A further tightening of emission legislation with Euro7 is expected. FEV’s hypothesis for the key challenges of the next European emission legislation consist of the following major topics:
- A general reduction of the gaseous emission limits CO: 500 mg/km . HC: 50 mg/km . NOX: 35 mg/km
- Non-allowance of auxiliary emission strategies that can lead to high emissions
- Particle number emissions measured down to 10 nm instead of 23 nm
- Incorporation of further emission components limits for the lab tests
- Extension of the RDE legislation framework to incorporate further emission components and short driving trips
FEV has investigated how ultimately even a zero-impact combustion engine could be achieved, causing less emissions than those contained in the ambient air. In particular the following targets have been set:
- Emissions in WLTC
- NOX: 40 µg/m³ (corresponds to approx. 0.03 mg/km)
- PM (2.5): 25 µg/m³ (corresponds to approx. 0.02 mg/km)
Compared to today’s Euro 6d legislation, this means a reduction of NOX emissions by 99.9 percent and PM emissions by 99.2 percent.
FEV has developed extensive patented and patent-pending development methods in the field of simulation, as well as testing and aging of emission-relevant components, which make it possible to demonstrate high robustness and forecast accuracy at an early stage of development.
RDE emission simulation and identification of worst case cycles
Emission simulation at FEV is an essential pillar in the frontloading of development. Presented
for the first time in 2016 at the Vienna Engine Symposium, and further refined since then, this modular FEV simulation toolchain based on the GT-Suite software environment is now an essential part of FEV development activities. Engine raw emissions are modeled based on stationary and transient measurement data from engine and roller test benches. The simulation models of the exhaust aftertreatment follow a map-based approach. Still, discretization of the catalyst monoliths allows a good description of the warm-up behavior to take into account individual, temperature-dependent conversion rates. Figure 1 depicts the all relevant variables which are included in the calculation of the conversion.
Knowledge of which vehicle- and powertrain-specific cycles can lead to the highest emissions is essential for reliable compliance with all emission limits under RDE conditions. FEV has realized an abstraction of such real driving conditions. The result is a derivation of a concise number of parameters. This parameterization allows machine learning techniques to be applied to identify the worst case RDE cycles based on an analysis of a few hundreds of simulated cycles. This methodology has meanwhile been successfully applied in many development projects.
Catalyst and gasoline particle filter characterization
In the course of the development of FEV’S RDE emission simulation methodology, it was identified that initially catalysts could hardly be modeled with sufficient precision. The reason for this lies in the mostly limited measurement data available from catalyst manufacturers and OEMs. However, for a precise prediction of the emissions under RDE boundary conditions, knowledge of the conversion rate at highest space velocities and in a wide temperature range is of high importance. FEV there-fore developed its own equipment that can be used to characterize catalysts under exactly these conditions. The system shown in Figure 2 is designed and proven for exhaust gas mass flows up to those produced by turbocharged V12 engines to measure the conversion efficiency at high mass flows and cold temperatures, such as they occur in a full load acceleration shortly after an engine start.
Catalyst and gasoline particle filter aging
FEV has established a method for rapid aging of catalysts and GPFs, as well. For GPF aging, the burner test bench is modified allowing oil to be burned in order to generate ash. Different methods have been investigated and finally oil injection was chosen. FEV generated a cycle and oil dosing strategy that is able to reproduce similar aging characteristics as they are found during vehicle durability testing.
Exhaust aftertreatment concept to achieve zero-impact emissions
Five building blocks form the exhaust aftertreatment concept for achieving zero-impact emissions.
- Optimization of NOX raw emissions during cat heating
- Exhaust aftertreatment with readiness immediately after engine start
- HC emission adsorption
- Increase of total catalyst volume
- GPF with improved filtration efficiency
The individual building blocks are discussed below.
NOX optimized catalyst heating
NOX aw emissions can be optimized by an adaptation of cat heating calibration. For very retarded ignition timings, a high amount of fuel is required to generate an IMEP that matches the FMEP. This results in dethrottling and a lower rate of internal EGR. The cylinder peak temperature increases and remains on a high level over a longer period of time. As a result, the NOX emissions increase. To achieve a drastic reduction in NOX emissions, an optimized cat heating calibration would therefore use only a mild spark timing retardation. As a consequence, HC raw emissions would increase, and additional measures need to be implemented to address this.
Electrically heated catalysts
Two electrically heated catalysts are integrated upstream of the main catalyst (4 kW per disc, 8 kW in total). The metallic substrate heats up rapidly achieving light-off after a few seconds. However, an engine start followed by cold exhaust gas flowing across the electrically heated catalysts would drop their temperature below the level needed for sufficient conversion efficiency. Therefore, a secondary air pump is used to flow air across the electrically heated catalysts prior to the engine start in order to heat up the main catalyst as well. Figure 4 illustrates the heat up process of the final system configuration. The convective heat transfer can clearly be seen in the lower half of the diagram. As soon as the engine is started the higher exhaust mass flow leads to even better convective heat transfer but at the same time also a reduction in the temperatures.
Emissions can be further optimized by ensuring that the catalyst system maintains a high temperature level. In a hybrid engine, this can be supported by the operation strategy and re-activation of the electrically heated catalysts.
Emission adsorption before catalyst light-off
One way to achieve emission adsorption is by dedicated coatings. In order to achieve a high adsorption efficiency, low temperatures are necessary. This matches with the lower incoming exhaust gas temperatures due to advanced ignition timings during cat heating. A metal substrate is considered since this allows high thermal inertia and thus low temperature increase in the first seconds of engine operation and an even distribution of the secondary air mass flow to the inlet face of the electrically heated catalyst. With a temperature limit of 850 °C the adsorption catalyst dictates the position of the exhaust aftertreatment system to be not closed coupled which in turn has benefit regarding thermal aging. Figure 5 shows a comparison of cat heating with and without HC adsorption, in this case downstream of the catalyst.
For exhaust aftertreatment systems targeting at catalyst pre-heating with a burner instead of electrically heated catalysts, the adsorption of the burner emissions via a small carbon canister positioned downstream of the catalyst might be a good solu-tion as well.
Increased catalyst volume
The catalyst volume is increase by 30 percent compared to the Euro 6d-TEMP base-line which is already using a bigger catalyst volume compared to former Euro 6b/c levels. This includes the volume of electrically heated catalysts. As a consequence, the space velocity at rated power is reduced to values at which high conversion efficiency can be maintained even in aged conditions.
GPF with improved filtration efficiency
Best-in-class Euro 6c and Euro 6d-TEMP engines without GPF already achieve PM emissions in WLTC of only 0.12 – 0.28 mg/km. Compared to the zero-impact target of 25 µg/m³ (approx. 0.02 mg/km), there is the need for a further PM emission reduction by 83 – 93 percent. This can well be achieved with a second generation GPF.
Final results and outlook
The outlined exhaust aftertreatment system is finally assessed in combination with a 2.0 l 4-cyl. turbocharged GDI engine in a plug-in hybrid configuration. Figure 6 shows the final exhaust aftertreatment system.
Extensive DoE investigations have been performed in order to achieve the zero-impact emission level while minimizing the fuel consumption penalty that arises from the electric pre-heating of the catalysts. Figure 7 depicts the correlation between the electrical pre-heating energy and all resulting gaseous emissions. Valid points fulfill the zero-impact target of NOX emissions lower than 40 µg/m³ as well as a balanced SOC of the battery at the end of the cycle. The optimum for meeting the zero-impact target at best possible fuel consumption is found slightly below 0.4 kWh. HC and CO emissions remain well below FEV’s anticipated Euro 7 limits. But, due to the concept, those emissions are not as drastically reduced as the NOX emissions.
The final results for the optimal operation strategy are depicted in Figure 8. The remaining NOX emissions – although hardly visible – mainly result from the first seconds after engine start. The oxygen storage capacity of the catalyst is completely filled at that time and initial rich operation is required to purge the catalyst before full NOX conversion efficiency is achieved. In the remaining part of the WLTC, NOX emission slips remain minimal. The electrically heated catalysts are re-activated for short intervals during the cycle to ensure the temperatures stay on a sufficiently high level at all times. Fuel consumption increases by 4.3 percent compared to the Euro 6d-TEMP baseline.
The zero-impact emission concept presented here is extremely biased towards achieving minimal NOX emissions. For the fulfillment of “just” the Euro 7 emission limit, several conceptual adaptations are possible, e.g. reduction of the number of electrical heated catalysts from two to one. Moreover, the adsorption catalyst could be eliminated, allowing the entire catalyst system to be re-located back to a closed coupled position.
The potential to achieve a simultaneous reduction in both NOx and CO2 emissions via fitting of a pre-turbine exhaust aftertreatment system (PT-EATS) in combination with a mild-hybrid concept was investigated via simulation. The main engine and hybrid system hardware were specified and thereafter, the operating strategies for recuperation and turbocharger control were determined to enable the system to meet a defined tailpipe NOX emissions value of 40 mg/km with a conformity factor (CF) of 1 over all real-world driving cycles. The performance and drivability of the demonstrator are defined to be equivalent to the base vehicle.
The main concept was to locate the exhaust aftertreatment system (EATS) directly downstream of the exhaust manifold, but upstream of the turbine as shown in Figure 1, so that the best CO2 reduction potential and the best aftertreatment performances can be achieved simultaneously.
The engine hardware and PT-EATS were designed and optimized via simulation to identify the best layout of the catalysts and to quantify the potential benefits for CO2 and NOX emissions reduction. The 48V system, made up of a belt starter generator (BSG) with the associated control components, an electric assisted turbocharger (e-TC) and the 48V battery as well as PT-EATS, were integrated to the existing engine model. The simulations optimized EATS component sizes to achieve successfully the integration within the engine bay. The e-Turbo was dimensioned in GT Power and moreover the EGR strategy was optimized to meet the extremely low engine-out NOx emission targets. Furthermore, the recuperation potential was established by using the simulation model. The original exhaust manifold was rotated 180 °C to enable the integration of the turbocharger and a larger EGR cooler was inserted to allow EGR to be used during full-load operation. Additional design modifications were made to the intercooler bracket, the water lines and air lines to allow the complete packaging within the engine bay of the chosen J-Segment demonstrator vehicle.
Placing the aftertreatment system upstream of the turbine results in an altered enthalpy and thermal inertia profile over the turbocharger compared to a conventional arrangement (Figure 2). The introduction of a 48V electric system to the vehicle enables the incorporation of a pre-turbine aftertreatment system via the integration of an e-TC which compensates the loss of pressure and temperature caused by the increased thermal inertia of the PT-EATS.
In the early phases of operation, the temperature before turbine is significantly lower than without the PT-EATS, owing to the increased thermal mass, but as the exhaust system heats up, a thermal lag and overall temperature offset is seen as a result of the higher thermal mass upstream of the turbine. The heat loss profile over the PT-EATS leads to a calculated cumulative enthalpy loss of ~ 4 percent over a WLTC (Figure 2). In order to maintain the boost pressure levels in such low enthalpy phases, the electric turbocharger generates additional boost pressure, it is also used to recuperate excess energy whenever possible.
The recuperation potential of the system was investigated at two part load operating points, shown in Figure 3. The comparison of the brake specific fuel consumption for 2 recuperation strategies was investigated and shown in Figure 3. Recuperation at the turbocharger via VGT is compared with the extraction of the same power over the BSG via an operating point load shift. The latter strategy shows a more energy-efficient path by up to 3.3 percent, as closing the VGT increases the pump losses, as seen in Figure 3, right. It should be noted however that increasing recuperation increases fuel consumption as additional power is needed to generate the same effective power.
The sizing of the e-TC was considered, as a larger turbine could reduce fuel consumption via optimized pumping losses, but, as typical passenger car driving scenarios are not significantly impacted by pumping loss based fuel consumption penalties, there was minimal benefit to upsizing the e-TC as the electrical boost required during transient operation would be increased so a smaller turbine was chosen. The key criteria in determining the size of the e-machine used is the transient response behavior of the vehicle. An acceleration from a standstill to 100 km/h was simulated with different sizes of e-TC to see which could achieve a comparable acceleration behavior to the base vehicle (8.7 s in the sprint to 100 km/h). These simulations are seen in Figure 4. Illustrating that without electrical boost support, the acceleration time increases to 13.0 s, confirming that an e-Turbo is required. An increase in the E-machine power above 11 kW showed no significant reduction in response time (9.0 s to 9.4 s) as acceleration was limited by the electrical machine speed to 180000 min-1.
Regarding the potential to reduce the cost and complexity of the EGR circuit, the use of an HP-EGR-only strategy was investigated which would consist of recuperating excess exhaust gas energy, while controlling the VGT position to achieve the required back pressure to drive higher EGR rates at comparable boost pressure. The results showed, that including the LP-EGR path reduces the required electrical energy over the WLTC by about 30 percent, therefore both HP- and LP- EGR are used. As the PT-EATS volume was found to show only minor impacts on the fuel consumption, compared to the EGR strategy influence, Figure 5, the EATS volumes were chosen to fill the available package space. All the above variables were combined to create an optimised air path strategy. Combining LP and HP-EGR, with a comparatively small turbine and an 11 kW electric motor allows for the lowest possible boost requirement during transient driving conditions. The additional energy requirement for this configuration over a WLTC was approx. 52 Wh when omitting recuperation at the BSG or e-Turbo.
Air path control
The electrical VGT turbocharger requires a dedicated control strategy to optimize the different operating states. For this concept configuration, the electric machine is mainly used for transient support during boost pressure build-up and recuperation during deceleration or in overrun operation. The conventional boost pressure control for the VGT was extended with an advanced model-based control for the power, respectively torque of the electric machine. In this approach the torque of the electric machine is calculated based on the difference between desired and actual turbine torque. An additional e-boost control factor is introduced to balance and adjust the responsiveness of the model-calculated torque demand to the e-machine against the electric energy consumption. The fuel consumption penalty and NOx engine-out emissions as a function of the e-boost control factor are shown, in Figure 6 for the WLTC.
When applying small e-boost control factors, the electric machine only supports during very high differences between desired and actual turbine torque, while for higher control factor values, the e-machine supports for smaller deltas in turbine torque. As such, the NOx engine-out emissions are reduced at higher e-boost control factor values, whereas the fuel consumption increases significantly in consequence of the increased electric power demand. These trends were combined to determine the target operation area.
Overall hybrid strategy optimization
The additional benefits of the 48V mild-hybrid system architecture shown in Figure 7 were evaluated. A 48V belt starter generator replaces the conventional 12V generator, a 48V battery with a capacity of 0.5 kWh, and the electrically supported electrical VGT turbocharger were integrated with the 12V on-board power supply provided via a bidirectional DC/DC converter. When optimizing the control of the electrical VGT turbocharger, a priority manager governs available power for the different consumers, based on the current state of the electric system. The simulation model uses a higher-level energy management strategy to ensure reliable supply for the on-board 12V network under all operating conditions and simultaneously maximizes the potential of the various 48V components to balance transient support during boost pressure build-up and recuperation potential at engine overrun operation and high enthalpy flow upstream turbine.
The balance of the electrical energy within the 48V system over the WLTC is shown in Figure 8. Recuperation takes place almost exclusively by the BSG, whereas the energy consumption is split in roughly equal parts between the supply of the consumers in the 12V network and to support the electrical boosting. As the recuperated energy exceeds the electrical energy consumption, approx. 30 percent is used to charge the 48V battery.
The second part of this paper in an upcoming magazin article will detail the PT–EATS system optimization and the overall system performance over key RDE cycles.