Liberalization of Electric Industries and Design of Electricity Markets
At the beginning of the second half of the 20th century, following the expansion and the interconnection of electricity networks at the national level, state-owned regulated monopolies were in charge of the entire value chain of the electricity industry in most developed countries. This value chain is split in four different parts: generation of electricity, transmission (which includes management of High Voltage networks and System Operations), distribution (Medium and Low voltage networks) and retail of electricity to final consumers. This type of architecture was the results of different factors:
• The need for large public investments at the end of Second World War to electrify countries to allow economic growth;
• The nature of electricity, which cannot be stored and requires high level of coordination between the different part of the value chain, in both short term and long term;
• Electricity sector was thought to be a natural monopoly, meaning that it would be more costly to have multi-firm production than to have a monopoly, due to high capital costs.
However, in the end of the 20th century, this architecture began to be criticized (Joskow and Schmalensee, 1988). Arrival of new and more efficient generation technologies such as CCGT questioned the fact that generation was a natural monopoly. Competition at the generation and retail level was believed to foster innovation and efficiency and would allow reducing retail prices for final consumers. There was a consensus that liberalizing this sector would benefit to society in the long-term (Joskow, 2008). Distribution and transmission levels, due to their high investment costs, have always been considered as natural monopolies. However, there was also a need to give them better incentives to be more efficient.
Opening of competition on the generation and retail businesses also required to unbundle legal monopolies by separating transmission and distribution structures from generation and retail, to avoid conflict of interests and allow new competitors to enter the market. Moreover, an independent body should be created, which will be in charge of creating the rules of new markets and supervise competition. This process was first put in place in UK in the 80s, followed by some US States and European Union.
However, there was still a need for coordination between the different elements of the value chain, as the physics of electricity remained the same. The process to allow this coordination was internalized before reforms in the vertical-integrated companies. With liberalization, this coordination is done through markets, by confronting supply of electricity and demand. These markets must be organized, meaning that rules should be set in order to allow coordination, from short-term to long-term, in an efficient way, while ensuring security of supply.
It is possible to distinguish different sequences in electricity markets. Electricity being a flow, it is impossible to store it in the form of electricity, (it is possible to store it in another form of energy – chemical, electrostatic, mechanical, thermal…). Every Watt produced at one point of the system should be consumed at another point. The main objective of electricity markets is to ensure a balance between supply and demand, at the most effective cost. Figure 1.14 gives an overview of these different sequences.
Market actors are called Balance Responsible Parties (BRP) in power markets. A BRP can have generation assets and/or retailer, but it can also be a pure energy trader, without any generation or consumers. A Balance Responsible Party has the legal duty to have a balanced perimeter, meaning that every generated or bought energy has to be either consumed by its clients or sold in the market. The different markets where BRPs can trade energy are in blue in Figure 1.14. There are three different type of markets: future energy markets, where BRPs can trade energy for the coming years, months or weeks; Day Ahead Market (DAM) where BRPs can trade energy for the next day; and intraday market, where BRPs can trade energy for the coming hours. On these markets, the price of energy for a certain delivery period will be determine by confrontation of offer and demand: BRPs wanting to sell energy (e.g. producers) will make offer bids, based on their costs, while BRPs wanting to buy energy (e.g. retailer) will make demand bids based on their willingness to pay. The price and the amount of energy traded for this delivery period is the point where demand curve (demand bids sorted in decreasing order) meets offer curve (offer bids sorted in increasing order). BRP will trade on these different markets based on the information they have on forecast of generation and consumption. The closer from real-time, the better will be this information. However, it will also be more costly to balance near real-time, as resources might be scarce.
In parallel, Transmission System Operator (TSO), in charge of balancing the system, procures reserve. Reserve is provided by flexible actors (Balancing Services Provider – BSP) able to adjust their consumption or generation output to rebalance the system in case of imbalance. It means these actors will not produce or consume at the maximum or the minimum output, in order to be able to change it when required. There are different types of reserve, depending on the reactivity of the resources to an imbalance and the duration of the deviation from its original output. BSPs can be paid based on the capacity (in MW) they offer and on the activated reserve (in MWh).
Challenges and Opportunities of Massive Diffusion of EVs for Electricity Systems
In addition to the massive penetration of renewable energy generation, a massive uptake of EVs will represent another challenge for System Operators. This new electricity consumption could in the coming years impact the long-term adequacy between offer and demand, for both energy and power and both in term of generation asset and network (distribution and transmission). It is possible to calculate some order of magnitude of these impacts. For example in France, the annual mileage of a personal car is about 17,000 km. With a consumption of 0.20 kWh/km, the annual consumption of 1 million EVs would be about 3.4 TWh, which represent an increase of the 2017 electricity consumption (475 TWh) by 0.7 %. With respect to the capacity increase, if every vehicle charge at the same time with a power plug of 3 kW, the instantaneous consumption of these vehicles would be 3 GW, which represent an increase of 3 % compared to the historical peak consumption in France and the generation output of about two nuclear units. Moreover, as people typically plug-in when they come back from home, the consumption from Electric Vehicles would be synchronized with peak consumption and could aggravate the so-called “duck-curve”.
These order of magnitude shows that EVs will represent a major challenge in term of capacity, which is the major determinant of electricity networks. Indeed, System Operator will size the network in order to fit the maximum peak consumption. In case of a major diffusion of Electric Vehicles and if these vehicles charge as soon as plugged (“dumb” charging), System Operators will have to reinforce the entire network, which would imply massive investments. It would be particularly true for distribution networks, in places where EVs would diffuse rapidly, such as cities. However, it is possible to pilot charging pattern of the vehicle in order to reduce stress to the network and provide flexibility services to System Operators.
Personal vehicles have on average very low utilization rates. According to a survey conducted by the French government in 2008 (Ministère de la Transition Ecologique et Solidaire, 2008), 42 % of vehicles are circulating less than two hours a week, which represents a utilization rate of 1.1% and 89 % of the vehicles are circulating less than 8 hours a week (utilization rate of 4.7%). Moreover, 75% of vehicles have a daily mileage of less than 42 km. For EVs, it means that, potentially, the car is plugged to the network 95 % of the time and need a daily recharge of less than 8 kWh (if we consider an average consumption of 0.2 kWh/km). Considering that the typical EVSE at home is about 3 kW, it means that it would take only 3 hours to charge completely the car for daily trips. We have seen in the previous paragraph that System Operators might face an issue of increasing peak consumption in case of “dumb charging”. Personal cars are often at home during the entire night, the vehicle might be plugged to the network for more than 8 hours. It might not be necessary, in most cases, to charge the vehicle as soon as plugged. It would be preferable to delay charging hour to charge the vehicle during period of low consumption. This process is called load-shifting as consumption is delayed from peak to off-peak period It is required to have a price signal to allow this delay of charging pattern. Different types of price signals have been studied in the literature (Gyamfi et al., 2013). From the simplest to the more complex price structure:
Time-of-Use (TOU) price: electricity is charged at predefined different prices within predefined time periods.
Critical Peak Pricing (CPP): electricity is charged at a very high price for a certain number of days or hours during the year with extreme peak situations.
Real-Time-Pricing (RTP): electricity is charged at a price reflecting real-time evolutions of electricity markets.
Which Flexibility Services for Fleet of EVs?
First and priority service provide by an Electric Vehicle is mobility. Flexibility provision should not endanger the delivery of this service to the customer. We can distinguish two time-scales in the provision of mobility services:
Short time scale: flexibility provision should not endanger capability of the user to do its next trip at the end of the charging session, meaning that battery should have enough energy to run for the desired mileage.
Long time scale: battery capacity is affected by the number of cycles that will occur during its lifetime. Car manufacturer guarantee the battery for a certain mileage of the vehicle and for a certain capacity fading. Provision of flexibility should not reduce extensively battery lifetime. (Wang et al., 2016).
It is necessary to manage smartly charging pattern of the vehicles to reduce these two constraints but also to find the appropriate services, which allow maximizing value while fitting with the characteristics of EVs. We will give an overview of main flexibility services in Europe and the associated technical requirement. Then we will analyze if these services fit with EVs characteristics (with unidirectional or bidirectional capabilities), and give order of magnitude on possible remuneration. These different services are not accessible by a single user. It will be necessary to pool a fleet of EVs to make offers on the markets. The party that will manage charging patterns of the vehicles and make offers on the markets is called aggregator.
The first way to use flexibility of EVs charging pattern is to do energy arbitrage: with unidirectional capabilities, the principle, explained in Paragraph 3.1 is to charge the battery at the least costs given volatility of electricity price. With bidirectional capabilities, it is also possible to feed electricity into the grid when electricity prices are high. Figure 1.22 shows average prices in Day Ahead Market for year 2017 in six different countries. The trend of price evolution during the day is the same in every countries: peak price around 8:00 and 18:00 and low prices during night. However, the range of variation is not the same in every countries and shows high seasonality, as can be seen if Figure 1.23.
Other Sources of Value
It exists other types of sources of value for unidirectional and bidirectional vehicles:
Self-consumption if user possess a generation asset. The value of self-consumption will depend on the spread between 1) the retail price of electricity and 2) the Feed-in-Tariff for renewable generation The higher the spread, the higher the value of maximizing selfconsumption.
Value on capacity markets, created in some European countries. EVs have to curtail consumption or reinject electricity during a certain number of peak hours. It reflects the value of investing in new generation assets to cope with peak consumption hours.
Value for distribution grids. By managing charging patterns of EVs, it is possible to avoid local congestion on distribution grids. Electricity Distribution Companies could avoid new investments on transformers to cope with peak consumption situations. EVs could also provide voltage regulation.
A Literature Review on Smart Charging of Electric Vehicles
We have seen in the previous paragraph that charging patterns of Electric Vehicles can be managed by aggregators to offer flexibility services and reduce charging costs. However, this process should take into consideration mobility needs of the user to ensure that he will have sufficient amount of energy for his next trip, through algorithms. Extensive reviews of the literature was performed in (García-Villalobos et al., 2014) and (Hu et al., 2016). According to these two reviews, algorithms can be categorized according to four main parameters: Characteristics of vehicles studied, control strategy of the operator, objective of the algorithm and optimization method.
Table of contents :
TABLE OF CONTENTS
LIST OF FIGURES
LIST OF TABLES
INTRODUCTION: ELECTRIC VEHICLES AT THE CONVERGENCE OF TWO
INDUSTRIES IN MUTATION
1 Mutation of the Automotive Industry: The Emergence of Electric Vehicles
1.1 Environmental Impacts of the Transport Industry
1.2 Involvement of Policy Makers in the Development and Diffusion of Electric Vehicles
2 Mutation of the Electricity Industry: Liberalization and Decarbonization
2.1 Liberalization of Electric Industries and Design of Electricity Markets
2.2 Towards Massive Penetration of Renewables and Distributed Energy Resources .
2.3 The Challenge of Increasing Flexibility Requirements
3 Using Electric Vehicles as Distributed Flexibility Assets
3.1 Challenges and Opportunities of Massive Diffusion of EVs for Electricity Systems
3.2 Which Flexibility Services for Fleet of EVs?
3.3 A Literature Review on Smart Charging of Electric Vehicles
3.4 Toward the Elaboration of Business Models for Aggregator
4 Thesis Organization
IMPACT OF MARKET RULES ON PROVISION OF FLEXIBILITY BY
DISTRIBUTED ENERGY RESOURCES: A QUALITATIVE ANALYSIS
1 Barriers to Entry for New Entrants in Flexibility Markets
2 Modular Analysis of Barriers to Entry
2.1 Module A: Administrative Rules Regarding Aggregation of Distributed Energy Resources
2.2 Module B: Definition of Products
2.3 Module C: Remuneration Scheme
2.4 A Tool for Investors and Policy Makers
3 Costs Associated with the Opening of Markets
4 Two Case Studies: Geographic Comparison and Evolution of Regulatory Framework
4.1 Comparison of Four Market Zones in 2016
4.2 Evolution of Regulation in France: Towards the Creation of a Single Market Zone in
5 Partial Conclusion
REVENUES AND PROFITABILITY ANALYSIS OF A FLEET OF ELECTRIC
VEHICLES PROVIDING FLEXIBILITY SERVICES
1 Simulation of Fleets Participating to Frequency-Containment-Reserve
1.1 Description of the Model
1.2 Validation of the Model
2 Revenue Analysis of a Fleet of Bidirectional EV Chargers Providing Frequency Containment Reserve
2.1 Market-Designs and Rated Power Scenarios
3 Net-Present-Value Analysis of an Investment in Bidirectional EV Chargers
3.1 Model and Base-Case Scenario
3.2 Sensitivity Analysis
4 Partial Conclusion
EVALUATION OF THE VALUE OF COOPERATION BETWEEN AGGREGATOR
AND CAR MANUFACTURER
1 Roles of the Aggregator And Value Chain Of Smart Charging
2 Presentation of the Model: Actors and Case Studies
2.2 Calculation of Net Present Value of Car Manufacturer and Aggregator
2.3 Presentation of Case-Studies
3.1 Smoothing of Revenues Function
3.2 Case-Study 1
3.3 Case-Study 2
3.4 Case-Study 2bis: Introduction of bargaining power of the Manufacturer
3.5 Case-Study 3
4 Analytic Model
4.1 Reference 1
4.2 Reference 2
4.3 Case-Study 1
4.4 Case-study 2
4.5 Case-Study 3
4.6 Validation of Analytic Model
4.7 Sensitivity Analysis
5 Partial Conclusion
CONCLUSION AND RECOMMENDATIONS