start-up time for a NOx sensor to an acceptable level.
1.1 Objectives
The goal of this thesis is to understand and investigate the effect of water blow out phenomena inside the silencer. There are four important objectives formulated in this thesis project. They are:
• Design an experimental test method to investigate the water blow out phenomena at different locations inside the silencer.
• Develop a mathematical model for water blow out phenomena.
• Verification of the model using the data obtained from experiments for a Scania CAS1 silencer.
• Create calibration guidelines on how to use the developed model.
Literature Study
It is essential to analyse the formation of water inside the silencer and its behaviour under different conditions. For this purpose, a literature study is carried out to understand the phase transfer phenomena (condensation and evaporation of water) inside the silencer. This knowledge is helpful to determine when and where the water condenses inside the silencer and when it starts to evaporate.
This literature study also includes the design analysis of the silencer mainly due to two reasons. The first one is to identify the probable locations where water can get trapped inside the silencer. The second one is to locate the exact positions of different exhaust gas sensors like NOx sensor, temperature sensor and urea dosing sensor in the silencer. This thesis project aims to develop a better dew point strategy for a NOx sensor in specific. Hence, the study is also focused on learning the basics of a NOx sensor and its principle of operation.
Comprehensive understanding of the new dew point strategy developed in Scania is also a vital part of this literature study. The effect of water blow out has to be investigated experimentally where a test method needs to be designed for this purpose.
In order to design a test method, various experimental test procedures which are formulated and reported before are analysed in this literature study.
2.1 Diesel emissions
Diesel is a non-renewable fuel whose composition can be assumed to be a hexadecane (C16H34) [5]. During an ideal combustion process, diesel fuel produces carbon dioxide (CO2), water (H2O), nitrogen (N2) and oxygen (O2) [3]. None of these diesel emissions are harmful to environment other than carbon dioxide (CO2) [3]. This is because of the fact that carbon dioxide (CO2) is a green house gas.
Non-ideal combustion of diesel fuel emits other pollutants along with normal diesel emissions. These pollutant emissions are harmful to the environment in many ways [3]. Some of the most common pollutant emissions from diesel combustion includes nitrogen oxides (NOx), particulate matter (PM), carbon monoxide (CO) and hydrocarbons (HC) [3]. The overall concentration of emissions from diesel combustion process is shown in Figure 2.1.1 [3].
Figure 2.1.1: Emissions from diesel combustion [3]
As it can be observed from Figure 2.1.1, the relative concentration of pollutant emissions formed during diesel combustion is low [3]. However, these pollutant emissions depend on many other factors and the composition shown in Figure 2.1.1 is a typical composition for a specific case. By using exhaust after treatment devices like particulate filters and NOx reduction catalysts, these pollutant emissions reduces by a larger extent and almost becomes zero [3].
2.2 Basics of NOx sensor
A NOx sensor is used to detect the amount of NOx level present in the exhaust gas stream. A typical NOx sensor is a combination of two or three electrochemical cells arranged adjacently as shown in Figure 2.2.1 [2].
Figure 2.2.1: Schematic of a NOx sensor [2]
There are three important elements in a NOx sensor, as given below:
• Oxygen pump cell - In this cell, oxygen (O2) gas which is present in the exhaust is pumped out into the atmosphere through a zirconium electrolyte [7]. The cell reaction for the O2 pump cell is shown in Equation 2.1 [7].
O2+ 4e− → 2O2− (2.1)
• Zirconia electrolyte - There are many materials available which can be used as electrodes like platinum, rhodium and palladium in a NOx sensor [2]. However, a solid state yttrium-stabilized zirconium (YSZ ceramics) is the most commonly used material for an electrolyte in a commercial NOx sensor [2]. The major advantage of using YSZ ceramics is that the conductivity of O2 ions is very high at high temperatures [2].
• NO sensing cell - In this cell, NOx which is left in the exhaust gas is decomposed to produce oxygen ions [7]. These oxygen ions provides an estimate of the NOx content in the exhaust. The overall cell reaction in the NO sensing cell is shown in Equation 2.2 [7].
N O + 2e−→ 1
2N2+ O2− (2.2)
A NOx sensor can also be used to detect the oxygen concentration in addition to the NOx concentration present in the exhaust [15]. The output from the oxygen pump cell can be used to detect the oxygen concentration thereby estimating the air-fuel equivalence ratio (λ) as well. The output from the NO sensing cell can be used to evaluate the NOx content in the exhaust.
The term λ is defined as the ratio of actual air-fuel ratio (AFR) to stoichiometric AFR.
For stoichiometric combustion, λ is equal to 1. For lean mixtures, the value of λ is less than one and for rich mixtures the value is less than one [16].
λ = AF R
AF Rstoich (2.3)
A commercial NOx sensor contains three major components: the sensor, a control module and a connecting cable which can be seen in Figure 2.2.2 [15].
Figure 2.2.2: NOx sensor [15]
2.3 Estimation of dew point temperature
The dew point temperature is the temperature at which the water vapour present in the exhaust starts to condense and form liquid water droplets. This dew point temperature depends on the concentration of water vapour in the exhaust and the total pressure of the exhaust [14]. The value of dew point temperature for gasoline and diesel engines are found to be different and varies with the value of λ [5].
2.3.1 Gasoline engines
Gasoline engines operate mostly at stoichiometric conditions with λ=1. By assuming the gasoline fuel to be isooctane (C8H18), the complete combustion reaction for a gasoline engine is shown in Equation 2.4 [5].
C8H18+25
2 (O2+ 3, 76N2)→ 8CO2+ 9H2O + 47N2 (2.4)
For 1 kg of gasoline fuel burnt, around 1.4 kg of water vapour is formed which can be calculated using Equation 2.4. This constitutes to around 10% weight ratio of water vapour formed in the exhaust.
The mole fraction of water vapour thus formed in the exhaust can be calculated using Equation 2.5 [5].
yH2O=
wH2O MH2O wH2O
MH2O + MwN2
N2 + MwCO2
CO2
(2.5)
From Equation 2.5, the mole fraction of water vapour is around 0.14 for one kg of gasoline fuel burned. This represents about 14% mole ratio of the water vapour formed in the exhaust.
Partial pressure of water vapour formed in exhaust can be estimated using Equation 2.6 [5].
pH2O = yH2O∗ pexhaust (2.6)
where, pexhaust is the total pressure of the exhaust. Assuming the total pressure of exhaust to be 1 bar for a gasoline engine, the partial pressure of water vapour formed in the exhaust is calculated to be 0.14 bar according to Equation 2.6.
The vapour pressure curve of water is as shown in Figure 2.3.1 [5]. From Figure 2.3.1, a partial pressure of 0.14 bar of water vapour corresponds to a temperature of approximately 53°C. Thus the dew point temperature for a gasoline engine can be derived to be around 53°C [5].
Figure 2.3.1: Vapour pressure curve of water [5]
2.3.2 Diesel engines
Diesel engines operate at lean conditions with λ>1 at most of the times. The dew point temperature for diesel engines is largely affected by the variations in λ value. However, assuming stoichiometric combustion with λ=1 and composition of diesel fuel to be hexadecane (C16H34), the combustion reaction for diesel engines can be formulated as shown in Equation 2.7 [5].
C16H34+49
2 (O2+ 3, 76N2)→ 16CO2+ 17H2O + 92N2 (2.7) By following the same procedure as discussed in section 2.3.1, the dew point temperature for a diesel engine is estimated to be around 52°C [5]. This value is slightly lower than the dew point temperature for a gasoline engine, because of the lower water concentration in the exhaust [5]. Since diesel engines operate with lean mixtures at most of the times, the water concentration is less when compared with that of gasoline engines [5]. However, the maximum dew point is always attained when the mixture is operated at stoichiometric ratio i.e., λ=1. For λ>1, the dew point temperature is reduced due to dilution effect. The dew point temperature reduces a bit for values of λ<1 as well, because of the unburnt fuel in the exhaust [5].
2.4 Cold start of the engine
During the cold start of an engine, the temperatures of the pipe walls and the solid substrates (DOC, DPF and SCR) are generally very low (< 50°C) at ambient conditions.
As discussed in the section 2.3, the combustion process of both diesel and gasoline fuels results in the formation of water vapour as one of the by-products. For stoichiometric combustion (λ=1), around 13% of water vapour is formed along with other gases [11].
Under the cold start conditions, the water vapour which is formed as a by-product can condense into liquid and form droplets on the pipe walls [11]. In addition to this, the condensed liquid can get adsorbed on to the solid substrates as well [13]. Hence, analysing the behaviour of silencer during cold start is a challenging task.
2.5 Condensation and evaporation of water
Water vapour formed in the engine exhaust condenses into liquid droplets when the temperature of the surface walls falls below the dew point temperature. These condensed liquid droplets evaporate when the wall temperature of the surface exceeds the dew point temperature [14]. Hence, it is essential to study the condensation and evaporation phenomena to understand the formation of water in different exhaust after treatment devices.
2.5.1 Condensation
The dew point curve for the condensation of water phenomena is shown in Figure 2.5.1. In the Figure 2.5.1, x axis indicates the temperature of the exhaust gas and y axis indicates the concentration of water in gas phase. During the starting phase (cold start of the engine), the exhaust gas temperature is more than that of the wall temperature.
Hence, heat transfer takes place from hot exhaust gas to the cold solid substrate or the pipe walls. This leads to reduction of the gas temperature below dew point where the water vapour is saturated and gets condensed into water droplets. During condensation, the latent heat of condensation is transferred from the condensed liquid onto the solid substrate or the wall. Hence, the heat transfer takes place between gas and solid keeping the liquid temperature constant [14] .
Figure 2.5.1: Condensation process [14]
2.5.2 Evaporation
Figure 2.5.2 shows the dew point curve for the evaporation of water in the exhaust.
After the initial phase, there is some condensed liquid on the solid substrate or the wall. The temperature of the exhaust gas is more than that of the wall temperature in this process. The condensed liquid on the walls or solid substrates starts to evaporate increasing the concentration of water vapour in the exhaust. Here, the latent heat of vaporisation is transferred from the condensed liquid to gas keeping the solid temperature constant. Both the condensation and evaporation process stops after the walls are completely dry [14].
Figure 2.5.2: Evaporation process [14]
2.6 Design of the silencer
In sections 2.3, 2.4 and 2.5, different ways in which water gets formed inside the silencer have been discussed briefly. This section focuses on understanding the design of the silencer at Scania, which is very important to identify the possible locations where water can get trapped inside the silencer. The schematic design of the silencer is shown in Figure 2.6.1.
Figure 2.6.1: Schematic of the silencer
Inlet tube
The inlet tube of the silencer consists of a temperature sensor, NOx sensor and urea dosing injector which are mounted immediately after the down pipe. The inlet of the silencer is shown with blue colour tube in Figure 2.6.1.
Catalyst 1 and DPF
Catalyst 1 and Diesel particulate filter (DPF) are located after the inlet tube. Catalyst 1 is shown in yellow-orange colour and DPF is shown in green colour in Figure 2.6.1.
There is only one temperature sensor located in this part near DPF which can be seen in Figure 2.6.1.
Evaporation tube
The evaporation tube is represented with blue colour in Figure 2.6.1. Before the evaporation tube, there is one temperature sensor (green) and one mid bed NOx sensor (red) mounted. In the evaporation tube, there is one urea dosing injector (yellow) mounted as well. Therefore, there are a total of three sensors mounted in this part of the silencer which can be observed in Figure 2.6.1. The long tube connecting the evaporation tube and catalyst 2 is called as Acoustic pipe.
Catalyst 2
Catalyst 2 is divided into two parts: The upper part and the lower part as shown in Figure 2.6.1. The exhaust gas which is coming into this part of the silencer is divided equally into the upper and lower parts of the catalyst 2. There are no sensors mounted at this location of the silencer.
Outlet
The outlet of the silencer for this engine can have a vertical tail pipe or a horizontal tail pipe as shown in Figure 2.6.1. If the engine has a vertical tail pipe, both the temperature sensor and the NOx sensor are located in the outlet tube. In case of a horizontal tail pipe, both the NOx sensor and the temperature sensor are mounted in the tailpipe which can be seen in Figure 2.6.1.
2.7 Dew point strategies for a sensor
Generally, dew point temperature estimation and evaluation of start-up time of a exhaust gas sensor depends on different models running in the engine control unit (ECU) [12]. These models can make use of several measurement data from exhaust system like mass flow rate of exhaust, exhaust gas temperature, lambda value etc. to estimate the dew point temperature. Once the ECU determines the start up time of a sensor and predicts that it is safe to turn on the sensor, a dew point status signal is transmitted through controller area network (CAN) bus for the sensor to operate from this point [12].
2.7.1 Dew point strategy at Scania
Scania’s present dew point strategy is to wait for a fixed amount of time to start the sensor after the exhaust temperature reaches a standard value. This strategy takes a lot of time for the sensor to start operating in the exhaust. To overcome these issues, a new dew point strategy model is developed at Scania.
In the new dew point strategy model, mass balance with the difference between saturated steam pressure and actual steam pressure is considered to estimate when it is safe to turn on the sensor [10]. This means that the strategy considers only the phase transfer (condensation-evaporation) of water phenomena in the start up time evaluation [10]. The strategy neglects any kind of water blow out effect i.e., the transfer of water in liquid phase from one part to another part inside the silencer. However, a model which includes both the phase transfer and water blow out is necessary for more accurate results.
2.8 Different experimental setups
Several experimental test methods which are already designed and tested in the past are discussed in this section. This is useful in designing a test method for the water blow out experiment in the silencer.
Setup-1
Schmeisser et al. conducted experiments to analyse the effect of H2O adsorption or desorption phenomena on a zeolite SCR catalyst under cold start conditions. The experiments were performed on a test rig level and an engine test bench level. For the test rig method, a sample of SCR catalyst was taken and heated up to 550°C for 5 hours. Later, O2 gas containing a H2O content of 10% was blown into the SCR at high speed[13]. For the engine test bench method, cold start phenomena under real exhaust conditions was tested. This was done by hydro thermally (air + H2O) conditioning the SCR catalyst for 5 hours at 600°C in an oven [13]. These two experimental methods were used to perform H2O adsorption or desorption experiment.
The experiment starts from a low temperature of 40°C feeding only N2 gas into SCR catalyst for 200 s [13]. After 200 s, water was included with a content of 8% v/v along with the N2gas for 400 s. Later, only N2gas was fed into the SCR catalyst for sometime.
The results obtained from this experiment was helpful in understanding the effect of water adsorption or desorption phenomena on SCR catalyst [13].
Setup-2
Ola Stenlåås et al. discusses two different test methods which are useful for the water blow out experiment. In the first test method, a number of cold starts were done continuously by having suitable time gap between them so that the engine will come back to ambient temperature each and every time. This leads to the formation of a definite quantity of water in several parts of the silencer [10]. Now, a hot driving process was performed such that the entire water in the silencer can be evaporated.
The time taken from the start of the experiment till formation of white smoke at the end of silencer was noted. Also, the time taken between the formation of white smoke till it ends was recorded [10]. These times are used in the dew point strategy model to estimate several different parameters. By the end of hot driving process, it was concluded that the entire water formed in the silencer is emptied [10].
In the second test method, an alternative way to introduce water inside the silencer was discussed. A fixed amount of water was poured into the silencer by disassembling different sensors present at different locations of the silencer [10]. After introducing water, the hot driving process which is discussed in the first test method was repeated [10]. However, the second test method was not tested as a part of the experiment. This is provided as a possible method to introduce water inside the silencer [10].
Setup-3
Arumugam Sakunthalai et al. focused on the analysis of cold start and idle behaviour of an engine at different cold ambient conditions. For this purpose, a test methodology was developed. In the methodology, an engine under investigation was soaked at a particular temperature for 8 hours in a cold cell maintaining intake air, coolant, oil and fuel temperatures at a constant level [1]. Later, the engine was started and run for 3 min and the exhaust emissions in the upstream of after treatment devices are calculated.
However, the temperatures of oil, fuel, coolant and intake air are still maintained at the
same constant level as before [1]. The cold start experiment was repeated for 3 times and the average value of exhaust emissions from these 3 cold start experiments was analysed for better results. The engine was cranked with fuel cutoff to take out all the residual gases at the end of each cold start experiment [1].
2.9 First order recursive filter
During this project, a first order recursive filter is used to delay an input signal by a certain amount of time. This is an exponential filter and the basic equation for this filter is as shown in Equation 2.8.
y[n] = x[n]∗ (1 − C) + y[n − 1] ∗ C (2.8) Where,
y[n] : Current or Filtered output signal x[n] : Current input signal
C : Filter parameter
y[n-1] : Output signal from previous iteration
The filtered output signal depends on the filter parameter, input signal and the output signal from previous iteration. Filter parameter is given by Equation 2.9.
C = e−hτ (2.9)
Where,
C : Filter parameter h : loop period (s) τ: Time constant (s)
Loop period (h) is 0.01 s when the filter is called from a 100 Hz loop and 0.05 s if it is called from a 20 Hz loop. Cutoff frequency for this filter is given in Equation 2.10.
f = 1
2πτ (2.10)
Where,
f : Cut-off frequency (Hz) τ: Time constant (s)
An example of the first order recursive filter is shown in Figure 2.9.1. In Figure 2.9.1, Y-axis is the filtered output value and X-axis is the number of iterations. The scale on X and Y axes are set as an example for demonstration purpose. At first, the filtered output is initialized to a known predefined value (Initial value). Later, the filtered output reduces to a desired value in a specific number of iterations. The most important parameter which can be tuned in Equation 2.8 is the filter parameter (C). This can change the filter behaviour and the result as well. By tuning the value of filter parameter (C), the number of iterations taken by the filter can be altered. It can increased or decreased to reach the desired value from initial value in a given number of iterations. For example, in Figure 2.9.1 the filtered output value reduces from to 10 (Initial value) to 3 (Desired value) in 32 iterations.
Figure 2.9.1: First order recursive filter
2.10 Shape of cells in substrates
Cells of substrates in Scania are manufactured in two different shapes. The first type is a square shaped cell and the second one is a wave shaped cell. The shape of the cell depends on the selection of the silencer for testing [8], [4].
Square shaped cell
A square shaped cell in a substrate with an inner side of x and wall thickness of h is shown in Figure 2.10.1. Generally for catalyst 1 substrate, the void inside a cell is around 45-55 % [9]. Hence, an average value of 50 % void is assumed for catalyst 1 which means the rest (50 %) is the coated surface for the substrate. Equation 2.11 refers to this assumption [10].
Figure 2.10.1: Square shaped cell in a substrate [10]
The two governing equations used to calculate the values of x and h for any square shaped substrate are shown in Equations 2.11, 2.12 [10].
x2 = 4∗ h ∗ x + 4 ∗ h2 (2.11)
(x + 2∗ h)2 = 1
ρcell (2.12)
Where,
x : Total height of cell in a substrate (mm) h : Wall thickness of a substrate (mm)
ρcell: Cell density of a substrate (Cells Per Square Inch)
Maximum amount of water in any substrate with square shaped cells is given by Equation 2.13 [10].
M axH20 = 0.3
x ∗ V (2.13)
Where,
V : Total volume of the substrate (L)
x: Total height of cells in the substrate (mm)
Wave shaped cell
In the technical report for substrates, the shape of cell is mentioned as a wave [9].
However, the shape of the cell is assumed to be a sine wave in this report. This assumption is made mainly for the calculations purpose i.e., to calculate the total area of a substrate.
A sine wave with amplitude A, length x and wall thickness h is as shown in Figure 2.10.2.
Figure 2.10.2: Sine wave
One half of the sine wave is shown in Figure 2.10.3. In Figure 2.10.3, the length of the arc pq is approximately equal to the length of line segment pq. Length of line segment pq is given by Equation 2.14.
pq =
…
A2+ (x
4)2 (2.14)
Figure 2.10.3: One half of a sine wave
(a) Stack of wave shaped cells (b) Single wave shaped cell Figure 2.10.4: Wave shaped cells in the substrate
Figure 2.10.4a shows the arrangement of a stack of sine wave shaped cells in the substrate. In this arrangement, AB is considered to be a single cell which can be observed in Figure 2.10.4b. From Figure 2.10.4b, there are six line segments with the length of each one equal to length of line segment pq. These six line segments makes one complete cell in the substrate. Hence, the total perimeter of substrate for a single sine wave shaped cell is given by Equation 2.15.
P = pq∗ 6 =
…
A2+ (x
4)2∗ 6 (2.15)
Area of substrate for a single sine wave shaped cell is given by Equation 2.16.
Ar = P ∗ h =
…
A2+ (x
4)2∗ 6 ∗ h (2.16)
The governing equation to calculate the total area of a substrate with sine wave shaped cells is given in Equation 2.17.
Atot = Ar∗ N =
…
A2+ (x
4)2∗ 6 ∗ h ∗ N (2.17) Where,
Atot : Total area of the substrate (mm2) A : Amplitude (mm)
x : Length (mm)
h : Wall thickness of a substrate (mm) N : Total number of cells in the substrate
Water blow out Experiment
An experiment is performed in order to observe and investigate the water blow out phenomena. This experiment is carried out in the AR1 engine test rig at Scania. The silencer used for testing is a Scania CAS1 silencer. This chapter deals with the experimental setup, procedure and the data obtained from water blow out experiment.
3.1 Setup
Before performing the water blow out experiment, it is important to locate the possible places where water can get trapped inside the silencer. This is done by analysing the CAD model of the silencer. The possible locations of water getting trapped inside the silencer is listed in Table 3.1.1.
Table 3.1.1: List of various places where water can get trapped in the silencer
S No Location
1 Inlet water pocket
2 Catalyst 1
3 DPF
4 Evaporation tube + Water pockets before and after it
5 Space before Catalyst 2
6 Catalyst 2
7 Outlet
In order to perform this blow out experiment, several modifications are made to the silencer. These modifications include drilling and welding inspection holes in some parts of the silencer. Inspection holes with a diameter of 20mm and a pitch of 1.5mm (M20*1.5) are welded at different locations inside the silencer. These locations are decided based on the CAD model of the silencer. All these modifications are done in a company called ’Dynamate’ in Södertälje.
There are seven inspection holes drilled and welded in the silencer. They are:
• Two inspection holes near the inlet water pocket which can be seen in Figure 3.1.1.
There is one inspection hole on the top of inlet tube and one inspection hole in the bottom where the inlet tube ends. This can be observed in Figure 3.1.1.
Figure 3.1.1: Inlet water pocket
• Two inspection holes near the water pocket after evaporation tube which can be seen in Figure 3.1.2. There is one inspection hole at the start of flex tube before the exhaust gas starts to climb up and one inspection hole at the end of flex tube.
These inspection holes are indicated with blue arrows in Figure 3.1.2.
Figure 3.1.2: Water pocket after evaporation tube
• Two inspection holes are located near the space before catalyst 2. One of the inspection hole is located at the bottom and the other at the top which can be seen in Figure 3.1.3.
Figure 3.1.3: Space before catalyst 2
• One inspection hole on the top of silencer near the outlet which can be seen in Figure 3.1.4.
Figure 3.1.4: Outlet
There are some important instruments and equipment used in the blow out experiment. They are:
• A IPLEX G Lite portable videoscope is used to record a video of the water blow out phenomena at different locations inside the silencer. The videoscope used in the experiment is as shown in Figure 3.1.5.
• A funnel and a hose pipe are used to pour water into the silencer. Water is poured through inspection holes or sensor holes in the silencer. Different sensors present at different locations in the silencer are dismantled before the experiment. This is done such that the sensor holes in addition to the inspection holes can be used to pour water into the silencer.
Figure 3.1.5: IPLEX G Lite portbale videoscope
There are two different types of hose pipes used in the experiment. This can be seen in Figure 3.1.6. The diameter of the white hose pipe is a bit larger when compared with the blue hose pipe. However, the length of the blue hose pipe is more than that of white hose pipe. The selection of hose pipe is done based on the location of blow out investigation inside the silencer.
In some of the locations inside the silencer, it is not possible to pour water using a funnel and hose pipe accurately. Hence, a syringe is used instead of a funnel to pour water at these locations. The funnel, hose pipes and syringe used in the experiment is as shown in Figure 3.1.6.
Figure 3.1.6: Funnel, Hose pipes and Syringe used in the experiment
• A weighing machine is used to find the amount of water poured into the silencer.
The weighing machine used in the experiment is shown in Figure 3.1.7.
Figure 3.1.7: Weighing machine used in the experiment
Finally, the water blow out experiment is carried out in AR1 engine test rig at Scania.
The speciality of this test rig is that it is possible to vary both the temperature and the mass flow rate of the air simultaneously. The control room for AR1 engine test rig is shown in Figure 3.1.8.
Figure 3.1.8: Control room of AR1 engine test rig
The complete setup of the modified silencer mounted inside the AR1 engine test rig is shown in Figure 3.1.9.
Figure 3.1.9: Complete setup of the silencer in AR1 engine test rig
3.2 Procedure
The water blow out experiment is performed differently for the substrates (DPF and Catalyst 2) and for other locations inside the silencer. Experimental procedures for the substrates and other locations has been discussed briefly in this section.
Blow out in substrates
For the water blow out experiment, it is assumed that catalyst 1 behaves in the similar manner as that of catalyst 2. Hence, the experiment is performed on catalyst 2 and the same data is used for both the catalysts.
The water blow out experiment is carried out on DPF and Catalyst 2 substrates. The step wise procedure for the blow out experiment in substrates is as follows:
• Similar to the other locations, here also a funnel and hose pipe is used to pour water into the silencer. Initially, known quantity of water is introduced into one of the substrates.
• Cold air at 25°C is blown into the silencer at different mass flow rates. In this case, the mass flow rate is increased until water starts to blow out from the substrate.
This blow out is observed through a videoscope and a video is also captured for further use. Whenever the water starts to blow out from the substrate, the corresponding mass flow rate is noted.
• Now, this process is repeated for different quantities of water poured into the silencer. The corresponding mass flow rates where blow out starts to occur and videos from videoscope are captured as data for this experiment.
• The same experimental procedure is adapted for both DPF and Catalyst 2 substrates and relevant data is acquired.
Blow out at other locations
The step wise procedure for the water blow out experiment for other locations is as follows:
• As mentioned in Table 3.1.1, there are 5 different possible places where water can get trapped inside the silencer apart from the substrates. To start with, one of these 5 possible locations is selected to perform the blow out experiment.
• A funnel and hose pipe is used to pour water at these places. Alternatively, a syringe and hose pipe can also be used to pour water at some places. The amount of water introduced into the silencer is different for different locations. This can be observed in Table 3.2.1. The amount of water is weighed using a weighing machine shown in Figure 3.1.7.
Table 3.2.1: Amount of water poured into the silencer at different locations Location Amount of water (L)
Inlet water pocket 0.2
Water pocket before evaporation tube 0.116 Water pocket after evaporation tube 0.5
Space before catalyst 2 0.1
Outlet 1
• All the other inspection holes and sensor holes are closed apart from those which are present at the selected location for blow out experiment.
• A videoscope is gently introduced into one of the inspection holes at the selected location. It is important to setup the position of before the flow starts.
• Cold air at approximately 25°C is blown into the silencer at a specific mass flow rate. A videoscope is used to record a video of this blow out phenomena. This video is saved for further use and data analysis purpose.
• The experiment is repeated for different mass flow rates with the same amount of water and the corresponding videos are captured through a videoscope.
• All the above steps are repeated for every location mentioned in Table 3.2.1 and the required data is obtained.
It is important to notice that for the blow out experiment, the velocity of the flow (v) is a vital factor than the mass flow rate ( ˙m) of the exhaust. The relation between mass flow rate and velocity of the flow in a stream is given by Equation 3.1.
˙
m = ρ∗ A ∗ v (3.1)
Where,
˙
m: Mass flow rate (kg/min) ρ: Density of cold air (kg/m3) A : Cross section area of a part (m2) v : Velocity of the flow (m/s)
Density of cold air at 25°C is around 1.1839 kg/m3 [17]. Cross-sectional area (A) will vary from part to part inside the silencer. Hence, the velocity of the flow depends on the mass flow rate ( ˙m) and cross-sectional area (A) in a stream. The whole blow out experiment is based on the variations in mass flow rate ( ˙m). This is because all the control systems in Scania are designed in a way that it is easy to vary mass flow rate rather than velocity of the flow. Hence, mass flow rate is considered as a basic parameter to vary for the blow out experiment.
3.3 Data
The data obtained from the blow out experiment at all locations mentioned in Table 3.1.1 is discussed in this section.
Inlet water pocket
As mentioned in Table 3.2.1, the quantity of water poured into inlet water pocket is 0.2 L. The complete data procured from the blow experiment at inlet water pocket is shown in Table 3.3.1.
Table 3.3.1: Data obtained from blow out experiment at inlet water pocket
Mass flow rate (kg/min) Volume of water (L) Total time of blow out phase
10 0.141 33
12.5 0.112 42
15 0.052 53
17.5 0 17
20 0 15
The second column in Table 3.3.1 represents the quantity of water remaining at the inlet water pocket after the completion of blow out phase i.e., after reaching the equilibrium or saturation point. The third column in Table 3.3.1 represents the time taken to reach the equilibrium quantity of water. This time is obtained from the video captured through a videoscope.
From Table 3.3.1, it can be observed that a minimum of 17.5 kg/min mass flow rate is required to completely blow out the water from inlet water pocket. For mass flow rates higher than 17.5 kg/min, a complete blow out is observed and the time taken for complete blow out decreases with increase in mass flow rate. For mass flow rates less than 17.5 kg/min, partial blow out is observed. This means only a fraction of water is blown out and this fraction is dependent on the mass flow rate of cold air blown into the silencer.
However, it is observed from the videos that the blow out is dominant only in the first few seconds. It is important to consider this behaviour for developing a model and calibrating it. Hence, the time taken for 90% blow out or major blow out phase is also noted from the videos recorded using a videoscope. For calibration purpose, this 90%
blow out time is used which will be discussed later in Model development chapter in this report. The data obtained for the 90% blow out is shown in Table 3.3.2.
Table 3.3.2: Data for 90% blow out phase
Mass flow rate (kg/min) Time taken for 90% blow out (s)
10 10
12.5 14
15 15
17.5 7
20 6
Water pocket before evaporation tube
As noticed in Table 3.2.1, the amount of water introduced into the water pocket before evaporation tube is 0.116 L. The entire set of data obtained from the blow experiment at this location is shown in Table 3.3.3. From Table 3.3.3, it is observed that low mass flow rates of 10 kg/min is enough to blow out water entirely from the water pocket before evaporation tube. This is because the water pocket here is very small and can trap very less water in it. Also, amount of water trapped in this location depends on the silencer design and gas flow properties. However, partial blow out is observed for mass flow rates less than 10 kg/min. The data obtained from this water pocket is not useful for developing a model. Hence, the time taken for 90% blow out is not acquired for this particular location.
Table 3.3.3: Data obtained from blow out experiment at water pocket before evaporation tube
Mass flow rate (kg/min) Volume of water (L) Total time of blow out phase
3.33 0.096 13
5 0.034 26
7.5 0.0152 13
10 0 11
Water pocket after evaporation tube
The quantity of water introduced into the water pocket after evaporation tube is 0.5 L.
This can be observed in Table 3.2.1. The data obtained from blow out experiment at this location is shown in Table 3.3.4.
At this location, water climbs up the flex tube and enter the acoustic pipe even for low mass flow rates of 7.5 kg/min. However, only a fraction of water climbs up the flex
Table 3.3.4: Data obtained from blow out experiment at water pocket after evaporation tube
Mass flow rate (kg/min) Volume of water (L) Total time of blow out phase
7.5 0.3 23
10 0.2 32
15 0 17
20 0 16
tube at low mass flow rates. This behaviour can be observed in the data obtained, in second column of Table 3.3.4. For mass flow rates greater than 15 kg/min, complete water blow out is observed. For mass flow rates less than 15 kg/min, partial water blow out is observed from the videos recorded using a videoscope.
The data for 90% blow out phase is shown in Table 3.3.5. This 90% blow out time is used for calibration purpose.
Table 3.3.5: Data for 90% blow out phase
Mass flow rate (kg/min) Time taken for 90% blow out (s)
7.5 10
10 13
15 9
20 8
Space before Catalyst 2
The vacant space before catalyst 2 cannot hold and trap water in it. The quantity of water poured at this location simply flows into the catalyst 2. Hence, it is sufficient to look at the blow out effect in catalyst 2, rather than in the space before it. This behaviour is observed through the videos captured using the videoscope.
Outlet
As shown in Table 3.2.1, the quantity of water poured into the outlet is 1 L. The complete set of data obtained from blow out experiment at outlet is given in Table 3.3.6.
The outlet volume is very big and a minimum of around 20 kg/min mass flow rate is
Table 3.3.6: Data obtained from blow out experiment at outlet
Mass flow rate (kg/min) Volume of water (L) Total time of blow out phase
20 0.85 26
25 0.6 64
35 0.25 32
required to observe blow out effect in the outlet. For any mass flow rates less than 20 kg/min, negligible blow out effect is observed. From Table 3.3.6, it can be seen that even a mass flow rates of 35 kg/min is not sufficient to completely blow out water from the outlet. The experiment could not be extended to mass flow rates higher than 35 kg/min as it is practically not possible to record a video using a videoscope. However, the partial blow out data is obtained for mass flow rates of 20, 25 and 35 kg/min.
The data for 90% blow out phase is shown in Table 3.3.7. This data is helpful to calibrate the water blow out model.
Table 3.3.7: Data for 90% blow out phase
Mass flow rate (kg/min) Time taken for 90% blow out (s)
20 16
25 23
35 24
DPF
Table 3.3.8: Data for blow out experiment in DPF
Volume of water (L) Mass flow rate when blow out occurs (kg/min)
0.5 None (At least > 35)
0.7 25
1.0 15
The results obtained from blow out experiment in DPF is shown in Table 3.3.8. The first column represents the quantity of water introduced into DPF and the second column represents the mass flow rate when water starts to blow out from the DPF. The mass flow rate when blow out occurs reduces with increase in amount of water poured
into the silencer. However, the blow out is observed to occur in the canning math which is located underneath the DPF. There is no blow out observed from the DPF substrate.
This is a problem since it does not represent the actual driving condition. To account for this, a new experimental test method is proposed in chapter 7.
Catalyst 2
As mentioned in sub-section 3.2, it is assumed that both the catalysts behaves in the same manner for water blow out experiment. Hence, the experiment is conducted only on the catalyst 2 alone. The results from the experiment is shown in Table 3.3.9
Table 3.3.9: Data for blow out experiment in catalyst 2
Volume of water (L) Mass flow rate when blow out occurs (kg/min)
0.15 20
0.20 15
0.25 10
From Table 3.3.9, it can be observed that the catalyst 2 starts to blow out even for low mass flow rates of around 10 kg/min. However, a complete water blow out effect is not observed since some of the water could have been adsorbed in the catalyst 2 itself.
Hence, only a partial water blow out effect is noticed in the catalyst 2.
Model Development
A mathematical model is developed using the data obtained from the water blow out experiment. The entire silencer is divided into seven different sub-volumes and the actual maximum amount of water in each and every sub-volume is calculated. This data is helpful in the development of a better dew point strategy at Scania. The seven different sub-volumes inside the silencer is shown in Table 4.0.1.
Table 4.0.1: Different sub-volumes inside the silencer Sub-volume number Sub-volume description
1 Inlet tube
2 Inlet water pocket + Catalyst 1
3 DPF
4 Water pockets before and after evaporation tube + Evaporation tube
5 Acoustic pipe
6 Catalyst 2
7 Outlet
Water blow out phenomena is neglected in sub-volume 1 and sub-volume 5 in the silencer. As mentioned in Table 4.0.1, sub-volume 5 is a long straight acoustic pipe.
There are no places where water can be trapped in sub-volume 5 apart from the formation of small water droplets on the walls. Hence, blow out is neglected in this particular sub-volume.
In addition to this, water blow out phenomena is neglected in sub-volume 1. This is mainly due to two reasons:
• Sub-volume 1 in the silencer is an inlet flex tube. This is the initial part of the silencer. In real case scenario, the initial parts of the silencer gets heated up very quickly. Hence, the water present in this sub-volume has more probability to evaporate rather than to blow out.
• Moreover, there are no water pockets and no places where water can get trapped in sub-volume 1 as well.
Apart from sub-volume 1 and sub-volume 5, blow out experiment is conducted at all the other sub-volumes in the silencer and the required data is obtained. The schematic representation of the water blow out model is shown in Figure 4.0.1.
Figure 4.0.1: Schematic of water blow out model
4.1 Maximum amount of water (No mass flow)
As it is observed in Figure 4.0.1, the first step in water blow out model is to calculate the maximum amount of water theoretically at zero mass flow rate. For this purpose, a detailed analysis is conducted on the CAD model of the silencer. The calculations of maximum amount of water for every sub-volume apart from sub-volume 1 and sub- volume 5 are as follows:
Sub-volume 2
Water pocket near inlet tube and catalyst 1 constitutes sub-volume 2 inside the silencer.
The sum of maximum amount of water at inlet water pocket and catalyst 1 yields the total maximum amount of water in this sub-volume. At the water pocket near inlet tube, the maximum amount of water is found out to be 0.4 L experimentally.
Total volume of first part of catalyst 1 estimated from CAD model of the silencer is approximately equal to 8 L. Cell density of first part of catalyst 1 (ρcell) is equal to 260 CPSI (Cells Per Square Inch) and 50% covered surface [10].The shape of the cells in first part of catalyst 1 is a square shaped cell with an inner side of x and wall thickness of h [10].
Substituting the value of ρcell in Equation 2.12 and solving Equations 2.11, 2.12 gives the value of x = 1.11 mm and h = 0.23 mm as result. Hence, the total height of cells for a square shaped catalyst 1 is calculated to be 1.11 mm. Substituting the Vf irstpartof catalyst1
= 8 L and xf irstpartof catalyst1= 1.11 mm in Equation 2.13 gives the maximum amount of water in a square cell shaped first part of catalyst 1 substrate to be 2.16 L.
Total volume of second part of catalyst 1 is found to be 10.04 L according to the CAD model of the silencer. The shape of every cell in second part of catalyst 1 is square in shape with an inner side x and a wall thickness of h [8]. The schematic diagram for a square shaped cell can be seen in Figure 2.10.1. However, wall thickness for second part of catalyst 1 (h) is noted to be 0.11 mm [8].
Substituting the value of h in Equation 2.11, the value of x is found to be equal to 0.531 mm. Hence, the total height of cells for second part of catalyst 1 is estimated as 0.531 mm. Substituting Vsecondpartof catalyst1 = 10.04 L and xsecondpartof catalyst1 = 0.531 mm in Equation 2.13 gives the maximum amount of water in second part of catalyst 1 is found out to be 5.67 L.
Total maximum amount of water in sub-volume 2 = Maximum amount of water at inlet water pocket (0.4 L) + Maximum amount of water in first part of catalyst 1 (2.16 L) + Maximum amount of water in second part of catalyst 1 (5.67 L) = 8.23 L. Hence, the total maximum amount of water in sub-volume 2 is approximated to be 8.3 L.
Sub-volume 3
Sub-volume 3 inside the silencer consists of DPF substrate. From experiments, it was found that the blow out takes place only in the canning math underneath DPF. There is no blow out observed in the DPF substrate. However, the maximum amount of water is calculated for both the DPF substrate and the canning math underneath it.
Total volume of DPF substrate is around 18 L (ϕ12.5′′ × 9′′). Total height of cells in DPF substrate is around 1.29mm [4]. Substituting these values in Equation 2.13, maximum amount of water in DPF is found out to be 4.18 L. However, the canning math underneath DPF can hold only upto 1 L of water. This value is found out experimentally by pouring water into DPF substrate.
Total maximum amount of water in sub-volume 3 = Maximum amount of water in DPF (4.2 L) + Maximum amount of water in canning math (1 L) = 5.2 L. Hence, the total maximum amount of water in DPF constitutes to be 5.2 L.
Sub-volume 4
Sub-volume 4 in the silencer consists of: the water pocket before evaporation tube, the evaporation tube and the water pocket after evaporation tube. The maximum amount of water in the water pocket before evaporation tube is found out to be around 0.12 L experimentally.
Evaporation tube is a cylindrical tube of length 434.2 mm and radius of 68.35 mm. The total volume of evaporation tube is calculated using Equation 4.1.
V = π∗ r2 ∗ l (4.1)
Substituting the values of length and radius for evaporation tube in Equation 4.1, total volume of evaporation tube is calculated to be 6.4 L. As it is observed from Table 3.3.4, a mass flow rate of 15 kg/min is enough to blow out the water completely from this sub- volume. Cross sectional area for the evaporation tube is given by Equation 4.2.
A = π∗ r2 (4.2)
Substituting the value of radius (68.35 mm) in Equation 4.2 gives the cross sectional area (A) of the evaporation tube to be 0.0146 m2. This is the entire cross sectional area of the evaporation tube. Substituting the value of cross sectional area (A), density of cold air at 25°C (1.1839 kg/m3) and mass flow rate for complete blow out (15 kg/min) in Equation 3.1 gives the velocity of the flow to be 14.395 m/s. Hence, a minimum velocity of cold air inside the evaporation tube for a complete blow out scenario is 14.395 m/s.
Maximum amount of water inside the evaporation tube occurs during the idling of the engine. Now, cold air with velocity 14.395 m/s and mass flow rate 5 kg/min are assumed to flow inside the evaporation tube during idling stage. Substituting the values for velocity (14.395 m/s) and mass flow rate (5 kg/min) in Equation 3.1 gives the value of area of cross-section (A1) to be 0.00488 m2. The total cross-sectional area of the evaporation tube (A) is 0.0146 m2. The portion of area filled with water is given by Equation 4.3. The value of A2calculated from Equation 4.3 is 0.00972 m2.
A2 = A− A1 (4.3)
The percentage of the area filled with water is given by Equation 4.4.
P = A2
A ∗ 100 (4.4)
On substitution of values for A and A2in Equation 4.4, the value of P is calculated to be 66.5%. Hence, the maximum amount of water which could be present in evaporation tube is around 66.5 % of the total volume of the tube. The total volume of tube is around 6.4 L. Hence, the maximum amount of water in the evaporation tube is approximately equal to 4.25 L.
The water pocket after evaporation tube can hold upto 0.51 L of water. This value is found experimentally by pouring water into the silencer. Finally, maximum amount of water in sub-volume 4 is equal to 4.88 L (0.12 L + 4.25 L + 0.51 L). The maximum amount of water in sub-volume 4 is assumed to be 5 L in the model.
