Building-to-grid flexibility: Modelling and assessment metrics for residential demand response from heat pump aggregations

Highlights

• Novel metrics are designed to quantify building-to-grid demand response flexibility.

• Power and energy paybacks following load curtailment can vary from 0% to 50%

• Dwellings with high thermal inertia show a smaller but longer payback.

• Hybrid heating system reduces power and energy paybacks to less than 20%

Abstract

Increased flexibility has been identified as a key requirement in future power systems. Much flexibility could be provided by energy vectors other than electricity. In particular heat may be a valuable source of flexibility, as electrification of space and water heating introduces highly flexible resources such as electric heat pumps. However, current methods for assessing aggregated demand side flexibility, particularly from residential buildings, may not be adequate given the variety of different grid services that flexibility may be used to provide to different stakeholders, and considering relevant comfort constraints. On these bases, in this work several metrics, relevant to different stakeholders, are introduced to quantify building-to-grid demand response flexibility from heat pump aggregations. Specific control algorithms for the aggregations are also proposed and tested through a multi-energy residential energy consumption tool. A number of case studies are carried out to demonstrate the value of the proposed metrics and algorithms, especially in relation to flexibility exploitation with long sustain times (e.g., reserve services), which can noticeably affect user comfort. Our results indicate that the payback behaviour of heating units following a demand response event can vary substantially with different types of dwellings. More specifically, the power payback is negligible in dwellings with high thermal inertia, while the power and energy payback can reach 10% and 50%, respectively, in dwellings with low thermal inertia. The benefits from hybrid (electric + gas) heating, which can reduce energy payback and comfort loss, are also demonstrated. For instance, in a cluster of dwellings with low thermal inertia, the energy payback following a DR event is reduced from 50% to 20% and the maximum comfort loss of the participants is decreased from 1.6 °C to 0.5 °C.

Introduction

Due to the threat of global warming and climate change, many jurisdictions have set ambitious targets for energy conversion from renewable technologies. However, generation variability and uncertainty from several renewable technologies mean that more flexibility¹ will be required to integrate these technologies successfully [1]. In a traditional power system, flexibility is contributed by flexible generators and large-scale storage (e.g., pumped hydro). Currently, with communication technologies advancing, significant amounts of flexibility may now be available from the demand-side, and in particular the residential sector, particularly given electrification of heating [2].

Multiple actors could benefit from this demand side flexibility. For example, system operators could be interested in the potential of residential resources (particularly electro-thermal thermostatically controlled loads) to provide various frequency control services, of which more will be required in the near future [3]. System services of interest include fast frequency response [4]/regulation services [5] and slower reserve products [6]. Besides services to the system operator, network operators may also need flexibility to address network congestion [7], or may consider flexibility from the demand side to minimise the cost of network expansion [8]. Retailers can also benefit from using residential flexibility to bid on energy markets to maximise their profits [9].

Understanding the potential of DR from the residential sector and particularly buildings to provide “building-to-grid” flexibility required by various power system actors requires two steps, namely, design of appropriate metrics and quantification of those metrics. Metrics may simply relate to the power that can be shifted/curtailed/increased, such as discussed in [10], [11], [12], [13]. In these examples the flexibility of specific types of load can be assessed using software with bottom-up methods. For instance, EnergyPlus is used to simulate the energy consumption of commercial buildings and apartments in [10]. In that work, sensitivity studies are carried out with different building parameters and target temperature ranges to demonstrate the numerical distribution of the potential flexibility of single building blocks for DR applications. On the other hand, there are many general metrics for flexibility quantification. Examples of other appropriate metrics are the appliance flexibility index and acceptable delay time metrics, as shown in [11], [12], respectively. These two metrics measure flexibility as the acceptable time shifting of appliances’ operation. Moreover, since the random behaviour of occupants may affect the individual appliance’s flexibility, it is also important to quantify the flexibility of loads at the aggregate level. Hence, in [13] a metric called flexibility index of aggregate load is introduced, which is used to indicate the probability of demand increase or decrease of a group of loads. A further metric denoted as percentage flexibility level, which determines the amount of flexible demand for DR applications, is also given in [13]. Recognising that provision of flexibility products with long sustain times (e.g., few hours) can have an adverse effect on user comfort, the comfort level satisfaction of occupants is also used as the metric to correspond to the DR capacity as shown in [14]. Following another approach, the comfort level is considered intact if the temperature indicator is within the pre-defined bounds [15].

Other publications have focused on the utilisation of flexibility for practical applications, such as renewable integration, network congestion and capacity support. In [16], the flexibility of generic electric heating units is used to increase the renewable penetration level at a microgrid scale, while a sensitivity study is performed with different occupancy profiles and weather patterns. The effectiveness of domestic DR based on real-time price signals has also been investigated, as shown in [17]. The flexibility of electric heating can also actively participate in ancillary service markets by means of a novel thermostat technology which takes into consideration the uncertainties from both demand and generation sides [18]. Case studies in the field can also provide valuable information. For example, a field trial undertaken in Belgium [19] found that potential upward and downward flexibilities per household are 65 W and 430 W respectively. Extrapolating this to the national level implies that domestic flexibility could equal 1.8% (upward) and 12.1% (downward) of installed generation capacity in the country. In a New York State study [20], it has been found that the system capability of integrating wind generation can increase to 5 GW, which is equivalent to a 33% wind penetration level in the grid of New York State, if the EHP penetration level reaches 20%. In the United Kingdom (UK) case [21], it has been found that system peak demand can be reduced by 7 GW (equivalent to 9% of the peak) by 2050 with the utilisation of the flexibility from domestic appliances and EVs. Despite the breadth of current research, a common gap in the above literature [10], [11], [12], [13], [14], [15], [16], [17], [18] is the neglect of the impact of DR in the period after service provision. Where this impact is considered, such as in the high-level case studies [19], [20], [21], there is no quantitative analysis on the payback effect of DR. In addition, continuous operation of DR can lead to the saturation of flexibility and the synchronisation of appliances, so that the service provision can eventually become unsustainable. Further, the recovery of flexibility in the post-DR period can lead to power spikes and additional energy consumption due to the synchronisation of appliances state, which can subsequently challenge system security. This behaviour is of significant interest to network and system operators. Hence, a comprehensive set of metrics for assessment of residential DR is clearly missing.

In order to cover the gaps mentioned above, this paper presents a unified framework to assess the flexibility of current and future residential heating technologies and their performance in providing building-to-grid demand response. Specific contributions include:

• Several general flexibility metrics which are specifically designed to assess the benefits and impact of building-to-grid DR activities to different stakeholders in both electricity and gas sectors. In addition, the potential comfort loss of occupants due to DR is specifically investigated.

• Two control algorithms of Electric Heat Pump (EHP) operation are introduced which are used to determine the on/off states of aggregated EHPs in the provision of ancillary services.

• Simulation and assessment of DR from residential EHP aggregation for services with long sustain times is carried out.

• The performance of EHP-only and hybrid (electricity-gas) heating systems is compared in different DR applications.

The rest of the paper is organised as follows: In Section 2, specific metrics are presented to assess EHP and hybrid heating systems’ performance on grid services applied to different stakeholders, alongside the implemented DR control mechanisms. In Section 3, an overview of a high-resolution (one-minute) residential energy consumption model is presented, which is used to simulate the operation of EHPs and the comfort level of occupants. In Section 4, a grid reserve service is selected and various case studies are carried out to investigate the effectiveness and impact of DR on technical and economic perspectives. Section 5 concludes the findings of this paper and future works.