1 Introduction

Wall heat losses represent a significant factor in energy loss, both at partial load and full load conditions [1]. Reducing these losses offers a significant opportunity to lower fuel consumption and emissions. One way to minimize wall heat losses is through the use of insulation layers on the combustion chamber side.

To reduce the heat transfer between the combustion chamber wall and the working gas to increase the thermal efficiency, several detailed potential analyses for insulating combustion chamber walls in combustion engines have been carried out since the 1980s. The work of Woschni et al. [2,3,4,5,6] focused on a directly injected single-cylinder diesel engine. The combustion chamber insulation was carried out using a modified piston crown made of Nimonic 80A, with the insulation potential being equivalent to a 5 mm thick zirconia layer, implemented using air spaces and long, narrow heat flow paths. While the process calculation for the built piston promised a fuel consumption reduction of 1% [2 g/kWh], the experimental investigations showed a fuel consumption increase of 7.5% [15 g/kWh], although the desired reduction in the temperature difference between the working gas and the combustion chamber wall during combustion could be demonstrated. The increase in heat transfer coefficient in the first part of the combustion with increasing surface temperature of the combustion chamber walls were identified as reasons of the increase of fuel consumption.

The scientific community commonly refers to this phenomenon as "convection vive" in their field. Concerning combustion engines, the phenomenon describes that heat transfer from the combustion chamber wall to the cold areas of the engine can be increased by natural convection. According to this effect, the thickness of the laminar sublayer near the wall decreases due to an increase in wall temperature. This layer of stationary working gas is created by the contact of the combustion chamber wall and the flowing working gas, with decreasing velocity of the gas towards the wall and becoming zero at the wall. The decreasing layer thickness of the stationary gas accelerates the heat transfer between the working gas and the wall, which results in an increase of the heat transfer coefficient. Furthermore, the flame burns closer to the wall, which can further increase the wall temperature and result in increased thermal component loading.

Another disadvantage of the "classic insulator" used by Woschni is the large thermal mass of the modified piston crown made of Nimonic 80A. As categorical seen in Fig. 1 (middle), this led, due to the constant increase in wall temperature level throughout the engine cycle, during gas exchange to warming of fresh air, which negatively affects cylinder filling and emissions behavior. Therefore, Woschni concludes that a constant increase in wall temperatures throughout the engine cycle is not productive.

Fig. 1
figure 1

Mean combustion chamber temperature Tg, Piston wall temperature TW and heat transfer coefficient α during an engine cycle for different piston surface materials. Phase with negative (red marked area) and positive (green marked area) temperature difference (TW–Tg). [10]

Full size image

In 2013, Wakisaka, Kawaguchi et al. [7,8,9] from Toyota showed a concept for a "thermal swing" coating with low thermal conductivity and low total thermal mass, represented by low heat capacity and a thin coating layer. This resulted in quick heat release while also providing insulation, causing wall temperatures to quickly follow gas temperatures (Fig. 1: right). The efficiency benefits come from rapid heating during combustion and fast cooling during gas exchange. The "Thermal Swing Heat Insulation Technology" (TSWIN) was first used in 2015 in a series-production diesel engine (vehicle: Toyota Hilux, engine: turbocharged 2.8-L four-cylinder engine with direct injection). Using a piston crown coating made of silica-reinforced porous anodized aluminum (SiPRA), at least a 2% fuel consumption advantage was achieved compared to the standard piston without insulation.

Although numerous studies examine the impact of insulating coatings on diesel engine combustion processes, only a limited number of examples can be found for gasoline engines. Therefore, the following experimental results are intended to contribute to an increase in knowledge in the field of “temperature swing” coatings and their influence on spark ignited engines.

2 Experimental investigations background

The experimental investigations embedded in the FVV-project "Influencing Wall Heat Losses in SI Engines", in which three research institutions were participated: Otto-von-Guericke-University Magdeburg, RWTH Aachen University, and the Forschungszentrum Jülich. As part of this project, a thermal insulation material was applied to the piston surface, providing a simple application and large impact during combustion process. Yttria stabilized zirconia (YSZ) was chosen for its low thermal conductivity, toughness, and mechanical stability. The selected coating material was further developed by the project partner Forschungszentrum Jülich with a focus on maintaining mechanical stability while increasing porosity. This reduced the density and thus the total thermal mass, as well as the thermal conductivity (from 2 W/mK to 0.88 W/mK) [12]. With increasing temperatures in the range of 20 to 450 °C, the thermal conductivity decreases from 0.88 W/mK to 0.57 W/mK. At Jülich, the coating material was finally applied to the piston by atmospheric plasma spraying. At RWTH Aachen University, a 1D simulation model in GT-Power was developed to predict coating efficiency. The model was validated using 3D-CFD and measurement results from the single-cylinder test bench. The simulation model predicted a maximum efficiency increase of 0.3% for the selected material with a layer thickness of less than 100 µm [12]. The experimental investigations, conducted at Otto-von-Guericke-University Magdeburg using a piston with a layer thickness of 62 µm, are presented in detail below.

3 Test bench preparation

3.1 Configuration

Experimental investigations were conducted on a single-cylinder test bench, with external devices providing oil, cooling water, fuel, and intake air at temperature and pressure-controlled conditions. In addition to standard measuring equipment, fuel and exhaust gas measuring systems were integrated into the tests. The engine was equipped with a piezoelectric cylinder pressure sensor and two piezoresistive sensors to record intake and exhaust gas pressures to determine combustion parameters. Engine and test bench specifications are listed in Table 1.

Table 1 Specifications of research engine and test bench
Full size table

The engine was equipped with a "compression ratio adjustment unit" which connects the lower engine, containing the crank mechanism, to the upper engine unit comprising the liner, cylinder head, and camshaft box. The adjustment mechanism, shown in the figure in Table 1, is a ring with an internal thread that can be rotated like a spindle drive.

3.2 Piston design

A flat piston surface would be optimal for the coating process, as it would minimize inhomogeneities and variations in coating thickness during application. However, valve pockets could cause under- or over-coverage during atmospheric plasma spraying, resulting in localized areas with uneven coating thickness. The challenge in piston design was to achieve uniformity and consistency in both the coating material and thickness for the coating process, while also considering the combustion engine requirements, such as an acceptable compression ratio (CR).

To address this issue, a trapezoidal piston roof was machined from the flat raw piston, allowing for a compromise between coating consistency and combustion engine requirements. The final dimensions of the piston roof were determined using a 3D-CAD model of the engine, with further fine-tuning accomplished through iterative valve clearance tests on the physical engine. The compression ratio was adjusted using the "compression ratio adjustment unit," and mechanical fine-tuning was performed on the piston roof. Figure 2 illustrates the final piston as a CAD model.

Fig. 2
figure 2

Trapezoidal piston roof design

Full size image

By considering the available valve spread and maintaining a valve clearance to the piston of at least 1.5 mm, the design achieved a compression ratio of 9.7, as shown in Table 1. The aluminum-alloyed piston (FM-S2N) is designed without cooling channels and has an average roughness depth of 6 µm towards the combustion chamber after machining.

3.3 High-dynamic temperature measurement

Highly dynamic temperature probes were inserted into two machined pistons with trapezoidal roofs to assess the influence of a piston coating on engine efficiency. The signal transmission was cable based to minimize susceptibility to interference on the measurement signal and required the integration of a lever system into the crank mechanism.

Each piston had five temperature probes inserted into identical positions, with one coated after probe insertion and the uncoated one remaining as a reference. Placing surface temperature probes directly on the coating was not possible, as inserting the sensors after the coating process would have resulted in strong inhomogeneities around the probes. Of the five temperature probes integrated into each piston, three were used to record the piston surface temperature (T1, T2, T3) and two were used to measure the temperature within the piston structure (T4, T5). T1 and T3 were placed on the surface towards the exhaust and intake sides, respectively, and T2 was positioned at the center of the piston surface. The two structure temperature probes were also placed at the center of the piston, whereby T4 was 0.5 mm and T5 was 2.5 mm below the piston surface. Figure 3 shows the exact positioning of the temperature probes.

Fig. 3
figure 3

Position of temperature probes in the piston –left: Top view manufactured/right: Side view CAD

Full size image

The temperature probes utilize fine wire-based NiCr-Ni thermocouples, with their legs connected via micro welding under a microscope. They were then embedded in eroded piston holes using high-temperature glue in a layer process with intermediate drying. Surface thermocouples were positioned to protrude from the piston surface, with only the welding points of surface temperature probes finally grounded flat under the microscope after installation to ensure flatness with the piston surface.

All thermocouples were connected by thermo-cables to the analog output amplifiers of type AD8495 from Analog Device, which were located outside of the engine. The amplifiers already contained an internal temperature reference junction compensation and were further connected to the KiBox.

3.4 Lever system for signal transmission

The lever system was designed to guide the thermo-cables from the piston to the outside of the engine with minimal mechanical stress. A CAD model of the engine was used to construct and implement the lever system, which included a swing guide with a Teflon stopper to limit transverse movement (seen in Fig. 4). The design of the lever system also considered installation and necessary engine modifications, including reworking the piston inner contour and connecting rod to accommodate both the connecting rod and swing arm, slotting the engine liner to provide clearance, and redesigning the crankcase side cover of the basic engine to allow integration of the lever system.

Fig. 4
figure 4

Lever system – left: CAD/middle: Manufactured / right: Swing guide

Full size image

To minimize interference with the crank mechanism, the lever system balanced stability and functionality, whereby integrating of more thermocouples would have required additional space for thermo-cables and increased mechanical stress at the joint feedthroughs, ultimately further limiting the durability of the system. By gluing the thermo-cables to the lever system, the piston, connecting rod, and lever system formed a single unit (see Fig. 4: middle) that could only be separated by damaging the cables. This resulted in two functional measurement units: one with a coated piston and one without.

3.5 Measurements using lever system

Six stationary operating points were defined for lever system measurements (Table 2) with avoidance of high engine speeds to prevent thermo-cable stress and high loads to avoid combustion anomalies, overstressing, and coating damage. At the measurements with lever system, the coated piston was measured first and secondly the uncoated piston, whereby the operating points for both application cases were approached in the same order, with a recording time of three minutes for each operating point. This time was determined as a compromise between measurement accuracy, based on preliminary tests without a lever system, the number of operating points, and a minimum estimated durability of 1.5 h for the thermo-cables.

Table 2 Stationary operating points: defined by Engine speed and Indicated mean effective pressure
Full size table

Data from the test bench (e.g., speed, torque) and exhaust gas emissions, were continuously recorded within the three minutes recording time at a frequency of 10 Hz, with gravimetric fuel measurement taken at 0.1 Hz. Mean values were calculated over the entire operating point measuring time for these variables. Kistler KiBox was used to record the intake-, cylinder- and exhaust gas pressures as well as the signal of the temperature probes at the piston with a resolution of 1°CA. Three recordings of 125 engine cycles were taken for each operating point at the beginning, middle, and end of a three-minute period. Pressures and temperatures were processed according to their resolution, and cycle-related data, such as the center of combustion, were averaged over the total number of measured engine cycles, resulting in an average value over 375 engine cycles.

The engine was operated with a stoichiometric air/fuel ratio (λ = 1) control, and an optimal center of combustion of 8°CA a.TDCF was preferred for each operating point. However, knock limitations prevented this from being achieved in all cases. For knock evaluation, the KiBox utilized the Siemens VDO algorithm and provided both a knock frequency (KFRQ) and a statistical knock amplitude (KPEAK_STAT). Knock threshold values were defined as a statistical knock amplitude of 0.6 bar and a knock frequency of 5%, which were not to be exceeded [11].

The total measuring time for each of the two applications, with and without coated piston, for the six operating points was around one and a half hours. Thereby, the 15 min of total measuring time per operating point were divided into three time steps: approaching the operating point, waiting for the engine to reach thermal steady state with any necessary fine adjustments, and finally recording. By maintaining constant boundary conditions at the inlet, environment, and engine conditioning, a basis for comparison for both application cases was established.

4 Measurement results

The results of the measurement series of both application cases are presented below. Two different types of diagrams are mainly used. One presents the parameters of interest as a setting or mean value calculated from measurement results over the operating points, where the operating points are described by set values of engine speed and indicated mean effective pressure (IMEP). For direct comparison, the application case of the coated piston with “grey square” and the application case with uncoated piston “green triangle” are together in one diagram. The other form of representation depicts the temporal discontinuity of certain sizes, such as the measured piston temperatures, and shows them in dependence of degree crank angle. In this representation format, zero degree crank angle always represents the TDCF. For the temperature representations, mean values for each signal of the temperature probes were calculated for each degree crank angle of the 375 engine cycles, so that as a result, a mean temperature profile over the engine cycle can be evaluated for each operating point. To include different operating points, each operating point is assigned a certain color in the time-dependent representations. Each color is represented by two different line types in the diagrams. The application case with coated pistons (“coated”) has a “full” line expression and the one with uncoated pistons has a “dotted” line.

In Fig. 5, the combustion parameters that are important for evaluating the coating, such as the center of combustion, fuel consumption, and the actual value of the IMEP, are first displayed over the operating points, along with the set ignition time. The ignition time was chosen for each operating point and each application case in such a way that the center of combustion was located efficiency optimized at 8°CA a.TDCF. For the operating points with a set IMEP of 10 bar and greater, the optimal center of combustion could not be reached due to knocking limitations. The operating points are marked "red" for clarity and both the knocking frequency (KFRQ) and the statistical knocking amplitude (KPEAK_STAT) are given as numerical values. A change in knocking tendency solely through the coating could not be determined. Based on the ignition time, it is very easy to see that for the uncoated application case, the ignition time was adjusted in the direction of late adjustment so that the centers of combustion were optimally efficient or knocking limits were maintained. Generally, the shift towards a later ignition adjustment also moves the center of combustion into later timing. However, a late adjustment of ignition was necessary in all non-knock-limited operating points to avoid of having the center of combustion too early, which suggests a longer burn duration for the coated application case. Since the actual value of IMEP only differs by a maximum of 0.04 bar in the non-knocking limited operating points, a load fluctuation can be ruled out as the reason. Regarding a lower fuel consumption, the measurement data do not show a clear trend for either of the two applications, but rather a result that depend on the operating point. Since the actual value of the IMEP in the operating point comparison is almost identical for both applications, the stated indicated efficiency reflects the trend of fuel consumption (Fig. 6). A higher fuel consumption, therefore, means a lower efficiency. An exception is the operating point of 2500 rpm and 6 bar. Despite lower fuel consumption, it shows a lower efficiency in the coated application, as the load, expressed by the actual value of the IMEP, is 0.11 bar lower than in the uncoated application. The maximum efficiency difference throughout all operating points considered is only 0.71% in the application comparison. This small influence is confirmed by the simulation results from RWTH Aachen University.

Fig. 5
figure 5

Setting and combustion parameters: Ignition time, Center of combustion, Fuel consumption, Indicated mean effective pressure

Full size image
Fig. 6
figure 6

Exhaust gas analysis and efficiency: Indicated efficiency, Exhaust gas temperature, HC and NOx-emissions

Full size image

Regarding exhaust temperatures (Fig. 6), there is a maximum application case difference of 9 K. Generally, the exhaust temperatures for the coated application case are higher than for the uncoated case. One reason for this could be the tendency for a later center of combustion in the coated application case, which reduces the peak temperature in the cylinder and made the engine less efficient due to the advanced expansion. The more heat energy is not converted into work, the more can be found in the exhaust gas. Another reason could be the coating and its functionality. With less heat absorption of the combustion chamber wall, more heat is available for the exhaust gas. However, this cannot be proven, as the average exhaust gas temperature increase for the coated application was only 4 K. Two important emission components, HC and NOx, are shown in Fig. 6 as indicators of the quality of combustion. Throughout all operating points, the coated application case shows lower HC emissions and higher NOx emissions. Lower HC emissions could indicate better combustion efficiency, and due to higher exhaust temperatures, could indicate post-combustion oxidation of unburned fuel in the exhaust gas system. Higher exhaust gas temperatures could also explain the higher NOx emissions, as well as indicate a higher peak temperature in the combustion chamber.

For a more precise analysis of the coating influence, the temperatures of the temperature probes placed in the vertical center axis of the piston are shown in Figs. 7 and 8. This includes the temperature probes at the surface (T2) and two within the structure, 0.5 mm (T4) and 2.5 mm (T5) below the surface, respectively. The representation of the three temperatures is shown in Fig. 7 over degree crank angle and in Fig. 8 as a cycle-averaged value over the operating points. For better comparison, the diagram for the surface element (T2) is shown twice in Fig. 7, but with different scaling (upper right). Similarly, for better visualization, a diagram with the difference between both measured structural temperatures (T4–T5) is included in Fig. 8 (upper right). The exact temperature values were used to form the difference, while the temperatures of T4 and T5 are rounded to integers for presentation. When evaluating the results, it should always be noted that the temperature probes for the coated and uncoated pistons were at the same position, meaning that the probes were below the coating in the coated application case.

Fig. 7
figure 7

Crank angle depending piston temperatures T2, T4 and T5

Full size image
Fig. 8
figure 8

Operating point depending piston temperatures T2, T4 and T5

Full size image

Figure 7 shows a maximum temperature swing of 160 K for the uncoated case of measured surface temperatures T2. The temperature swings tends to increase with higher engine load and speed, which is also shown as a tendency in the cycle-averaged temperatures for T2 in Fig. 8. This increase could be attributed to the slowly decreasing temperature level after the swings and to a thermal inertia in the uncoated piston. There are also temperature swings at the piston surface T2 for the coated case, which decrease with increasing engine speed and load, whereas the averaged surface temperatures in the engine cycle increase (Fig. 8). In general, the temperature swings for the coated case are lower and time-delayed compared to the uncoated case, which could be due to the coating shielding the temperature probe T2 from the combustion chamber. This shielding is also reflected in the cycle-averaged temperatures for T2, as these tend to be smaller for the coated case. An exception to this tendency is the operating point of 1250 rpm and 3 bar, as the surface temperature is higher throughout all degree crank angle and thus also on average for the coated application. Regardless of the application case, the structural elements T4 and T5 show almost stationary temperature curves over the engine cycle, with the temperature difference showing that temperatures for the coated case are about 1–2 K below those of the uncoated case for almost all operating points.

A comparative representation of the three surface temperature probes (T1, T2, T3) cannot be made throughout all operating points, as there were fatigue failures of the thermocouples during the measurement campaign. Only in Figs. 10 and 11 all three measured surface temperatures for coated and uncoated application case are shown for one exemplary operating point.

The findings on temperatures in the piston center axis allow of conclusions to be drawn about the existing thermal potential. To illustrate this, the measured temperatures were used to calculate a heat quantity or a heat flow normalized to the existing piston area and shown in Fig. 9. The heat flow from T2 to T4 and from T4 to T5 is shown as a time dependent and as a cycle-averaged value. Since the heat flows directly depend on the temperature differences of the temperature probes, the heat flow curves are predictable. Thus, the heat flow between the structural elements (T4 to T5) directly reflects the trend of the temperature difference from Fig. 8. With the exception of one operating point (1250/3), the coated application case always has a lower heat flow. Depending on the operating point, the difference in the applications ranges between 100 and 1000 W when looking solely at the structural elements.

Fig. 9
figure 9

Crank angle depending and mean values for heat flows in the piston: T2 to T4, T4 to T5

Full size image

For heat flows that include the surface temperature, relatively high heat flows are observed due to the high surface temperatures in the uncoated application. However, the order of magnitude of these heat flows should not be overemphasized at this point because surface temperature probes towards the combustion chamber naturally always offer the possibility of strong measurement uncertainty as a mixture of surface and working gas temperature may be recorded. When considering the cycle-averaged heat flows, it is noticeable that the heat flows for the coated application have a smaller dependence on engine speed and load than in the uncoated. This underscores the insulating functionality of the coating.

Figure 10 is intended to classify the measured temperatures on the coated piston. It illustrates a comparison between surface temperatures on the piston obtained from both measurement and simulation, over degree crank angle for the exemplary operating point of 1500 rpm and 6 bar. This operating point serves as a comparison between measurement and simulation, as for this one, all three surface temperatures (T1, T2, T3) could be recorded by the measuring technique. The colored curves represent the simulation data while the black ones show the measurement results. The simulations were conducted at RWTH Aachen University [12, 13]. The 1D simulation is characterized by dividing the piston and coating into multiple discrete thermal layers, with each layer only providing temperature information identical over the entire piston area. This means that, for example, there is only one temperature on the piton surface that is the same throughout the entire piston surface. The fact that the boundary conditions within the layers are constant and thus simplified is reflected in the uniform temperature levels outside the temperature swings in the simulated curves.

Fig. 10
figure 10

Piston temperatures for operating point: 1500 rpm/6 bar—Measurement (T1,T2, T3) and Simulation for coated piston

Full size image

Hence, the temperature level cannot serve as a comparison parameter for measurement and simulation, but only the amplitude of the temperature swings triggered by the combustion. Based on this criterion, the measured T2 and T3 curves can be compared well with the simulated swing 62 µm below the coating surface (blue curve), which is directly on the base piston surface. It should be noted that the coating thickness of the measured piston is 62 µm. The measured temperature swing of T1 is comparable to the simulated swing 50 µm below the coating surface (green curve), which is 12 µm above the base piston surface. However, as the piston was coated after the temperature probes were applied, inhomogeneities in the layer thickness or structure of the coating, as well as possibly inaccurate flatness of the temperature probes on the base piston surface, could be the cause of any discrepancies. In addition, the fact that the temperature probe T1 became defective shortly after measuring the example operating point and no longer provided plausible temperatures, supports the assumption of the mentioned inhomogeneity.

The question about the reasons for the different measured temperature levels outside the temperature swings should not go unanswered, as it includes both the results shown in Fig. 10 and those in Fig. 11. First, it is worth noting that the surface temperatures level on the exhaust side (T1) appear to be lower than those on the intake side (T3). The surface temperature probe T2, located in the middle of the piston, exhibits the highest temperature of the three surface elements. These distinct temperature variances can be attributed to the structural configuration of the engine. The oil injection nozzle for piston cooling is located beneath the piston on the exhaust side, which could account for the comparatively lower temperatures on that side, especially since the piston doesn't have built-in cooling channels. Additionally, the thick piston crown and its proximity to the roof of the combustion chamber contribute to higher temperatures at the middle surface temperature probe (T2).

Fig. 11
figure 11

Piston surface temperatures for operating point: 1500 rpm/6 bar—Measurement for uncoated piston (T1, T2, T3) and Simulation for coated piston

Full size image

Taking into account explicable deviations between simulation and measurement, Fig. 10 shows that simulation and measurement can be brought into agreement. For this reason, a certain credibility can be attributed to the simulated surface temperature for the coated application case (simulation: red curve in Fig. 10). Based on this consideration, the simulated surface temperature for the coated application case (Simulation) was taken up and compared to the measured surface temperatures for the uncoated application case (T1, T2, T3) for the operating point of 1500 rpm and 6 bar, which was already used as a basis in Fig. 10. The result is shown in Fig. 11, with color and line style following Fig. 10. Since temperature probes could not be integrated on the coating surface, the simulated surface temperature is intended to enable a comparison of surface temperatures for both application cases. The simulated temperature swing on the coating surface, at 80 K, is about twice as large as the largest measured temperature swing without coating (at T2 with 45 K). Furthermore, the uncoated measurements (T1, T2, T3) show a strongly delayed temperature decrease after the swing, indicating a large thermal inertia that can be attributed to the difference in thermal mass between the aluminum base piston and the coating. In addition to the coating thickness, the specific heat capacity and thermal conductivity are essential for this. Overall, the surface temperatures for the coated piston show a significantly better temperature response behavior, which means that the wall temperature can better follow the gas temperature.

5 Conclusion and outlook

In the course of the “Experimental investigation of thermal swing piston insulation at single cylinder gasoline engine”, a test bench environment was first established for a single cylinder engine at Otto-von-Guericke-University Magdeburg. With the help of a CAD model of the single-cylinder engine, a trapezoid piston design was developed and iteratively integrated into the engine so that combustion engine and coating-related requirements could be met. Furthermore, it was possible to implement a highly dynamic piston temperature measurement for the engine. The main tasks included the application of temperature probes and the implementation of signal transmission through a lever system, which required structural modifications to the engine drive unit and basic engine. For the experimental investigation, the piston surface was coated with 62 μm thick yttria-stabilized zirconia (YSZ).

The material and coating process used represent a further development that occurred at Forschungszentrum Jülich. While maintaining mechanical stability, the thermophysical properties were optimized to increase porosity and reduce thermal conductivity. To estimate the coating's potential, a 1D simulation model was developed by research partners at RWTH Aachen University and validated using 3D CFD.

Given the limitations on total measuring time imposed by the use of the lever system, measurements were conducted successfully using the lever system for both coated and uncoated pistons. The experiments were carried out at 6 stationary operating points, ranging in engine speed from 1250 to 2500 rpm and from 3 to 12 bar IMEP in the load range. Based on the measured fuel consumption and the resulting indicated efficiency, no clear statements about the potential of a piston coating could be made. This is due to the very small differences in indicated efficiency (maximum 0.71%) between the two applications, which may be associated with an advantage or disadvantage of the coating used depending on the operating point. The measurement results match the simulation results from RWTH Aachen, which also predicted a very small impact on efficiency. The exhaust gas analysis showed a higher exhaust gas temperature for the application with coated piston, which averaged 4 K. In addition to the coating itself, the tendentially later center of combustion in the coated application could also be a reason that can lead to a peak temperature reduction in the combustion chamber. The tendentially higher NOx and lower HC emissions can be attributed to the higher exhaust gas temperature level and possible after-oxidations.

The measured surface temperatures at the piston T2 for both application cases show temperature swings (max. 160 K), which are triggered by the combustion. Naturally, these swings are higher for the application without coating due to the damping effect of the coating. Regardless of the application, the temperature probes in the piston structure T4 and T5 show almost stationary temperature profiles over the engine cycle, with the temperatures for the coated application being approximately 1–2 K below those of the uncoated for almost all operating points. Based on these structural temperature measurements, there is a tendentially smaller heat flow for the coated piston. Depending on the operating point, the difference between the applications based on the piston surface is between 100 and 1000 W. This difference results primarily from the operating point-dependent heat flow differences in the uncoated application. This dependence is less pronounced in the coated application, which can be attributed to the effect of the coating. However, it must be noted that the coating only influences the heat flows in the piston structure as well as the fuel consumption and indicated efficiency quantitatively in a marginal way.

Finally, it could be shown that the measured and simulated surface temperatures can be well compared for the coated application case, if the influences caused by practical application, such as inhomogeneities in the material, are taken into consideration. In addition, a comparison of the simulated surface temperature of the coating with the surface temperature measured on the uncoated piston showed that the surface temperature of the coating can better follow the gas temperature, which can be attributed to the low thermal mass and low thermal conductivity of the coating.

Overall, the investigated influence of thermal piston insulation is minimal, but the effect of the coating could be established through the measurements. However, this statement only applies to the coating material and process used in this study with the specified layer thickness. Further investigations with other materials, including the extension of stationary engine operation to cold start and catalyst heating conditions, could uncover potentials. Additionally, exploring the impact of other fuels could broaden the knowledge horizon. Particularly for hydrogen-powered combustion engines, wall heat losses increase due to faster combustion, so that a higher efficiency of the coating could be expected.