New Insights into Abnormal Combustion Phenomena Induced by Diesel Spray-Wall Impingement under Engine-Relevant Conditions

: High altitude and low temperature is the common extreme environment for internal combustion engines. Under such operating conditions, heavy-duty diesel engines often suffer from serious abnormal combustion, such as knocking combustion, which results in piston crown break-down and cylinder head erosion. Spray-wall impingement and pool fires are considered potential causes; however, the detailed mechanism remains poorly understood owing to the lack of research data. In this study, for the first time, the destructive abnormal combustion induced by diesel spray-wall impingement was identified using an optical rapid compression machine under engine-rele-vant conditions at high altitudes. Combining instantaneous pressure and temperature measurements with simultaneously recorded high-speed photography gives useful insights into understanding the detailed combustion processes. The experimental results show that depending on the extent of diesel spray-wall impingement, supersonic detonation-like reaction fronts featuring bright luminosity can be observed. The propagation of these reaction fronts in-cylinder results in severe pressure oscillations with an amplitude approaching hundreds of atmospheres, which is like the super-knock events in boosted direct-injection spark-ignition engines. Further parametric analysis indicates that the interplay between the diffusion combustion controlled by diesel spray and the premixed combustion dominated by attached film evaporation results in the formation of abnormal combustion. Destructive reaction fronts tend to occur at a prolonged ignition delay time, which facilitates the mixing between diesel evaporation and hot air


Introduction
The combustion performance in an extreme environment is an important criterion for internal combustion engines.In the frigid plateau region, heavy-duty diesel engines often suffer from serious abnormal combustion, which results in piston crown breakdown and cylinder head erosion.Spray-wall impingement and pressure field inhomogeneity are considered as potential causes for erosion as pool fires and pressure oscillations are measured in actual engines [1][2][3][4].However, little is known about the mechanism for such abnormal combustion, motivating fundamental investigations.
Fuel spray impingement on the engine wall in a direct-injection setting is a fundamental issue that affects mixture formation, combustion performance, and emissions [5][6][7].Regarding the aspect of combustion, Dec et al. [6] found that spray-wall impingement deteriorated combustion and increased particulate emissions.Li et al. [7] found that appropriate impinging distance (longer than liquid-phase penetration) enhanced fuel-air mixing and combustion processes.The ignition and combustion characteristics after spray-wall impingement are significantly affected by injection parameters and engine operating conditions.Ma et al. [8] found that small spray angles were necessary to create concentrated combustion at elevated ambient temperatures.Liu et al. [9] found that small wall distances and high wall temperatures accelerated spray vaporization, which in turn shortened ignition delay time and promoted flame development.Moreover, spray-wall impingement can also lead to the formation of fuel film and pool fires [10].Drake et al. [11] found that the vaporization of fuel film was slow compared to flame propagation, and pool fires were a likely source of particulate emissions.These studies have given useful insights into the fundamentals of constrained spray combustion.However, investigations into the influence of fuel spray-wall impingement on abnormal combustion remain scarce, especially for heavy-duty diesel engines.
The abnormal combustion involving pressure oscillations often refers to engine knock, which has been extensively studied in spark-ignition (SI) engines [12].Literature shows that knock severity is determined by reaction front propagation modes [13].Based on the Zel'dovich gradient theory [14], Bradley and coworkers [15,16] observed different autoignition propagation modes and proposed an operational peninsula to determine the critical conditions of detonation development.Based on thermal and concentration stratifications, Savard et al. [17] identified spontaneous propagation and deflagration in premixed turbulent combustion.Chen and coworkers [18,19] observed supersonic reaction fronts, developing detonation, and subsonic reaction fronts.Using optical rapid compression machines (RCM), Qi et al. [20] confirmed that super-knock in SI engines was a result of detonation.Tanoue et al. [21] found that a smaller gradient of ignition delay time caused faster flame propagation and heavier knock severity.Like SI engines, diesel engines may also suffer from engine knock owing to rapid heat release of the premixed combustion stage [22].However, to the best of the authors' knowledge, the role of spray-wall impingement in the abnormal combustion with destructive pressure oscillations remains poorly understood.As spray-wall interactions become enhanced in direct-injection engines with elevated injection pressures, whether the developing detonation depicted in Bradley detonation peninsular [15,16] could be observed under spray combustion conditions is worth exploring.
The objectives of this study are to investigate the fundamental mechanism for the abnormal combustion induced by diesel spray-wall impingement.Using an optical RCM platform, instantaneous pressure and temperature acquisition and high-speed photography were synchronously employed for combustion measurements under engine-relevant conditions.Correlations were done between visualization images and pressure profiles to gain knowledge to interpret these.Injection parameters and ambient conditions were comprehensively investigated to clarify the fundamental mechanism.This study will provide new insights into the combustion control of spray-wall impingement and the nature of knocking combustion.

Experimental Setup
The RCM used in this study has been described in [23].Briefly, the RCM is a singlepiston, pneumatically driven, and hydraulically stopped machine.The schematic of the experimental platform and the layout of the optical measuring system is provided in Supplementary Material S1.To obtain optical access, the metal end-wall was replaced by a stepped quartz window.The piston features a crevice design to avoid roll-up vortices, which ensures a homogenous charge in the combustion chamber.A fuel injector was installed flush with the cylinder wall at the top of the combustion chamber.After being injected out of the nozzle, fuel spray is broken up into ligaments and droplets.Depending on the injection parameters and ambient conditions, the fuel spray impingement on the wall forms the primary impingement region, and the collision of splashing droplets with each other forms the secondary impingement region, as shown in Figure 1.Subsequently, spontaneous ignition and a diffusion flame with bright luminosity can be observed after a period of fuel-air mixing and ignition delay.

Methodologies
In combination with a charge amplifier (Type 5064C; Kistler, Winterthur, Switzerland) and data acquisition device (USB-6366; National Instruments, Austin, TX, USA), a piezoelectric pressure transducer (Type 6045A; Kistler) was employed for instantaneous pressure measurements.The transducer, with a sampling frequency of 100 kHz, was installed flush with the cylinder wall.A temperature transducer at the microsecond level (Type E123CU; NANMAC, Milford, MA, USA) was employed for instantaneous temperature measurements.The transducer had a response time of 20 µs, with an accuracy of 0.5% F.S. To cross-check the variations in charge temperature, an "adiabatic core" hypothesis [24] for isentropic compression was employed and the results of the comparison can be found in Supplementary Material S2.A high-speed camera (Photron FASTCAM SA-Z, San Diego, CA, USA), with a 105 mm lens (AF Micro-Nikkor 1:2.8 D, Nikon, Tokyo, Japan), was employed for capturing the natural luminosity of combustion processes.The camera frame rate could reach as high as 220,000 fps, which corresponded to a resolution of 256 × 256 pixels and a shutter speed of 3.15 µs.The lens aperture was maximized to improve definition and underexposure.The optical diagnostics and temperature measurements were synchronously triggered by the instantaneous pressure signal with a defined threshold.
In this study, a straight-run diesel with a cetane number of 52 was employed, which is commonly used in frigid plateau regions.The detailed fuel properties are shown in Table 1.The composition of the diesel was analyzed based on the method of SH/T0606-2005 and was provided in Supplementary Material S3.In the experiments, high-purity oxygen and nitrogen (>99.99%) were prepared in a mixing tank equipped with magnetic stirring and heating systems such that the variation of oxygen content in atmospheric air at different altitudes could be considered.The amounts of oxygen and nitrogen were calculated according to the quantity of injected diesel.Liquid diesel was injected through an injector with a single nozzle capable of repeatable injections.Diesel spray was controlled by an injector driver box and the LabView VI was employed to accurately regulate solenoid forces.Fuel was delivered to the injector through an electronically controlled common rail.A TTL pulse was employed to control injection timing and pulse width.Figure 2 shows the calibration of injection mass versus injection pressure and spray penetration versus time.Table 2 gives the experimental conditions of diesel spray-wall impingement in the combustion chamber with a diameter () of 70 mm.Further combining sidewall injection, it can mimic the situations of heavy-duty engines with a larger combustion chamber.The injection pressure ( ) was varied from 20 to 100 MPa to achieve diesel spray-wall impingement at different ambient conditions.Such injection pressures are lower than highpressure common-rail systems, but still provide insights into the fundamental combustion induced by spray-wall impingement.The nozzle diameter (  ) of the injector maintained 0.32 mm to mimic the actual scenarios of heavy-duty diesel engines.Injection timing was triggered when the instantaneous in-cylinder pressure reached 5.0 atm to simulate low background pressure while completing the targeted fuel quantity within a limited injection duration.In combination with the quantity of diesel injection, two groups of injection pulse width () were adopted (i.e.,  = 1.5 and 2.0 ms).Regarding the ambient conditions, intake temperature ( ) was maintained at 343 K while two groups of intake pressure ( , i.e.,  = 0.65 and 1.3 atm) were performed to allow for the role of intake pressures.At the current compression ratio, the target pressure ( ) could approach 22-44 atm at the top dead center (TDC), with the target temperature ( ) close to 835 and 860 K, respectively.The experimental repeatability of nonreactive and combustion cycles can be found in Supplementary Material S4.

Pressure Characteristics of Abnormal Combustion
To illustrate the abnormal combustion induced by diesel spray-wall impingement, Figure 3 first shows the pressure traces and the amplitude of pressure oscillations (with 4-25 kHz band-pass filtering) under different injection parameters and ambient conditions.Comparing experimental results, several observations can be obtained.First, for a given ambient pressure, with the elevation of injection pressure, diesel spray-wall impingement increases; consequently, the abnormal combustion with high-frequency pressure oscillations becomes prevalent.Depending on the injection pressure, the amplitude of pressure oscillations can reach dozens and even hundreds of atmospheres.Such combustion characteristics are very similar to the super-knock events encountered in boosted SI engines.Second, for the given injection pressure, with the elevation of ambient pressure, the increased mixture density shortens the diesel spray penetration, thereby restraining spray-wall impingement; meanwhile, ignition timing is advanced, and destructive pressure oscillations are alleviated.However, severe pressure oscillations are still observed at high injection pressures.Third, the quantity of diesel injection also shows an important influence on abnormal combustion.With the increase of injection pulse width, the amplitude of pressure oscillations is significantly enlarged; the strongest one occurs at  = 60 MPa,  = 20 atm, and  = 2.0 ms, where the amplitude goes up to 260 atm (eight times higher than the target pressures).Therefore, both injection parameters and ambient conditions play decisive roles in the formation of the abnormal combustion induced by diesel spray impingement.

Optical Diagnostics on Abnormal Combustion
To identify the detailed combustion processes, Figure 4 shows the results of synchronous measurements on pressure, temperature, and flame development, which correspond to case (A) shown in Figure 3.It is observed that, after the diesel injection, there is no combustion heat release until  = 0.6 ms aTDC, when combustion starts and rapid pressure rise is observed.During this period, the target pressure and temperature at the TDC reaches  = 22 atm and  = 835 K, respectively, and are then maintained almost constantly before the main combustion.The visualization images show that there are two combustion processes involving the entire combustion process.The luminosity with yellow color by soot radiation represents diesel diffusion combustion while the blue luminosity denotes the premixed combustion of vaporized diesel-air mixtures.Specifically, before the diesel diffusion combustion, a blue reaction front starts to appear in the nearwall region adjacent to the location of spray-wall impingement.Then, new autoignition occurs in the front of the blue reaction front, which induces a secondary reaction front propagation.This reaction front quickly passes across a diesel spray beam at a speed of up to 1348 m/s (higher than the local sound speed but lower than the detonation speed at identical operation conditions).At this moment, only a slight increase in pressure trace is observed, without pressure oscillations.After passing across the spray beam, tertiary autoignition occurs ahead of reaction fronts at t = 0.821 ms aTDC, which results in new reaction fronts with faster propagation speeds and brighter luminosities.As these reaction fronts approach the end-wall, rapid pressure rise and severe pressure oscillations are observed, with an amplitude beyond 100 atm.   3. It is observed that, differing from case (a), the diffusion combustion starts before the TDC at high intake pressures; meanwhile, the flame kernel first appears within the beam of diesel spray, behaving with bright yellow luminosity.During this period, diffusion combustion leads to an obvious increase in pressure and temperature.With the progress of combustion processes, blue reaction fronts start to appear in the near-wall region of the combustion chamber.This blue reaction front quickly consumes the remaining diesel-air mixture at a propagation speed approaching 1405 m/s (still lower than the detonation speed at identical operating conditions).Meanwhile, no new autoignition and combustion mode transition is observed within this period.Again, rapid pressure rises, and severe pressure oscillations are induced as the reaction front approaches the end-wall.

Fundamental Mechanism for Abnormal Combustion
To identify the underlying mechanism for abnormal combustion formation, Figure 6 summarizes the experimental results in a diagram consisting of normalized ignition delay time and injection pressure.Herein, the ignition delay time is defined as the time interval between the start of fuel injection and the formation of the flame kernel; the normalized ignition delay time is defined as the ratio of actual ignition delay time to that of the baseline at  = 20 MPa.Depending on the injection pressure, distinct combustion behaviors can be observed for the given operating conditions.It is observed that, for the baseline cases, diesel atomization is so poor at low injection pressures that only a small amount of fuel is burned, producing weak luminosity.With the elevation of injection pressure to  = 30 MPa, diesel atomization is significantly improved and ignition delay time is reduced.Consequently, standard spray combustion starts to be observed in the combustion chamber.The visualization images within pressure regions A and B depict the combustion characteristics of free spray combustion.
As injection pressure is elevated to  = 40-50 MPa, ignition delay time is still reduced, but the trend becomes weakened.This is because a small amount of diesel spray impinges on the cylinder wall at high injection pressures.Then, the evaporation of fuel film promotes the mixing between diesel vapor and hot air.Eventually, premixed combustion is observed just after diesel spray combustion, characterizing supersonic reaction front propagation.When  = 60-70 MPa, diesel spray-wall impingement increases and ignition delay time is prolonged owing to the cooling effect of the cylinder wall.This promotes the evaporation of fuel film and the mixing of diesel vapor and hot air.Consequently, supersonic reaction fronts featured by blue luminosity are observed before diesel spray combustion.The visualization images within pressure regions C and D depict the combustion characteristics of the abnormal combustion induced by diesel spray-wall impingement.
As injection pressure is elevated to  = 80 MPa and above, the increased diesel spray-wall impingement in turn promotes the evaporation and mixing processes of diesel spray and hot air in the near-wall region [25].Consequently, there is some reduction in ignition delay time, which inhibits the evaporation of fuel film.Spray combustion and premixed combustion almost occur at equivalent timescales, and the abnormal combustion with pressure oscillations is suppressed.The visualization images within pressure region E depict the combustion characteristics of attached film combustion.
In summary, the interplay between the diffusion combustion controlled by diesel spray and the premixed combustion dominated by attached film evaporation results in the formation of abnormal combustion.Destructive reaction fronts tend to occur at a prolonged ignition delay time, which facilitates the mixing between diesel evaporation and hot air.

Conclusions
The current work aims to clarify the abnormal combustion of heavy-duty engines operated at high altitudes.Fuel spray-wall impingement was investigated within an optical rapid compression machine.Instantaneous pressure and temperature measurements combined with high-speed photography were employed for the spray combustion of straight-run diesel under engine-relevant conditions.Different injection parameters and ambient conditions were considered.
First, diesel spray-wall impingement is the precondition for the formation of abnormal combustion.Diesel spray impingement results in prolonged ignition delay time owing to the cooling effect of the wall, which promotes the evaporation of fuel film and the mixing of diesel vapor and hot air.Depending on the extent of diesel spray impingement, supersonic detonation-like reaction fronts can be observed, which results in severe pressure oscillations with an amplitude up to hundreds of atmospheres.
Second, abnormal combustion is mainly dominated by the interplay between the diffusion combustion controlled by diesel spray and the premixed combustion controlled by attached film evaporation.In turn, excessive spray-wall impingement inhibits abnormal combustion by modifying ignition delay time and mixing time to a considerable level.Destructive reaction fronts tend to occur at a prolonged ignition delay time, within which the mixing of diesel film evaporation and hot air is facilitated.
Finally, injection parameters play a first-order role in abnormal combustion compared with ambient conditions.With the elevation of injection pressure, diesel atomization is significantly improved, manifesting reduced ignition delay time.Further elevating injection pressure, spray impingement becomes significant and ignition delay time is prolonged, which promotes the evaporation of fuel film and the mixing of diesel and air.The non-monotonic behavior are the main reasons for abnormal combustion.
The current study only addresses the formation conditions and combustion characteristics of abnormal combustion, and further quantitative investigations on the spraywall impingement using different kinds of fuels are worth exploring in future studies.

Figure 1 .
Figure 1.Schematic depicting (a) spray impingement; and (b) fuel film formation in the combustion chamber.

Figure 2 .
Figure 2. Calibration results of injection mass versus injection pressure and spray penetration versus time.

Figure 4 .
Figure 4. Traces of pressure, pressure rise rate, and temperature and visualization images of abnormal combustion at  = 20 atm,  = 1.5 ms, and  = 60 MPa (corresponding to case (a) shown in Figure 3).(a) Pressure and pressure rise rate and (b) visualization images.

Figure 5
Figure 5 further shows the results of synchronous measurements on pressure, temperature, and flame development at  = 40 atm, corresponding to the case (b) shown in Figure3.It is observed that, differing from case (a), the diffusion combustion starts before the TDC at high intake pressures; meanwhile, the flame kernel first appears within the beam of diesel spray, behaving with bright yellow luminosity.During this period, diffusion combustion leads to an obvious increase in pressure and temperature.With the progress of combustion processes, blue reaction fronts start to appear in the near-wall region of the combustion chamber.This blue reaction front quickly consumes the remaining diesel-air mixture at a propagation speed approaching 1405 m/s (still lower than the detonation speed at identical operating conditions).Meanwhile, no new autoignition and combustion mode transition is observed within this period.Again, rapid pressure rises, and severe pressure oscillations are induced as the reaction front approaches the end-wall.

Figure 5 .
Figure 5. Traces of pressure, pressure rise rate, and temperature and visualization images of abnormal combustion at  = 40 atm,  = 1.5 ms, and  = 60 MPa (corresponding to case (b) shown in Figure 3).(a) Pressure and pressure rise rate and (b) visualization images.

Figure 6 .
Figure 6.Correlations between normalized ignition delay time and injection pressure under different injection parameters and ambient conditions.The cases of  = 20 MPa are taken as the baseline.

Table 1 .
Fuel properties for spray-wall impingement experiments.

Table 2 .
Experimental conditions of diesel spray-wall impingement.