Plug-in Electric Vehicle Impact on Ontario's Grid

Size: px

Start display at page:

Download "Plug-in Electric Vehicle Impact on Ontario's Grid"

Suzanna Fisher
6 years ago
Views:

1 Plug-in Electric Vehicle Impact on Ontario's Grid Claudio Cañizares Department of Electrical & Computer Engineering Power & Energy Systems ( WISE ( With some slides from Dr. Amir Hajimiragha and PhD student Isha Sharma. 1

2 Outline Sponsors Motivation Objectives Generation and Transmission Studies: Deterministic model Probabilistic model Plug-in hybrid vehicles (PHEVs) Ontario studies Distribution Systems: Optimization Model Test System Case studies Conclusions 2

3 Supporting Organizations 3

4 Motivations Energy: Global demand for energy is projected to increase by 50% over the next 25 years. This significant demand increase along with the dwindling supply of oil has raised concerns over the security of the energy supply as well as the environment. 4

5 Motivations Transportation sector: It s one of the largest and fastest growing contributors to both energy demand and greenhouse gases. In Canada, represents almost 35% of the total energy demand and is the highest source of greenhouse gas emissions. In view of these issues and the problems with the supply of oil, the subject of alternative fuels for meeting the future energy demand of the transport sector has gained much attention. 5

6 Motivations Optimal utilization of existing infrastructure: It s desirable for both economic and environmental reasons. Ontario s electricity infrastructure is designed to meet the highest expected demand, which only occurs a few hundred hours per year (at most 5% of the time). This system is underutilized for the remainder of the time, in particular during off-peak hours (namely, 12 to 7 am). During these time intervals, the system could generate and deliver a substantial amount of energy to other sectors such as transport. Surplus generation capacity of Ontario s electricity network during off-peak periods could be utilized for charging the batteries in Plug-in Vehicles (PEVs). 6

Motivations Given Ontario s emerging electricity supply mix and a low GHG footprint, enabling PEVs and displacement of gasoline by electricity would be a major contribution to GHG reduction from the

7 Motivations Given Ontario s emerging electricity supply mix and a low GHG footprint, enabling PEVs and displacement of gasoline by electricity would be a major contribution to GHG reduction from the transport sector Greenhouse Gas Emissions (Megatonnes) Source: OPA Source: OPA Range of Results for Scenarios Case A Case B 7

8 Motivations Opportunity also exists to dovetail into smart grid development plans for Ontario. 8

9 Motivations Major auto manufacturers are committed to bringing PEVs to the market. In the US, government, universities, utilities and the automotive industry are actively partnering to make PEVs a reality in the next 3-10 years. PEV energy storage capacity presents unique opportunities to better integrate intermittent energy resources such as solar and wind power. A significant penetration of these vehicles will have important positive ramifications for power systems by introducing energy storage capacity for the grid. The economics and technical considerations appear promising, but R&D is needed to address specific technical and system questions and identify policy instruments. Penetration of PEVs is contingent on availability, cost and consumer acceptance. Integration into the existing Ontario power system needs to be understood and addressed from infrastructure, planning and regulatory perspectives. 9

10 Objectives Investigate the technical and economic feasibility of improving the utilization of Ontario s power grid during off-peak periods. Reduce the demand on fossil fuels for transportation in Ontario. Consider the existing Ontario s electricity infrastructure and future plans to derive the maximum level of PHEV penetration into Ontario s transport sector, with minimum impact on the grid. 10

11 Generation and Transmission (G&T) Deterministic Model In order to find the maximum level of extra load in the form of PHEVs that can be added to an electricity network in a given period of time: A multi-interval DCOPF model with loss factor considerations was developed based on a zonal model of the electricity system during base-load time intervals. PHEVs are treated as additional discrete zonal loads that depends on number of vehicles. Based on a piecewise linearization of power losses, the resulting optimization model is a Mixed Integer Linear Programming (MILP) problem. 11

12 G&T Deterministic Model The objective function in this optimization model consists of electricity generation and imported/exported power cost and revenue components from 12 am to 7 am, as well as CO 2 credit components which are assigned to each PHEV that can be added to the transport sector: Year Penetration levels Emission credit ω ω ω { 8 365( Pgiy + Pimiy Pexiy ) HEP λi N PHE V iy EC} y i Zone Zonal generation, imported and exported power Hourly Energy Price Number of PHEVs 12

13 G&T Deterministic Model CO 2 credit of PHEVs: The net GHG emission reduction is influenced by the share of fossil fuel in the marginal generation mix. PHEVs result in lower emissions compared to gasoline vehicles even for regions with high CO 2 levels from electric generation. Based on base-load generation mix in and typical social cost values of CO 2 emission, a CO 2 credit can be assigned to each PHEV: Based on 30 km/day all electric operation of a PHEV, total CO 2 cut by one PHEV in a populated area in Ontario was found to be 1.5 ton/year. Maximum penetration levels are achieved for credit values larger than CAD. These values do not necessarily represent an actual CO 2 emission credit to be traded in the market. 13

14 G&T Probabilistic Model The uncertainty in some parameters such as electricity prices and Light Duty Vehicle (LDV) growth affect the optimal planning. These uncertainties were studied using: Monte Carlo simulations to evaluate and rank the effect of parameters in the model. Robust optimization to determine optimal plan for a given uncertainty budget (no interval solutions). 14

15 G&T Probabilistic Model Robust optimization: Uncertain MILP model: It can be solved using the robust formulation: 15

16 G&T Probabilistic Model This model can be transformed into a standard MILP problem: 16

17 G&T Probabilistic Model i 2 [0, J i ] represents the budget of uncertainty, with i = 0 corresponding to no protection against uncertainty, and i = n (J i n) yielding a very conservative solution (all uncertain parameters in constraint e taking their worst value). 17

18 PHEVs LDVs consist of passenger vehicles and light trucks with gross weight below 4.5 tons. Based on the Canadian Vehicle Survey: More than 95% of the vehicles on Canadian roads fall into the LDV category. Number of LDVs by vehicle type for Canada (2005 base): Compact sedan: 29% Mid-size sedan: 29% Mid-size SUV: 4% Full-size SUV: 4% Van: 16% Pickup truck: 18% 18

PHEVs Per capita number of vehicles in Ontario (0.55) is assumed to be valid during the planning period of 2008-2025.

19 PHEVs Per capita number of vehicles in Ontario (0.55) is assumed to be valid during the planning period of The total number of light vehicles in each zone of Ontario between 2008 and 2025 was found based on Ontario s zonal population in this period and per capita number of vehicles. The objective is then to find the maximum percentages of these light vehicles that can plug into the power grid for some or all of their energy needs. 19

20 PHEVs Maximum numbers of PHEVs in Ontario s transport sector considering 2 different adoption rates (transitions) and based on ideal penetration by 2025 (100%): 20

21 PHEVs Main assumptions: Average daily drive per LDV: ~50 km All electric daily trip: 30 km (referred to as PHEV30; it covers at least 60% of the average daily drive per light vehicle in Ontario) Maximum allowable depth of discharge: 70% Connection power level (120 V/15 A): 1.4 kw Charging efficiency: 85% 21

22 PHEVs Battery charging requirements for different types of PHEV30 vehicles: Charging times fit in the 8 hours of off-peak time periods in Ontario from 12 am to 7 am. 22

23 PHEVs Hourly power requirements (kw) for different types of PHEV30 vehicles at 1.4 kw connection power level: Average hourly power consumptions for each type of vehicles were considered in the model. 23

Ontario System Model Electricity network model: Based on zonal representation of Ontario s network, a 10-bus simplified model for this network was developed.

24 Ontario System Model Electricity network model: Based on zonal representation of Ontario s network, a 10-bus simplified model for this network was developed. This is mostly a 500 kv network, with a 230 kv interconnection between northern zones. Transmission capacity enhancements were considered based on the existing projects and future developments published by the Ontario Power Authority (OPA). 24

Ontario System Model Electricity generation model: A zonal pattern of generation capacity procurement from 2008 to 2025 contributing to baseload energy in Ontario was developed.

25 Ontario System Model Electricity generation model: A zonal pattern of generation capacity procurement from 2008 to 2025 contributing to baseload energy in Ontario was developed. This is based on the Ontario Integrated Power System Plan (IPSP) and the general information published by the OPA and IESO. The model specifies the total effective generation capacity which is available in each zone during base-load time periods from 2008 to

26 Ontario Results Maximum power requirement of PHEVs in Ontario considering the assumed transition curves and based on ideal penetration by 2025 (100%): 26

27 Ontario Results Assuming different penetrations for the various zones based on population density for Transition 1: 27

28 Ontario Results For Transition 2: 28

29 Ontario Results In spite of somewhat higher penetrations by 2025 based on Transition 1, the total costs for the system during the planning span is the same as for Transition 2. 29

30 Ontario Results Maximum grid and electricity market potential: Max Penetration (PHEV 30km valley filling ) Staggered penetration Max. penetration % Share Penetration [%] Optimal Yearly Penetration (PHEV 30km off-peak charging) HOEP ($/MWh) July

31 Ontario Results Parameter uncertainty ranking (54 parameters): 31

32 Ontario Results Comparison of deterministic (DM) and robust (RM) optimization results (no generator emission constraints): 32

33 Distribution System Model Conductors/cables, transformers, LTCs and switches are modeled using ABCD parameters and are constants except for LTC: A and D matrices for LTC (continuous variable): 33

34 Distribution System Model 34

35 Distribution System Model 35

36 Distribution System Model Equality constraints: Feeder. Transformers. LTCs. Capacitors. Inequality constraints: Bus voltage limits. Feeder current limits. Charging socket limits. 36

37 PHEVs All the case studies have been carried out considering PHEV30 mid-size sedan (8.14 kwh battery capacity). Due to life cycle considerations, SOC at the start of charging is 20% and is charged to 90% of its full capacity. PEV is not available from 7A.M. to 5 P.M. for charging. Level 2 charging: MaxW=4.8 kw( v/40-100a). Level 1 charging: MaxW=1.4 kw(120v/15a). 85% charging efficiency. 37

38 PHEVs 100% penetration implies that every house has one PHEV. Number of PHEVs added at each load bus L are integers: where: x = penetration level in p.u. PD L = kw load at bus L AVHL = Average hourly kw load 38

39 Test System Base system Load profile Total Demand (kw) 23 a b c Time (hour) 39

40 Case Studies Minimizing the total energy drawn by the substation at different penetration levels: Penetration Levels 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Energy (MWh) Ploss (MW) Cost ($)

41 Case Studies (Min. Energy 90% Penetration) kw A System load with & w/o PEV 7 Feeder (Bus ) current Time (hour) Time (hour) a w/o PEV b w/o PEV c w/o PEV a with PEV b with PEV c with PEV a b c Imax kw Demand at Bus 675 with & w/o PEV A Time (hour) Feeder (Bus ) current Time (hour) a w/o PEV b w/o PEV c w/o PEV a with PEV b with PEV c with PEV a b c Imax 41

42 Case Studies (Min. Energy 90% Penetration) Phase voltages for buses with lowest voltage magnitude at peak load: Bus voltage per unit a 646-b 675-c Time (hour) 42

43 Case Studies Minimizing the total cost of charging PHEVs at different penetration levels: Energy (MWh) Ploss (MW) Penetration Levels 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Cost ($)

44 Case Studies (Min. Cost 90% Penetration) System load with & w/o PEV Demand at Bus 675 with & w/o PEV kw Time (hour) a w/o PEV b w/o PEV c w/o PEV a with PEV b with PEV c with PEV kw Time (hour) a w/o PEV b w/o PEV c w/o PEV a with PEV b with PEV c with PEV A Feeder (Bus ) current Time (hour) a b c Imax A Feeder (Bus ) current a b c Imax Time (hour) 44

45 Case Studies (Min. Cost 90% Penetration) Phase voltages for buses with lowest voltage magnitude at peak load: 1.04 Bus voltage per unit a 675-c Time (hour) 45

46 Case Studies Maximize PHEV charging: The quadratic term adds a high penalty value to the objective function at no-integer solutions, controlled by the parameter K. MINLP problem is converted into an NLP. 46

47 A kw Case Studies (Max. PHEV Charging) System load with & w/o PEV Time (hour) Feeder (Bus ) current 15 Time (hour) a w/o PEV b w/o PEV c w/opev a with PEV b with PEV c with PEV a b c Imax kw A Demand at Bus 671 with & w/o PEV Time (hour) Feeder (Bus ) current Time (hour) a=b=c w/o PEV a with PEV b with PEV c with PEV 23 a b c Imax 47

48 Case Studies (Max. PHEV Charging) Phase voltages for buses with lowest voltage magnitude at peak load: Bus voltage per unit a 611-c Time (hour) 48

49 Case Studies Total energy drawn (kwh) No PEV Case 1 (Min Energy) (Min energy) (16.13%) Case 2 (Min losses) For 90% penetration (16.45%) Case 3 (Min cost) (22.49%) Case 4 (Max PEV charging) (60.58%) Total losses (kw) (Min losses) (34.93%) (29.77%) (55.04%) (138.46%) Total cost for charging PEVs ($/day) (1.36%) (1.96%) (0%) 2, (243.44%) 49

50 Conclusions It will take anywhere from 3 to 5 years for PEVs to begin to assume any noteworthy share of the market and longer for a critical mass to emerge. For the timeframe, charging needs can be managed with existing options without significant disruption. Beyond 2015, the planning process, informed by emerging data on consumer acceptance, would be expected to address future needs. Detailed assessments of market potential will be required and coordination of activities amongst planning agencies, utilities and auto manufacturers will be necessary to ensure the requisite infrastructure is in place when needed. 50

51 Conclusions Charging levels: Level 1 overnight charging requires no significant installation and operation challenges and costs. Level 2 upgrades to home garage charger could be provided either at a small cost (or as an incentive) to the first adopters. Level 3 fast charging capability will be necessary for those customers who opt for it, but as a premium service option. 51

52 Conclusions Workplace charge stations: Necessary to develop consumer acceptance of PEVs. The cost of installation and electricity use can be recovered in a number of ways such as payroll deductions or within the employee s benefits package. Public charge stations: Need to be installed in high traffic zones providing all three options for charging but at different prices. The public installations could be led by either a utilitymunicipality partnership or private sector entity investment. 52

53 Conclusions Region-specific or neighborhood specific maps of vehicle purchases and demand for charging stations need to be developed. The utility should install override controls (with customer agreement) and entice the customer to charge at times when it is best from a utility operations perspective (e.g. Peak Saver DR program). 53

54 Conclusions Charging of PEVs during on-peak hours will have a significant effect on the grid that will have to be planned for, especially in highly populated areas such as the GTA due to clustering. Significant PEV penetration levels: Will definitely impact the grid and its associated electricity market. Even if all charging takes place at night, there will be upward pressure on electricity prices. Prices should be conveyed in a smart manner to make optimal charging decisions. To mitigate and manage the impact that high penetration and concentration of PEVs, smart charging strategies and technologies will have to be developed and deployed in the context of smart grid technologies. 54

55 Conclusions Vehicle-to-Grid (V2G) as well as Vehicle-to-House (V2H) technologies: Present many potential advantages to the grid such as regulation and energy storage for wind and solar power. These technologies will not be economically feasible until several of the issues with batteries are resolved. Since technologies will likely improve in the long-term, R&D is needed now to be ready for deployment. Standards for PEV charging devices: Installations and communications are currently under development by a variety of institutions (e.g. SAE, UL, CSA). These communication standards will be very much dependent on the standards finally adopted for smart grid applications (e.g. ZigBee and WiFi). 55

56 Conclusions A champion agency should be identified and empowered to ensure the policy goals can be attained. This lead agency would work with all stakeholders to develop a clear set of regional plans for implementing electric mobility initiatives. To promote sustainable mobility, the planning efforts must also address social science issues such as urban land use and others. 56

57 Further Reading A. Hajimiragha, C. A. Cañizares, M. W. Fowler, S. Moazeni, and A. Elkamel, A Robust Optimization Approach for Planning the Transition to Plug-in Hybrid Electric Vehicles, IEEE Transactions on Power Systems, vol. 26, no. 4, November 2011, pp A. Hajimiragha, C. A. Cañizares, M. Fowler, and A. Elkamel, Optimal Transition to Plug-in Hybrid Vehicles in Ontario-Canada Considering the Electricity Grid Limitations, IEEE Transactions on Industrial Electronics, special issue on Plug-in Hybrid Electric Vehicles, vol. 57, no. 2, February 2010, pp C. A. Cañizares, J. Nathwani, K. Bhattacharya, M. Fowler, M. Kazerani, R. Fraser, I. Rowlands, and H. Gabbar, Towards an Ontario Action Plan For Plug-In-Electric Vehicles (PEVs), Waterloo Institute for Sustainable Energy (WISE), University of Waterloo, Report, May 2010, 165 pages. Available at: 57

R e p l a c i n g D i e s e l G e n e r a t i o n w i t h R e n e w a b l e S o u r c e s i n N u n a v u t C o m m u n i t i e s : F e a s i b i l i

R e p l a c i n g D i e s e l G e n e r a t i o n w i t h R e n e w a b l e S o u r c e s i n N u n a v u t C o m m u n i t i e s : F e a s i b i l i t y S t u d i e s Indrajit Das Claudio Canizares Dept.