Potential Analysis and Optimal Management of Winter Electric Heating in Rural China Based on V2H Technology

Abstract

In order to improve the air pollution problem in northern China in winter, coal-to-electricity (CtE) projects are being vigorously implemented. Although the CtE project has a positive effect on alleviating air pollution and accelerating clean energy development, the economic benefits of electric heating are currently poor. In this study, a system based on vehicle-to-home (V2H) and photovoltaic power generation that can effectively improve the benefits of CtE projects is proposed. First, a V2H-based village microgrid is proposed. The winter temperature and direct radiation of the Beijing CtE project area are analyzed. Extreme operating conditions and typical operating conditions are constructed for potential analysis. After that, a bi-layer optimization model for energy management considering travel characteristics is proposed. The upper layer is a village-level microgrid energy-dispatching model considering meeting the heating load demand, and the lower layer is a multi-vehicle energy distribution model considering the battery degradation. The results show that the distribution grid expansion capacity of the electric heating system based on V2H and PV generation is reduced by 45.9%, and the residents’ electricity bills are reduced by 68.5%. The consumption of PV can be completed. This study has effectively increased the benefits of electric heating in northern China during winter. This helps the CtE project to be further promoted without leading to large subsidies from the government and the State Grid.

1. Introduction

1.1. Background

China has made a commitment to reach the peak of CO₂ emissions by 2030 and achieve carbon neutrality by 2060 [1]. With the continuous and steady development of China’s economy in recent years, air pollution has become increasingly serious. Especially during the heating season, foggy weather will continue to appear for a long time and remain at its peak [2]. The issue of heavy air pollution, exemplified by haze, significantly impacts air quality [3]. There are 6.5 billion square meters of floor space in rural China that need to be heated in winter, and about 200 million tons of loose coal are burned every year [4]. Compared to centralized urban heating, this type of heating has more serious problems [5]. Based on this background, the State Grid of China has launched the development strategy of “electric energy substitution” [6]. Electricity is used to replace coal, and gas is used to replace coal. Among these measures, the coal-to-electricity(CtE) project is the most important part [7]. Since the conversion of CtE for rural heating is the main purpose of the project, rural residents are the main target of the project [8].

With the support of the "double carbon" policy, the Chinese government is vigorously developing the new-energy vehicle industry [9,10,11]. A series of plans have been issued, making the development of new-energy vehicles in China unprecedented [12,13]. According to statistics, by 2025, sales of electric vehicles (EVs) will account for about 20% of total new car sales. By 2035, EVs will dominate new car sales [14,15]. Energy systems, when deeply integrated with renewable energy sources and vehicle electrification, confront substantial obstacles concerning system stability and reliability. Considering the intermittent and volatile nature of photovoltaics (PV) and the huge impact of large numbers of EVs charging simultaneously, this stochastic and dynamic nature will pose a challenge to the grid [16].

Vehicle-to-Grid (V2G) technology has been proposed as an important way to address renewable energy consumption and maintain power system stability. V2G stands for Vehicle-to-Grid. It refers to the use of electric vehicles as storage devices for energy. The vehicle’s battery is utilized to store electrical energy. This electrical energy can be retained in the vehicle or returned to the grid [17]. This bi-directional technology allows EVs to be used both as loads and as distributed energy sources [18,19,20]. However, the viability of V2G requires major infrastructure upgrades to the grid, which are currently difficult to achieve [21]. Vehicle-to-Home (V2H) is an alternative technology to V2G that requires more major infrastructure upgrades to the grid to implement [22,23]. It allows EVs to store and release energy between the battery and the home. In recent years, more and more homes have been installed with distributed power generation systems, such as small photovoltaic systems and wind power systems [24]. Distributed generation systems, together with load and energy storage devices in buildings, make up the building microgrid [25]. Combining this with V2H technology, EVs can store energy. The ability of the building to consume distributed energy is improved, while increasing the reliability of the grid. Therefore, it is important to use V2H technology to control the behavior of EVs parked in or around the home, and to optimize their charging and discharging strategies [26].

1.2. Related Work

China’s CtE project is a rural energy-renovation project. The project works mainly through the transformation of transformers, laying cables, overhead lines, and other measures. Some scholars have conducted studies on coal-to-electricity projects in rural China. Long et al. [27] proposed a two-stage construction scheme evaluation method including preliminary selection and Pareto-based selection to obtain the optimal construction scheme for the associated distribution network and regional integration under electric and thermal loads. Li et al. [28] leveraged life cycle assessment, equivalent annual cost, and multi-attribute decision-making techniques to assess the environmental ramifications, financial outlay, and user experience of eight typical heating solutions for rural areas. Following extensive field evaluations, three superior methods emerged, specifically, biomass pellet heating, air source heat pumps, and air conditioning systems. Su et al. [29] proposed a spatial analysis methodology. This approach employed six metrics, including the efficiency of the heating system, the primary energy use intensity of the heating system, the discounted payback period, the internal rate of return, CO₂ emission intensity, and PM2.5 emission intensity, to assess various heating techniques. Zhang et al. [30] employed a life cycle assessment (LCA) approach to evaluate the environmental benefits of geothermal heating in comparison to other methods. The analysis focused solely on the environmental advantages of geothermal heating, without taking into account other indicators. Zhang et al. [31] investigated the economic benefits of implementing the CtE project in grid enterprises. The findings suggest that the project, undertaken by the grid company, exhibits subpar economic performance throughout its lifecycle. This can be attributed to factors such as high initial investment, an irrational electricity pricing mechanism, and significant seasonal variations in peak and off-peak power demand. As described in the previous study, the study of CtE in rural China now includes a comparison of economics. It has been demonstrated that the economics of current CtE projects are poor, but not yet resolved.

Various studies have been conducted to optimize capacity allocation and power dispatch in microgrids, and there are studies that consider EVs as energy storage devices. Different topologies and algorithms are used to maintain economics, reliability goals and other objectives. Venegas et al. [16] developed an open-access agent-based EV simulation model. Analysis showed that non-systematic insertion reduces the available flexibility, especially given the current trend towards larger cell sizes. Hu et al. [32] developed a real-time optimization model for EV rolling, considering wind power forecast errors and EV forecast information to minimize grid load fluctuations and maximize wind power consumption. The simulation results show that the model can achieve peak regulation and wind power consumption. Wei et al. [33] integrated a physical economic model with a data-driven model to efficiently and accurately determine the optimal sizes of microgrids. This approach took into account nonlinear battery degradation and optimized power dispatch considering varying EV charging profiles. Tookanlou et al. [34] proposed an electric vehicle day-ahead scheduling model that considers the benefits to electric vehicles, charging stations, and retailers of performing electric vehicle scheduling. The results show that the scheduling model reduces the charging cost of EVs and increases the revenue of charging stations and retailers. There are also several studies on the use of V2G technology for heating. Daramola et al. [35]. combined natural gas and fuel cell cogeneration technologies with renewable energy and BESS. BESS supported V2G operation and stores excess power from cogeneration and renewable energy sources. Significant reductions in CO₂ emissions can be achieved. A dynamic tariff subsidy approach for distribution networks with a high penetration of PV, heat pumps, and V2G-enabled EVs was proposed by Huang et al. [36]. Wang et al. [37]. developed a planning model using V2G techniques to minimize the daily scheduling cost, considering the patio-temporal characteristics and room temperature demand of electric vehicles.

Optimization algorithms used in previous work can be divided into three categories: data-driven models, software-based optimization, and physical economic models [33]. Data-driven modeling relies on the ability to accumulate experience from historical datasets [38]. A framework is to use machine learning (ML) methods to estimate near-optimal solutions [39]. However, generating training datasets efficiently remains challenging. Chen et al. [40]. also emphasized the problem of a lack of training data, which can lead to unreliable data-driven results. The computer simulation software HOMER (https://www.homerenergy.com/ accessed on 30 June 2023) is one of the most commonly used software for analyzing renewable energy systems [41]. However, The software’s interfaces, optimization objectives, and algorithms are inflexible, and cannot be customized to suit various application scenarios, particularly in terms of accommodating electric vehicle charging needs.

Physical economic models include both rule-based and optimization-based approaches. Rule-based methods have a clear structure but usually fail to achieve economic optimality [42]. Optimization-based methods should be more adaptable to optimization trends in different scenarios, and can be classified as global planning, real-time feedback, heuristic algorithms, and hybrid methods. Linear programming [43] or its subcomponent, mixed-integer linear programming [44], can efficiently compute the best possible solution and guarantee optimality, but they are not suitable for nonlinear constraints and require a large amount of time for a large number of variables. Heuristic algorithms, including genetic algorithms (GA) [45] particle swarm optimization (PSO) [46], and whale optimization (WOA) [47], make them easy to comprehend and have robust global search capabilities, but they do not guarantee the optimality of the results, and take a relatively long time to compute. In order to solve the optimal solution of the nonlinear system and to speed up the solution and avoid falling into dimensional explosion, the improved physical economic model is chosen in this work, and a bi-layer optimization model is constructed.

1.3. Motivation and Contribution

Based on the research studies investigated, there is still a lack of proposals for making electric heating economical in winter. Although some analysis has been carried out on the economics of the heating method, and the planning of the distribution network renovation has been optimized, the State Grid and the government still need to invest money, and it is difficult to recover the costs. In terms of V2H, there is also no analysis on the potential of EVs as distributed energy storage devices for optimizing electric heating. Therefore, an innovative way to make it more economical is still needed for rural electric heating systems in northern China.

To address the research gaps, this study has built a village microgrid based on V2H. The focus of this work is on the implementation of rural microgrids using V2H- and PV- based systems, and analyzing the potential. The EV owners and their travel behaviors were analyzed. Based on these behaviors, extreme and typical operating conditions of electric heating scenarios were set. A bi-layer optimization model was proposed to optimize the energy management of the system. The potential of the electric vehicle to participate in the vehicle–grid interaction was calculated by combining the peak and valley electricity prices in the electric heating area. The goal of this work is to improve the economics of electric heating systems in northern China, which includes the government, grid companies, and electricity-using residents. The research idea is illustrated in Figure 1.

The subsequent sections of this paper are structured as follows. The V2H-based microgrid structure and scenario analysis is established in Section 2. The bi-layer optimization model is built in Section 3. Section 4 describes the optimization results and economic analysis. Finally, conclusions and future work are discussed in Section 5.

2. Microgrid Topology and Scenario Analysis

In this section, the structure of the microgrid for this work is described. The climatic conditions of Beijing for the heating season 2012–2022 are introduced. The residential electricity load using an air source heat pump for heating is analyzed.

2.1. Microgrid Topology

The objective of the work on the optimal management of electric heating systems considering EV travel is to discuss the impact on the economics of winter heating after users build a rooftop PV system while EVs are connected to the distribution grid. Therefore, for these objectives, this work constructs a village microgrid, as shown in Figure 2, in which each household constructs PV power generation equipment and all have EVs as a means of travel. EVs meet daily travel needs and can be used as energy storage devices to participate in V2H when EVs are not traveling.

In this work, a microgrid is established to serve 100 households, with each household equipped with a 12.7 kW PV installation on their roof. Additionally, all EVs in the microgrid have a battery capacity of 60 kWh, which is considered a typical and representative value.

2.2. Scenario Analysis of Electric Heating

In this work, Beijing city is used as the study target. This work is based on environmental data and resident load data for the heating season 2021–2022 from the real world. This subsection analyzes the typical daily loads of the environment and residents for heating, during the winter heating period.

Beijing experiences a warm temperate semi-humid and semi-arid monsoon climate [48]. Winters are cold and dry, and daylight is abundant. Residents have heating needs, and the heating season in Beijing lasts for four months, from mid-November to mid-March [49]. This work is carried out against the background of environmental data for the heating season 2021–2022, and the temperatures for the whole heating season are presented in Appendix A (Figure A1, Figure A2, Figure A3, Figure A4, Figure A5 and Figure A6). The temperature data are at the hourly level. It can be seen that the daytime temperatures are high and the nighttime temperatures are low throughout the heating season, with daytime temperatures above 0 °C and dropping below 0 °C at night. The temperature difference between day and night is large, reaching 15 °C, meaning that the heating demand is greater at night than during the day.

The average daily temperatures during the heating season are shown in Figure 3. As can be seen, the highest average daily temperatures are around 10 °C around the beginning and end of the heating season. Starting in late December, the average daily temperature is below 0 °C, but generally above −5 °C until late February. The lowest average daily temperature was on December 24, when the average daily temperature decreased to −10 °C, due to the snowy weather that occurred two days before.

There was one 160 kVA transformer in a village in the Miyun District, Beijing, before and after the implementation of the “CtE” project, which was not replaced. The load factor of the transformer was extracted on 31 August 2021 (non-heating season) and 7 November 2021 (heating season). The power factor was taken as 0.9, due to the increase in the heating load in winter, and the active power of the load before and after the implementation of the CtE project was obtained. As shown in Figure 4, there were nine households under this transformer after the research.

Based on the weather data, k-means clustering analysis was carried out for the highest, lowest and average temperatures for each day of the heating season. k-means clustering is a technique for clustering data into a specified number of classes [50]. It reveals the intrinsic properties and patterns of the data. k-means clustering is a divisional clustering method that divides the data into k clusters. Each cluster has a center, called the center of mass, and the k-value needs to be given. The k-means clustering algorithm proceeds as follows [51]:

• Step 1. Randomly select k data points as initial clustering centers;
• Step 2. Divide each data point by the nearest center of mass, measuring the distance between two sample data points;
• Step 3. Recalculate the center of mass of each cluster as the new cluster center, so that its total squared distance is minimized;
• Step 4. Repeat step 2 and step 3 until convergence.

The results of clustering into two and three clusters are shown in

Table 1. Temperature clustering result with the number of 2 clusters.

	Cold Weather	Warm Weather
Average temperature	−2.74	5.69
Highest temperature	3.2	12.3
Lowest Temperature	−8.7	−0.9

and

Table 2. Temperature clustering result with the number of 3 clusters.

	Cold Weather	General Weather	Warm Weather
Average temperature	−3.06	2.61	8.14
Highest temperature	2.6	9.9	14.2
Lowest Temperature	−8.9	−3.7	1.1

. The number of clusters of two means that the weather is divided into cold weather and warm weather. A cluster number of three means that the weather is divided into cold weather, average weather, and warm weather. The results show that November 7 falls in the cold cluster and the temperature drops significantly from November 6 to November 7. Therefore, it can be judged that November 7 was an extreme cold weather day, and the usage of heating devices by users was high. As can be seen in Figure 4, the residents’ electricity consumption increased significantly during the extremely cold weather after the adoption of electric heating. The power increased nearly 10 times, and the power consumption fluctuated insignificantly throughout the day. This is due to the increased heating load in winter. This puts a higher demand on the distribution network capacity to meet the electricity demand of users.

3. The Bi-Layer Optimization Model

This study focuses on refining an energy management model that considers travel characteristics. The approach considers the integration of electric vehicles as energy storage units within the distribution grid and the installation of rooftop PV systems by residents. The primary objectives are to explore the capacity requirements of the distribution grid and assess the potential for reducing residents’ heating costs. The upper layer of the optimization model addresses microgrid energy dispatch at the village level, ensuring the fulfillment of heating load demands. The lower layer concerns the allocation of energy among electric vehicles, considering battery degradation.

The optimization variables of the optimization-based bi-layer optimization algorithm include both the parameters within the system at the upper level and the system power allocation at the bottom level. The energy management strategy of the system affects the economic operating cost of the system, which in turn affects the selection of the parameter configuration of the system. Therefore, the bi-layer optimal configuration of the parameters based on the energy management strategy considering the optimal power allocation is consistent with the actual operation of the system after it is put into use, and can also save the operating cost of the system in the whole-life cycle. At the same time, the two-layer solution model can avoid the problem of dimensional explosion during the solution process, which can easily fall into local optimum.

3.1. Upper-Layer Model

The upper-level model is used to optimize the energy management of the village microgrid. Based on the completion of PV consumption, the objective is to minimize the cost of electricity for residents, so the objective function is as follows:

$$\min f=\min {\displaystyle \sum }_{t=1}^n{\mathrm{P}}_{grid}\left(t\right) \mathrm{R}\left(t\right)\Delta t$$

(1)

where P_grid(t) denotes the power drawn from the distribution network at hour t. R(t) denotes the hourly tariff at hour t. n is the total duration of the model calculation. t is the optimization step, and the optimization step in this study is 1 h.

The constraints include the power–energy balance constraint, the power limitation constraint, and the SOC limitation constraint of the EV battery; the equation constraints are as follows:

$${\mathrm{P}}_{load}(t)={\mathrm{P}}_{grid}(t)+{\mathrm{P}}_{pv}(t)+{\mathrm{P}}_{EV}(t)$$

(2)

where, P_load(t) denotes the load power at time t. P_PV(t) denotes the PV power at time t. P_EV(t) denotes the EV cluster charging or discharging power at time t.

The power limitation constraint is shown in Equations (3) and (4). Due to the current two-way charging point power limitation, the transmission power between the EV and the grid needs to be limited to a certain range. In order to prevent the large multiplier fast charging from damaging the internal structure of the battery and thus accelerating the battery aging, the discharging and charging power of the EV also cannot exceed its maximum power limit. In this work, the maximum power of charging and discharging of a single EV is set to 7 kW. The total power of the EV cluster is the sum of the charging and discharging power of the non-traveling vehicles.

$$0\le {\mathrm{P}}_{grid}\left(t\right)\le {\mathrm{P}}_{grid\max }$$

(3)

$$-{\mathrm{P}}_{EV\max }\le {\mathrm{P}}_{EV}\left(t\right)\le {\mathrm{P}}_{EV\max }$$

(4)

$${\mathrm{P}}_{EV\max }\left(t\right)=N_i{\mathrm{P}}_{EV\max }$$

(5)

The SOC limitation constraint for electric vehicles is shown in Equation (6). The SOC of the electric vehicle must be limited within a certain range to prevent the battery from over-charging or over-discharging, which can lead to safety accidents with the battery. At the same time, considering the daily travel characteristics of EVs, the battery SOC also needs to be controlled within a reasonable range at the end of the optimization phase, to ensure that the EV can operate and work normally on the next day.

$${\mathrm{SOC}}_{\mathrm{lower}}\le \mathrm{SOC}\left(t\right)\le {\mathrm{SOC}}_{\mathrm{upper}}$$

(6)

SOC_lower and SOC_upper are the lower and upper bounds of the EV SOC at the end of the optimization phase.

3.2. Lower-Layer Model

The lower-level model is used to develop a scheduling strategy for each EV, based on meeting the travel demand of EVs and the energy scheduling of the upper-level village micro-grid to minimize battery degradation. The objective function is as follows:

$$\min f=\min {\displaystyle \sum _{\mathrm{i}=1}^{100}{Q}_{loss,i}}(t)$$

(7)

where Q_loss,i(t) denotes the decay of vehicle i at moment t. In this case, the battery decay model uses the Arrhenius decay model, and the Arrhenius model considers the battery temperature, multiplier, and ampere-time throughput, as shown in Equation (8). The battery multiplier is power dependent, as shown in Equation (9).

$$Q_{loss}^{}=A\exp (\frac{-E_{\mathrm{a}}+k{C}_{\mathrm{rate}}}{RT})A{h}^{\mathrm{z}}$$

(8)

$$C_{\mathit{rate},i}=\frac{{\mathrm{P}}_{ev,i}}{Q}$$

(9)

The constraints include the power balance constraint and the power limitation constraint. The equation for this constraint is as follows:

$${{\displaystyle \sum }}_{\mathrm{i}=1}^m{\mathrm{P}}_{ev,i}\left(t\right)={\mathrm{P}}_{EV}\left(t\right)$$

(10)

where m denotes the number of non-traveling vehicles and P_ev,i(t) denotes the power of the ith vehicle at time t.

The parameters used in the optimization model are presented in

Table 3. Parameters used in the model.

Parameters	Value
Photovoltaic capacity per household (kW)	12.7
Maximum EV charging power (kW)	7
EV battery capacity (kWh)	60
SOC upper boundary of the EV (%)	100
SOC lower boundary of the EV (%)	0
Number of residential households	100

.

3.3. Methodology and Process of Solving

The optimal scheduling model of the electric heating system considering electric vehicle travel optimizes the energy management of the village-level microgrid and the charging and discharging behavior of EVs. The objective function is nonlinear, and the solution of the optimization model in this study is a nonconvex optimization problem, which cannot be solved by traditional mathematical models. The optimization objectives of the upper optimization model and the lower model are different, and suitable algorithms need to be selected for solving the upper and lower layers.

The upper-level model uses a particle swarm optimization algorithm to solve this problem. This algorithm is one of the heuristic algorithms for solving non-convex optimization problems. It has the advantages of less setting for parameters, strong optimization capability and fast solution speed [52]. In the particle swarm model, the feature space of each particle is defined by its position vector and velocity vector. The position vector represents a potential solution to the optimization problem, while the velocity vector indicates both the direction of movement and the particle’s speed at that particular position. During each iteration of the calculation, the velocity and position of any particle are calculated as shown in Equations (11) and (12).

$$V_m^{q+1}=\omega V_m^q+c_1{r}_1\left[X_{gbest}-X_m^q\right]+c_2{r}_2\left[X_{pbest,m}-X_m^q\right],m=1,2,\dots ,M$$

(11)

$$X_m^{q+1}=X_m^q+V_m^{q+1},m=1,2,\dots ,M$$

(12)

where $V_q^m$ and $X_q^m$ are the velocity and position vectors of the mth particle in the qth iteration, respectively. m is the total particle size, and, in this case, M is taken as 500 and the total number of iterations Q is set to 100, thus increasing the probability of obtaining the global optimal or near-global optimal solution.

The lower model is computed using a genetic algorithm. A genetic algorithm is a computational model of a biological evolutionary process that simulates the natural selection and genetic mechanisms observed in Darwinian biological evolution, serving as a method for searching for optimal solutions by mimicking the natural evolutionary process [53]. Its main feature is the direct operation on structural objects without the qualification of derivation and function continuity. It has an inherent implicit parallelism and a better global optimal search capability. By utilizing a probabilistic search method, the optimal search space can be automatically determined and guided without rigid rules, allowing the search direction to be adaptively adjusted.

Combining the above descriptions, the process of solving the optimal scheduling model of an electric heating system considering electric vehicle travel is shown in Figure 5. The optimization of the model was solved by MATLAB 2020b software. The computing tool is a computer with a Microsoft operating system, equipped with a Rui Long R7-5800H processor.

4. Case Study

4.1. Basic Data

The aim of optimizing electric heating systems while incorporating EV charging was to investigate the effects on distribution grid capacity and resident electricity bills, particularly when residents install rooftop PV systems and connect their EVs to the grid for bidirectional V2H operations. This research aimed to compare this with the traditional CtE system.

This work was calculated in the context of climate data for the 2021 heating season. The load conditions, EV travel intentions and PV generation were determined by the environment, as shown in Figure 6. Temperature affects load and PV generation. Weather affects load and PV generation. Direct radiations affect PV generation.

The capacity of the distribution grid should meet the energy supply of the load when dealing with extreme working conditions. Therefore, the capacity expansion potential of the distribution grid should be calculated under extreme working conditions. In electric heating systems, EVs are used as energy storage devices or to provide a small amount of energy. The load is mainly supplied by the PV and the distribution grid. Therefore, the extreme operating conditions occur when the load is large and the PV generation is small. The weather from 21 to 27 December 2021 was selected as the extreme working condition, as shown in Figure 7. These 7 days experienced a round of snowfall. This work considered that EVs prepare power for the extreme weather one day in advance, because the weather forecast now allows residents to know about the rain and snow in advance. So, the optimization calculation included the day before and two days after the snowfall, for a total of four days.

The reduction potential of resident heating costs should be calculated by selecting typical continuous sunny weather; 1–7 December 2021 was selected as a typical working condition, as shown in Figure 8.

After selecting the weather data, the load data of the extreme weather load and literature load data were interpolated, to obtain the load on different days, according to Figure 4. In this work, the heating power of the inhabitants was obtained proportionally, from the data of a typical day. In Section 2.2, our analysis found that 7 November 2021 was extremely cold. Based on this finding, we obtained the heating power for different days by comparing the temperature on different days with the temperature on November 7. The PV power generation data for different days were obtained by interpolating the data based on the rooftop PV data from the Automotive Research Institute of Tsinghua University. The real power generation data of the rooftop photovoltaic system of the microgrid at the Institute of Automotive Research, Tsinghua University, on 2 November 2020, was obtained. These data were for a typical sunny day in winter. The PV generation power for the studied date was obtained proportionally, based on the comparison of the direct radiation on the studied date and November 2. The operational data of the Tsinghua University microgrid on 7 November 2020 was added in Appendix B. This project constructed a village micro-grid for 100 households with an average household rooftop PV of 12.7 kW. The PV generation and load are shown in Figure 9 and Figure 10.

The CtE project area was in Beijing, China. After being recognized as CtE residents by the government, residents implemented peak and valley tariffs during the heating season (1 November–31 March): valley (8:00 a.m.–20:00 p.m.) 0.4883 CNY/kWh, peak (20:00 p.m.–8:00 a.m. the next day) 0.3 CNY/kWh. During the heating season, CtE residents used electricity in the valley section (8:00 p.m.–8:00 a.m. the next day) at a government subsidy of CNY 0.2 per kWh.

Electric vehicle trips refer to the annual report on the development of transportation in Beijing in 2021. The EV daily travel mileage in this work was 50 kM. EV power consumption was 13 kWh per 100 km. Travel was influenced by weather, with snow travel being one-third of that on sunny days. The number of vehicles that did not travel at each moment during sunny weather is shown in Figure 11 [54]. EV travel patterns were based on the 2021 Beijing Transportation Development Annual Report. This report provides an overview of Beijing’s transportation development in 2021, which includes travel patterns for private EVs.

4.2. Results and Analysis

The analysis was carried out from the two perspectives of energy flow and power flow, as shown in Figure 12 and Figure 13. The energy flow perspective shows that during the 4-day extreme weather round, 100 residential loads consumed more than 55,000 kWh of electricity. A total of 1059.5 kWh of electricity were consumed by EV cluster trips, and 12,279.98 kWh were generated by PV. During this extreme weather round, the vast majority of PV power was consumed directly by the load, and a small portion of the 786.68 kWh was required by the EV system to complete the energy transfer. In this round of weather, we set the EV cluster vehicle to give a small amount of power: 30% SOC, that is, 1800 kwh. The total can be calculated to need a distribution grid power extraction of 42,548.79 kwh. We can deduce that for the distribution grid to have continuous full power extraction, we need a household average of 4.43 kW distribution network capacity.

The above shows an analysis from the perspective of the power flow. The middle blue line indicates the power difference between the load and PV power generation: above 0 indicates the need for PV consumption, below 0 indicates the need for energy supply. The analysis is performed by combining energy flow and power flow. From the power flow, we can see that the power difference is 147.59, 337.53 kw, and 380.45 kw during PV dissipation on day 1, day 3, and day 4, respectively. In the energy flow analysis, we mention that the distribution network continuously reaches full power, with an average of 4.43 kW per household. This means that the maximum charging power of the EV cluster will be exceeded if the distribution network reaches full power when the PV system is stronger, during the daytime. In summary, a round of extreme weather requires a distribution grid power of 4.6 kW. For the purpose of this analysis, the distribution grid power mentioned in this work refers to the active power, not the apparent power, as it is usually referred to, and the capacity margin is not considered.

The results of energy distribution at 4.6 kW active power of the distribution grid are shown in Figure 14. It can be seen that the power is taken from the grid at full power almost all the time during a round of extreme weather. The daytime grid power withdrawal decreases because of the increase in daytime PV generation and EV cluster charging power limitations. The replenishment time for the EV clusters is concentrated in the daytime, which is due to sufficient daytime light intensity. Superimposed on the distribution grid energy, the excess energy is given to EV clusters in total, on the basis of meeting the load demand.

The SOC variation of the EV system is shown in Figure 15. There are four days in total, with the second day being snowy weather. This work prepares the EV system for the day before the snowy weather, in order to cope with the upcoming round of extreme weather. The third and fourth days are two days of cold weather after the snow. As you can see, on the first night, the energy of the traveling vehicle is transferred to the non-traveling vehicle. The non-traveling vehicles are able to have more capacity for daytime energy storage. On the first day, the non-traveling vehicles also completed their capacity. Therefore, the first day’s energy reserve includes the energy transfer from the non-traveling vehicle to the traveling vehicle at night and the energy reserve of the non-traveling vehicle during the daytime. The first day’s energy reserve reached its maximum at 19:00, with an average SOC of 95.51%.

4.3. Benefit Analysis

The results of a typical continuous seven-day operating condition optimization are shown in Figure 16. The constraint EV-cluster starting SOC is the same, so EVs only transfer energy. The distribution grid needs to provide the electric vehicle travel power consumption. It can be seen that the electric vehicle does not travel using the cluster daytime consumption of PV, and that the night-time power supply to the load grid for it to take electricity is mostly concentrated at night, because of the low price of electricity at night.

The benefit analysis is organized around three scenarios. Scenario 1 is without PV power generation and without EVs in the network. Scenario 2 is with PV but without EVs in the network. Scenario 3 is with both PV and EVs in the network. The power taken from the grid for the three scenarios is shown in Figure 17. Greater than 0 means that power is taken from the grid, and less than 0 means that power is fed to the grid. Scenario 2 has no energy storage devices such as electric vehicles, which need to return energy to the grid, and feeds 0.4 CNY/kWh. Scenario 3 can be completed without feeding power to the grid.

The benefit analysis is carried out in three aspects. The first point is the distribution network capacity, where the electricity company can gain benefits. The second point is the residents’ electricity bill, where the residents can gain benefits. The third point is government subsidies, where the government can benefit. This is shown in Figure 18.

The perspective of the average household distribution capacity is shown in Figure 18a. Scenario 3 is 45.9% lower than scenarios 1 and 2. This is beneficial to the power companies, because the expansion of the distribution network for “coal-to-electricity” residents is a huge investment, and is the main reason for the power companies’ losses.

The cost of heating for residents is shown in Figure 18b. With the addition of the rooftop PV system, the resident’s electricity bill is significantly lower, because rooftop PV can reduce the electricity taken from the grid, thus reducing the electricity bill. The difference in electricity costs between the Scenario 2 and Scenario 3 residents is not significant, because Scenario 2 considers the benefits of daytime feed-in to the grid. Compared to Scenario 1, electricity costs are 68.59% lower for Scenario 2 residents and 68.5% lower for Scenario 3.

The government subsidizes nigh-time electricity consumption, as shown in Figure 18c. The government subsidy is slightly reduced after building rooftop PV systems and V2H. This is due to the low electricity price at night, and the optimization result of Scenario 3 draws more electricity at night.

5. Conclusions

This study focuses on the implementation of a rural microgrid. It explores the potential analysis and energy management strategies for microgrids specifically designed to meet electric heating demands. The main contributions of this work are:

(1)

Analysis of the current situation and issues associated with the Coordinated to CtE project in Beijing. The study proposes a V2H-based system as an effective solution for addressing the issue of PV consumption and electric heating in rural areas of China.

(2)

To tackle the issue of dimensional explosion and the tendency to fall into local optima during the solution process, a bi-layer optimization model for energy management is proposed. The upper layer comprises a village-level microgrid energy-dispatching model aimed at fulfilling heating load demands. The lower layer focuses on a multi-vehicle energy allocation model that considers battery degradation.

(3)

The energy distribution results are obtained by solving the cases of typical and extreme conditions separately. In extreme weather, EVs store electrical energy one day in advance, which includes the transfer of electrical energy from vehicles that do not travel at night to vehicles that do travel, and the storage of electrical energy in vehicles that do not travel during the day. Under typical working conditions, the capacity of EVs as an energy storage device is sufficient when traveling is considered. EVs are able to complete the energy dissipation of PVs and release the electrical energy at night.

(4)

Optimization demonstrates substantial benefits, including a 45.9% reduction in the distribution capacity of the electric heating system based on V2H and PV power generation. Furthermore, residents can enjoy a significant reduction of 68.5% in their electricity bills. Additionally, the internal consumption of PV can be fully utilized.

This work addresses the problem of rural energy use in northern China. In combination with the ongoing CtE project, it makes full use of rural spatial resources and proposes an energy use system based on PV discovery and V2H technology. The proposed system is able to reduce the cost of distribution network renovation for the CtE projects now being implemented. Reducing the heating costs of the residents is attributed to the construction of the PV. In the future development process, in order to ensure the sustainable development of distributed photovoltaic systems, the main body of its development should be headed by enterprises such as power generation enterprises, supplemented by residents. A reasonable investment may be for the government, enterprises, and residents to participate in the completion of the roles of the three parties: government guidance, enterprise investment, and resident participation.

Efforts should be devoted to V2H-based microgrid energy management. However, the heating load is a flexible load that allows for building heat storage and participation in demand response, which is not considered in this work. In addition, different models of vehicles are not considered, and future work should consider the variability of EVs.