Combined provision of primary frequency regulation from Vehicle-to-Grid (V2G) capable electric vehicles and community-scale heat pump

Highlights

• Different demand side units deliver a combined frequency regulation service.

• Fast-regulating EVs with small storage capacity enable primary frequency regulation.

• Heat pump with heat storage enables EVs to bid larger capacity into reserve market.

• Frequency response from electric vehicles is experimentally validated.

• Economic synergy effect may be obtained by combined service provision.

Abstract

A combined delivery of frequency containment reserve (FCR) from a fleet of electric vehicles (EVs) and a large-scale heat pump is proposed to exploit synergies between the fast regulating batteries and the large heat storage capacity of the thermal system. The feasibility of the proposed strategy is assessed with a state of charge model for the EVs delivering FCR, based on one year of system frequency measurements including the conversion losses in the charger and a dynamic model of the heat pump system. Both models were previously validated against experimental data. The proposed operation of the EVs was experimentally validated as part of this study. The heat pump offsets the energy content of the frequency deviations for 25 EVs and thereby enables them to bid a larger power capacity on the FCR market without violating energy constraints of the battery capacity. For the case of 2018 an additional income from capacity and power price payments was obtained that could not be generated by the EVs or the heat pump alone. This synergy effect was shown to be highly dependent on the achievable capacity payments, which were relatively high for the 2018 case.

Introduction

The future energy system will have to accommodate a higher share of distributed energy resources (DER) than today. At the same time, synergies can be found among the different sectors to secure the integration of DER and achieve a coordinated operation. An integrated energy system consisting of electricity, heat and transportation is increasingly recognised as the research paradigm to address these challenges [1], [2]. It involves increasing interactions across energy infrastructures and optimising this integration to provide services at supply and demand level. Due to the increased share of intermittent DER that needs to be accommodated in the energy system, and the replacement of conventional thermal power plants, ancillary services have to be delivered by new actors in the system. One option is to use demand side management to supply frequency containment reserve (FCR) to the grid. Demand side units often provide a link to a neighbouring energy sector with certain requirements and constraints [3]. This requires a well coordinated operation of these demand side units in order to be able to supply both the ancillary service and the primary service of the unit. In order to characterise the capability of different systems to provide operational flexibility, the following parameters can be used, as proposed by Makarov et al. [4] and modified by Ulbig & Andersson [5]:

• The power provision capacity, i.e. the load by which the current power uptake or supply of the unit may be decreased or increased

• The power ramp rate, which defines how fast a unit is able to change the power uptake or supply

• The energy provision capacity, which defines for how long the change in power uptake or supply can be maintained and thereby is a measure of the storage capacity of the system.

They show that combining different types of units results in combination of the strengths of the individual units. In the current study, it is proposed to deliver FCR from a combination of different demand side units. This approach is demonstrated for a combination of electric vehicles (EVs), which are able to regulate their power within seconds and alternate between production and consumption, and large-scale heat pumps, which can access a large storage capacity. It is expected that in this way it may be possible to create a combined system, which can regulate fast and has a large storage capacity.

The transportation sector accounts for 25% of the globalenergy-related CO${}_2$ emissions of which light-weight passenger vehicles account for over half [6]. Electrifying the private transportation sector allows to use renewable electricity as fuel and thereby it is expected to considerably reduce the transport related CO${}_2$ emissions. Vehicle-to-Grid (V2G) research may serve to limit the self-induced adverse effects of EVs in terms of additional grid loading, but also to make the EV an active asset in supporting a stable, economic power system based on renewable energy. By controlling the charging power, it is possible to minimise the energy costs through adaptive or predictive charging [7]. The use of EVs for providing ancillary services to the power system can be an additional revenue for EV owners and can assist the integration of larger amounts of renewable sources [8], [9]. A high speed response of distributed battery-based energy resources can cover the regulation requirements of significantly larger thermal power plants with slower ramp rates, which are decreasing in numbers, as the generation is transitioning towards renewable production [10]. EVs that charge using power electronics can regulate their power uptake quickly during the grid connected hours. However, an EV battery has a limited energy provision capacity and thus availability for frequency regulation delivery. Thingvad et al. [11] showed that it is necessary to reserve part of the maximum charging power capacity to be able to re-balance the state of charge (SOC) of the battery when delivering frequency regulation. This is necessary, as the grid frequency is often too high or too low during consecutive hours, leading to continuous charging or discharging of the battery as response to the frequency deviation. This may be avoided by bidding only a part of the capacity for frequency regulation and using the remainder for maintaining the SOC within a bounded range This would however result in a reduction of the possible revenues from capacity payments.

The heating and cooling sector is responsible for ca. $\text{50\%}$ of the final energy consumption in Europe, of which the largest share is still supplied by fossil fuels [12]. Different strategies have been proposed on how the European target of 80% reduction in annual greenhouse gas emissions in 2050 compared to 1990 levels can be reached [13], [14]. According to these, large-scale and domestic heat pumps are expected to play a key role in the future heat supply, as they enable the exploitation of low-temperature ambient or excess heat sources and provide a link to the electricity sector. Thereby, they unlock the thermal storage potential for providing flexibility, e.g. as ancillary services to the power sector. Large-scale heat pumps can access a large thermal storage potential that can be found in district heating networks [15], thermal storage and the buildings’ thermal mass [16]. This may be used to provide flexibility to the electricity sector [17] by shifting the time of operation of the heat pump or by adapting the heat pump load. However, large-scale heat pumps are typically optimised to reach maximum energy efficiency and operate continuously in base-load, not to be able to change their load quickly. In order to secure safe operation, most large-scale heat pumps will take minutes to start up or change load. In Denmark, large-scale heat pumps typically use ammonia as refrigerant [18], as it provides a relatively high coefficient of performance (COP) while having no ozone depletion potential and no global warming potential [19]. In order to deliver district heating at forward temperatures of $60$ °C to $90$ °C, ammonia heat pumps are typically built as two-stage heat pumps with flooded evaporators. A previous study showed that the achievable ramping rates in this kind of system is limited by waiting times of the compressors which ensure a stable operation of the heat pump, and by the risk of condensation in the suction line during ramp-down [20]. Without further adaption of the system design and the control strategy, this kind of heat pump is expected to regulate too slowly to provide FCR directly, i.e. the regulation time is larger than $150$ s. However, as they are coupled to a large storage capacity on the thermal side they may still have a large potential to provide energy flexibility.

The aim of this study was to assess the feasibility of combined supply of FCR using two different technologies, each with their individual characteristics with respect to the system demand. This approach was applied to a specific case of a fast reacting pool of EVs with low energy capacity, and a slower large-scale heat pump integrated with the district heating network and a thermal storage tank, which provides a large storage capacity. The service supplied by the combined system was the Danish frequency containment reserve — normal operation (FCR-N), which is a frequency controlled service that requires ramping times below $150$ s and is described in more detail in Section 2. The EV pool consist of $25$ vehicles that are charged in an orchestrated way to allow for FCR-N delivery. The batteries of the V2G EVs should react quickly to changes in the grid frequency, while the heat pump should outbalance the energy bias of the frequency deviations, $e_n^{\text{bias}}$, to avoid that the EVs become fully charged or depleted. Thereby, the combination with the large-scale heat pump is expected to allow to bid the full EV power capacity for FCR-N, despite the heat pump not being fast enough to deliver FCR-N by itself. The aim was to identify whether a combined operation of these units may lead to increased provision of FCR-N, what constrains the combined operation, and whether a combined operation is economically feasible. In order to answer these questions, a model of the charging of $25$ EVs and a model of a large-scale heat pump system were used to assess the feasibility of providing a combined service for one year of operation using $2018$ data. The work was conducted in five steps:

• Analysis of the consumption of a district heating (DH) network in Copenhagen to calculate the number of hours that the heat pump is available to deliver offset electricity to the EV batteries, i.e. to balance out the energy bias of the frequency deviation.

• One year of grid frequency measurements and a stochastic driving consumption was used to calculate the SOC of 25 EVs. A SOC control strategy was implemented to change the base line power on an hourly basis according to SOC constraints. The control strategy was used to generate the offset schedule for the heat pump.

• The yearly heat pump operating cost was calculated using a thermodynamic system model and electricity spot market prices. It was compared to the cost of operating the heat pump according to the optimised schedule used by the operators.

• The income from supplying FCR-N service was calculated for delivering FCR-N from the EVs alone and combined with the heat pump. Additional operation cost and battery degradation were considered.

• The control method for maintaining the SOC was experimentally validated by delivering FCR-N with a single EV during 15 h.

A description of the market framework for FCR-N, an overview of the different models used and the simulation approach is given in Section 2, together with a description of the assessed EV and heat pump system, model descriptions and experimental set-up. The results for the available offset power from the heat pump, the experimental test of following a given frequency pattern with an EV, and the economic analysis for the combined delivery of FCR-N are presented in Section 3, the discussion may be found in Section 4 and the conclusion in Section 5.