INGAS PROJECT: SECOND YEAR MAIN RESULTS

INGAS PROJECT: SECOND YEAR MAIN RESULTS SUB PROJECT SPA1 Thanks to the support of simulation activities carried out by IFP on compression ratio determination and on the turbomatching optimization, the first preliminary version of the 1.4 turbocharged Multiair stoichiometric engine has been upgraded, increasing compression ratio by 1,2 point and adopting a variable geometry gas turbine as shown in Fig. A1_1. Fig. A1_1: Final engine configuration with dedicated CR and VGT Under this configuration a maximum torque of 230 Nm is attained at 1900 rpm, corresponding to a specific value of 164 Nm/dm3, and a maximum power output of 104 kw (141 CV) at 5000 rpm corresponding to 74 kw/dm3; the Ingas project target of 130 Nm/dm3 and 65 kw/dm3 have been so achieved. Experimental activity at the engine test bench in CRF facilities has been mostly devoted to the engine control unit SW calibration, dealing to the final calibration data set needed to let experimental activities start also on the validator vehicle. This phase has been also supported from the scientific point of view by the results obtained by TU GRAZ and CNR-IM concerning the study of air/fuel mixing phenomena and combustion process. In the first case experimental activities carried out on the optical tools (Figure A1_2) have point out interesting parameters determining the final air/fuel distribution inside the combustion chamber, first being the injection timing and its duration. INGAS 218447 Second Year Publishable Summary 1

Fig. A1_2: Optical Intake Manifold with Injector: CAD Model On the other side, CNR IM has completed the evaluation on multi cylinder engine of the influence of natural gas / hydrogen blends on combustion process and pollutant formation; results confirmed that the main effect is related to the reduction in the ignition delay while the combustion duration, under stoichiometric conditions, is less affected. Combustion process, anyway, results more complete and a reduction in THC formation is observed over the simple reduction of the carbon fraction of the fuel (Figure A1_3). THC, upstream catalyst, [g/kwh] THC, upstream catalyst, [g/kwh] THC, upstream catalyst, [g/kwh] 140 Experimental Points 140 Experimental Points 140 Experimental Points 120 NG 120 11 NG/H2 20% 120 11 NG/H2 40% 11 100 100 8.4 8.4 100 8.4 Torque, [Nm] 80 Torque, [Nm] 80 5.9 Torque, [Nm] 805.9 5.9 60 60 3.3 60 3.3 3.3 40 40 40 0.8 0.8 0.8 20 1000 1500 2000 2500 3000 3500 4000 4500 5000 Speed, [rpm] 20 1000 1500 2000 2500 3000 3500 4000 4500 5000 Speed, [rpm] 20 1000 1500 2000 2500 3000 3500 4000 4500 5000 S peed, [rpm] Fig. A1_3 Total hydrocarbon emissions for NG, NG/H2 20% and NG/H2 40% blends at different engine speed and torque. Concerning the development of the aftertreatment system, ECOCAT has completed the design and the internal evaluation of all the catalyst formulations based on different Pd-Rh ratio completed by a Pd-only one. Samples have been realized using an advanced and innovative substrate able to increase flow turbulence and heat exchange for enhancing fast light off operations (Figure A1_4). INGAS 218447 Second Year Publishable Summary 2

Relative heat transfer factor = 1.0-1.8 Fig. A1_4 Ecocat substrate self-locking structure on the left and the curve for the relative heat transfer factor over the grooves Samples have been supplied to IFP after having been aged at ECOCAT test facilities and they have been tested at the IFP engine test bench in terms of conversion efficiency windows and light off behaviour using the 1.4 turbocharged prototype engine, being the same as CRF s one but for the Multiair system. Experimental activities have determined two promising catalyst formulations with a wider conversion efficiency window and claiming a reduction in light off temperature by 40 C compared to the baseline configuration. These two formulations will be prepared to build up the exhaust pipe line to integrate onto the validator vehicle. At the same time, ECOCAT has also internally continued the evaluation of other catalyst formulation adding also Pt to enhance durability, almost with regard to THC and CO conversion: tests have been carried out both at lab scale and on a Euro 4 bifuel vehicle available in ECOCAT showing a significant reduction of the deterioration factors on emissions; for this reason, other samples will be realized and tested in order to be in measure of selecting the best solution to integrate on the validator vehicle. Finally, during the second year activities related to the study for the definition and integration of the storage system onto the validator vehicle have been launched under strong collaboration with partners from SP B1. A Fiat Bravo has been retained as representative of the C-segment class and the procured vehicle has been modified in the rear part to integrate the 5 type IV gas vessels that have been designed to guarantee the required vehicle range of 500 km (Figure A1_5). INGAS 218447 Second Year Publishable Summary 3

Fig. A1_5: Study for the integration of the gas storage system The vehicle preparation has been completed in all the parts composing the storage and high pressure feeding line. Also on the engine side, the updated version of the 1.4 turbocharged Multiair stoichiometric engine has been installed on the vehicle and the modifications to the wiring harness / engine management system completed, the vehicle being ready to start the experimental activities on the chassis dyno and on road. SUB PROJECT SPA2 SPA2 project objectives: The overall target of SPA2 is to meet the EU6 emission standards with a Turbo DI CNG engine in a segment D vehicle, optimized for mono-fuel CNG operation. The CO2 emission target is 140g/km in the NEDC at same or better driving performance. The performance targets are defined with 300Nm and 130kW. The concept for the powertrain should enable all these targets using an innovative approach. Summary of work performed: In the second project year a lot of work was put on the development of a sufficiently reliable DI injector by Siemens. Despite the expectations at the end of the first year the Gen2 injector concept was not reliable enough. Whereas the new needle group proved its reliability the hydraulic lift transformer had some problems. The newly developed injector (called Gen4) is now based on a lift transformer with metal bellows. The multi cylinder engine generation 2 was setup by Daimler. A suitable ECU environment from Conti provides the basis for the advanced SW development. In parallel the combustion system development went on using the relatively short times with working DI injectors to investigate low-end-torque performance and catalyst heating strategies. Also the investigations on the transparent single cylinder engine went on. The method of Laser Induced Fluorescence (LIF) enables the visualization of the mixing process of air and gas in the combustion chamber. A second measurement campaign with optimized settings was done. Main results achieved: The newly developed Gen4 DI CNG injector is reaching now the required performance regarding flow, repeatability and durability is used at the engine test bed as well as in the INGAS 218447 Second Year Publishable Summary 4

demo vehicle. The developed DI-CNG injector represents an important milestone because it is the keycomponent for CNG-DI. Fig. A2.1 Gen4 Ingas DI-CNG injector prototype The engine and EMS HW was improved step by step and has reached now a status that is suitable for all further investigations. Based on the transparent engine investigations some very promising injection strategies were identified. Especially for catalyst heating the post injection of a small amount of CNG shows some potential for a faster catalyst light-off at lower raw emissions. Also the 3D CFD simulation of the injection and mixing process made good progress. A simulation model was established and optimized that is able to show and prove the effects seen on the transparent engine. On the MCE some of the injection strategies from the optical engine were already tested. The catalyst heating strategy with post-injection shows sufficient combustion stability. The spark retardation demand for same exhaust gas temperature is 15 degrees CA less compared with conventional single injection, homogeneous catalyst heating. On raw emission side HCs are in a similar range whereas NOx emission is reduced by more than 80%. Particulate number is at an acceptable level. INGAS 218447 Second Year Publishable Summary 5

The full load target at low-end-torque is already fulfilled with actual setup, the rated power target was tested up to now with MPI only, 120kW were reached with the actual Turbo setting. The Ingas target of 130kW with direct injection seems to be tested. Outlook: Further testing regarding injection strategies are planned. Full load capability of the DI engine has to be demonstrated. Main task within the 3 rd period is function development and calibration of the engine on the dynamic test bed and in the test-vehicle at Conti to demonstrate emissions and drivability. SUB PROJECT SPA3 Engine & Vehicle The procurement of the 2 stage turbocharger including rig test measurements have been done. Rig testing considered the separate measurements of the splitted stages (high pressure and low pressure) as well as the measurements of the 2 stage system. Using the single stage turbocharger, improvements by the implementation of a preswirl device have been demonstrated on the turbocharger test rig. Utilization of these positive effects on the engine will be investigated within next reporting period. In principle preswirl devices also have potential concerning improvements also within the 2 stage application. Due to the remarkable influence of the natural gas injection window on the engine behaviour, that has been observed on stationary engine testing, CFD calculations regarding the analysis of the homogenisation quality level have been initiated. First results will be reported in the next period. Multicylinder engine (MCE) testing has been done using the DoE methodology due to the complexity of three main influencing operation parameters (ignition timing, relative air/fuel ratio and natural gas injection timing). NEDC testing points out, that the CO2 emission level can be brought close to the target, if adequate gear transmission ratios are applied. Concerning the targets for NOx and HC catalyst efficiencies of 92,7% for NOx respective 88,4% for HC will be required. These CAT performances are very demanding. Further procedure Due to the high demand on the EATS a decrease of the compression ratio (ε=13 ε=12) is intended. The engine investigations then will focus on the evaluation of the trade-off exhaust gas temperature increase vs. loss of efficiency. As a reference for the multi point injection a spot test using a central natural gas injection (single point) is planned. This way the performance check (emissions) decoupled from the EOI impact is possible. The potential of the variable compressor on the engine will be investigated. Concerning vehicle, the delivery of the OPEL ZAFIRA is expected in November 2010. After delivery the packaging activities will start immediately as well as the procurement of a second engine management system (DSPACE Rapid Pro) and the preparation of the cable harness. INGAS 218447 Second Year Publishable Summary 6

EATS There are two main issues to solve for the aftertreatment system: One is the removal of unburned hydrocarbons (HC) and carbon monoxide (CO) from the exhaust; another is the removal nitrogen oxides (NOx) from the exhaust. Since the operating conditions for the CNG engine in Subproject A3 with respect to air-fuel stoichiometry is ultra-lean, i.e. combustion in substantial excess of air, known three-way catalyst technology cannot be used to remove NOx emissions. Removal of unburned HC and CO, the most likely technology to be used is catalytic oxidation to form CO2. Methane (CH4) which is the most abundantly present one is one of the most stable hydrocarbons and hard to oxidize. High exhaust temperature (>450 C) and as low as possible sulphur content (ppb level) in the engine out gas are identified as the most critical design parameters. Currently, the most active catalyst for methane oxidation appears to be highly dispersed palladium (Pd) particles supported on alumina (Al2O3). Thus, a catalyst screening program has been initiated at HT in order to identify catalyst compositions which are more resistant to sulphur poisoning and, preferably, also less inhibited by water. Initial test results and NICE engine emission/temperature data show that Euro 6 HC levels are difficult to meet for a state-of-theart Pd/Al2O3 catalyst. Concerning the second issue, removal of NOx, the mainstream technology will be selective catalytic reduction (SCR) to form N2 applying use of stoichiometric reductant mixed into the exhaust gas upstream to the catalytic converter. The SCR converter will be used in conjunction with an ammonia gas injection system to reach the emission targets in the A3 subproject. There are several advantages with NH3 compared to urea as reduction media. The main advantage to use NH3 is the earlier start temperature of reduction and volume of storage tank. SUB PROJECT SPB0 The Subproject B0 deals with the fuels aspects of the InGAS project, supporting the development of advanced gas engines. In the second year of the project, the survey of gas compositions in Europe, carried out in cooperation of E.ON Ruhrgas GDF SUEZ, has been completed (task B0.1.1); the report - comprising also the conclusive choice of limit gases - was delivered. The second finished task in the responsibility of E.ON Ruhrgas was the study on H2 embrittlement of CNG steel tanks (B0.1.3). It was concluded, that at the moment, steel tanks are not suited for higher H2 contents. Although the material is the same as for hydrogen tanks, current processing manufacturing parameter of the CNG steel tanks do not allow hydrogen admixture in a higher degree. An approach for a solution of this problem has been identified. Another task of E.ON Ruhrgas, the WTT calculations for the relevant gas supply chains (B0.1.4), was started. The figures obtainted will later on serve as input data for the WTW calculations by CRF. The E.ON Ruhrgas task B04.4 is concerned with engine tests on fuel impact on CNG components in the fuelling system. The gas train of the test engine is equipped with an special device injecting oil into the gasflow, thus simulating the existence of compressor oil in the CNG. At the moment, the intended long term tests with oil admixture are in execution. Their objective is to determine the sensitivity of CNG fuelling components, especially of the injectors, with respect to the oil content in the gas. INGAS 218447 Second Year Publishable Summary 7

GDF SUEZ and Prague University (Josef Bozek Research Center JRBC) cooperate in the investigation of the fuel composition influence on engine operation (WP B0.2 and WP B0.4). Respective engine tests are allocated to JRBC, whereas simulation activities are split up to both partners. During the second year the efforts of GDF SUEZ were focused upon description of engine response to fuel composition variations in combination with control interventions. In this manner the possibilities for achieving fuel-flexibility of engine was assessed. By means of numerical simulations the maps were compiled each of them surveying influence of presence (content) of a single fuel component in a fuel blend in combination with a single control parameter. Impact on engine power and efficiency as well as on mechanical and thermal load of engine parts is described within wide range of fuel composition, operating condition and control adjustments. A computational procedure was developed dedicated to generate the Heat Release patterns taking into account (beside the engine geometry and its operational conditions) the influence of fuel composition. The procedure is able (under certain conditions) to be included into simulation tools used for engine development / optimization. At JBRC, a computational procedure was compiled, tuned and calibrated in the most generalized form dedicated to generate Heat Release patterns depending on fuel composition. The procedure can be incorporated as a part of simulation tools used for optimization of engines fuelled by blends of methane with various components. Comprehensive set of maps was generated by simulations which describes engine response to fuel composition variations and to control interventions (and to combination of both of them). The maps are helpful to find the most effective way to compensate impact of fuel composition variation and to determine the inevitable residual drawbacks associated with use of extremely weak fuel blends. Simulation activities of GDF SUEZ and JRBC were continuously supported by delivery of experimental data dedicated for calibration and verification of models. Certain aspects were described by direct evaluation of acquired experimental data, namely exhaust gas composition and cycle-to-cycle variability. In the subsequent period the complete set of guidelines for compensation of fuel composition variation by appropriate control parameters settings will be compiled. At the same time the definition of limits for acceptable range of fuel composition / quality will be recommended. WP B0.5 is dedicated to the development of a low cost gas quality sensor by MEMS. The feasibility of the gas quality sensor concept of has been shown from a technical as well as an economical aspect. Several tests with lab prototypes have been initiated and results are promising. Discussions within the InGAS community must clarify within which respect the sensor system still needs to be optimized during the 3rd year of the InGAS project. INGAS 218447 Second Year Publishable Summary 8

SUB PROJECT SPB1 InGAS sub-project B1 deals with the development and integration of an advanced CNG storage system on vehicles designed for passenger mobility and/or goods transportation in an urban context. WPB1.1 which was completed M18 as scheduled, focused primarily on defining the specifications and requirements for advanced storage system and its constituent sub-systems and components. For this activity specific reference was made to two prototype demonstrator vehicles, to be developed and realised during the course of this project by CRF in cooperation with the other partners involved: Fiat Bravo passenger car for which a current production, commercially available CNG version does not exist, was realised during the period M12-M24 and corresponds to one of the main Deliverables of SP-A1. Fiat Grande Punto to be developed virtually and realised within SP-B1 (see WPB1.5) to demonstrate that the achievement of the technical requirements specified within the DoW, in particular as regards expected vehicle range, and the integration on the vehicle of an advanced, high capacity, low-weight CNG storage system using innovative Type IV pressure vessels tested according to the ECE R110 regulations. In WPB1.2, the development of the advanced hybrid pressure vessels by Xperion proceeded with the design evolution and testing of three different generations of vessels. In the final generation, a series of 16 different improved fibre lay-ups were designed and tested during the period M13-M24 in order to optimize their behaviour and avoid failure during the critical extreme-temperature cycle test. In comparison to previous vessel generations, the final design solution exhibits the following principal characteristics: applying carbon fibres in the cylindrical as well in the dome sections, a slightly higher content of carbon fibres and less glass fibres, an improved gravimetric and volumetric storage density, same cost level despite increased material costs, due less production costs. The final achieved design promises the best compromise between weight saving, economics and volume. As regards WPB1.3, and specifically the development of the electronic pressure regulator, since the basic design work had already been completed in the period M1-M12, the task during M13-M24 was to prepare prototype hardware for final evaluation of the functionality. Virtual validation was undertaken and hardware testing was performed on a test facility and with a test vehicle. On the basis of the functionality which was proved during these tests, the electronic pressure regulator was provided for further tests on the engine test bench within the context of InGAS SP-A1. The results confirm the complete functionality of the electronic pressure regulator for incorporation into the FIAT Bravo demonstration car being realised by CRF. As concerns the development of the advanced in-tank shut-off valve within WPB1.3, during the period M13-M24 the internal tank valve which Ventrex adopted from Bosch (in 2009 Ventrex bought the CNG High Pressure Components Business Division of the Robert Bosch GmbH) was completely re-designed in terms of its functionality and optimised according improved functionality and lifetime. Also during this period a considerable number of tests were performed both in house and externally based on the requirements of INGAS 218447 Second Year Publishable Summary 9

the ECE- R110 and the functionality and misuse tests required by different OEMs. In this context, Ventrex collaborated closely with Xperion and BAM. During the period M13-M24, the principal developments in WPB1.4, which regards the development of the advanced CNG storage module, have focused on production issues and the potential impact in terms of the layout of a typical vehicle production plant; the main conclusion which can be drawn from the analysis is that the storage module concept can be made to be compatible with standard assembly process operations and material flow assuming it is contemplated early in the process design analysis (for the pallet and towveyors solution which is proposed). WPB1.5, initiated Month 19, concerns the virtual design concept development of the CNG vehicle integrating the advanced rear storage module developed in WPB1.4: The aim is to develop an optimised mono-fuel CNG version of the baseline vehicle, in this case the current production Grand Punto vehicle (Bsegment), in order to achieve with extended range (of the order of 500km) with only minor modifications to the upper body: According to the preliminary analysis which has been conducted during the period M13- M24, it will be feasible to integrate the storage module into the baseline vehicle with only limited modification to the lower section. Furthermore, given the relatively compact configuration of high-pressure Type IV vessels in order to maximise vehicle range, of particular importance is the assessment of crashworthiness. Preliminary results from numerical simulations performed in the period M13-M25 confirm that it should be feasible to optimise the structural resistance of the body-in-white and ensure integrity of the CNG vessels with respect to the different crash load cases under investigation. SUB PROJECT SPB2 Sub-Project B2 deals with the development of an aftertreatment technology for natural gas vehicles, with special focus on the reduction of methane emissions under stoichiometric and lean conditions. An assessment of the potential of a lean-nox trap to remove NOx under lean conditions is also part of the objective. Two approaches are considered for methane abatement. The first one consists in the development of new catalytic formulations for improving the light-off temperature of current high loaded Pd-formulations and to ensure a high long-term durability. The objective is to improve this temperature from 400 C (state of the art) to 350 C after long-term ageing. The second approach consists in the development of a catalytic coated heat exchanger able to keep the catalyst temperature in an active range for ensuring high methane conversions. The overall objective is to fulfil the Euro6 emission standards in the NEDC driving cycle but also to demonstrate the efficiency of the technology outside of the cycle. The first field of activity dealing with advanced catalyst development focuses on the development of Pdbased catalysts as well as on mixed oxide formulations. The second year has mainly been devoted to the improvement of the reference catalyst and the new formulations based on Pd-coatings as well as on mixed oxides impregnated with Pd. For the reference catalyst a small amount of Platinum has been incorporated in the wash-coat formulation. Engine tests carried out on a bi-fuel vehicle showed a significant reduction of the deterioration factors for all the measured emissions (see figure B2.2). INGAS 218447 Second Year Publishable Summary 10

Relative emission 2 1,8 1,6 1,4 1,2 1 0,8 0,6 0,4 1,9 1,8 1,8 1,7 1,7 1,6 1,5 1,5 1,4 1,4 1,3 1,1 1,1 1 1 1,0 1 1,0 Fresh 40 h 80 h Fresh 40 h 80 h Fresh 40 h 80 h 0,2 0 THC CO NOx Pd-Rh/ Fresh Pt-Pd-Rh/ Fresh Pd-Rh/40 h Aged Pt-Pd-Rh/40 h Aged Pd-Rh/80 h Aged Pt-Pd-Rh/80 h Aged Fig. B2.1 Engine test results (relative emissions) for the 40 and 80 hours aged Pd-Rh and Pt-Pd- Rh samples Additionally, earlier developed Pd-based catalysts supported on Al2O3 or Al2O3/CeO2 have been tested with a higher precious metal content corresponding to that of the reference catalyst. An improved activity for methane conversion could be reported, e.g. 95% conversion at 350 C compared to 70% conversion achieved with the reference catalyst at same precious metal loading. Due to the high sensitivity of Pd-based catalysts to sulphur poisoning, the Pd-based systems have been evaluated towards their resistance to sulphur poisoning and their regenerability. As expected the catalysts deactivate strongly after exposure to sulphur but can be regenerated after treatment under stoichiometric or rich conditions and high temperatures around 750 C. To improve the performance of the precious metal based systems, different strategies based on lambda variations have been investigated. It appears that for each temperature of the catalyst, the lambda value can be adjusted towards slightly rich conditions in order to improve and optimize the light-off temperature. Furthermore, continuous lambda variation between lean and rich enables a further improvement of the methane conversion. First investigations performed on the engine test bench with the reference catalyst have confirmed that the methane light-off can be improved by adjusting the lambda value to slightly rich conditions. As an alternative to cost intensive Pd-based catalysts new formulations based on mixed oxides have been developed. Catalyst formulations based on spinels perovskites or hexaaluminates and different preparation methods like co-precipitation or sol-gel were developed and tested. Selected most active formulations were also modified by insertion of some amount of Pd. Although the performance of the developed catalysts (e.g. CuMn oxides) surpasses that of mixed oxides described in the literature, the activity is worse than that of the reference material. The second technical approach for improving the methane conversion consists in the development of an integrated heat exchanger able to keep the exhaust temperature in an appropriate range under transient conditions. The heat exchanger concept has been designed and tested through detailed simulation work and lab scale experiments. Dynamic simulations studies have been performed in the NEDC with the main INGAS 218447 Second Year Publishable Summary 11

objective of improving the cold start phase. Several cold strategies assuming a fuel burner, a bypass of the heat exchanger inflow channels and an electrically heated catalyst have been compared and assessed for further implementation on the engine test bench. Manufacturing and coating of a laboratory and an engine bench prototype have been completed within the second year. The bench prototype corresponds to an assembly of four laboratory prototypes, which characteristics are represented in figure B2.3. After canning and adding of sensors the system has been implemented on the engine test bench for detailed investigation in the third year of the project. In parallel and based on the laboratory results, the main design modifications for the second heat exchanger generation have been defined and implemented in the manufacturing process. 311 Fig. B2.2: Final hex-setup for bench scale prototype Concerning the NOx removal under lean conditions, it has been demonstrated previously on a laboratory scale that a NOx storage catalyst can be regenerated even at low temperatures if hydrogen is present in the exhaust gas during a rich spike. Experiments performed on the engine test bench with a CNG engine have shown that under rich conditions the hydrogen content of the exhaust gas is higher than 1% for lambda values < 0.96. Therefore, with regard to its regenerability, the NOx storage catalyst is a suitable technology for the abatement of NOx in the exhaust of a lean CNG engine. INGAS 218447 Second Year Publishable Summary 12