MEV Journal of Mechatronics, Electrical Power, and Vehicular Technology 11 (2020) 45-54 Journal of Mechatronics, Electrical Power, and Vehicular Technology e-ISSN: 2088-6985 p-ISSN: 2087-3379 www.mevjournal.com doi: https://dx.doi.org/10.14203/j.mev.2020.v11.45-54 2088-6985 / 2087-3379 ©2020 Research Centre for Electrical Power and Mechatronics - Indonesian Institute of Sciences (RCEPM LIPI). This is an open access article under the CC BY-NC-SA license (https://creativecommons.org/licenses/by-nc-sa/4.0/). Accreditation Number: (RISTEKDIKTI) 1/E/KPT/2015. A study effects of injection pressure and wall temperature on the mixing process of NOx and NH3 in Selective Catalytic Reduction system Muhammad Khristamto Aditya Wardana a, b, *, Ocktaeck Lim a a Department of Mechanical Engineering, University of Ulsan Ulsan, 680-749, Republic of Korea b Research Centre for Electrical Power and Mechatronics, Indonesian Institute of Sciences Komp. LIPI Bandung, Gd 20, Lt 2, Bandung, West Java, 40135 Indonesia Received 10 May 2020; accepted 16 May 2020; Published online 30 July 2020 Abstract Diesel engines are commonly used for public transportation on-road and off-road applications. Growth production of the diesel engine is very significant from year to year. Nitride Oxide (NOx) from diesel engine was one of the major sources of air pollution. Selective Catalytic Reduction (SCR) has been successfully used to reduce NOx from a diesel engine with a chemical reaction from ammonia (NH3). The mixing reaction between NOx and NH3 reaction can produce steam (H2O) and Nitrogen (N2). However, ammonia uniformity pattern usually not homogenization and the ammonia was difficult to mix with NOx. The constant air flows incomplete to assist the spray injector to spread NH3 to all corners of SCR. The impact study of turbulent phenomena and standard k-epsilon Low-Reynolds Number model to the mixing process in the SCR system using STARCCM+. The simulation studies are conducted under different pressure (4 to 6 bars), the injection rate (0.04 g/s) and temperature (338 K – 553 K) and the high pressure and high velocity magnitude creating turbulent swirl flow. The ammonia decomposition process and mixing process with NOx were investigated using a box with optical access. The simulation and numerical study results validated using back pressure value and the distribution of NOx concentration value from the catalyst outlet. The wall temperature will increase the urea evaporation to generate ammonia and gas pressure will increase the mixing process and chemical process in the SCR system. These reactions enable to optimize the SCR system technology which eventually able to reduce the NOx quantity from a diesel engine. ©2020 Research Centre for Electrical Power and Mechatronics - Indonesian Institute of Sciences. This is an open access article under the CC BY-NC-SA license (https://creativecommons.org/licenses/by-nc-sa/4.0/). Keywords: diesel engine; wall temperature; wall impingement; urea water solution (UWS); urea injection; selective catalytic reduction (SCR). I. Introduction Diesel engines are commonly used for public transportation on-road and off-road applications. Growth production of the diesel engine is very significant year to year [1]. The diesel engine is used for commercial and passenger vehicle [2][3][4]. Some of the major advantages of diesel engine over another fuel engine are the higher durability and increased fuel efficiency. On the other hand, the main disadvantage is related to exhaust emissions from the diesel combustion process. The emission of the diesel engine is one of the major sources of air pollution which need to be controlled by the after- treatment system. Government has released a regulation for automotive industry related to the exhaust gas emission to further reducing the nitrogen oxide emissions. The emission rules have reached the pollution limit of the EURO VI and the US TIER 2 regulation [1][5][6]. The ammonia in the Selective Catalytic Reduction (SCR) as the main solution to control specific emissions from the engine. The technology of SCR used the injection of a urea liquid into the exhaust system. The hot temperature from exhaust makes a urea-water solution (UWS) evaporates and decomposes to be ammonia. Ammonia concentrate used as reducing nitrogen oxides (NOx) in the SCR system. Many researchers study around the injector for spraying urea (NH3), gas temperature, and catalyst * Corresponding Author. Tel: +82-01074637823 E-mail address: m.k.aditya.w@gmail.com https://dx.doi.org/10.14203/j.mev.2020.v11.45-54 https://dx.doi.org/10.14203/j.mev.2020.v11.45-54 http://u.lipi.go.id/1436264155 http://u.lipi.go.id/1434164106 http://mevjournal.com/index.php/mev/index https://dx.doi.org/10.14203/j.mev.2020.v11.45-54 https://creativecommons.org/licenses/by-nc-sa/4.0/ https://crossmark.crossref.org/dialog/?doi=10.14203/j.mev.2020.v11.45-54&domain=pdf https://crossmark.crossref.org/dialog/?doi=10.14203/j.mev.2020.v11.45-54&domain=pdf https://crossmark.crossref.org/dialog/?doi=10.14203/j.mev.2020.v11.45-54&domain=pdf https://creativecommons.org/licenses/by-nc-sa/4.0/ https://crossmark.crossref.org/dialog/?doi=10.14203/j.mev.2020.v11.45-54&domain=pdf https://crossmark.crossref.org/dialog/?doi=10.14203/j.mev.2020.v11.45-54&domain=pdf M.K.A. Wardana and O. Lim / Journal of Mechatronics, Electrical Power, and Vehicular Technology 11 (2020) 45-54 46 substrate in the SCR system, but the ammonia still difficult to mix with NOx. Based on our previous study [7], it has been mentioned that the urea is difficult to evaporate and easily attached to the wall and making the solid deposit in the system. It is consistent with the study carried out by Hasan [8] which found the solid deposited effect to the mixing process at the SCR system. The most important concept for the mixing process is the turbulent airflow from the exhaust gas in the SCR system. Exhaust airflow was assisting the spray injector to spread NH3 to all corners in the SCR system. To solve that situation, this study presents a simulation and experimental study of the turbulent phenomena impact and the standard k-epsilon Low- Reynolds Number model to mixing process in the SCR system using STARCCM+. To create turbulent in the SCR system, this study uses pressure from the exhaust gas and wall temperature from the SCR system. The simulations study based on the exhaust gas system with a high pressure to create turbulent swirl flow. The ammonia decomposition process and mixing process with NOx were investigated using a box with optical access. The data obtained from the optical box can assist the identification of the actual condition in the SCR system. The effects of wall impingement on deposit formation in diesel SCR system has been investigated the wall temperature (338 K) and (573 K). In the previous research carried out by Auvray [9], it was shown the temperature (473 K) and (598 K) can be used to analyze kinetic modelling on NH3 in the SCR system. Furthermore, our previous study [4] also used a high temperature of 536 K to study the ammonia uniformity in SCR system technology. A similar study from Smith [2], it has used a low temperature of 473 K and a high temperature of 608 K for analyzing the deposit formation reaction. Based on those investigations, the temperatures used in this system are 338 K for low wall temperature and 553 K for high wall temperature. The simulation and numerical study results validated using back pressure value and the distribution of NOx concentration at the SCR system. II. Materials and Methods A. Mathematical model The numerical study with 3D simulation for understanding the mixing process in the SCR system was shown with STARCCM+ software version 11.04. In another paper from Fischer et al. [10], compare the Reynolds-averaged k-epsilon-models and Reynolds-stress-model (RSM) to accounting the anisotropic character of turbulence in the swirl flow. Fischer et al used pipe with oval/flat crosses section for assist mixing process in the SCR system. The RSM describe the anisotropic turbulence between the primary swirl core and the outer secondary. This method is useful for understanding the TKE values in the pipe cross section and dissipation value in the swirl core. However, the RSM difficult to predict the highly turbulent reaction at the inner and outer swirl core. Yi [11] has analyzed the ammonia homogenization from two different mixers with renormalizing of Navier-Stokes equations (RNG) and the k-epsilon model for turbulence study. The computational fluid dynamic (CFD) model of their research simple SCR system installation. The straight pipe with urea injector and leading into the catalyst. The urea injector is located at the wall with a 90° angle relative to the main flow. Without a mixer system, the low quality of ammonia uniformity has resulted. Integrating a helix swirl mixer in the system leads the ammonia to raise the good quality of uniformity. This study was applying straight pipe and optical box with fully SCR system. The urea injector is located at top of the optical box. That urea can release the UWS spray was from top to bottom. This observed is focus on the distribution of exhaust gas with pressure to get the turbulence quality without the mixer. The eddy viscosity from the k-epsilon equation, determined from a single turbulence length. The calculation of turbulent diffusion can achieve by the specified scale. Although in the real system, all motion scales will contribute to the turbulent diffusion. The numerical computation with a realizable k- epsilon model for resolved the viscous layers and viscous sublayer in the SCR system. These equation functions of wall treatment can be extended and solving the multiphase flow phenomena and motion of gasses from NOx and ammonia wherein the gasses is characterized by properties that are aggregated over a large number of individual molecules [12][13]. Each particle in the system has associated with a physical equation, such as density, velocity, vorticity and temperature. The vector quantity in the numerical properties assumed to be observable and hence there are ‘streamlines’ at the local velocity vector. Streamlines can never cross except at point sources or sinks of fluid. The transport equations for the Low-Reynolds Number with k-epsilon model equation can solve the computational domain: 𝑑 𝑑𝑡 ∫ 𝜌𝜀 𝑑𝑉𝐴 + ∫ 𝜌𝜀�𝑣 − 𝑣𝑔�. 𝑑𝑎𝐴 = ∫ �𝜇 + 𝜇𝑡 𝜎𝜀 �∇𝜀. 𝑑𝑎 𝐴 + ∫ 1 𝑇 �𝐶𝜀𝑙(𝐺𝑘 + 𝐺𝑛𝑙 + 𝐶𝜀3𝐺𝑏) − 𝐶𝜀2𝜌(𝜀 − 𝜀0) + 𝜌𝛾𝑦 + 𝑆𝜀�𝑑𝑉𝑉 (1) where 𝑆𝑘 and 𝑆𝜀 are specified source properties, 𝜀0 is the ambient turbulence value for turbulence decay counteracting, 𝜌 is the source density, and 𝐺′ is the additional product from source term that is given by: 𝐺′ = 𝐷𝑓2 �𝐺𝑘 + 2𝜇 𝑘 𝑑2 � exp�−𝐸𝑅𝑒𝑑 2� (2) where 𝐷 is the distance to the nearest wall of the source terms, 𝑑 is the distance to the wall, 𝑘 is turbulent kinetic energy, 𝜇 is dynamic viscosity, and 𝑓2 is a damping function that is defined as: 𝑓2 = 1 − 𝐶 exp(−𝑅𝑒𝑡2) (3) where 𝑅𝑒𝑑 = √𝑘𝑑 𝑣 (4) and M.K.A. Wardana and O. Lim / Journal of Mechatronics, Electrical Power, and Vehicular Technology 11 (2020) 45-54 47 𝑅𝑒𝑡 = 𝑘2 𝜀𝑣 (5) The coefficients 𝐶 and 𝐸 are using the default values of 0.3 and 0.00375. and the value for coefficient 𝐷 in the system is in Equation (6): 𝐷 = 𝐶𝜀2 𝐶𝜀𝑙 ≈ 1.3 (6) However, direct numerical simulation (DNS) of low- Reynolds number channel flow suggests that better results are obtained with 𝐷 = 1, which is the default value that is used in STARCCM+. DNS was the presence of free stream turbulence, inaccuracy computing method the 𝐷 = 1 was used for minimizing the error result from turbulence equation. The momentum coupling model for flow sections used the vertices scale, the turbulent viscosity 𝜇𝑡 has a relation on getting the turbulent kinetic energy 𝑘 and dissipation 𝜀, in equation (7): 𝜇𝑡~ 𝑘2 𝜀 (7) With regards to the scalar transport model equation, a direct proportionality of turbulent diffusive momentum and scalar transfer is assumed. This method can explain the turbulent scalar diffusion is directly linked to the turbulence models viscosity prediction in Equation (8): 𝜌𝜇𝐽𝛾 = − 𝜇𝑡 𝑆𝑐𝑡 𝜕𝑦 𝜕𝑥𝐽 (8) The constant equation to get the proportionality value is used the turbulent Schmidt number 𝑆𝑐𝑡. An underestimated turbulent viscosity value in the equation model can be stable by reducing the Schmidt number value (0.7 – 0.9). Throughout the following investigations by Wardana [7], the turbulent Schmidt number for Standard k-epsilon Low-Reynolds Number model used the default value of 1.0 for described in this literature review. 1) Thermolysis and evaporation of UWS droplets Despite numerous experimental studies have been conducted [4][14][15], theoretical understanding for the decomposition and evaporation of UWS and Adblue droplets are still not fully investigated. A theoretical study [9][12] was implemented using STARCCM+ for simulation which then compared by experimental works to get the best accuracy. This method assumes two processes of the UWS evaporation until the UWS droplet is only composed of urea. It uses a spherical method for the evaporation and decomposition processes without using urea crystallization in the process. Rapid mixing model is used to evaluate the dissolved urea on the evaporation process of urea water solution. Because this model can identify the high transport coefficients from the liquid phase. The result can explain the homogenous distribution temperature in the system, concentration particle value and fluid properties in the UWS droplet [9]. The variation model in UWS concentration droplet's can be evaluated by Equation (9): 𝑑𝑌𝑢 𝑑𝑡 = − 𝑚𝑣𝑎𝑝 𝑚𝑑 𝑌𝑢 (9) where 𝑚𝑑 is the mass of droplet particle, 𝑑𝑌𝑢 is urea concentration value and 𝑚𝑣𝑎𝑝 is a vapour value of mass flow in the system. The evaporation rates in this equation are calculated by the Abramzon- Sirignano model [12]. That model useful for low computational to observe the UWS droplet particle and suitable predict the spray modelling [12]. The urea particle will easily melt at 400 K [12] and the Arrhenius model can describe that chemical kinetics reaction by Equation (10): 𝑑𝑌𝑢 𝑑𝑡 = −𝐴𝑓𝑆 exp �− 𝐸𝑎 𝑅𝑇𝑑 � (10) where 𝑆 is droplet surface area in the system, 𝐴𝑓 is the frequency factor [kg/(sm2)], 𝑅 is the universal value, 𝑇𝑑 is the droplet temperature for gas constant and 𝐸𝑎 is the activation of energy in the system. 2) Urea water solution injection Urea-water solution is the most commonly used for ammonia precursor in the SCR system because this material was safe for the environment. The commercial name for UWS was Adblue, this aqueous was a combination from 67.5 to 70 % deionized water and 30 to 32.5 % urea [12]. Gas emission from diesel engine usually produces more than 90 % of NOx and 5 to 10 % of NO2. That quantity depending on the diesel engine type. NO2 from diesel engine emission is useful for fast SCR reaction process. That chemical is produced from engine-emitted NO when oxidation reaction in the catalyst. In order to describe the UWS injection phenomena in the SCR systems, this study used a mathematical description to explain all processes [3][9][12]: • Interaction momentum between Exhaust gas and UWS droplets • Thermolysis and Evaporation of UWS droplets • Heat transfer from exhaust temperature to the wall and droplets. This section presents the mathematical basics model to implement the CFD modelling conduct the numerical simulations of the SCR system by STRACCM+ software. The flowchart model for this simulation is shown in Figure 1. Figure 1. The flowchart of simulation process M.K.A. Wardana and O. Lim / Journal of Mechatronics, Electrical Power, and Vehicular Technology 11 (2020) 45-54 48 B. Experimental setup The experiment condition used a single cylinder with 498 cm3 of displacement, naturally aspirated engine with 4-cycle, and SOHC 4 valves system to get the precision of the engine test. The engine specification is shown in Table 1. The pressure at the exhaust was 4 to 6 bar [5][6][16] and leading into the optical box with straight pipe. A schematic diagram for the optical box test is shown in Figure 2. The pressure from UWS injector was ranging from 4 to 6 bar, the ammonia injection rate (0.04 g/s) and exhaust temperature constant in 338 K – 553 K [5][6][16]. The injection specification is presented in Table 2. The optical box for measuring position and dimension was realized with silica glass. The positions and dimension were chosen to minimize the effect of the gas flow in the system. The optical box gas flow distribution is shown in Figure 3. The mixing process between NOx and ammonia is observed in the optical box. The Horiba MEXA- 7100DEGR is used to identify the emissions value from the system (hydrocarbon and NOx). C. Validation of simulation To verify the Standard k-epsilon Low-Reynolds Number model simulation results from STARCCM+, the observation from ammonia uniformity at the SCR system have been carried out on the experiment test [5][17]. The exhaust line directly connects with the SCR system. The Horiba MEXA-7100DEGR can determine the gas distribution and the hydrocarbon value from the catalyst outlet. The Standard k-epsilon Low-Reynolds Number equation model is suitable for computing the convection flows. The situations to have a Low- Reynolds Number version of the Standard k-epsilon model, and Non-linear Constitutive Model as shown in Table 3. The setups have two-equation models for the simulation of the measured ammonia uniformity. The error value for the velocity magnitude model is a good agreement to observe the validated literature cases for the k-epsilon model in the SCR system simulation [2][11]. By considering the complete process from UWS spray to the ammonia distribution in the optical box. Table 1. Engine specification Engine parameter Value Displacement 498 cm 3 Bore 83 mm Stroke 92 mm Compression ratio 19.5 Con. rod length 145.8 mm Crank radius 43.74 mm Valve system SOHC 4 valve Fuel system Electronic common rail Table 2. Ammonia and exhaust injection Parameter Value Exhaust inlet 10, 15, 20 m/s Injector inlet 10, 15 m/s Injection rate 0.04 g/s Temperature 338 K – 553 K Pressure 4, 5, 6 bar Figure 2. Schematic diagram of test engine and measurement setup Table 3. Analyzed turbulence models and boundary conditions Setup Turbulance model Boundary condition a 1 Standard Low Reynold Number k-epsilon PI - VO aV = Velocity; P = Pressure; I = Inlet; O = Outlet M.K.A. Wardana and O. Lim / Journal of Mechatronics, Electrical Power, and Vehicular Technology 11 (2020) 45-54 49 III. Results and Discussions A. Exhaust simulation gas Simulated distribution of the exhaust pressure streamlines in an optical box system, as it has been predicted by STARCCM+. This simulation using the Standard k-epsilon Low-Reynolds Number model with the prediction on the effect of exhaust gas for ammonia decomposition phenomena can be calculated with the Schmidt number value. This energy value used to determine the turbulent viscosity and diffusion constant in the equation model. Besides that, this simulation model is implying the reduction of the anisotropic information for turbulent energy quantity at least for the scalar transport value. Figure 4 shows the difference turbulence streamline and scalar distribution in the optical box. Gas flow in the optical box is the most important indicator to mix ammonia and NOx. This flow strongly influences the homogenization of ammonia vapour, even without a sufficient temperature from Figure 4. The difference turbulence streamline and scalar distribution in the optical box: (A) Exhaust turbulent phenomena and (B) Exhaust scalar phenomena Figure 3. The measurement position of exhaust flow and ammonia injector spraying to the optical box: (a) Urea injection spray pattern; (b) Exhaust gas pattern; (c) Combination pattern from urea and exhaust gas M.K.A. Wardana and O. Lim / Journal of Mechatronics, Electrical Power, and Vehicular Technology 11 (2020) 45-54 50 the ammonia injection. The amount of pressure for mixing process of NOx and ammonia is shown in Figure 4(A). This picture described a pressure ratio between the lowest (4 bar) and the highest (6 bar). The colour line explains the velocity value in the simulation. However, the quality of the mixing process from Figure 4 was explained by the quantity of gas distribution in the simulation. It affects the different movement of air in the optical box. When the higher pressure was used, this condition makes the more diffuse flow of air generated and it was a good result for mixing ammonia and NOx. But instead, when the smallest pressure is used, ammonia and NOx mixing is difficult to be created as airflow having difficulties to spread close to the wall and makes ammonia on the wall could not be parsed by airflow. The airflow and pressure quantity also increasing the wall temperature value in the optical box. Figure 4(B) explains the temperature distribution in the simulation. Heating wall temperature propagates by flowing heat of exhaust gas from inlet to the outlet, more high gas flows from the exhausts more heat of the wall temperature, and its effects on the reduction ammonia droplet in the wall. This statement also has been explained in our previous study [7] which is observed the wall temperature value for predict the urea injection process in the SCR system. Figure 5 shows the velocity magnitude quality. As can be seen from the figure, the constant exhaust flow mass value achieve at 0.045 kg/s, the maximal flow mass value achieve at 3.15 kg/s and minimal flow mass at -3.5 kg/s in the optical box. The free gas at the optical box makes flow mass in an optical box unstable in the first time. Flow mass is difficult to spared before free gas come out from the optical box. That case also happens in constant velocity magnitude. Maximal velocity occurred due to the empty condition when the exhaust gas inside the optical box. Figure 6 shows the pressure and force value. As can be seen from Figure 6, constant pressure occurred in 20 s after the exhaust come from the inlet. An empty condition in the optical box makes pressure decrease significantly to -102.15 bar. Exhaust gas occupies empty space in the optical box to encourage the free gas came out from outlet to generate exhaust force gas and make pressure increase constantly after declining. Maximal pressure happens in 13.2 bar and decrease after 12 s exhaust came from the inlet. That was contradicted with exhaust gas force value when the exhaust gas pressure value decreases in an optical box, exhaust gas forces increase until 19.78 kN and the force value decrease on -2.63 kN. Massive pressure was counted by volume of the optical box, the extent of affecting the increase and decrease of the exhaust gas pressure against time. With a dimension of 90 x 30 x 30 cm3 of the optical box, the exhaust pressure needs 20 s for stabilizing into 1 bar. If using the small volume that can easily stable lowest than this simulation. B. Injector effect on exhaust simulation gas The distribution of the exhaust pressure can predict the streamline reaction in the optical box. At Figure 5. Velocity and flow mass effect at the optical box M.K.A. Wardana and O. Lim / Journal of Mechatronics, Electrical Power, and Vehicular Technology 11 (2020) 45-54 51 the first step of the exhaust inside the optical box, the pressure of exhaust was dropped to -102.15 bar, that condition also happen in the real SCR in public vehicles. The amount of the pressure drop was depending on the size of the exhaust muffler (pipe and SCR system). When the SCR system has a pressure drop, the exhaust flow is difficult to spear the adblue from urea injector. The urea from adblue attached in the wall will settle and harden because the wall will heat up due to the creeping temperature of the engine heat. That reaction has an effect after long period application of SCR system in the diesel engine. The timing for spraying adblue in commercial diesel engine has been analyzed in exhaust simulation gas. For injector condition, the constant pressure and velocity value can set the turbulent flow in the system. The distribution of gas and liquid were accounted, but it does not give an insignificant impact from the spray characteristic in the flow field. That reaction occurring because the injector has low pressure and momentum for spraying the UWS. Nonetheless, the flow field during the transient condition keeps valid. Figure 7 shows the difference turbulence streamline from UWS injector and exhaust gas in the optical box. This simulation shows that the NOx and ammonia mixing can be observed and analyzed in the optical box. Nonetheless, interdependencies from the liquid and gas phase with swirl fan mixing are avoided in this study. The ammonia vapour concentrations got the averaged value by each computational cell when one-time injection duration. The resulting quasi-steady spatial for ammonia distributions value is observed on the cross sections as long as in the steady flow quantities. The simulation was plausible with visualizing and evaluates the ammonia homogenization process in the optical box. The turbulent kinetic energy (TKE) is the main parameter to determine turbulent viscosity and turbulent diffusion in this simulation [5]. Figure 8 shows the TKE value on the optical box. This TKE result got the maximum curves when the exhaust gas comes to the inlet (2.5 s). The gas flow distribution creating the steep velocity gradients by the source of turbulence. The kinetic energy structure divided into small- scale structures turbulence energy and transferred into dissipates rate of energy. The turbulent dissipation value in the system is shown in Figure 7. Though just using the k-epsilon model, the turbulent scheme showed very clearly. Beside that for detail result, it still needs more deeply parameter using deference mathematic model. C. k-epsilon model validation The ammonia uniformity in this study resulted that without a mixer system, the low quality of ammonia uniformity has resulted. Integrating a helix swirl mixer in the system leads the ammonia to raise the good quality of uniformity. That result is explained on this validation purposes [10]. 2 results from this study will be compared to find the best values by the Standard k-epsilon Low-Reynolds Number model [10]. Figure 9 shows the back pressure value between simulation and experiment in the optical box. This result explains the mass flow quantity from exhaust temperature and the amount of gas distribution. The Standard k-epsilon Low-Reynolds Number model has been calculated by heat temperature from exhaust inlet with the increasing wall temperature parameter. In that case, the temperature was the Figure 6. Pressure and force effect at the optical box M.K.A. Wardana and O. Lim / Journal of Mechatronics, Electrical Power, and Vehicular Technology 11 (2020) 45-54 52 most decisive to mix ammonia and NOx, besides the turbulence in the optical box. The back pressures value occurring at low volume condition can determine the magnitude of measurement precision. This parameter condition useful for computation in the validation discussion. The particular range in this simulation has offset value of approximately 1.0 bar and that value better 0.2 bar from RSM simulation condition. The measuring value for back pressure at high volume flow occurs when the measurement has precise and reliable data [10]. That parameter can predict simulation by comparing the Figure 7. Exhaust and injector turbulent phenomena of (A) with low exhaust pressure and (B) with high exhaust pressure Figure 8. Turbulent kinetic energy (TKE) and turbulent dissipation rate (TDS) effect in optical box M.K.A. Wardana and O. Lim / Journal of Mechatronics, Electrical Power, and Vehicular Technology 11 (2020) 45-54 53 different turbulent model, which is expected to get similarities and accuracy data measurement and simulation. IV. Conclusion Simulated distribution of the exhaust pressure streamlines in the optical box system, as it has been predicted by STARCCM+. Gas flow in the optical box is the most important indicator to mix ammonia and NOx. This flow strongly influences the homogenization of ammonia vapour, even without using a sufficient temperature from ammonia injection. Heating wall temperature propagates by flowing heat of exhaust gas from inlet to the outlet, more high gas flow from exhausts more heat the wall temperature, and the effect on the reduction ammonia droplet in the wall. The free gas at the optical box makes a mass flow in the optical box unstable in the first time. With a dimension of 90 x 30 x 30 cm3 of the optical box, the exhaust pressure needs 20 s for stabilizing into 1 bar. If using the small volume that can easily stable lowest than this simulation. An empty condition in the optical box makes the pressure decrease so deeply until -102.15 bar when exhaust came from the inlet. Massive pressure was counted by volume of the optical box used, the extent of affecting the increase and decrease of the exhaust gas pressure against time. For injector condition, the constant pressure and velocity value can set the turbulent flow in the system. The distribution of gas and liquid were accounted, but it does not give an insignificant impact from the spray characteristic in the flow field. That reaction occurring because the injector has low pressure and momentum for spraying the UWS. Nonetheless, the flow field during the transient condition keeps valid. The turbulent kinetic energy (TKE) is the main parameter to determine turbulent viscosity and turbulent diffusion in this simulation. In comparison TKE curves and TDS curve, the result has a different path to describe the turbulent energy based on the region section. The high momentum in this result can validate the strong energy from the exhaust to increase the distribution of ammonia in the system. Although this study has different parameter and CFD models with others study, this result can be good knowledge for the researcher to understand the mixing process of ammonia and NOx in SCR system technology. Acknowledgement This work was supported by a research program supported by the Department of Mechanical Engineering (Generation Fuel and Smart Power Train Laboratory), University of Ulsan, Republic of Korea. Nomenclature CFD : Computational Fluid Dynamic MARS : Monotone Advection and Reconstruction Scheme NOx : Nitrogen Oxide K : Kelvin temperature scale SCR : Selective Catalyst Reduction TKE : Turbulent Kinetic Energy TDS : Turbulent Dissipation Rate UWS : Urea Water Solution RSM : Reynolds Stress Model Declarations Author contribution M.K.A. Wardana as the main contributor of this paper. All authors read and approved the final paper. Funding statement This research did not receive any specific grant from funding agencies in the public, commercial, or not-for- profit sectors. Conflict of interest The authors declare no conflict of interest. Additional information No additional information is available for this paper. References [1] R. Conway, S. Chatterjee, M. Naseri, and C. Aydin, “Demonstration of SCR on a Diesel Particulate Filter System on a Heavy Duty Application,” SAE Int., 2015. [2] H. Smith, T. Lauer, V. Schimik, and K. Gabel, “Evaluation and Prediction of Deposit Severity in SCR Systems,” SAE Int. J. Engines, vol. 9, no. 3, pp. 2016-01–0970, 2016. [3] J. Baleta, H. Mikulčić, M. Vujanović, Z. Petranović, and N. Duić, “Numerical simulation of urea based selective non-catalytic reduction deNOx process for industrial applications,” Energy Convers. Manag., 2016. [4] M.K.A. Wardana, G.M.H. Shahariar, K. Oh, and O. Lim, “Ammonia uniformity to predict NOx reduction efficiency in an SCR system,” Int. J. Automot. Technol., vol. 20, no. 2, pp. 313– 325, 2019. [5] H. Smith, M. Zöchbauer, and T. Lauer, “Advanced Spray Impingement Modelling for an Improved Prediction Accuracy of the Ammonia Homogenisation in SCR Systems,” SAE Tech. Pap., 2015. [6] P. Gaynor, B. Reid, G. Hargrave, T. Lockyer, and J. Wilson, “An Experimental Investigation into DEF Dosing Strategies for Heavy Duty Vehicle Applications,” SAE Int. J. Engines, vol. 8, no. 3, pp. 1196–1206, 2015. [7] M.K.A. Wardana, K. Oh, Y. J. Lee, Y. M. Woo, and O. Lim, “Effects of urea injection timing on predicting NOX conversion in SCR systems,” Int. J. Automot. Technol., vol. 21, no. 1, pp. 137–145, 2020. [8] G.M.H. Shahariar, M.K.A. Wardana, and O. T. Lim, “Investigation of urea-water solution spray impingement on the hot surface of automotive SCR system,” J. Mech. Sci. Technol., vol. 32, no. 6, pp. 2935–2946, 2018. Figure 9. Comparison pressure loss at mixing device https://doi.org/10.4271/2015-01-1033 https://doi.org/10.4271/2015-01-1033 https://doi.org/10.4271/2015-01-1033 https://doi.org/10.4271/2016-01-0970 https://doi.org/10.4271/2016-01-0970 https://doi.org/10.4271/2016-01-0970 https://doi.org/10.1016/j.enconman.2016.01.062 https://doi.org/10.1016/j.enconman.2016.01.062 https://doi.org/10.1016/j.enconman.2016.01.062 https://doi.org/10.1016/j.enconman.2016.01.062 https://doi.org/10.1007/s12239-019-0031-x https://doi.org/10.1007/s12239-019-0031-x https://doi.org/10.1007/s12239-019-0031-x https://doi.org/10.1007/s12239-019-0031-x https://doi.org/10.4271/2015-01-1054 https://doi.org/10.4271/2015-01-1054 https://doi.org/10.4271/2015-01-1054 https://doi.org/10.4271/2015-01-1054 https://doi.org/10.4271/2015-01-1028 https://doi.org/10.4271/2015-01-1028 https://doi.org/10.4271/2015-01-1028 https://doi.org/10.4271/2015-01-1028 https://doi.org/10.1007/s12239-020-0014-y https://doi.org/10.1007/s12239-020-0014-y https://doi.org/10.1007/s12239-020-0014-y https://doi.org/10.1007/s12239-020-0014-y https://doi.org/10.1007/S12206-018-0550-9 https://doi.org/10.1007/S12206-018-0550-9 https://doi.org/10.1007/S12206-018-0550-9 https://doi.org/10.1007/S12206-018-0550-9 M.K.A. Wardana and O. Lim / Journal of Mechatronics, Electrical Power, and Vehicular Technology 11 (2020) 45-54 54 [9] X. Auvray et al., “Kinetic modeling of NH3-SCR over a supported Cu zeolite catalyst using axial species distribution measurements,” Appl. Catal. B Environ., vol. 163, no. 2, pp. 393–403, 2015. [10] S. Fischer, R. Bitto, T. Lauer, C. Krenn, J. Tauer, and G. Pessl, “Impact of the Turbulence Model and Numerical Approach on the Prediction of the Ammonia Homogenization in an Automotive SCR System,” SAE Int. J. Engines, vol. 5, no. 3, pp. 1443–1458, 2012. [11] Y. Yi, “Development of a 3D Numerical Model for Predicting Spray, Urea Decomposition and Mixing in SCR Systems,” SAE Int., no. 724, 2007. [12] J. Baleta, M. Vujanović, K. Pachler, and N. Duić, “Numerical modeling of urea water based selective catalytic reduction for mitigation of NOx from transport sector,” J. Clean. Prod., vol. 88, pp. 280–288, 2015. [13] M. Colombo, I. Nova, E. Tronconi, V. Schmeiβer, and M. Weibel, “Mathematical modelling of cold start effects over zeolite SCR catalysts for exhaust gas aftertreatment,” Catalysis Today, vol. 231. pp. 99–104, 2014. [14] B. Opitz et al., “Simulation study of SCR catalysts with individually adjusted ammonia dosing strategies,” Chemical Engineering Journal, vol. 264. pp. 936–944, 2015. [15] H. Nishiyama, Y. Tanaka, and T. Adachi, “A study on the improvement of NOx reduction efficiency for a Urea SCR System,” SAE Tech. Pap., vol. 46, no. 3, pp. 589–595, 2015. [16] T. Lockyer, B. Reid, G. Hargrave, P. Gaynor, and J. Wilson, “Optical investigation on the ability of a Cordierite substrate mixing device to combat deposits in SCR dosing systems,” SAE Tech. Pap., vol. 2015-April, no. April, 2015. [17] M. Zöchbauer, H. Smith, and T. Lauer, “Advanced SCR flow modeling with a validated large Eddy simulation,” SAE Int., vol. 1046, no. 3, p. 14, 2015. https://doi.org/10.1016/j.apcatb.2014.08.003 https://doi.org/10.1016/j.apcatb.2014.08.003 https://doi.org/10.1016/j.apcatb.2014.08.003 https://doi.org/10.1016/j.apcatb.2014.08.003 https://doi.org/10.4271/2012-01-1291 https://doi.org/10.4271/2012-01-1291 https://doi.org/10.4271/2012-01-1291 https://doi.org/10.4271/2012-01-1291 https://doi.org/10.4271/2012-01-1291 https://doi.org/10.4271/2007-01-3985 https://doi.org/10.4271/2007-01-3985 https://doi.org/10.4271/2007-01-3985 https://doi.org/10.1016/j.jclepro.2014.06.042 https://doi.org/10.1016/j.jclepro.2014.06.042 https://doi.org/10.1016/j.jclepro.2014.06.042 https://doi.org/10.1016/j.jclepro.2014.06.042 https://doi.org/10.1016/j.cattod.2014.01.044 https://doi.org/10.1016/j.cattod.2014.01.044 https://doi.org/10.1016/j.cattod.2014.01.044 https://doi.org/10.1016/j.cattod.2014.01.044 https://doi.org/10.1016/j.cej.2014.11.114 https://doi.org/10.1016/j.cej.2014.11.114 https://doi.org/10.1016/j.cej.2014.11.114 https://doi.org/10.4271/2015-01-2014 https://doi.org/10.4271/2015-01-2014 https://doi.org/10.4271/2015-01-2014 https://doi.org/10.4271/2015-01-1039 https://doi.org/10.4271/2015-01-1039 https://doi.org/10.4271/2015-01-1039 https://doi.org/10.4271/2015-01-1039 https://doi.org/10.4271/2015-01-1046 https://doi.org/10.4271/2015-01-1046 https://doi.org/10.4271/2015-01-1046 Introduction II. Materials and Methods A. Mathematical model 1) Thermolysis and evaporation of UWS droplets 2) Urea water solution injection B. Experimental setup C. Validation of simulation III. Results and Discussions A. Exhaust simulation gas B. Injector effect on exhaust simulation gas C. k-epsilon model validation IV. Conclusion Acknowledgement Nomenclature Declarations Author contribution Funding statement Conflict of interest Additional information References