1. Introduction
In general, the conventional electric grid is a centralized system that connects the power output from many fossil and non-fossil fuel-based power plants to one and transmits the power steadily to electricity users. The most commonly seen electricity users are residential ones. The electric grid (EG) transmits power from remote areas (where the power plants are located) over a long distance to the residential electricity demand centers. This allows power flow only in one direction from high voltage to a low voltage level, and then it is distributed to the electricity consumers, who are typically called residential electricity users (REUs), but in today’s trend, the role of renewable energy (RE) has become very crucial in EG. RE, in most cases, is localized, and the power generation capacities vary from small to megawatt-scale. In most cases, the generated voltages from the RE-based power plants are small relative to those of conventional power plants. Even though RE-based power plants are feasible for local power generation, their integration into a large and highly centralized grid poses a limitation to their use and overall, it is understood that this is challenging [
1]. This is due to the frequent fluctuation and, as a result, frequent loss of generation occasioned by the intermittent nature and variability of RE. Besides, it is further compounded by the reverse power flow that may occur in a conventional EG that was originally designed to allow power flow only in one direction, but later, with the advancements seen in distributed energy resources (DERs) and grid integration regulations, the deployment of RE into EG has become a feasible option [
1,
2]. Today, the trend for the use of DERs, particularly RE for electric power generation both at the off-site and on-site load centers, has increased exponentially across the globe. The growth in RE use is due to the need for decarbonization in the power sector. Besides, environmental friendliness and energy security, as well as consistent improvement in technology and falling cost of RE are also favoring this growth [
3].
Now with the support of DERs and other improvements seen in energy-related technologies like power electronics and power system control and operations, the conventional EG has taken a paradigm shift towards modernization. Even after modernization, the EG continues to experience uncertainties that directly or indirectly affect the energy consumption patterns of REUs [
4]. The uncertainty in supplying power to REUs is more often seen and is mainly due to power outages. The reasons for power outages might be any of the following: extreme weather events, asset failure, natural disasters, power surges, acute accidents, and even the operational errors by the workforce [
5]. Such uncertain situations are permitting us to think of this as a resilience problem. In most cases, the power outages may last from a few minutes to weeks, depending on the nature of the resilience issue and the power supply system (PSS) configuration. Thus, the power outage situations result in the complete unavailability of the power to REUs. From the resilient PSS point of view during the power outages, power availability has to be ensured, and it can be done in many ways.
Indeed, there are numerous ways to understand resilience, and these would depend on the nature of the system. Recent studies, for this reason, have called for engineering greater resilience especially in the power and other industrial systems. A few studies have employed network analysis to understand the resilience and sustainability of industrial symbiosis system that facilitates energy, water and material flows [
6,
7]. The network analysis is further extended and applied in few critical infrastructures in the United States of America (USA) and United Kingdom (UK) (for e.g.,: energy resources and power sector, information technology and communication, finance, healthcare and public health, transportation, and food and agriculture) to understand the implications of interconnectedness and interdependencies on resilience [
8,
9]. But when it comes to power sector, metrics-based approaches are used [
10]. More often the mitigation strategies for enhancing the resilience of power systems are based on short term and long-term measures adopted [
11].
However, in the literature, many authors have varying opinions on the way to ensure resilience, and it is often considered as an issue of reliability. Few studies exist in the literature where researchers have discussed the difference between reliability and resilience of the power supply system (PSS). Reliability more often deals with component failure, but the resilience of PSS is different, and it often deals with the capability of a PSS to sustain and bounce back to normal operation after any unseen or unexpected uncertainty [
12]. This means the resilience assessment admits the possibility of PSS failure and focuses on its recovery and adaptation, thereby ensures continuous power supply to the REUs. The U.S. National Academy of Sciences defines resilience as the ability to plan for, recover from, and adapt to adverse events over time [
13], but the most recent definition of resilience from the literature is the ability of a system to sustain, to rapidly recover, and learn to adapt its structure to unexpected disruptive events [
14]. Based on the above definitions of resilience, it is clear that, for any sought of disruption, the PSS should be able to recover soon and provide the power to REUs, but ensuring resilience is quite difficult with a conventional PSS. Even with modern electricity infrastructure, the disruptive events are still occurring, and resilience is a big question. For instance, irrespective of whether a nation is developed or developing, the power outage incidents are happening across the globe. From 1960 to 2019, across the globe, thousands of power outage events were recorded, and among these, a few are very massive events that lasted for more than a month. In
Figure 1, a heat map showing the locations where power outages events have occurred across the globe is indicated [
15].
From the heat map shown in
Figure 1, it is understood that the USA.’s EG experiences these power outages most often. A recent study, on significant power outages across the USA, also reported the data on power outages events and the affected population [
16]. A list of power outage events due to numerous resilience issues across the USA are shown in
Figure 2 [
17].
From
Figure 2, it can be seen that the number of power outage events occurred are different for each state in the USA The least number of power outage events occurred are eight in Rhode Island, and the maximum number of power outage events occurred are 537 in California. As per the United States Department of Energy (US DoE) statistics, during one such major power outage, at least 50,000 customers have been impacted, and approximately 300 MW unplanned firm load loss is experienced [
17]. Moreover, these power outages will have a significant impact on society, residential, and industrial operations. The effect would depend upon the frequency of power outage occurrence at a particular location. From
Figure 2, it is clear there is a significant impact in most states of the USA, and hence there is scope for researching on improving the resilience of PSS, so that, the REUs are ensured with adequate power supply.
In order to ensure and improve the resilience of PSS, many solutions have been proposed in the literature. The most suggested and preferred solution in the literature is the use of a backup power facility, either as a storage or generation option [
18]. For residential houses, microgrids are mostly recommended. For instance, a microgrid is modeled for powering 100% of electrical loads using a renewable-based power system. It is suggested that the RE-based power system is only capable of powering the houses based on the nature of resilient issues and also depends on the intensity of the power outage [
19]. On the other hand, diesel generator (DG)-based studies are also presented by a few researchers, and they suggest that continuous power supply is possible during power outages [
20]. Later with the advancement seen in DERs, the use of battery energy storage (BES) systems has become more popular. Few studies have shown any evidence on ensuring resilience; if the PSS configuration has the combination of BES with RE or DG based backup facilities or any RE-based hybrid configuration [
21,
22]. In a study conducted for the USA, for providing improved power resilience, Anderson et al. suggested the use of hybrid renewable energy-based microgrid composed of solar photovoltaics (PV), DG, and BES [
22]. In addition, few studies conducted for the USA were mainly concentrated methods to assess and enhance resilience. For instance, a probabilistic method is developed to assess the resilience of the PSS considering the disruptive event caused due to a hurricane [
23]. In another study, the concept of survivability through microgrids is introduced for enhancing the power resilience of PSS.
From the brief literature review, it is understood that there is a thrust to carry out research on power resilience, but so far, although many different PSS configurations have been proposed, none have been compared based on their feasibilities [
24,
25]. In addition, the studies related to the resilience of PSS, highlighting of the techno-economic and environmental indicators is very limited. Hence, different PSS configurations that are location-specific were proposed by considering a resilience framework embedding techno-economic and environmental indicators. These PSS configurations include; electric grid + battery energy storage, electric grid + battery energy storage + natural gas power generator, electric grid + battery energy storage + photovoltaics. The proposed resilience framework is based on being prepared, sustaining, recovery, and learning to adapt. Based on the proposed framework, realistic and meaning full indicators are explored from the techno-economic-environmental point of view. The main contributions of this manuscript are as follows:
A four-component resilience framework with techno-economic and environmental indicators to understand the resilience of residential electricity user (REU) power supply system (PSS).
A battery energy storage (BES) as a preparedness measure that is not considered in most of the literature is considered here while modeling the proposed PSS configurations.
The proposed three different REUs are modeled considering power outage duration as well as the electric load conditions of the New York-based residential multi-story building as a case study.
Evaluation of unmet and compensated electric loads for resilience comparison between the three PSS configurations of an REU.
This paper has a total of six sections;
Section 2 presents the four-component resilience framework and the considered indicators; in
Section 3, the description of the proposed three PSS configurations for REUs along with modeling is given; in
Section 4, data collection, techno-economic and environmental modeling and simulation for the proposed three PSS configuration is shown along with control strategies. In
Section 5, the results are presented, and a thorough discussion is made, and in
Section 6, conclusions are provided.
2. Resilience Framework
For understanding the resilience of the PSS configuration, a well-structured framework is necessary. The proposed framework should ensure that the system can sustain, recover, and adapt to the power outages or any other disruptive events. In this study, the proposed framework has four components, namely; preparedness, robustness, recovery, and adaptation [
13,
14]. These four components are clearly presented and depicted in
Figure 3.
From
Figure 3, it is understood that the highlighted depictions of the disruptive events (Level-1 and Level-2) and the variation in PSS functionality as per the disruptive events matches the resilience definition. Each component of the proposed resilience cycle is different based on its nature, and the four components are briefly explained below:
Preparedness: This component suggests the preparedness level of the PSS configuration for power outages. Here, an assumption is made that each REU is already prepared for power outages by employing backup energy storage.
Robustness: This component suggests the level to which the PSS configuration can sustain and will be able to supply power to REUs in the event of a power outage.
For understanding this component, an indicator, i.e., an increase in unmet electric load, is considered.
Recovery: This component suggests the level to which the PSS configuration was recovered and able to supply power back to the REUs.
Adaptation: This component suggests the level of learning by the PSS configuration based on the experienced power outages. More or less, it will give information on how the preparedness levels are improved based on the learning from previous power outages incidents.
Here, based on the above four-component resilience cycle, and for each component, indicators were identified, which were further used as constraints during techno-economic and environmental optimization of PSS configurations. These indicators include the; unmet electric load, compensated loads, power supplied by the PSS configuration, the cost of energy, net present value, initial capital investments, and the emissions (carbon dioxide; sulfur dioxide; nitrogen oxides).
5. Results and Discussion
The results of three PSS configurations under grid outages are presented briefly along with the resilience assessment. In addition, the three PSS configurations are compared based on techno-economic and life cycle based environmental indicators. Grid outage modeling is the common result of the three PSS configurations. Based on the power outage data, the disruption in the power supply to the REUs is evaluated and presented in
Figure 11.
From
Figure 11, it is understood that a total of three power outages occurred, and the duration of the outages lasted from 1.21 h to 311.95 h. The first power outage was an intentional attack and is shorter, and within 1.21 h, the grid restoration has taken place. The second power outage is due to system operational error, and however, the grid restoration happened within 3.18 h. These two power outages were shorter compared to the third incident that took almost 311.95 h to restore. The third power outage was the major one, and it happened due to the shortage of fuel resources on the grid side. Due to the power outages, the unmet electric load at the REUs is increased in each power outage incident, see in
Figure 12.
From
Figure 12, it is understood that the unmet electric load will increase if the EG is not restored within the specified time. The cumulative unmet electric load for the duration of power outage events is 31.211 kW (outage-1), 98.2 kW (outage-2), and 7676.08 kW (outage-3), respectively. If no backup power facility is available (either energy storage or generation), the REUs have to face the emergency, and this clearly shows that the preparedness towards grid outages from the REUs side is nil. As a result, REUs will have to experience deficit energy that is around 7805.49 kWh/y. In this case, the PSS configurations are not resilient. The indicators for resilience and the techno-economic and environmental feasibility of PSS configuration without any backup energy storage and generation are presented in
Table 5.
However, when it comes to a practical situation, the REUs will be prepared for a certain level of emergencies, and aligns with the assumption made under the preparedness component of the proposed resilience framework. Here, the REUs are already prepared for power outages, and as a result, in each PSS configuration, BES is considered. Hence, by considering the grid uncertainty, the proposed three PSS configurations for REUs were studied to understand which configuration is more resilient. In the below subsections, the results of proposed PSS configurations are briefly presented.
5.1. REU with EG + BES Based PSS Configuration
In EG + BES based PSS configuration, the REUs load is directly connected to the EG. Since REUs are connected to the grid, the power supply will obviously be disrupted in the event of a power outage. But as per the proposed resilience framework, preparedness measure is already considered. So, the REUs are already prepared for power outage emergencies with backup power facilities using a BES. In general, the REUs will not have sufficient information to decide the battery storage capacity specific to unmet electric load during power outage events. Here, to have an understanding of the feasibility of EG + BES configuration, irrespective of the REUs ability to afford the BES, we modeled the battery capacity considering the maximum possible power outage duration as battery autonomy, i.e., roughly 13.18 days. Simulation is carried out with the power outage events during the usual conditions, where the REUs draw power from the EG, and at the same, the batteries are fully charged. Here, when the power outage occurred, the battery starts discharging for meeting the REUs electric load requirements. As batteries are charged to 100%, they can power the REUs for the complete blackout duration and thus meet the required load. The PSS configuration with EG + BES was able to supply power and can be one of the resilient solutions. The indicators for resilience and the techno-economic and environmental feasibility of PSS configuration are presented in
Table 6. However, when it comes to the practical situation, an individual REU may not be able to afford a BES with 13.18 days of battery autonomy.
5.2. REU with EG + NGPG + BES Based PSS Configuration
In EG + NGPG + BES based PSS configuration, the REUs electric load is connected to the NGPG and a BES. These two back up options will meet the electric loads in the case of emergencies. This configuration is considered under the prosumer category, but the choice of operation of the NGPG and power selling to the grid is more dependent on the REUs. The simulation is carried out, assuming that the REUs will operate the NGPG only in emergency situations. As per the PSS configuration, a simulation model is developed in the HOMER tool. The model based on sensitivity analysis showed that the NGPG was able to generate the electricity for the gird outage durations. The compensated load by NGPG during the power outage events is presented in
Figure 13 and
Figure 14.
In
Figure 15, the power produced by the NGPG and along with the power purchased from the EG, is shown.
From
Figure 15, it is clear that in the event of a grid outage, the NGPG able to restore the PSS thereby provides continuous power supply to REU. The indicators for resilience and the techno-economic and environmental feasibility of PSS configuration are presented in
Table 7.
5.3. REU with EG + PV + BES based PSS Configuration
5.3.1. Weather Data for PV Modelling
For the studied residential location, throughout the year, there exists a significant amount of solar radiation potential but with varying capacities. The data on solar radiation potential is collected from the National Aeronautics and Space Administration (NASA) databases using the data access provision provided by the HOMER tool and presented in
Figure 16.
From
Figure 16, at the REUs, the observed solar radiation power potential varies from 0.02 kW/m
2/day to 0.48 kW/m
2/day, which is sufficient to power the REUs. In terms of energy potential, the daily potential for the given month is varied between 1.67 to 5.67 kWh/m
2/day recording maximum in June and minimum in December. The annual average daily solar radiation is 3.80 kWh/m
2/day, and the observed clearness sky index is 0.4906, which is suitable for solar power generation. Apart from solar radiation, in the PV system modeling, we used wind speed and ambient temperatures to estimate the impact of module temperature on the total power produced. In the considered locations, the observed annual average wind speed for the last ten years was around 5.7 m/s. The wind speeds are ranged between 4.60 to 6.68 m/s, with an average of 2.75 m/s. The annual average temperature is observed at 11.68°C. The monthly average daily temperatures were observed to vary between −0.99 to 23.89°C recording the minimum in January and maximum in July.
5.3.2. Power Performance of EG + PV + BES based PSS Configuration
In EG + PV + BES-based PSS configuration, the electric load is connected to an on-site solar PV power plant as well as the EG. This configuration is considered under the prosumer category, where the REUs sell the power to the grid. The simulation is carried out by considering the local weather parameters, as discussed in
Section 5.3.1 and power outage conditions, as shown in
Table 1. In our simulation, the effect of sensitive parameters on the overall power generation is also observed. From the investigation, it was understood that the PV plant was able to meet the electrical load requirements of REUs, and at the same time, it was able to sell power to the EG. The power generation profile under EG + PV + BES-based PSS configuration is depicted in
Figure 17. The indicators for resilience and the techno-economic and environmental feasibility of PSS configuration are presented in
Table 8.
5.4. Comparison of PSS Configurations
In this section, the investigated three PSS configurations were compared based on their resilience, and techno-economic and life cycle-based environmental feasibility indicators. The detailed comparison of the indicators is presented in
Table 9. In
Table 10, the renewable-based prosumer category of the PSS configuration (i.e., EG + PV + BES) results are compared considering the two conditions that are with and without sales to EG.
Based on the obtained comparative results shown in
Table 9 and
Table 10, it is understood that the three configurations (EG + BES; EG + NGPG + BES; EG + PV + BES) were observed to be resilient and were able to supply power during power outages. Here, the resilience was achieved with optimal planning of energy infrastructure under each PSS configuration. In addition, based on the nature of the PSS category, i.e., consumer or prosumer, we observed a significant effect on the techno-economic and environmental indicators. For example, the initial capital cost was observed to be very high for EG + BES configuration, and this is because of the battery capacity needed to meet the massive power outage that lasts for almost 316 h. The initial capital cost for EG + NGPG + BES-based PSS configuration is quite less, as it is only planned for emergency backup. If the NGPG is made to operate continuously considering the grid sales, the capital cost will increase; also, the operating and maintenance cost would be very high. The third PSS configuration, i.e., EG + PV + BES, can be operated in prosumer mode, and in this configuration, the initial cost would be a little high but with almost zero investment on fuel resources, which is better when compared to NGPG. In addition, the payback is also viable. PV, when operated considering the REUs load consumption alone, has lower amounts of emission than the operation with grid sales. Even from the CO
2 emission point of view, the PSS with PV and BES seems to have lower emissions. In addition, the observed CoE is lower for the PSS with PV and BES. Overall, after a thorough quantification of the techno-economic and life cycle based environmental indicators, the EG + PV + BES was able to perform well by satisfying the resilient condition and ensuring the power supply.
5.5. Discussion on Ensuring Resilience and Future Research Directions
In this section, a brief discussion is provided on how the studied three PSS configurations ensured the power resilience considering the four-component resilience framework. Besides, based on the provided discussion, few unaddressed issues were identified, and accordingly, the future research directions are proposed. In this study, the proposed three PSS configurations were resilient enough in supplying power to the REUs in the grid outage events, but each configuration differed in their techno-economic and life cycle-based environmental indicators as discussed in
Section 5.4. The resilience is investigated based on PSS functionality variation, which is the function of the available power and the electricity demand at the REUs.
In the case of REU that is dependent only upon the EG (where there is no facility for on-site generation and storage), due to grid outage, there is a sudden fall in PSS functionality to zero. The fall to zero would happen immediately as there is no backup, thus low preparedness. The value of PSS functionality will be maintained as zero until the grid restoration has taken place, making the system not robust to grid outages. However, the recovery is observed to be 100%, but only after a dedicated amount of time, i.e., the time taken for grid restoration. The fourth component, i.e., adaptation, is more important. Based on the experiences, the REUs must learn and must adapt existing approaches that make the system more resilient. The adaptation is more related to improving the system efficiency and associated assets, which involves capital investment. With this, the PSS can adapt to previously experienced situations and is able to enhance preparedness to deal with the outage in the future.
In the case of EG + BES-based PSS configuration, during the grid outage, the PSS functionality is observed to vary. Here the resilience of PSS configuration is dependent upon the size of BES. If the considered BES capacity can manage the unmet loads, there will not be any drop in PSS functionality, and it is always maintained at maximum until and unless there is another disturbance at BES. The size of BES would again influence capital investment. The four-components of resilience will be affected; for instance, preparedness can be achieved even with low capacities of BES, which needs a lower investment, whereas the PSS may not be robust enough if the grid restoration times are longer. The recovery component is again dependent upon the capacity of the BES and grid restoration times.
The other two considered PSS configurations can generate on-site power generation. During the grid outages, they provide continuous power supply making the PSS configurations more resilient. In the case of EG + NGPG + BES, the PSS functionality is more dependent upon the NGPG operation time and natural gas fuel availability at the REUs, whereas, in the case of EG + PV + BES, the PSS functionality is more dependent upon the intermittent nature of the solar irradiance. In the considered location, these two resources are available, hence the impact on PSS functionality and resilience components is not much.
Based on the above-provided discussion, a few future research directions can be proposed. These include:
- ∙
Investigation of the investments made in energy infrastructure and their impacts on improving resilience can be considered as one of the research directions, as in our study. We observed the variations in economic indicators based on the proposed energy infrastructure.
- ∙
Resilience framework incorporating islanding and other grid disturbance detection approaches could be considered as one possible future research direction.
- ∙
Tools that support grid restoration can be beneficial in optimal scheduling of on-site power generation facilities at REUs side, and in fact, they could influence the robustness component of the resilience cycle.
- ∙
Modernization and re-design of PSS infrastructure by using advanced technologies like Blockchain [
42,
43,
44], Internet of Things [
45], Energy Internet [
46], Blockchain-based Internet of Things (B-IoT) [
47], and Artificial Intelligence (AI) techniques such as Machine Learning (ML), and Deep Learning (DL) [
48] could favor in tracking the power outage events and thereby allows us to have a data-driven solution. Hence, power sector digital transformation from the context of resilience can be one of the possible future research directions.
- ∙
Location and ecosystem specific studies can be further modeled to evaluate the feasibilities of PSS configurations in the context of the proposed resilience framework.