ENERGY AND ENVIRONMENTAL CHARACTERIZATION OF PLUG-IN HYBRID ELECTRIC VEHICLES

Transcription

1 ENERGY AND ENVIRONMENTAL CHARACTERIZATION OF PLUG-IN HYBRID ELECTRIC VEHICLES Ricardo Lopes Instituto Superior Técnico, Technical University of Lisbon; Avenida Rovisco Pais, Lisboa May 213 Abstract: Modern society faces challenges in terms of energy impacts and pollutant emissions and should search for sustainable alternatives. The transportation sector has several challenges for emissions and transportation costs reduction. Plug-in Hybrid vehicles are one answer from the automobile industry to this issue. Through this work, the most representative vehicles of this technology (Toyota Prius Plug-in and Opel Ampera) were studied in terms of their energy impact and pollutant emissions. For that, on-road monitoring using a portable laboratory was performed. The results for both vehicles were studied according with the Vehicle Specific Power methodology. Taking into account the risks and the difficulties of obtaining electricity flows, a methodology that allows calculating electricity consumption from these vehicles as well as estimating electric range was developed. When compared to measured data, the maximum errors for the electric range were 4.2% for the Toyota and -.2% for the Opel. In terms of fuel consumption the maximum error verified was -4.1%. Applying the developed methodologies to both vehicles and for two driving profiles (Lisbon metropolitan area and a sample of American driver), it was verified that the Toyota in charge sustaining mode is more efficient then in charge depleting mode when compared to the Opel, which has a behavior close to an electric vehicle in charge depleting, but it is not as efficient in charge sustaining mode. It was observed that the Toyota consumption is highly dependent on the aggressiveness with which it is driven; fuel consumption reductions can reach 8% when driven softly. Key-words: Plug-in Hybrid Vehicles, global emissions, energy characterization, Vehicle Specific Power, portable laboratory for emissions, on-road monitoring 1. Introduction In recent years, modern society has been facing several economic and environmental challenges such as the reduction of pollutant emissions and improving the efficiency of energy use. Both these challenges are related to each other as most of the world energy comes from burning fossil fuels, consequently emitting pollutants that can be noxious for human health and for the environment. The world s energy use has been increasing significantly in the last decades, with the transport sector and in particular the road passenger sector the biggest responsible for this growth. For this reason, in the transportation sector, the challenges faced are the same as referred above: improve overall efficiency, reduce pollutant emissions and reduce operational costs. Therefore the effort from the manufacturers is to produce more efficient vehicles that can comply with regulations ( (1) for European Union), and are also able to provide a cost effective way of transport (cost per km). In order to reduce operational costs per kilometer and emissions from vehicles, manufacturers started to adopt different solutions as it is the case of Hybrid technology in automobiles, which means the use of two distinct power sources. The most common type of Hybrid combines an internal combustion engine (ICE) with an electric motor (EM). Consequently, besides having a gas tank to store liquid fuel to power the ICE, it also uses batteries to store energy to power the EM. 1

2 Plug-in Hybrid Electric Vehicles (PHEV) are a natural evolution from HEV. These vehicles have characteristics similar to HEV but have the possibility of being plugged-in, which allows them to be externally charged, making it possible to increase full electric driving range, as well as the number of situations in which the electric motor is solicited. PHEV are considered to be a good solution to reduce CO 2 emissions from the transport sector. Studies like the one made by Doucette et al (2), in which CO 2 emissions from electric and plug-in Hybrid vehicles are modeled and compare with published values for CO 2 emissions from conventional ICE vehicles (CV), show the advantage of such solution for the future CO 2 emissions reduction. Methods for evaluation of the impacts from these vehicles include simulation tools, which can provide estimates of vehicle performance (in terms of energy use and pollutant outcomes). Reference simulation tools include Copert, MOVES and ADVISOR (3) (4) (5). However, it is usually necessary to have on-road or laboratorial inputs of a given vehicle technology. Regarding PHEV, there are still few studies that characterize Plug-in Hybrids under real-world operation. Data collecting from vehicle on-road monitoring can be analyzed with the Vehicle Specific Power methodology, as was done by Frey et al (6). The particularity of this study was that the vehicle used was altered from a conventional HEV, meaning that this was not a production PHEV and therefore not projected to work as one. There is a lack of studies for Plug-in Hybrid Electric Vehicles (PHEV) developed from start to work as such and which only started to be available to the consumer recently. Therefore, the focus of this work is to perform an energy and environmental characterization, under real-world operation, of the most representative PHEVs available. For that reason a methodology to characterize the working parameters of such vehicles should be developed in order to provide the necessary information for the simulation tools as well as a systematic way of evaluating future models yet to be released. The main objectives of this research work are: Quantify fuel consumption and pollutant emissions for different operating modes based on measured data; develop a methodology to estimate energy flows in and out of the battery based on Vehicle Specific Power (VSP) and battery State Of Charge (SOC) information; develop a methodology to estimate autonomy in Charge Depleting (CD) mode based on a given drive cycle; and estimate fuel and electricity consumption as well as pollutant emissions from these vehicles when driven by any driver based on drive cycle data. In a PHEV there are two driving modes, which are Charge-Depleting (CD) and Charge-Sustaining (CS) (7) (8). Charge-Depleting is an operating mode in which the battery state of charge (SOC) may fluctuate, but, on average, decreases while driving. Charge-Sustaining consists on an operation mode in which the battery SOC may fluctuate but, on average, is maintained at a defined level while driving. 2. Methodology In the present work it was necessary to develop a methodology to study Plug-in Hybrid electric vehicles. A portable laboratory was used to collect several parameters during normal use of the two PHEV vehicles studied following an approach similar to Frey (6). Parameters like altitude, speed, exhaust gas concentration and battery state of charge are mandatory to fully characterize these vehicles in terms of energy and environmental impact. A laboratory as used by Gonçalves (9) made possible to monitor the referred parameters. During these trips measured on-board, a broad range of driving conditions should be addressed in order to fully characterize all the power spectrum of the vehicle. After gathering this data at 1 Hz, a subsequent data analysis was made using the VSP methodology which groups points of similar power demands in 14 modes. 2

3 The electricity consumption of these vehicles is very important because of the capability of relying exclusively on electricity stored on-board for long distances (more than HEV and less than EV). For this reason and because measuring directly the energy flows on these vehicles is dangerous due to high currents and voltages, it was necessary to develop a methodology to indirectly measure the electricity consumption that allows calculating the electricity consumption under CD mode. The VSP methodology has a particularity that allows simulating a certain trip only by knowing the time spent on each VSP mode. This will allow the results to be validated by simulating a real trip measured on board. Case studies are made in order to simulate the use of these vehicles by typical drivers. 2.1 Measuring apparatus and procedures: For the present work, the VE-LAB portable laboratory for vehicle on-board monitoring was used following similar procedures to (9). The measurements were performed during normal use of the vehicles on Portuguese roads and real time data was collected using a Gas analyzer, to measure exhaust gases concentrations, an OBD reader, to obtain data from vehicle sensors as speed, battery SOC, engine RPM, etc, a GPS receiver with barometric altimeter, to measure altitude with more accuracy, and a Laptop running LAbVIEW to read and store measurements information. These procedures are based on Gonçalves (9) and Frey et al. (6). 2.2 Vehicles used: For this work, two vehicles were used and monitored, an Opel Ampera (Chevrolet Volt in some markets), and a Toyota Prius Plug-In. These two vehicles were studied due to their differences on the powertrain configuration and because they represent the majority of available PHEV. Both these vehicles are Plug-In Hybrid vehicles (PHEV) meaning that both of them have an ICE as well as an EM. The fact that both are PHEV means that both can be charged from the electric grid (from, for example, a home electric installation) allowing the vehicles to increase its full electric range or the use of the EM instead of the ICE. 2.3 Roads and routes: The selected routes were chosen in order to allow the widest possible range of driving conditions. All the trips were performed in the Lisbon metropolitan area, comprehending urban and highway driving. The selected routes were repeated several times with different driving characteristics in order to have the most possible number of driving situations and measured points. The car was always driven by the same driver (as suggested by Frey et al. (1)), in different days and consequently different traffic conditions. A load of around 17 kg (accounting for the 2 passengers and equipment) was always present in all the tests and for every vehicle tested. The Toyota Prius Plug-in was driven for more than 2 km through more than 5 hours of on-board measurements. The Opel Ampera was driven for more than 316 km in a total of about 5 hours and 4 minutes of monitoring. 2.4 Data analysis Vehicle Specific Power Methodology: In order to characterize the power demand for vehicle motion, a useful definition is the Vehicle Specific Power (VSP) methodology which is a simplification of all the forces present on the vehicle. This methodology is a road-load model which allows to estimate instantaneous tractive power per unit vehicle mass, making it very useful to use VSP to perform energy and environmental characterization of vehicles (11), (9). Equation 1 represents the power demand at every second of driving during a trip and valid for light duty vehicles (11). VSP = v (1.1 a grade +.132) +.32 v 3 Eq. 1 A modal analysis is used to group points of similar power per mass (W/kg) demand. Thus, the points collected during on-road measurements are grouped in bins or modes, where each mode has assigned 3

4 the correspondent fuel consumption and pollutant mass emission rates. A 14 mode analysis is used for light-duty vehicles (12) 2.5 Methodology for electricity consumption prediction: During normal driving of the vehicle, constant speed and slope situations were measured at different speeds, namely, 4, 5, 8, 1 and 12 km/h as well as constant speed in downhill conditions to measure power regeneration conditions. This allowed the measurement of battery consumption/regeneration at approximately constant power demands (VSP) as can be seen in Figure 1 a). By measuring the battery SOC variation at any time interval of approximately constant the battery consumption is characterized depending on VSP value. This allows calculating the battery consumption throughout a certain trip as can be seen in Figure 1 b). y = -.69x x a) b) Figure 1 a) SOC variation per VSP value (Toyota Prius Plug-in) b) Comparisson betwen measured and estimated battery SOC (Toyota Prius Plug-in) 3. Results During road tests, data was collected at 1Hz in order to be analyzed afterwards. This data was subsequently treated in a MatLab program developed during this work. The results for the two vehicles will be presented separately, starting by the Toyota Prius Plug-in. These results will be subdivided in CD and CS driving situations. If CS driving conditions are compared to CD ones, in the Toyota Prius Plug-in, average consumption and CO 2 emissions per VSP mode are as present in Figure 2. It can be seen that especially in high VSP modes (especially from mode 11 on) that fuel consumption is practically coincident, which is coherent with the electric power available and vehicle mass (~32 W/kg VSP mode 12). Fuel (g/s) a) CS CD 15 b) 1 CO 2 (g/s) CS CD Figure 2- Comparison between CD and CD a) Fuel consumption, b) CO 2 emissions (Toyota Prius Plug-in) 4

5 This presented fuel consumption in CD is weighted taking into account the time distribution with ICE OFF. This time distribution is present in Figure 4 a). This time distribution was obtained during on-road measurements with rather aggressive driving. This behavior can be explained by the fact that the battery capacity of this vehicle is not very high. Therefore the vehicle powertrain management will prefer saving battery in higher power solicitations (VSP modes 1 to 14), to use it in lower power solicitations (VSP modes 1 to 9). In Figure 3 the pollutant emissions per VSP mode for the Toyota Prius Plug-in are represented. All these emissions are very low, for the Hydrocarbons (HC) and Nitrous Oxide (NO), higher levels of emissions, especially at high powers, were observed under CD when compared to CS. This can be explained by cold starts, more often under CD mode. CO (g/s),4,3,2,1 CS CD,3,2 CS CD CS CD a) b) c),15,2 HC (g/s),1 NO (g/s),1, Once again, it should be said that the results for CD driving are strongly dependent of the driver behavior, as it influences the number of situations with ICE ON. If this vehicle is driven with a more soft, less aggressive behavior, this time distribution with ICE OFF will be considerably different. For an urban situation without aggressive driving, also measured on-board, this distribution was computed and adjusted for the fuel consumption measured. The time distribution of engine OFF for the same vehicle but with a soft driving behavior is present in Figure 4 b) Figure 3 - Pollutant emissions under CD and CS (Toyota Prius Plug-in) a) CO emissions; b) HC emissions; c) NO emissions % time with ICE OFF 1% 8% 6% 4% 2% % % time with ICE OFF 1, a) b),8,6,4,2, Figure 4 Comparison between percentages of time with ICE OFF for a) aggressive driving, b) soft driving The electricity consumption modeled by the developed methodology originated the results in Figure 5. 5

6 Electricity consumption (Wh/s) Figure 5 Electricity consumption per VSP mode in CD (Toyota Prius Plug-in) For modes 12 and 13, despite the fact the graph shows a bar, the number of point were very small, having just four and two seconds respectively accounting for 4% and 6% of all the points measured in these modes in CD. This shows that in those modes the ICE will be solicited to provide the necessary power. The same analysis as presented for the Toyota Prius Plug-in was also done for the other vehicle studied (Opel Ampera) and its results are now presented. Fuel (g/s) 2,5 2 1,5 1,5 8 a) b) 6 CO 2 (g/s) Figure 6 a) Fuel Consumption, b) CO2 emissions under CS mode (Opel Ampera) Figure 6 a) shows Fuel consumption and Figure 6 b) shows CO 2 emissions. The almost constant Fuel consumption in the last 4 modes is explained by the ICE maximum power. Data from the manufacturer and vehicle weight corresponds to a maximum of 36 W/kg or VSP MODE 13, the remaining power is, therefore, supplied by the battery. This was verified in practice, electricity consumption being observed in CS from mode 11 on. Was not possible to verify what would happened if these high power situations were maintained under CS for longer periods of time. CO (g/s),8,6,4, HC (g/s),15,1, NO (g/s),7,6,5,4,3,2,1 a) b) c) Figure 7 - Pollutants emission under CS mode for the Opel Ampera a) CO emissions; b) HC emissions; c) NO emissions 6

7 In Figure 7 the pollutant emissions under CS for the Opel Ampera are presented and were found to be very low. Only CO emissions are a bit high at high power which is probably because these points were obtained right after the end of the CD mode, meaning that the ICE was cold, and consequently having more emissions. The Electricity consumption during CD mode (under CD this vehicle just uses electricity from its battery) is present in Figure 8. 4 Electricity Consumption (Wh/s) Figure 8 Electricity consumption per VSP mode under CD mode (Opel Ampera) The negative values for Electrical consumption are present to account for the regenerative capacity of the vehicle. This means that when the vehicle is in these modes it stores energy, which will permit a reduction on global energy usage. 4. Case studies During the present work, the main goal was always kept in mind, which was to be able to predict what will be the energy consumption and emissions from the measured vehicles when they are used by different users in different regions. Therefore four case studies were established, including a typical driver in Lisbon metropolitan area (LMA) (Portugal), a North Carolina State University (NCSU) sample driver, and two real trips measured onboard, one with an Opel Ampera and another with the Toyota Prius Plug-in in order to validate the results. 4.1 Lisbon metropolitan area driver: Data of this Lisbon metropolitan area driver resulted from monitoring 49 drivers during one week each. It represents almost 5 hours of driving with an average speed of 52.7km/h, which represented a total of 2629 km driven (13). This data was used to simulate the use of both vehicles with a driving profile typical from this geographical area. For the Opel Ampera, full electric range (in Normal and Sport Mode) was calculated and the results are present in Table 1. As the vehicle runs in full electric mode there are no pollutant emissions: Table 1 - Simulation results for typical Lisbon metropolitan area driver in CD (Opel Ampera) State of charge 1% 75% 5% 2% 1% CD driving (km) Electricity consumption (kwh/km).36 WTW CO 2 emissions (g/km) 11 7

8 The 1% of charge corresponds to the full charge reported to the driver although it does not correspond to battery SOC of 1% measured by OBD. In these cases % of charge corresponds to the start of the CS mode, which happens around 24% of battery SOC. Simulation for the Toyota Prius Plug-in was not as straightforward as for the Opel Ampera case in which CS and CD are clearly separated. This vehicle during CD can use only the EM, only the ICE or it can combine both. For this reason, the contribution of each consumption factor should be addressed during CD. For this time distribution of VSP, the results for consumptions, pollutant emissions and electric range in CD mode for this vehicle are present in Table 2. Table 2 -Simulation results for typical Lisbon metropolitan area driver in CD (Toyota Prius Plug-in) State of charge 1% 75% 5% 2% 1% CD driving (km) Electricity consumption (kwh/km).125 Fuel consumption (l/1km) 3.8 WTW CO 2 emissions (g/km) 145 TTW CO emissions (g/km).13 TTW HC emissions (g/km).16 TTW NO x emissions (g/km).4 At CS, both vehicles use fuel. The results obtained for the average consumption in CS for a driver in LMA and its consequent emissions are present in Table 3. Table 3 - Simulation results for typical Lisbon metropolitan area driver in CS (Opel Ampera and Toyota Prius Plug-in) Opel Ampera Toyota Prius Plug-in Fuel consumption l/1km WTW CO 2 emissions g/km TTW CO emissions g/km TTW HC emissions.3.1 g/km TTW NO x emissions.35.2 g/km 4.2 North Carolina State University sample driver: Using a sample of driving profiles collected in North Carolina State University (NCSU), in United States of America, it was possible to calculate a time distribution per VSP mode for this sample American driver. The simulation of the Opel Ampera with the NCSU sample driver results on the data present in Table 4 for all CD driving autonomy. Again as the vehicle is driven in pure electric mode there are no pollutant emissions. Table 4 - Simulation results for a NCSU sample driver in CD (Opel Ampera) State of charge 1% 75% 5% 2% 1% CD driving (km) Electricity consumption (kwh/km).163 WTW CO 2 emissions (g/km) 59 8

9 When compared to the LMA driver, the American driver has an increase on full electric autonomy. This is basically due to the fact that the American driver spends more time in negative VSP values, being able to regenerate more energy. The Toyota Prius Plug-in is simulated for this sample of American drivers; the results are presented in Table 5 for electric range and consumption. Table 5 - Simulation results for a NCSU sample driver in CD (Toyota Prius Plug-in) State of charge 1% 75% 5% 2% 1% CD driving (km) Electricity consumption (kwh/km).121 Fuel consumption (l/1km) 2. WTW CO 2 emissions (g/km) 16 TTW CO emissions (g/km).1 TTW HC emissions (g/km).12 TTW NO x emissions (g/km).27 Because the NCSU sample driver has a less aggressive behavior, and because time distribution per VSP mode is more close to the one verified for soft driving behavior, for this simulation the time distribution with ICE OFF was chosen in accordance. A small increase on electric range and a decrease on average fuel consumption during CD were verified when compared to the LMA driver. This is explained by the reason stated on the previous section for the Opel Ampera case and because this driver has a less aggressive driving When in CS mode, and for the same driver, both vehicles will have the results present in Table 6, for fuel consumption and emissions. Table 6 - Simulation results for a NCSU sample driver in CS (Opel Ampera and Toyota Prius Plug-in) Opel Ampera Toyota Prius Plug-in Fuel consumption l/1km CO 2 emissions g/km CO emissions g/km HC emissions.3.8 g/km NO x emissions.25.1 g/km 5. Conclusions The objective of this work was to assess the energy consumption and associated emissions from the use of Plug-in Hybrid Electric Vehicles (PHEV). It was necessary to establish if the Vehicle Specific Power methodology (VSP) was a valid methodology to characterize these vehicles and also to estimate the electricity consumption and electric driving range, for which indirect battery energy flows methodologies were developed and validated. Two different PHEV where monitored with the use of a portable laboratory in order to perform an energy and pollutant emissions characterization. Operational points were divided into CD (Charge Depleting) and CS (Charge Sustaining), which represent the two main driving conditions. A methodology to predict electricity consumption was developed which provides a solution to estimate electricity consumption without having to measure current flows, a possibly dangerous procedure, due to 9

10 the high voltage present in these systems. This methodology was also validated to calculate CD driving range. The VSP methodology proved to be an adequate tool to characterize these vehicles, being the only problem of it the lack of VSP modes for negative VSP. These modes are the ones where these vehicles regenerate energy to its batteries. The fact that there are only 2 modes for negative VSP can originate overestimations of the regenerative capacity as it groups a large amount of points with different characteristics. The results show that using the VSP methodology it is possible to simulate the energy impact and pollutant emissions from both the vehicles tested when driven by any type of driver. For both simulated drivers (LMA and NCSU driver) the pollutant emissions are low and always bellow European regulations, even EURO 6. This is just not verified in terms of CO emissions for the Opel Ampera which can be explained with the fact that most of the high VSP points in CS were obtained with a cold engine running at high loads, and also, because on normal use all VSP range can be achieved, what does not happened during certification. This will provoke higher CO emissions due to the use of the entire VSP spectrum. For these results to be more accurate, more driving time under CS should be performed for this vehicle because most of the measured trips during this work were under CD mode. Future work includes addressing situations that need to be better characterized during on-road testing. References [1] EUROPEAN PARLIAMENT AND THE COUNCIL, REGULATION (EC) No 443/29 OF THE EUROPEAN PARLIAMENT AND OF THE COUNCIL, 29 [2] Reed T. Doucette, Malcolm D. McCulloch, Modeling the prospects of plug-in hybrid electric vehicles to reduce CO2 emissions, Applied Energy, Volume 88, issue 7, p , 211 [3] COPERT, (Last accessed May 213) [4] United States Environmental Protection Agency, (Last accessed May 213) [5] ADVISOR Advanced Vehicle Simulator. Bigladder software, (Last accessed May 213) [6] H.C. Frey, H.W. Choi, E. Pritchard and J. Lawrence, In-Use Measurement of the Activity, Energy Use, and Emissions of a Plug-in Hybrid Electric Vehicle, Paper 29-A-242-AWMA, Proceedings, 12nd Annual Conference and Exhibition, Air & Waste Management, 29 [7] M. Ehsani, Y. Gao, A. Emadi, Modern Electric, Hybrid Electric and Fuel Cell Vehicles, Boca Raton, CRC Press, 21 [8] T.H. Bradley, A.A. Frank, Design, demonstrations and sustainability impact assessments for plug-in hybrid electric vehicles, Renewable and Sustainable Energy Reviews 13, , 29 [9] Gonçalo Gonçalves, Energy and Environmental Monitoring of Alternative Fuel Vehicles, PhD Thesis, Lisboa, 29 [1] H. Frey, A. Unal, J. Chen, Recommended Strategy for On-Board Emission Data Analysis and Collection for the New Generation Model, Prepared for Office of Transportation and Air Quality, U. S. Environmental Protection Agency, 22 [11] J. L. Jimenez-Palacios, Sensing, Understanding and Quantifying Motor Vehicle Emissions with Vehicle Specific Power and TILDAS Remote, PhD Thesis, Massachusetts Institute of Technology, 1999 [12] H.C Frey, A. Unal, J. Chen, S. Li and C. Xuan, Methodology for Developing Modal Emissions Rates for EPA's Multi-scale Motor Vehicle Equipment emission System, Prepared by North Carolina State University for Office of Transportation and Air Quality, EPA, Ann Arbor, 22 [13] Nuno Pereira, Eficiência energética no sector dos transportes rodoviários: Metodologia para quantificação do excesso de energia consumido devido ao factor comportamental na condução de veículos automóveis ligeiros, MSc Thesis, Lisboa, 211 1