Experimental analysis of NOx reduction through water addition and comparison with exhaust gas recycling

Accepted Manuscript Experimental analysis of NOx reduction through water addition and comparison with exhaust gas recycling J. Serrano, F.J. Jiménez-Espadafor, A. Lora, L. Modesto-López, A. Gañán-Calvo, J. López-Serrano PII: S0360-5442(18)32347-8 DOI: https://doi.org/10.1016/j.energy.2018.11.136 Reference: EGY 14234 To appear in: Energy Received Date: 23 August 2018 Revised Date: 10 November 2018 Accepted Date: 28 November 2018 Please cite this article as: Serrano J, Jiménez-Espadafor FJ, Lora A, Modesto-López L, Gañán-Calvo A, López-Serrano J, Experimental analysis of NOx reduction through water addition and comparison with exhaust gas recycling, Energy (2018), doi: https://doi.org/10.1016/j.energy.2018.11.136. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT EXPERIMENTAL ANALYSIS OF NOx REDUCTION THROUGH WATER ADDITION AND COMPARISON WITH EXHAUST GAS RECYCLING J. Serrano a* , F. J. Jiménez-Espadafor a , A. Lora a , L. Modesto-López b , A. Gañán-Calvo b , J. López- Serrano b a Department of Energy Engineering, Seville University, Spain, Camino de los Descubrimientos, s/n, 41092 Sevilla, Spain b Department of Aerospace Engineering and Fluid Mechanics, Seville University, Spain, Camino de los Descubrimientos, s/n, 41092 Sevilla, Spain * Corresponding author. Tel.: +34 954482005. E-mail address: jserr[email protected] ABSTRACT The diesel engine requirements regarding the reduction of exhaust emissions, especially nitrogen oxide (NO x ) and particulate matter (PM) are becoming more stringent year by year. A current method used for NO x control is exhaust gas recirculation (EGR). However, this approach significantly increases the production of soot for medium and high rates. Water addition can also be used for NO x control, achieving 50% NO x reduction rates in relation to EGR with a lower production of soot and without additional carbon monoxide (CO) and hydrocarbons (HC) substantially. This paper analyses the weaknesses and the strengths of adding water into the intake manifold with a flow-blurring injector for NOx reduction on a current technology diesel engine with multiple injection thorough a proprietary tool for heat release rate that considered real gas properties. A reduction of NO x emissions around 60 - 70% was achieved with water injection at different loads and speeds. Besides, a clear relationship was established between the minimum attainable NO x emission and the thermal capacity of the load (air plus water), this result stablish the hard relationship between NO x formation and combustion chamber temperature and therefore shows the strong dependence of the Zeldovich mechanism. Keywords: Diesel engine; Water injection; EGR; Pollutant emissions; Heat release; Engine performance 1. INTRODUCTION To better protect the environment, Europeans have implemented more restrictive emissions norms (EURO-6) on passenger diesel cars [1]. These regulations have forced the automotive industry to explore substitutes to conventional fuel injection systems and develop other fuel injection strategies. Researchers have investigated various techniques and processes to control and reduce the nitrogen oxide (NO x ) and particulates emissions prior to post-treatment devices. Diesel engines follow a well-known trade-off between soot and NO x in the sense that any strategy for NO x reduction is accompanied by an increase in the soot. Most of the efforts for reducing this behaviour attempt to reduce NO x and soot at the same time. MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT Some techniques are related to the engine setting [2, 3], such as multi-injection [4, 5], start of the injection angle [4], the increase in the air mass or the cooling of the charge [6]. Others need auxiliary equipment or are not related to the typical settings, such as cooled exhaust gas recirculation (EGR) or water addition. One of the technologies for this purpose is multi-injection that replaces a single fuel injection shot with multi-injection events of a reduced size and improves the management of the dwell time [7, 8]. There are three types of multi-injection: 1) pilot + main injection (one or more pilot injections), 2) main + post-injection and 3) pilot +main injection + post-injection.The use of pilot injection produces a reduction in the fuel ignition delay limits and reduces the maximum heat release rate during the main injection. As a consequence, the process is principally mixcontrolled. When the amount of pilot injection was increased, the ignition delay was shortened to advance the start of combustion (SOC) [9]. The NO x emissions decreased due to the shortened premixed duration and prevented a rapid heat release rate [10, 11]. In the case of the post-injection use, part of the unburned fuel was oxidized and a decrease in the soot engine-out was obtained. The EGR approach is a pre-treatment technique, which is used widely to reduce and control the NO x emission from diesel engines. Re-circulated exhaust gas displaces fresh air entering the combustion chamber with carbon dioxide (CO 2 ) and water vapour present in engine exhaust. As a consequence of the air displacement, a lower amount of oxygen in the intake mixture is available for combustion (dilution effect). The reduced oxygen available for combustion lowers the effective air-to-fuel ratio [15]. This effective reduction in the air-to-fuel ratio affects exhaust emissions substantially. In addition, the mixing of exhaust gases with intake air increases the specific heat of the intake mixture, which results in a reduction of the flame temperature (i.e. a thermal effect) [14, 16]. However, the use of EGR leads to a trade-off in terms of soot emissions. The use of long-term EGR causes problems inside the engines such as higher carbon deposits, lubricating oil degradation and enhanced engine wear [12]. Besides the need to increase boost pressure at middle and high loads when using EGR while maintaining air-fuel ratio at a suitable level in order to minimize its negative impact on soot emissions especially at high engine load were the EGR flow rate and exhaust temperature are high [13]. Water addition is other method to control the NO x emission through the reduction of the combustion temperature [17]. The decrease of the in-cylinder temperature as a result of the increase in the heat capacity of the mixture (thermal effect), the delay of the SOC (for reduction of the load temperature and a small chemical effect on the engine combustion due to OH radicals of water [18]) and the decrease in the oxygen concentration (dilution effect) leads to a NO x emissions reduction. There are different methods for water addition: waterdiesel fuel emulsion (DWE), direct water injection (DWI) and water injection into the intake manifold (WI) [19, 20]. The three methods are: • The DWE is defined as an emulsion of water in standard diesel fuel with specific additives, mainly surfactants, to stabilize the system [21]. The presence of water in diesel fuels brings about a considerable reduction in NO x and in particulate matters (PM) emission. The DWE approach shows two effects: the first is the reduced peak MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT temperature in the cylinder, resulting in a lower level of NO x formed and the second is an improvement in the atomization and mixing with the surrounding air for faster and more efficient combustion characteristics [21, 22]. These improvements are attributed to water droplet micro-explosions that are due to the volatility difference between the water and the fuel [23, 24]. An important aspect is that diesel emulsions can be used without engine modifications. • In DWI, water is supplied directly into the combustion chamber by a separate injector. This method provides the option of controlling the water-to-fuel ratio [24]. The DWI approach enables changing the water-to-fuel ratio while varying engine parameters (speed and load) or can occur during engine warm-up (cold start), as opposed to DWE [25 - 27]. The ignition delay can be maintained at low levels, even if a large amount of water is injected and a large NOx reduction is obtained with this method [20]. The drawbacks of this method are: the complexity of the integration of additional components to the existing engine system and the additional requirements of a redesign of the fuel supply system [23]. • Various strategies to supply water into the inlet air (WI) are presented in the literature: multipoint water injection in the intake pipes close to the inlet valves [28 - 31], singlepoint water injection upstream of the compressor [30 - 32] or downstream of the compressor [32, 33]. The advantages of the WI include an increase in the volumetric efficiency due to the cooling effect, uniform on-line variation of water quantity and a nearly homogeneous water distribution in the combustion chamber. Some studies concluded that a higher water flow rate contributed to a longer ignition delay, higher peak heat release and lowered NO x emission. However, the increased rate also resulted in deteriorated production of carbon monoxide (CO) and hydrocarbons (HC) [20]. WI is the easier way to inject water in the engine. In the present work, the effect of water injection into the intake manifold, both upstream and downstream of the compressor, for a modern four-cylinder, turbocharged, water-cooled, common rail direct injection automotive diesel engine was studied in contrast to the application of EGR on the engine performance, combustion and exhaust emissions. To increase the surface-to-volume ratio of the water drops, and therefore diminish the water deposition on intake manifold, the water was injected into the intake air by a flow-blurring® (FB) injector [34-37]. An unexpected back-flow pattern in the injector’s interior produces brief small-scale perturbations in the FB mechanism, thus resulting in an efficient mixing between the gas and liquid phases. The FB injectors use a simple, yet robust, design in which the interaction of the fluids is controlled by a geometrical parameter, . This parameter is the ratio G/D, where G is the gap between the tip of the liquid feeding tube and the discharge orifice and D is the diameter of the discharge orifice [37, 38]. To atomize a liquid, an energy input is required to break its surface into smaller droplets, though only a small fraction of the input may be transformed into surface energy. The energy source may be mechanical, electromechanical or purely electrostatic. Unfortunately, in atomization processes, most of the input energy is dissipated so that the atomization efficiency (the ratio of surface energy generation to input energy) is typically low. This is also MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT true for pneumatic atomizers, in which the mechanical energy supplier is the gas stream that interacts with the liquid stream. The energy exchange taking place during this interaction involves several fluidic mechanisms leading to small droplet sizes at the expense of low efficiency due to the high energy dissipation [36]. Generally, an optimal efficiency may be achieved by maximizing surface production, minimizing droplet coalescence and minimizing gas expense. Qualitatively, the efficiency of an atomization device is proportional to its complexity and size [36]. Conversely, the atomization mechanism in FB injectors gives rise to a gas-liquid interaction that yields high atomization efficiency despite the relatively simple, yet robust, design of the device. Indeed, the flow geometry of FB atomizers surpasses the efficiency of the so-called pre-filming air-blast atomizers, a highly efficient yet complex and costly technological approach. This improvement is generated by an unexpected back-flow pattern in the FB injector’s interior that produces small-scale perturbations that result in an efficient mixing between the gas and liquid phases (see Figures 2(b) and 3 in [36]). In this study, the effects of three different water ratios on engine performance and exhaust emissions were investigated at two different engine speeds, two injection angles and two preinjection angles. Also, three different loads were chosen for every engine speed and water ratio. Nomenclature Ar argon T Temperature (k) c p specific heat at constant pressure (J/kg K) TDC top dead center CO carbon monoxide u internal energy (J) CO2 carbon dioxide V volume (m 3 ) c v specific heat at constant volume (J/kg K) W work (J) D diameter of the discharge orifice (m) WI water injection into the intake manifold DWE water-diesel fuel emulsion WID WI downstream the compressor DWI direct water injection WIU WI upstream of the turbo-compressor EGR exhaust gas recirculation x species FB flow-blurring FSN filter smoke number Greek symbols G gap between the tip of the liquid feeding tube and the discharge orifice (m) α thermal capacity of the engine charge and fuel power ratio h specific enthalpy (J/kg) θ crank angle HC hydrocarbon  the gap and the diameter of the discharge orifice ratio HP Low heat value (J/kg) HRR heat release rate (J/rad) Subscripts m mass (kg) c relative to the gas properties in the chamber NOx nitric oxides e relative to the evaporation O2 oxygen f relative to the fuel OH hydroxyl radical f,inj relative to the liquid fuel at injection conditions p In-cylinder pressure (Pa) fg relative to the gaseous fuel Q heat transfer to the wall (J) fl relative to the liquid fuel R specific gas constant (J/kg K) i realtive to the species RPM revolutions per minute j realtive to the species SMD sauter mean diameter (µm) m relative to the number of hydrogen atoms SOC start of combustion n relative to the number of carbon atoms SOI start of injection q realtive to the species ref relative to the reference conditions 2. EXPERIMENTAL SETUP AND TESTING MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT The experiments for this work were done on current production diesel engine 1.5 dCi k9k 766, which was made by Renault, and whose main characteristics are shown in Table 1. The original diesel engine (Fig. 1(a)) was modified to measure the combustion pressure and introduce sprayed water at two different positions, namely at the compressor inlet (Fig. 1(c)) and in the inlet manifold (Fig. 1(b)). The engine configuration and timing were performed through the Commander 4DE of Sybelle that allows the management of the following engine sub-systems: the start of injection angle, pre-injections (up to two), dwell angle between consecutive injections, fuel quantity of pre-injections, EGR ratio (through a map with RPM and injection quantity) and the attainable turbocharger pressure (through a map with RPM and injection quantity). Fig. 1. Experimental apparatus with water injection at turbo-compressor inlet, a) original engine configuration, b) water injection post-compressor, c) water injection at compressor inlet. Sensor numbers are defined in table 2. Table 1. 1.5dCi k9k 766 Engine characteristics. Type dCi 4 stroke 4 cylinders Engine displacement 1.461 cm 3 Bore 76 mm Stroke 80 mm Compression ratio 17.9:1 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT Maximum power 63 kW to 4000 RPM Maximum torque 220 Nm to 1800 RPM Fuel Diesel EN590 Lubrication Gear oil pump Injection pump Distributor injection pump with Common rail Injector Solenoid valve EGR Yes, water cooled Turbocharger Water cooled and controlled by waste gate valve Refrigeration Water cooled Tests were done with a commercial diesel fuel compatible with European Standard EN590 that is cooled with the engine cooling system. The recirculated gases (EGR) were cooled with the engine cooling system, so exhaust gas could achieve an almost stationary water temperature at the point of mixing with the inlet air. The EGR rate was evaluated through the change in the air mass flow when EGR was introduced for every load condition according to Equation (1): (1) where is measured online and comes from previous measurements of inlet air flow for every engine load without EGR. The airflow measurement was done with a hot wire flow meter by Bosch. Exhaust gas emissions of NO x , CO, HC, CO 2 and O 2 were measured with a Fisher Rosemount NGA 2000 model that used the following methods. The NO x was measured by chemiluminescence; HC by a flame ionization detector; and CO, CO 2 and O 2 by an electrochemical cell. The gas line from the engine exhaust to the measurement equipment was heated to 133ºC to avoid condensation of the heavier HC. Exhaust gas opacity was measured with an AVL Smoke Meter 415SE that used the filter paper method to determine the Filter Smoke Number (FSN), which is defined by ISO 10054. In this measurement, a variable but exactly defined sampling volume is sampled from the engine exhaust pipe and passed through clean filter paper inside the device. The filtered soot causes blackening of the filter paper, which is measured by a photoelectric measuring head and the result is analysed by a microprocessor. The value determined value is the FSN. The measurements from the AVL Smoke Meter demonstrated excellent repeatability and reproducibility due to the available manual or automatic adjustment of the sample volume. Fuel flow was measured with a Coriolis effect system. According to Figure 1, Table 2 provides all the measurement parameters of the system. Table 2. Sensors used in experimental apparatus. Position Sensor Units Measurement accuracy Comments 1 Airflow-meter ºC 0-480 kg/h (≤3%) Air inlet before compressor 2 Temperature ºC (–40+120)°C (≤3%) Air inlet before compressor 3 Temperature ºC 0-1000°C ± 1.5 ºC Air inlet before compressor        −⋅= EGRwithoutair EGRwithair m m EGR & & 1100(%) EGRwithair m & EGRwithoutair m & MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 4 Temperature ºC 0-1000°C ± 1.5 ºC Air inlet after compressor 5 Temperature ºC 0-1000°C ± 1.5 ºC Air inlet after cooler 6 Pressure bar [absolute] 0-10 bar (≤0.5%) Outlet compressor pressure 7 Temperature ºC 0-1000°C ± 1.5 ºC Air inlet after water injection 8 Pressure bar [relative] 0-250 bar ± 0.5% Combustion pressure inside cylinder 9 Temperature ºC 0-1000°C ± 1.5 ºC Exhaust flow before turbine 10 Pressure bar [relative] 0-250 bar ± 0.5% Combustion pressure inside cylinder 11 Temperature ºC 0-1000°C ± 1.5 ºC Exhaust flow after turbine 12 Injector Excitation voltage A 0-30 A ± 1% Ammeter 2.1 Droplet Size Characterization The droplet size distributions were produced with a FB atomizer that was operated at varying liquid flow rate and gas pressure conditions and recorded with a Malvern Spraytec® system in an ambient atmosphere. The Spraytec® is a light-scattering instrument equipped with a 632.8- nm He-Ne laser with a 15 mm diameter and a 300 mm range lens system capable of measuring particles in the size range 0.1-900 µm. The detection system consisted of 36 elements on a logspaced silicon diode array and had a maximum acquisition rate of 2.5 kHz (rapid mode). In a FB injector, water is fed through the central tube while air is introduced through the external tube [34-37]. For =/ < 0.25, a turbulent motion occurs in the interior of the liquid feeding tube that produces small-scale perturbations and breaks up the liquid surface, thus generating droplets. Figure 2 shows a typical size distribution of water droplets produced with a FB injector with =1.2mm and=0.2mm (see inset), where all droplets had sizes below 50. In this particular case, the distribution was bimodal with a relatively small peak at approximately 600 nm and a larger peak centered at ∼10. Figure 3 shows digital images of a typical cloud of water droplets formed with a FB injector. These results indicate that, unlike other pneumatic atomization devices, FB injectors may provide aqueous aerosols with a suitable droplet size distribution and water flow rate to achieve both cooling of the engine and complete water evaporation in a relatively short period of time. Note that according to the   - law of vaporization, a droplet’s lifetime decreases linearly with its surface area, which is represented by the square of the droplet diameter [38]. The current results and previous results from the authors are in line with observations by [39], where a FB injector produced droplets with a smaller Sauter mean diameter (SMD) and narrower size distribution than an air-blast injector for similar air-to-liquid mass ratios. Furthermore, the pressure drop generated across the device was also lower in the case of the FB injector [39]. MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT Fig. 2. Droplet size characterization. Fig. 3. Water injection spray apparatus. 3. EXPERIMENTAL TESTS A battery of load conditions was chosen to characterize the engine. Two engine speeds (1500 and 2000 RPM), three load conditions (40, 80 and 120 Nm) and two main injection angles (2º and 10º) were selected as basic conditions. In addition, tests with other conditions were carried out, such as a pre-injection with a dwell of 15º and water injection into the intake manifold both upstream and downstream of the compressor. Figure 4 gives a summary of the tested conditions that covered the main operational engine range. MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT In addition, the efficiency in WI upstream of the turbo-compressor is lower by about 3% - 5% for all the conditions (Figs. 10 - 11). The passage of the injected water through the turbocompressor in the WIU cases produces the erosion of the blades, as shown in Fig. 12. The cause of the erosion is the passage of the injected water through the turbo-compressor, mainly because some drops have not had enough time to vaporise and impact with the blades at high speed. This problem does not exist in the WID cases, because the water is injected after the passage of the air through the turbo. For this reason WIU it cannot be considered as a permanent solution for NO x control. Fig. 10. Water quantity – Efficiency at different conditions. Fig. 11. Water quantity – Efficiency at different conditions. Fig. 12. Eroded blades. 5.2 WI downstream of the compressor (WID) vs. the EGR technique Figures 13, 14, 16 and 17 compare the NO x and soot emissions between the WID and EGR cases. The figures also show the dependence of emissions with SOI. 0.320 0.340 0.360 0.380 Efficiency 2000 RPM - 10º - 80 Nm - WID 2000 RPM - 10º - 80 Nm - WIU 2000 RPM - 10º - 120 Nm - WID 2000 RPM - 10º - 120 Nm - WIU Larger quantity of water 0.240 0.290 0.340 0.390 Efficiency 2000 RPM - 2º - 40 Nm - WID 2000 RPM - 2º - 80 Nm - WIU 2000 RPM - 2º - 80 Nm - WID 2000 RPM - 2º - 80 Nm - WIU Larger quantity of water MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT (a) (b) (c) (d) Fig. 13. NO x emissions at different RPM, load and main injection angle: (a) 1500 RPM and 2 o , (b) 1500 RPM and 10 o , (c) 2000 RPM and 2 o and (d) 2000 RPM and 10 o . (a) (b) (c) (d) Fig. 14. NO x emissions at different RPM, load, main injection angle and pre-injection angle: (a) 1500 RPM, 2 o Main and 17 o Pre, (b) 1500 RPM, 10 o Main and 25 o Pre, (c) 2000 RPM, 2 o Main and 17 o Pre and (d) 2000 RPM, 10 o Main and 25 o Pre. MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT With a water flow of 12 kg/h, a greater reduction in NO x emissions than the “No Water/No EGR”. A reduction of 50% is achieved with a flow of 4 kg/h, but if that flow increases to 12 kg/h the reduction in NO x emissions is 70% or higher for cases at 1500 and 2000 RPM for low and medium load. At high load, the reduction is 50-60% for the maximum water-flow tested (12 kg/h). This can be observed for all injection strategies (main injection and main+pre-injection) in figures 13.a – 14.d. As might be expected, the NO x concentration is lower for 2000 RPM and 2º main injection. With a water flow of 12 kg/h, the NO x emissions are lower if the WID and EGR cases are compared whether it is a single main injection or pilot injection, except for the case 2000 RPM and 10º main injection (figure 13.d). The soot emissions in WID with main injection are similar to the “No Water/No EGR” cases, as it can be seen in the figures 16.a – 16.d. An increase of 50% of the level of the particles is observed in the cases with pre-injection at 1500 RPM and 12 kg/h (figures 17.a – 17.b). In EGR case, the soot emissions increased up to 20 times the “No Water/No EGR” cases (Figs. 16.a – 17.d). This phenomenon is due to the water vapour having a significantly higher specific heat capacity than dry air. Adding water to the intake manifold charge involves an increase in specific heat capacity. Because of that, the increment of temperature due to combustion is lower, as shown in Figure 18. All the temperatures were obtained from the HRR model. An additional cooling effect occurred by the evaporation of the water. The heat for evaporating the water was taken from the intake air and, consequently, the charge air temperature decreased. A high combustion temperature favours the formation of nitrogen oxides during combustion. Therefore, the reduction of combustion temperature by water injection led to a lower generation and emission of NO x , as can be seen in Figures 13.a – 14.d. Figure 15 shows how the NO x emissions grow with the temperature and the strong dependence of the NO x with SOI, where the emissions values are smaller for SOI = 2º. The highest NO x values were obtained as the torque increased at different RPM. Re-circulated exhaust gas displaces fresh air entering the combustion chamber with carbon dioxide and water vapour present in engine exhaust. As a consequence of this air displacement, a lower amount of oxygen in the intake mixture is available for combustion (dilution effect). The reduced oxygen available for combustion lowers the effective air-to-fuel ratio and this effective reduction affects exhaust emissions substantially. In addition, the mixing of exhaust gases with the intake air increases the specific heat of the intake mixture, which results in a reduction of the in-cylinder temperature (Fig. 18). This technique controls the NO x because it lowers the oxygen concentration and flame temperature of the working fluid in the combustion chamber. However, the use of EGR leads to a rise of soot emissions due to the reduction of the flame temperature. MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT Fig. 15. Mean temperature. (a) (b) (c) (d) Fig. 16. Soot emissions at different RPM, load and main injection angle: (a) 1500 RPM and 2 o , (b) 1500 RPM and 10 o , (c) 2000 RPM and 2 o and (d) 2000 RPM and 10 o . (a) (b) 0 200 400 600 800 1000 1200 700.0 800.0 900.0 1000.0 1100.0 1200.0 1300.0 NOx (ppm) Mean temperature (K) WID 10º - 2000 RPM WID 2º - 2000 RPM WID 10º - 1500 RPM WID 2º - 1500 RPM MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT (c) (d) Fig. 17. Soot emissions at different RPM, load, main injection angle and pre-injection angle: (a) 1500 RPM, 2 o Main and 17 o Pre, (b) 1500 RPM, 10 o Main and 25 o Pre, (c) 2000 RPM, 2 o Main and 17 o Pre and (d) 2000 RPM, 10 o Main and 25 o Pre. (a) (b) (c) (d) Fig. 18. In-cylinder temperatures: (a) 1500 RPM, 10º Main and 40 Nm, (b) 2000 RPM, 10º Main and 120 Nm, (c) 1500 RPM, 10º Main 25º Pilot injection and 40Nm and (d) 2000 RPM, 10º Main, 25º Pilot injection and 80Nm. Figure 19 shows the variation of the heat release rate of “No Water/No EGR”, exhaust gas recirculation and water injection cases. When the intake oxygen was replaced by water or exhaust gases, the ignition delay became longer. A long ignition delay promoted the pre-mixed stage of the combustion process, which decreased the diffusive burning and hence also contributed to the reduction in soot and NO x emission. MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT (a) (b) (c) (d) Fig. 19. Heat release rate:(a) 1500 RPM, 2º Main and 40 Nm, (b) 2000 RPM, 10º Main and 40 Nm, (c) 1500 RPM, 10º Main, 25º Pilot and 40 Nm and (d) 2000 RPM, 10º Main, 25º Pilot and 80 Nm. The water addition into the intake air did not significantly impact the peak-cylinder pressure values for the same quantity of fuel, RPM and load, as shown in Figures 20.a - 20.d. (a) (b) (c) (d) Fig. 20. In-cylinder Pressure: (a) 2000 RPM, 2º Main and 80 Nm, (b) 2000 RPM, 10º Main and 120 Nm, (c) 2000 RPM, 2º Main, 17º Pilot and 80 Nm and (d) 1500 RPM, 10º Main, 25º Pilot and 80 Nm. MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT The integral value of the area of the HRR from the SOC (maximum peak) as a function of the NO x is shown in Figure 21. This result shows a minimum in the NO x emissions when this value is between 60 and 90. The combination of SOI, RPM and water injection enabled the low values for the NO x and indicated the optimum of the shape of the HRR to minimize NO x emissions. Fig. 21. Integral value of the area of HRR from the SOC (maximum peak) as a function of NO x . The analysis of the test data determined the limit for the minimum attainable NO x emissions related to the thermal capacity of the engine charge (in this case, integrated by air plus water) when it is referred to the fuel power. This parameter, α, is defined in Equation (11): T= D0U VWXV ∙P Y,VWXV Z0U V[\ ∙P Y,V[\ F[^] 0U ,X\` ∙4a ,X\` [)^] (11) Figure 22 shows the relationship of NO x in ppm versus the parameter α for all test conditions given in Fig. 4 when water was injected downstream of the compressor (WID). For low α, in the range of 500-700, there is relevant dispersion in NO x emissions; this means that there were several agents controlling NO x formation from residence time to SOC. However, as the parameter α increases, the dispersion reduces. When α reaches ∼1200, the NO x emission is at minimum in an asymptotic way and obtains a value between 150-250 ppm. This means that the thermal mechanism clearly dominates. This results in a minimum value for the NO x emissions with regard to the thermal capacity of the engine load that was enhanced by the water (Cp water ∼ 4· Cp air ). Fig. 22. NO X as a function of the parameter α [W/MW]. 6. CONCLUSIONS 0 500 1000 1500 400 600 800 1000 1200 1400 1600 NOx (ppm) α WID MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT In the present study, the effects of water addition to the intake air on combustion, performance and emission characteristics were investigated experimentally and compared with EGR in a modern high-speed turbocharged common-rail DI automotive diesel engine installed at the Thermal Engines Laboratory. Based on the experimental results, the main effects of the water addition are summarized as follows: • The amount of water required in WIU was twice that for WID to obtain the same effects in emissions at 2000 RPM and lower effects at 1500 RPM. • When WI occurred before the turbo-compressor, it produced erosion of the blades as shown in Fig. 12 and should not be considered any more. • The NO x emission was reduced by adding water in comparison to the EGR. A similar tendency was observed at selected engine speeds and loads. • A 70 % reduction of NO x emissions was achieved with high WI ratios at 1500 RPM and 2000 RPM for low and medium loads when the experiments were conducted downstream of the compressor. This reduction is 50-60% at high load. This tendency is valid for all the injection strategies. • A reduction of NO x emissions was achieved with WID and without a soot penalty for all injection strategies, except in the pre-injection strategies at 1500 RPM with high water quantity. • The flow-blurring injectors supplied a higher water mass flow than saturation without any engine penalty for WID. • The water addition into the intake air did not result in any significant change of the peak cylinder pressure values and indicated power. • The water addition into the air did not result in any significant change in the specific fuel consumption and effective efficiency of this engine at selected engine loads and speeds. • The huge influence of the thermal engine charge on NO x emissions was stated up to the point to determine the maximum water required for the minimum attainable NO x . An amount of water much greater than 12 kg/h would have no the same impact on NO x emissions. • A minimum NO x was achieved. Through the experiment has been stated that there is a limit for the minimum achievable NO x with the water addition method. It is worth to mention that this result allows to define not only the minimum NO x level but also the fuel flow compatible with it. In addition, this result stablish the hard relationship between NO x formation and combustion chamber temperature and therefore shows the strong dependence of the Zeldovich mechanism. Based on the above conclusions, it can be stated that WI into intake air can be effectively employed in the existing diesel engine to reduce NO x emission and simultaneously keep the soot level low without loss of power and any negative effect on fuel consumption. ACKNOWLEDGMENTS This work is part of item DPI2013-46485-C3-3-R within the I+D+I National Plan in the period 2013-2016 and has been backed by the Spanish Government (Ministry of Economy and MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT Competitiveness). The authors are grateful to the Ministry of Science and Education of Spain for their financial support. REFERENCES [1] 〈http://www.dieselnet.com/standards/〉. Last access done 17 April 2018. [2] C. Vanesa Durán Grados, Zigor Uriondo, Manuel Clemente, Francisco J. Jiménez Espadafor, Juan Moreno Gutiérrez, Correcting injection pressure maladjustments to reduce NOX emissions by marine diesel engines, Transportation Research Part D: Transport and Environment, Volume 14, Issue 1, 2009, Pages 61-66, ISSN 1361-9209, https://doi.org/10.1016/j.trd.2008.09.012. [3] Qiang Zhang, Na Li, Menghan Li, Combustion and emission characteristics of an electronically-controlled common-rail dual-fuel engine, Journal of the Energy Institute, Volume 89, Issue 4, 2016, Pages 766-781, ISSN 1743-9671, https://doi.org/10.1016/j.joei.2015.03.012. [4] Balaji Mohan, Wenming Yang, Siaw kiang Chou, Fuel injection strategies for performance improvement and emissions reduction in compression ignition engines—A review, Renewable and Sustainable Energy Reviews, Volume 28, 2013, Pages 664-676, ISSN 1364-0321, https://doi.org/10.1016/j.rser.2013.08.051. [5] Su Han Park, Hyung Jun Kim, and Chang Sik Lee, Effect of Multiple Injection Strategies on Combustion and Emission Characteristics in a Diesel Engine Energy & Fuels 2016 30 (2), 810- 818 DOI: 10.1021/acs.energyfuels.5b02121 [6] Walter Knecht, Diesel engine development in view of reduced emission standards, Energy, Volume 33, Issue 2, 2008, Pages 264-271, ISSN 0360-5442, https://doi.org/10.1016/j.energy.2007.10.003. [7] A. Ferrari, A. Mittica, E. Spessa, Benefits of hydraulic layout over driving system in piezoinjectors and proposal of a new-concept CR injector with an integrated Minirail, Applied Energy, Volume 103, 2013, Pages 243-255, ISSN 0306-2619, https://doi.org/10.1016/j.apenergy.2012.09.039. [8] Ferrari A, Paolicelli F, Pizzo P. Solenoid injectors equipped with pressure balanced pilotvalve layout. Appl Energy 2015; 151:367–76. [9] Ehleskog, Rickard & L. Ochoterena, Raúl & Andersson, Sven. (2007). Effects of Multiple Injections on Engine-Out Emission Levels Including Particulate Mass from an HSDI Diesel Engine. SAE Technical Papers. 10.4271/2007-01-0910. [10] K. Akihama, Y. Takatori, K. Inagaki, S. Sasaki and A. M. Dean, Mechanism of the smokeless rich diesel combustion by reducing temperature, SAE paper No. 2001-01-0655 (2001). [11] Helmantel, Arjan & Golovitchev, Valeri. (2009). Injection Strategy Optimization for a Light Duty DI Diesel Engine in Medium Load Conditions with High EGR rates. 10.4271/2009-01-1441. MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT [12] Deepak Agarwal, Shrawan Kumar Singh, Avinash Kumar Agarwal, Effect of Exhaust Gas Recirculation (EGR) on performance, emissions, deposits and durability of a constant speed compression ignition engine, Applied Energy, Volume 88, Issue 8, 2011, Pages 2900-2907, ISSN 0306-2619, https://doi.org/10.1016/j.apenergy.2011.01.066. [13] D.T. Hountalas, G.C. Mavropoulos, K.B. Binder, Effect of exhaust gas recirculation (EGR) temperature for various EGR rates on heavy duty DI diesel engine performance and emissions, Energy, Volume 33, Issue 2, 2008, Pages 272-283, ISSN 0360-5442, https://doi.org/10.1016/j.energy.2007.07.002. [14] Alain Maiboom, Xavier Tauzia, Jean-François Hétet, Experimental study of various effects of exhaust gas recirculation (EGR) on combustion and emissions of an automotive direct injection diesel engine, Energy, Volume 33, Issue 1, 2008, Pages 22-34, ISSN 0360-5442, https://doi.org/10.1016/j.energy.2007.08.010. [15] Abd-Alla GH. Using exhaust gas recirculation in internal combustion engines: a review. Energy Convers Manage 2002; 43: 1027–42. [16] Aguilar, Francisco José & Torres García, Miguel & A. Correa Herrero, José & Villanueva, José. (2009). Effect of Turbulence and External Exhaust Gas Recirculation on HCCI Combustion Mode and Exhaust Emissions. Energy & Fuels - ENERG FUEL. 23. 10.1021/ef900330x. [17] Radloff E, Gautier C. Diesel engine NOx reduction using charge air water injection. ASMEICEF 2005-1235. [18] Samec, Niko & Dibble, Robert & Y. Chen, Jyh & Pagon, Andrej. (2000). Reduction of NOx and Soot Emission by Water Injection During Combustion in a Diesel Engine. [19] Xavier Tauzia, Alain Maiboom, Samiur Rahman Shah. Experimental study of inlet manifold water injection on combustion and emissions of an automotive direct injection Diesel engine. Energy 35 (2010) 3628-3639 [20] Z. Sahin, T. Mustafa, D. Orhan. Experimental investigation of the effects of water adding to the intake air on the engine performance & exhaust emissions in a DI automotive diesel engine Fuel, 115 (2014), pp. 884-895 [21] A. Lif and K. Holmberg, “Water-in-diesel emulsions and related systems,” Advances in Colloid and Interface Science, vol. 123–126, no. SPEC. ISS. pp. 231–239, 2006. [22] Subramanian KA. A comparison of water–diesel emulsion and timed injection of water into the intake manifold of a diesel engine for simultaneous control of NO and smoke emissions. Energy Convers Manage 2011; 52:849–57. [23] B. Tesfa, R. Mishra, F. Gu, and A. D. Ball, “Water injection effects on the performance and emission characteristics of a CI engine operating with biodiesel,” Renew. Energy, vol. 37, no. 1, pp. 333–344, 2012. [24] Wei Zhang, Zhaohui Chen, Yinggang Shen, Gequn Shu, Guisheng Chen, Biao Xu, Wei Zhao,