A Review of Distribution Grid Consumption Strategies Containing Distributed Photovoltaics

Abstract

With the growing energy crisis and environmental problems, distributed photovoltaic (PV), as a clean and renewable form of energy, is receiving more and more attention. However, the large-scale access to distributed PV brings a series of challenges to the distribution network, such as voltage fluctuation, frequency deviation, protection coordination, and other problems. In order to solve these problems, this paper provides a research overview of distribution network consumption strategies containing distributed PV. Firstly, this paper introduces the characteristics of distributed PV and its impact on the distribution grid. Then, the difficulties and challenges of distributed PV consumption are analyzed from the technical and economic levels. On this basis, a comprehensive consumption strategy is proposed, including the following aspects: firstly, it is mentioned to optimize the access method of distributed PV and to reduce the impact on the distribution grid by reasonably choosing the access point and access capacity of distributed PV. Secondly, it summarizes the ways to improve the regulation capacity of distribution networks and improve the consumption capacity of distribution networks for distributed PV through the introduction of energy storage equipment, flexible loads, and other means. Finally, advanced control strategies are summarized to improve the stability and reliability of the distribution network by implementing measures such as power prediction, voltage control, and protection coordination for distributed PV. Overall, this paper synthesizes the research on distribution network consumption strategies containing distributed PV from various aspects, which provides certain theoretical guidance and practical reference for solving distributed PV consumption problems.

1. Introduction

With the proposal of a “double carbon” target [1], new energy generation technology has set off a new wave of development. The use of renewable energy to replace fossil energy with large carbon emissions is one of the important ways to realize sustainable development. In this context, photovoltaic (PV) power generation technology has been developed in an unprecedentedly superior way. Solar energy is a clean energy source that can be utilized in a sustainable manner, and although the non-recyclability of PV panels is a potential pollution, the overall damage to the ecological environment is weak. In addition, solar energy is characterized by abundant reserves, a good economy, and a long supply time, so it has been highly valued in the development of clean energy. China is a vast country with abundant solar energy resources [2], and according to the data shown in Figure 1, it can be seen that the installed capacity of photovoltaics in China has achieved significant growth.

Not only that but in addition to China, other countries are also vigorously developing solar energy resources; according to the data released by the European Photovoltaic Association, in 2022, the EU 27 countries’ new PV installed 41.4 GW, an increase of nearly 50%, optimistic that the annual installed capacity of PV is expected to be close to 120 GW in 2026. The European countries’ PV installed situation is shown in Figure 2, 2022e; the figure represents the additional capacity in 2022. According to Fitch and Brazil’s Ministry of Mines and Energy statistics, Brazil’s total installed capacity of PV in 2022 reached 22 GW, with a new capacity of 9.0 GW and a new installed capacity year-on-year growth of 73.3%. According to research data from JMK, an Indian PV consultancy, 13.96 GW of solar PV systems were installed in India in 2022, a year-on-year increase of nearly 40%. This includes 11.3 GW of utility-scale PV, a year-on-year increase of about 47%, and some rooftop distributed PV and off-grid/distributed capacity. According to Fitch and the U.S. Energy Information Administration (EIA), in 2022, Japan’s installed PV capacity reached 77.6 GW, up 4.4% year-on-year, with 3.1 GW of new PV capacity. As major economies strive to develop PV, it is worth noting that the U.S. PV market is one of the few to see a recession.

With the continuous development of photovoltaic (PV) power generation, solving the problem of distribution grid consumption [3] containing distributed PV has become a key link. In this paper, we will discuss the main technologies and strategies for PV consumption. This includes distributed PV power generation, energy storage technology, microgrids, load-side management, and many other methods. These technologies and strategies have their own advantages and limitations in different application scenarios and need to be selected and combined according to the actual situation. This paper will provide useful insights and references for researchers and practitioners in related fields by reviewing and analyzing PV consumption.

2. Principle of Photovoltaic Power Generation and Consumption Mode

2.1. Principle of Photovoltaic Power Generation

The principle of photovoltaic (PV) power generation refers to the process of directly converting solar energy into electricity by utilizing the photovoltaic effect of semiconductor materials (e.g., silicon, germanium, etc.) [4]. The most important original in this process is the photovoltaic cell; when the sunlight irradiates the surface of the photovoltaic cell, the photons therein excite the electrons in the semiconductor, causing them to jump into the conduction band, forming free electron and hole pairs. These free electrons and holes combine near the P–N junction to form an electric current, which generates electricity. Figure 3 shows the equivalent circuit of a photovoltaic cell [5]. A complete photovoltaic power generation system consists of photovoltaic modules, racking, inverter, monitoring system, etc. The inverter converts the DC power to AC power to meet the demand of the grid or the user. The monitoring system is responsible for real-time monitoring of the operational status of the system to ensure power generation efficiency and safety.

As can be seen from Figure 3, the main components of a PV cell are the current source, the external load, and so on. In practice, a single photovoltaic cell is not enough to be utilized, so it is necessary to form a photovoltaic array by series or parallel connection to obtain a large enough output power.

The output characteristics of PV cells are generally divided into several parts: maximum power point, peak power, open circuit voltage, short circuit current, etc. The output characteristic curves of PV cells are shown in Figure 4 and Figure 5, which are shown in Figure 4 P–U characteristic curves with light and Figure 5 P–U characteristic curves with temperature, respectively [6].

2.2. Photovoltaic Power Generation Systems and Their Consumption Methods

2.2.1. Introduction to Photovoltaic Power Generation System

The importance of photovoltaic (PV) power systems in the renewable energy sector cannot be overstated. It is a zero-emission way of obtaining energy, which helps to reduce dependence on fossil fuels, lower greenhouse gas emissions, and combat climate change. In practice, PV power generation systems are not only widely used to generate power on the rooftops of residential and commercial buildings but are also utilized in the construction of large-scale PV power plants to provide large-scale clean energy for the public power grid. Photovoltaic power generation systems are mainly composed of photovoltaic modules, controllers, inverters, batteries, and other accessories (grid-connected does not require batteries). According to whether it depends on the public grid or not, it is divided into off-grid and grid-connected, in which the off-grid system is operated independently and does not need to depend on the grid [7]. The standalone system is shown as A in Figure 6, and the grid-connected system is shown as B in Figure 6.

2.2.2. Consumption Methods for Distributed PV Access to Distribution Grids

At this stage, for distributed PV access to the distribution network, the main elimination methods are as follows:

First, grid planning, optimization of access methods [8], and the capacity accessed in advance to do a good job of calculation can be based on the results of load forecasting or transformer and other capacity planning.

The second is energy storage technology [9], by storing the power generated by distributed PV systems, which can be used at night or low load to store the power generated by PV power generation through batteries or other energy storage devices for emergency use. This approach can improve energy utilization, but the cost is high, and the life and maintenance of the energy storage equipment are issues to be considered.

Third, smart grid technology [10]: smart grid technology can realize real-time monitoring and management of distributed PV systems, as well as dynamic adjustment of grid load.

Fourth, microgrid technology [11]: microgrid technology can combine distributed PV systems with traditional power grids to form a small independent power system to meet the energy demand in the region. This approach can improve the self-sufficiency and reliability of distributed PV systems, and also reduce the dependence on the traditional power grid.

Subsequent chapters of this paper will also focus on the above aspects to explore the consumption methods of distribution grids containing distributed PV.

3. Distribution Grid Consumption Strategy with Distributed PV

3.1. Study on the Strategy of Siting and Capacity Determination of Distribution Grids Containing Distributed PV

3.1.1. Objective Function

Consumption refers to the process of delivering the electricity generated by distributed PV power systems to the grid and coordinating its operation with other power systems. Distributed PV siting and sizing is the process of determining the size and layout of distributed PV power generation systems by taking into account various factors, such as topography, climate, building shading, and electricity load, in order to select the appropriate location and capacity for installation [12]. In practical application, distributed PV siting and capacity sizing can effectively solve the problem of consumption. By reasonably selecting the location and capacity size and coordinating the operation with the power grid, it can ensure that the distributed PV power generation system can operate in coordination with the power grid to avoid problems such as voltage fluctuations and frequency changes so as to guarantee the safe and stable operation of the power grid.

Before determining the appropriate distributed PV siting and capacity model, it is necessary to plan the objective function and constraints to ensure that the final optimization model is more in line with the actual, commonly used objective function as follows.

(1) Node voltage deviation: reflecting the voltage distribution before and after adaptive reactive power control, the lower the deviation value indicates that the voltage distribution is more reasonable.

$$f={{\displaystyle \sum _{i=1}^N(\frac{U_i-U_{ref}}{U_{i\max }-U_{i\min }})}}^2$$

(1)

where f is the sought voltage deviation, U_i and U_ref are the actual measured and reference values of voltage at node i, and N is the set of load nodes.

(2) Power loss: power loss reduction index.

$$f\prime =P_{Loss}-P_{Loss}\prime$$

(2)

where f′ is taken to its minimum value and is the best choice to reduce power loss, the equation represents the total power after access minus the total power before access.

(3) Minimum active network loss and minimum voltage deviation.

$$\left\{\begin{array}{l}\min f_1={\displaystyle \sum _{i=1}^N{R}_i\frac{P_i^2+Q_i^2}{U_i^2}{k}_i}\\ \min f_2={\displaystyle \sum _{t=1}^k\frac{(U_{ts}-U_{tN}{)}^2}{{U_{tN}}^2}}\end{array}\right.$$

(3)

where N is the total number of branches; P_i and Q_i are the active and reactive power of the ith branch, respectively; U_i is the voltage of the terminal node of the ith branch; R_i is the resistance of the ith branch; k_i is the switching state of the ith branch, with 0 representing the disconnection and 1 representing the closure; f₁ is the total active loss; t is the node number; k is the total number of the nodes; U_ts and U_tN denote the actual and rated voltage of the ith node, respectively; and f₂ denotes the voltage offset.

In addition to the above typical objective functions, more objective functions can be added in the subsequent optimization process to make the results more realistic.

3.1.2. Constraints

(1) Current constraints

$$\Delta P=P_{Gi}-V_i{\displaystyle \sum _{j\in i}{V}_j(G_{ij}\cos {\theta }_{ij}+B_{ij}\sin {\theta }_{ij})}=0$$

(4)

$$\Delta Q=Q_{Gi}-V_i{\displaystyle \sum _{j\in i}{V}_j(G_{ij}\sin {\theta }_{ij}-B_{ij}\cos {\theta }_{ij})}=0$$

(5)

where P_Gi and Q_Gi denote the active and reactive power injected by the distributed power source at the ith node, respectively; V_i denotes the voltage magnitude at the ith node, and G_ij, B_ij, and θ_ij denote the conductance, conductivity, and voltage phase angle difference between two nodes i and j, respectively.

(2) Node voltage constraints

$$U_{i\min }\le U_i\le U_{i\max }$$

(6)

where U_imin and U_imax are the minimum and maximum voltage amplitude allowed at node i, respectively; U_i is the node i voltage value.

(3) Branch current constraints

$$I_l\le I_{l\max }$$

(7)

where I_lmax is the upper limit of the branch l-carrying capacity.

The above constraints are also more typical constraints, and as many constraints as possible can be added when dealing with multi-objective problems to ensure that the model is more realistic to achieve the most stable access state.

3.1.3. Distributed PV Siting and Capacity Strategy

In order to rationally achieve the siting and sizing of distributed PV, most researchers have introduced intelligent algorithms for optimization. In order to improve the voltage quality of the system and reduce network losses, the GA-SA algorithm [13], which is a combination of genetic algorithm and simulated annealing algorithm, can be used, which can solve the distributed PV optimal allocation problem, but the objective function and constraints of the model are relatively homogeneous, and the conclusions obtained deviate from the actual results. The improved convergent multi-objective optimization algorithm [14] can be applied in practice; the researchers use the capacity limit as the main constraint to establish the siting and capacity model, which has obvious advantages in terms of network loss reduction rate and cost. However, the same problem is that the algorithm only has obvious advantages for single objective optimization.

In order to fit the reality further, the single-objective intelligent algorithm is no longer a good choice, and the multi-objective optimization needs to be further improved or combined with other algorithms. Optimization for multi-objective can be achieved by improving the non-dominated solution sorting genetic algorithm [15], which ensures that the harmonic distortion rate of each node does not exceed the limit and can effectively improve the system voltage distribution and reduce the system network loss. One of the problems is that the voltage and network loss targets determined by the model itself are too high.

In fact, in the current study of distributed power supply selection and capacity determination, the application of the particle swarm algorithm has been more mature, and many improved particle swarm algorithms [16,17,18,19,20,21,22] can obtain good optimization results, and some authors have introduced the Delphi method and fuzzy delta function to improve the algorithm’s weights, in order to enhance the reasonability and objectivity of the weights when selecting weights. Some authors also integrate the annealing principle into the optimization process, which can give the algorithm the ability to make sudden jumps, thus obtaining better optimization results. Based on the idea of interactive iterative nesting, author Fang Hui adopts a double-layer iterative hybrid particle swarm algorithm embedded in AC current calculation, which reduces the operational network loss of the distribution network system, improves the node voltage minimum, and reduces the possibility of voltage crossing the lower limit.

Overall, the particle swarm algorithm has been studied more maturely in the establishment of PV siting and capacity modeling, and because it is applicable to the multi-objective optimization problem, it has significant advantages for the network loss, voltage quality, harmonics and other problems in the distribution network. So, the authors mostly choose to carry out optimization of the particle swarm algorithm for PV siting and capacity determination; the basic flow of the particle swarm algorithm is shown in Figure 7 [23].

Particle swarm algorithm certainly has its superiority, but other intelligent algorithms also have unique advantages in siting and sizing. In the current case of particle swarm algorithms, research is more mature; for other intelligent algorithms, research innovation will be a new way of thinking in siting and sizing. The advantages and disadvantages of intelligent algorithms that have been applied to distributed photovoltaic and microgrid siting and sizing are shown in

Table 1. Comparison of advantages and disadvantages of intelligent algorithms for site selection and capacity determination.

Algorithm Name	Sophistication	Accurate	Algorithm Advantages	Algorithm Disadvantages
Improvement of GA-SA optimization algorithm [24]	High	Moderate	Adaptable, parallel capability, simple algorithm.	High complexity.
Particle swarm optimization algorithm [25,26]	Low	High	Algorithm is general, has few parameters, a simple principle, and a fast convergence speed.	Easy to fall into local extreme points.
Optimization algorithm for magnetotropic bacteria [27]	High	Moderate	Easy to implement, with few parameters, and applicable to multidimensional non-linear problems.	Complex model with uncertainties.
Improved genetic algorithms [28]	Medium	Higher	Introduces non-dominated sorting, high computational efficiency, and high diversity of populations.	Computationally time-consuming, loss of satisfactory solutions, need to specify the shared radius.
Pollen pollination optimization algorithm [29]	Low	Moderate	Strong exploitation, fast search, and continuous optimization.	Prone to local optimization and shared radius needs to be specified.
Butterfly optimization algorithm [30]	Low	High	Easy to implement, good convergence, and applicable to various types of optimization problems.	Convergence stagnation phenomenon and low processing efficiency.
Firefly optimization algorithm [31]	Medium	Higher	Can optimize single-peak and multi-peak functions, strong local search capability, simple implementation, and few parameters.	Strong dependency and easy to oscillate.
Optimization algorithm for manta ray foraging [32]	High	High	Strong search capability, good solution accuracy, and robustness.	Insufficient post-search capability and insufficient interval localization accuracy.
Ant colony optimization algorithm [33]	Moderate	Moderate	Easy to implement, can handle complex problems, non-linear problems, and large-scale problems, and can adaptively adjust parameters.	Slow convergence, premature convergence, and “stagnation”.

.

Nowadays, the research on intelligent algorithms is very rapid, and innovative algorithms are constantly being proposed, like the chemotaxis bacterial optimization algorithm, butterfly optimization algorithm, and firefly optimization algorithm mentioned in Table 1 are new meta-heuristic algorithms, which, compared to other traditional algorithms, have obvious optimization improvements in convergence, feature extraction, clustering, and other issues, so they have a relatively wide range of applications at this stage but are less widely used in the distributed PV siting and capacity determination, and the siting and capacity determination of microgrid operation and maintenance construction are less applied, which provides some algorithmic innovations for distributed PV consumption in the direction of capacity determination. In addition, the manta ray foraging optimization algorithm [34] mentioned in Table 1 is a new type of intelligent bionic population algorithm, which was only proposed by Zhao et al. in 2019 [23], which mimics the foraging process of manta rays in the ocean, mathematically models different feeding strategies, and mathematically describes the way of position updating of manta rays’ individuals, so as to achieve the search for the optimal solution in the complex solution space. Due to the uniqueness of the position updating method, the solution accuracy and robustness of the algorithm are also significantly improved compared with the traditional population intelligence bionic algorithm. The algorithm has been applied in the siting and capacity determination of distributed power supply, but due to the novelty of the algorithm, there are still a lot of places that can be innovated, which also provides an idea for the research in this direction.

The aforementioned PV siting and capacity-setting strategies have been extensively studied in China, and similar studies are being conducted in various countries around the world, especially in European countries where PV development is more advanced. In Germany, the European PV giant, the impact of PV plant siting on the existence of power system flexibility requirements has been studied [35], which shows that the location of the PV system significantly affects the amount of reserve required, and distributed PV systems can relatively reduce this requirement compared to centralized PV systems so decentralized siting is particularly important. In addition, during the access process, PV can cause a little shock to the grid. In order to improve the PV carrying capacity of the low-voltage grid, it is proposed to apply a voltage sag control method [36], which improves the effective control of the transformer during PV power generation, which can increase the accommodating capacity of the grid and eliminates the need for centralized controllers. In Germany, small PV systems are widely used in homes. How to design their optimal size and select the optimal capacity is a question that deserves to be investigated; MPPT charge controllers are a promising solution [37]. This approach can effectively reduce the cost of the system and assist in selecting the optimal capacity. MPPT charge controllers can be used in the selection of PV for a single household, but for collective residential buildings, it is not possible to establish the size of the PV in this way. At this point, an integrated solar system combining solar PV with collector heat storage has been proposed [38] to allow the subsystems to operate in concert to meet the building loads while exploring the power-to-heat process. A ray-tracing method for spatial and temporal calculation of large-scale PV capacity factors has also been proposed [39], which has the advantage of determining PV capacity factors for different environments as a basis for capacity establishment. This method not only determines the capacity of a city but also extends the environment to the whole world.

PV siting is also an important aspect of Spanish PV research. In order to improve the voltage quality and power loss, the active network loss and voltage drop can be established as the objective function, and the particle swarm algorithm can be used to solve the access capacity [40]. The results show that this method can significantly reduce the power loss and alleviate voltage drop throughout the whole network, which can provide guidance for the actual site selection. There are more solution algorithms available today, and there are numerous innovative ways to use them. Some authors have used the modified Jaya algorithm to find the optimal PV capacity and bus location [41] and proposed a new discriminative method to meet the requirements of improving voltage distribution and reducing network losses at high penetration rates. Some scholars zoomed out to the whole world to evaluate the combination of technology and geographic distribution [42], which ultimately resulted in the most environmentally friendly strategy, identifying Kosovo, South Africa, and Australia as the best locations for PV system installation. For building PV, multi-scale roof characterization using LiDAR data and aerial orthophotos [43] allows for rapid calculation of PV capacity. Combining LiDAR and orthoimage data processing for geometrically characterizing roofs at the slope level and calculating their PV solar potential, the method is a product of the combination of new energy and control fields. A new approach for off-grid PV DC microgrids has been proposed by T. Castillo et al. [44]. By comparing some of the existing methods, the proposed method is highly adaptable, allowing the design of installations and the ability to adjust them after the fact for new buildings.

Among the experimental and numerical analyses of the retrofitting of photovoltaic systems in Polish cities [45], there are results that show that changing the direction of solar radiation can increase the annual energy production by about 20–45%, effectively reducing the operating costs of buildings, as well as having environmental benefits. The possibility of transition from centralized to distributed energy systems in large Polish cities has been analyzed [46], and the results show that the application of micro PV systems will contribute to more energy-efficient residential buildings and a significant reduction of fossil-fuel-based energy consumption and that the development of micro PV systems seems to be one of the most effective options for the rapid transition from centralized to distributed energy systems based on individual renewable energy sources in large Polish cities. One of them. A greedy algorithm based on voltage deviation minimization was proposed to determine the appropriate PV storage capacity [47]. The results show that the level of PV capacity can be changed by time series analysis. Furthermore, incorrect capacity and location selection may lead to power quality deterioration problems in high-penetration power systems.

By analyzing the PV connected to the Dutch access power system, it was concluded that the magnitude of voltage fluctuations depends on the location of the PV in the grid, the capacity, and the grid configuration [48]. In order to solve this problem, researchers have proposed the introduction of electric vehicle technology to reduce voltage fluctuations, which is effective but raises the cost of charging for electric vehicle owners. An optimization method based on interwoven thermodynamic building models with genetic algorithms has been utilized in the Netherlands [49]. This method can effectively reduce the energy loss at the access point, increase the PV access capacity, and also reduce the cost of microgrid users, which has a good economy.

PV systems are also being studied in other European countries. In France, someone combined the PV carrying capacity of low-voltage grids with user voltage sensitivity and reliability analysis [50]. A voltage-based approach was proposed, and by analyzing the voltage reliability, it can be seen that when the power supply network is heavily loaded, the reliability indicators all improve with the increase in PV penetration, and conversely, when the power supply network is lightly loaded, the reliability indicators deteriorate instead. Italian researchers, on the other hand, in combination with their own water resources, have studied solar PV systems adapted to Mediterranean islands [51], using GIS to deal with the technical characteristics of each location and to calculate the potential PV capacity and the corresponding PV electricity production. In Sweden, some scholars are evaluating solar PV for rural electrification in the north of Ghana [52], concluding that PV is not the best option for the local population due to wiring problems that affect the economy.

At this stage, the research of distributed photovoltaic siting and capacity is a broader topic involving a number of fields. Among them, the research on siting and capacity optimization based on intelligent algorithms is the most extensive and the most popular research direction. Through the introduction of intelligent algorithms, the process of distributed photovoltaic system access to the distribution network is optimized and controlled in order to find the optimal siting and capacity solution. In fact, based on mathematical modeling and simulation technology, siting and capacitation research can also be carried out, and the combination of various models and cross-research will be the focus of the development of this field in the future.

3.2. Optimization of Distribution Grid Energy Storage with Distributed Photovoltaics

With the growing global demand for sustainable energy, solar photovoltaic (PV) power generation has become an important source of clean energy. However, due to its intermittency and instability, PV power generation often needs to be consumed and optimized for use through energy storage technologies. This chapter will provide a systematic summary of the current status of distributed PV using energy storage for energy consumption.

3.2.1. Distributed PV Energy Storage Methods

The distributed photovoltaic energy storage method refers to the combination of a photovoltaic power generation system and energy storage equipment to achieve optimal management of the power system and stable power supply. The distributed photovoltaic energy storage system access location is flexible, mainly in the medium- and low-voltage distribution network, microgrid, and user excess power into the power supply network. Reasonable planning of distributed energy storage, not only through the “peak shaving to fill in the valley”, plays a role in reducing the capacity of the distribution network but also in making up for the distributed randomness of the grid security and economic operation of the impact. An energy storage control flow based on solar photovoltaic power generation control is shown in Figure 8 [53].

The study of distributed PV energy storage systems mainly focuses on the above methods.

Table 2. Comparison of advantages and disadvantages of common distributed PV energy storage methods.

Energy Storage Method	Advantages	Drawbacks
Lithium-ion battery energy storage systems [54]	High energy density, large energy stored per unit volume, long cycle life, can be charged and discharged many times, small self-discharge rate, no self-discharge phenomenon, no memory effect, can be charged at any time, green and environmentally friendly, does not contain lead, mercury, and other harmful substances.	The production cost is slightly higher; although the energy density of lithium-ion batteries is high, its performance will be reduced in high-temperature environments. The cycle life of lithium-ion batteries is relatively short.
Compressed air energy storage systems [55]	Fast start-up time (<15 min), high energy and power density, black start capability, large energy storage capacity, low cost, long life, safety, and environmental protection.	Energy storage density is relatively low; energy storage efficiency is low, generally around 70%; energy storage equipment is large and needs to occupy more space.
Supercapacitor energy storage system [56]	Fast charging speed, long discharge time, small size and lightweight, easy to install and transport, long life, can be used hundreds of thousands of times, high power density, can provide high power output.	Energy storage density is relatively low; energy storage efficiency is low, generally around 70%; energy storage equipment is large and needs to occupy more space.
Liquid-flow battery energy storage system [57]	Ultra-long cycle life, high safety and stability, green, can be deeply discharged.	Batteries are too large, ambient temperature requirements are too high, and more expensive and complex systems.
Hydrogen storage systems [58]	Hydrogen storage is clean, efficient, and flexible in production; high density and large storage scale; small cost, which can effectively improve energy utilization.	Low energy conversion efficiency, high investment costs, and higher energy losses than other commonly used energy storage technologies.
Pumped storage systems [59,60]	Mature and reliable technology, large energy storage capacity, strong regulation capability, strong economy, long life cycle.	High construction costs, geographic constraints, ecological impacts, and dependence on water resources.

is a comparison of the advantages and disadvantages of several common distributed PV energy storage methods.

The current research on energy storage technology is mainly divided into the following directions. No equalization management topology photovoltaic lithium battery energy storage system [61], to solve the traditional design process of lithium battery equalization management, single overshoot problem, enhance the safety of the system, the system can be uploaded through the controller and then the prediction of the relevant data, but only theoretically feasible, there is no actual prediction with the counterpart. For the problem of slow response speed of compressed air energy storage, a virtual inertia coordinated control strategy between different optical storage units is proposed [62], which combines compressed air energy storage and supercapacitors, greatly improving the response speed of the system and also accurately realizing the crossover frequency control and so on.

The hybrid energy storage system based on a supercapacitor [63] suppresses the voltage fluctuation when distributed PV is connected to the distribution network, which is of practical significance for the consumption of distributed PV. However, the shortcoming is that only a PV plus an energy storage system is studied, and whether the energy storage system can be applied for multiple power supply configurations remains to be studied. The liquid-flow battery energy storage optimization configuration method [64], which is applicable to PV systems, can suppress the intermittency and volatility of the power generation system through rational planning and configuration. The method pin mainly on the economic problems of the energy storage system and the configuration of the energy storage system in the local area network (LAN), which improves the economy of the LAN power supply and the new energy consumption rate.

Hydrogen energy storage is a new type of energy storage technology, which is based on the principle of converting electric energy into hydrogen energy and then generating electricity by burning hydrogen. For the study of distribution network systems containing hydrogen energy storage [65], nine working mode judgments can be proposed, through which hydrogen energy storage can be utilized to participate in peak shaving and valley filling of the distribution network and to improve the consumption capacity of the distribution network. However, there are many simplifications in the calculation process of this study, and the actual conclusions are still debatable, but hydrogen energy storage has a good development prospect for photovoltaic distribution systems.

As an emerging technology, photovoltaic energy storage has great application prospects and development potential. With the continuous development of technology and cost reduction, it is believed that PV energy storage will become a very important research direction in the future renewable energy field.

3.2.2. Energy Storage Optimization for Distributed PV

The optimization of energy storage for distributed PV is also based on a variety of intelligent algorithms, and the intelligent algorithms applied are roughly the same as those in the siting and capacity determination part of distributed PV, but when optimizing the energy storage, the model mostly adopts a two-layer structure, and in the case of determining the optimal configuration and operation strategy of the energy storage, the economic issues also need to be considered, so in the literature, the economic constraints are usually considered as a layer. In the literature, economic constraints are usually integrated into the energy storage model as a separate optimization.

A multi-objective model can be established for economy and environmental protection to maximize the consumption of distributed PV [66], and the lower layer considers the friendly interaction of distributed energy storage devices to establish the daily net load deviation per unit of contact line power, which is used to establish a layered optimal allocation method for distributed energy storage in distribution networks. A distributed PV community energy-sharing optimization strategy based on a two-tier structure can also be proposed [67], where the upper tier of the strategy is operated for the energy storage price using the master–slave game of multi-community shared centralized energy storage, and the lower tier achieves the improvement of factors such as net output magnitude and net output correlation by improving the Shapley-valued distributed PV community inter-subjects’ energy-sharing revenue distribution mechanism. It is also a strategy to merge the two models under the premise of high-penetration rate [68] by considering the energy storage equipment using second-order cone relaxation to establish a system day-ahead scheduling linearization model and then based on the CVaR risk assessment model to establish a try scheduling model with an optimized loss function. Finally, the two models are merged to establish a day-ahead real-time multi-timescale scheduling model for regional integrated energy systems.

The establishment of a two-layer model for energy storage optimization is a more comprehensive approach, which can not only consider important factors such as the location and capacity of energy storage but also focus on the economic issues because the current stage of energy storage technology has just started, so the consideration of the cost of energy storage is the key to the success of a project. In this paper, by summarizing the model of the literature [69], the two-layer structure of the literature is drawn, as shown in Figure 9.

The process shown in Figure 9 is similar to most two-layer model structures, which can visualize the process of building a two-layer model and can provide a basic idea for readers who build a two-layer structure for research.

In addition to the establishment of the two-layer structure, there are also studies that only apply intelligent algorithms for optimization. By improving the particle swarm algorithm [70,71], the hybrid planning model for the optimal allocation of photovoltaic and energy storage is established by comprehensively considering the economic benefits, peak shaving, and valley filling issues. In order to improve the global search ability, the firefly algorithm [72] can be improved to quickly determine the optimal energy storage capacity.

Incorporating energy storage devices for distributed PV increases the construction cost but is relatively the most effective way to consume PV at present, both in terms of reducing the impact on the grid when PV is connected and increasing the capacity of the PV to be connected to the grid, so PV energy storage has been widely studied around the world. Standalone PV energy storage systems with positive pulse current battery charging have been studied in India [73], which can charge battery-powered road vehicles and help to reduce the future burden on the grid. Hybrid energy storage standalone solar PV system energy coordination management strategy [74], also an in-depth study of PV energy storage, this control strategy optimizes the flow rate of the storage system as well as the energy system of the discharge and charging cycles using the meta-heuristic Jaya algorithm to properly coordinate the PVs and ultracapacitors. Indian researchers verified the feasibility of a photovoltaic flywheel energy storage system [75], which improves the conversion efficiency of the solar converter and protects the loads from possible power outages and limitations, thus increasing the reliability of the power supplied to the loads. Pradosh Kumar Sharma et al. investigated an energy storage system based on wind and solar complementary technology [76] with the objective of utilizing the available wind and solar resources to provide electricity. The system is not as economical as the existing system, but it has a significant improvement in the financial deficit caused by the existing system; not only can it provide a stable and efficient power supply, but it also increases the access capacity of photovoltaic and wind power.

In Libya, PV energy storage is also one of the most popular research studies. Ali O.M. Maka et al. studied the performance of solar PV power generation systems with integrated battery storage [77], which showed that energy efficiency can be improved by using PV systems with integrated battery storage to obtain higher performance and availability. PV energy storage has also been studied in Thailand, where Oluwaseun Olanrewaju Akinte et al. investigated the energy storage management of a hybrid solar PV–biomass power system [78]. The study showed that the system can increase the efficiency of energy use, effectively facilitate storage management, minimize energy losses, and improve the lifetime of the microgrid network. Malasian scholars proposed an energy management and capacity planning for a PV–wind–biomass system considering hydrogen battery storage for hydrogen battery energy storage system [79]; this planning applied an SSA optimization algorithm to determine the optimal configuration of the system and the reliable technology, which also meets the economic aspects. Australian scholars proposed a control strategy for community battery energy storage systems in high PV penetration microgrids grid-connected [80], which does not require the prediction of net power and reduces the required battery capacity by about 16%. Thus, the scheme is more robust when the community battery takes on the power buffering task. Luccas Tadeu Farnezes Soares et al. from Brazil designed an auxiliary-side-based PV energy storage system using the functionality of a PV inverter [81], which helps to mitigate disturbances related to power quality, improve the difference between charging and discharging control and thus maintain the lifetime of the storage system. Alireza Kermani et al. from Norway investigated the optimal capacity and economic performance of a microgrid based on PV and battery systems [82]. The results showed that the method achieves optimal energy management in grid-connected mode, minimizes the power exchange between the microgrid and the grid, reduces energy costs, and improves the efficiency of PV power generation.

In conclusion, photovoltaic energy storage, as an emerging technology, has great application prospects and development potential. With the continuous development of technology and cost reduction, it is believed that PV energy storage will become a very important research direction in the future renewable energy field. The main prospects for it are as follows.

(1)

Continued technological innovation: more efficient, low-cost, long-life energy storage technologies are expected to be introduced, such as solid-state batteries, metal-air batteries, and so on.

(2)

Further market expansion: with the further decline in the cost of energy storage and continued policy support, the market size of PV energy storage systems is expected to expand further.

(3)

Intelligent management enhancement: through integration with technologies such as Internet of Things, big data analysis and artificial intelligence, the management of PV energy storage systems will become more intelligent and improve energy utilization efficiency.

(4)

Distributed power generation and micro-grid development: PV energy storage will play an important role in the construction of distributed power generation and micro-grid, enhancing the stability and flexibility of the grid.

(5)

Significant environmental benefits: with the popularization of PV energy storage technology, it can reduce dependence on fossil fuels, help reduce greenhouse gas emissions, and promote environmental protection.

(6)

Convergence of electric vehicle charging network: the combination of PV energy storage systems and electric vehicle charging infrastructure will be one of the future development trends, realizing the complementarity of green transportation and energy.

3.3. Voltage Control Strategies for Distribution Grids Containing Distributed Photovoltaics

3.3.1. Voltage Control using PV Inverters

Distributed PV power access will change the grid structure, and when high-penetration PV power access is provided, it may exceed the dissipation capacity of the distribution grid, resulting in the active and reactive power balance of the power system being broken, which in turn affects the stability of the grid voltage and frequency. Therefore, the voltage overrun problem of distributed PV access also belongs to the consumption problem.

The PV inverter is one of the very key devices in the grid-connected PV system, which can realize the inverter function of DC power through the control of power electronic switches, thus ensuring that the PV power meets the requirements of grid-connected. Some inverters also integrate functions such as maximum power point tracking to further improve the reliability and stability of the grid-connected system [83]. The topology of one of the PV inverters is shown in Figure 10 [84].

In view of the key role of PV inverters, it is possible to formulate a scheme for PV inverters to participate in the reactive voltage control of the grid [85,86,87], utilizing the principle of absorbing and releasing reactive power by PV inverters, using external inverters to achieve filtering function and joining reactive power compensation devices to achieve the absorption of reactive power, which effectively improves the quality of electric power. The PV inverter is used to realize the regulation and control of the grid-connected voltage, and the distributed PV voltage coordination control strategy for traction power supply is proposed considering the reactive power capacity margin of the inverter, which makes full use of the PV’s own voltage regulation capability. According to the principle of variable sag control combined with the energy storage system, the voltage stratification control method of PV plus energy storage can be designed. All of the above can make full use of the function of the PV inverter itself, but the common problem is that the inverter extremely easily fails a device; in the case of maximizing the use of its own function, design the corresponding inverter protection methods.

In addition to the voltage control function of the PV inverter itself, the PV inverter also has the function of maximum power point tracking (referred to as MPPT) in order to obtain the maximum power output in a variety of situations. At this stage, there are many research methods for the MPPT, but the tracking accuracy is still to be improved because the accuracy can not be improved, so the power of the PV system has the same characteristics as that of the PV system. Because the accuracy cannot be improved, the power of the PV system has a considerable loss. Therefore, the MPPT technology for PV systems should be studied in depth in the future. The structure of the PV MPPT system is shown in Figure 11 [88].

In order to improve the accuracy of PV MPPT, various researchers have also carried out optimization studies from various aspects, and the theoretical accuracy is indeed significantly improved, but whether it can be applied to the actual is still to be tested.

Through intelligent algorithms [89,90,91,92] to improve the accuracy of tracking, the application of intelligent algorithms is mostly similar to those summarized in energy storage; various types of intelligent algorithms for different defects, after the optimization of algorithms, have to reduce the search range, there is a change in the perturbation mechanism, and also to improve the dynamic performance of the optimization of the PV MPPT from a variety of perspectives to provide ideas. For example, the proposed MPPT algorithm based on multivariate universe optimization [93], which mixes the PV–temperature difference system and puts forward five kinds of arithmetic examples for verification, the verification results are relatively good but limited by hardware experiments and did not verify the practical feasibility, but the method is relatively novel, breaking through the limitations of previous intelligent algorithms, and providing a new way of thinking for other subsequent research.

With the development of deep learning in recent years, many scholars have focused on deep learning algorithms and proposed MPPT control of PV systems based on expert demonstration of deep reinforcement learning [94], which gives the experience of the traditional method directly to the experience pool of the algorithm and solves the problem of low dynamic efficiency in complex environments. A reinforcement learning-based MPPT control method for PV power generation systems is also proposed in [95], and the algorithm is verified through simulation to enable the system to find the maximum power point quickly. However, the deep learning algorithm gives the same question: in the process of simulation optimization, its accuracy can be more than ninety-five percent, the environmental problems are not controllable, and how to solve the uncertainty of the perturbation factors has to be studied.

Traditional photovoltaic MPPT optimization methods continue to be studied, such as the constant voltage tracking method, conductivity increment method, and interference observation method, etc., for the traditional methods in this paper will not do too much detail. In this paper, the advantages and disadvantages of each algorithm are summarized in

Table 3. Summary of advantages and disadvantages of common PV MPPT control methods.

Control Methods	Complexity Theory	Accurate	Advantages	Drawbacks
Constant pressure tracking method [96]	Lower (one’s head)	Lower (one’s head)	Simple to implement and low complexity.	Inability to adapt to environmental changes and low tracking efficiency.
Conductivity increment method [97,98]	Conveniently situated	Conveniently situated	Simple structure and easy to implement.	High hardware requirements and difficult step-size selection.
Perturbation observation method [99]	Relatively low	Lower (one’s head)	High tracking accuracy, easy to realize, and simple control structure.	Maximum point fluctuation is large, and misjudgment occurs.
Intelligent algorithm [100,101,102]	Your (honorific)	Your (honorific)	Good accuracy and good adaptability to disturbances.	Theoretically feasible but not proven in practice.
Deep learning algorithm [103]	Your (honorific)	Your (honorific)	Less computationally intensive, more accurate, and faster learning from experience.	Training samples are heterogeneous and need further optimization.
Fuzzy control [104]	Conveniently situated	High	Fast point-finding and anti-interference capability.	More computationally intensive and difficult to implement.
Complex algorithm [105]	Your (honorific)	Your (honorific)	High precision, fast response, and strong anti-interference ability.	High algorithmic complexity and more complex model building.

.

3.3.2. Voltage Coordination Control Methods

The voltage coordinated control method refers to determining the parameters such as the voltage of the middle bookstore, the tap position of the on-load voltage regulator transformer, and the input amount of the reactive power compensation device through optimization calculation under the premise of meeting the system operation constraints, so as to realize the minimization of active network loss. Common methods include active control, tap device control, and coordinated control of multiple devices. Among the existing studies, researchers have mostly adopted the voltage control methods of sub-area, sub-cluster, and joint with energy storage.

For example, the distribution network overvoltage responsibility sharing method with a high percentage of distributed PV access [106], whose central idea is to use the Shapley value method to calculate the percentage of overvoltage responsibility in the model, dividing the 33-node distribution network into five groups, and when overvoltage occurs, according to the different Shapley values, each group bears the corresponding overvoltage responsibility. There is also an energy storage voltage regulation control strategy based on cluster division [107], which firstly divides the distribution network into clusters and then, under the consideration of energy storage, allocates energy storage according to the cluster where the overrun node is located, and determines the power of the storage timing action at the time of the maximum gain, so as to achieve the suppression of voltage fluctuations. Joint energy storage has also been proposed to control the voltage [108], which first establishes the voltage stability model of the distributed photovoltaic storage system, then establishes the coordination model between PV and energy storage, and then achieves reasonable control of voltage.

In order to visualize the partition control of PV, this paper draws on the partition diagram shown in the literature [109] to show the reader, and the PV partition structure diagram is shown in Figure 12.

Figure 12 shows that partitioning is just an idea for subsequent studies, which can be performed both horizontally and across zones in combination with power electronic devices such as energy routers.

Some authors have utilized this similar horizontal zoning approach to propose a network partitioning and voltage coordination control method for distributed PV distribution networks containing high penetration [110], improving the electrical distance and intra- and inter-zone coupling metrics to achieve optimal zoning. Subsequently, voltage optimization is carried out with the objectives of node voltage deviation and network loss. Other authors utilize controllers for information transfer, and partitions are randomly divided into a number of them to solve the voltage overrun problem by adopting a cluster voltage control strategy to autonomously regulate the reactive power output from the distributed PV inverters while considering the scenario of communication deficiency [111]. Directions for future work on the voltage control strategy include further optimizing the design of the control system to improve the response speed and accuracy, exploring more intelligent and adaptive voltage control strategies to cope with complex and changing grid environments, and carrying out large-scale empirical studies to verify the effectiveness and feasibility of the strategy.

The above are some of the results of voltage control strategies by Chinese researchers; worldwide, voltage control strategies have been studied by a considerable number of people. Maneesh Kumar et al. proposed a robust decentralized voltage control strategy for remote microgrid systems [112], and overall, the proposed AFPID controller in this strategy performs better than the other methods in distributed voltage control. The literature [113] proposed an adaptive power ramp rate control method for photovoltaic (PV) power generation systems, which achieves fast dynamics and low power oscillation by adaptively calculating voltage steps. Experimental and simulation results confirm the performance of the method in various operating environments, and the method can be applied to large-scale PV power plants. The literature [114] designed a low-voltage ride-through control strategy for a solar PV grid-connected inverter; this scheme minimizes the harmonic and reactive power injected into the grid and also reduces the voltage oscillations at the DC bus. In addition, this control strategy provides dynamic control of active and reactive power.

In order to solve the problems of voltage fluctuations and voltage losses due to intermittent solar energy, Nasir Rehman et al. from India proposed a new local dynamic sag control strategy for providing reactive auxiliary support ports using decentralized [115]. Combined with a meta-heuristic reptile search algorithm technique, it greatly improves the rationality of the system voltage distribution and fully utilizes the capacity of the PV inverter. Tushar Waghmare et al., also in India, illustrated how to efficiently regulate higher-order sliding mode controllers in PV systems [116], and the proposed HOSM control strategy outperforms the conventional PI control technique in a perturbed environment, thus allowing the PV system to provide constant voltage to the power line. The literature [117] proposed an adaptive control framework for the flexible and effective management and control of clustered DC nanonetworks in islanded DC microgrid systems, where the adaptive model predicts the controller to provide better transient response and less current ripple.

Naser Vosoughi Kurdkandi et al. from the USA proposed a novel transformerless grid-connected inverter based on a switched capacitor [118]. The boosting capability was obtained by utilizing the series-parallel switching technique of switched capacitor modules. Higher overall efficiency and leakage current elimination capability were obtained by the common ground technique. French scholars Ehsan Jamshidpour et al. comparatively analyzed the equivalent circuits of two non-isolated Buck/Buck-Boost converters under synchronous control and applied them to a stand-alone PV battery-loaded system [119]. It was shown that the single switch converter has a lower size and cost, but its control strategy is limited. Sanjida Moury et al. from Canada proposed a multi-port converter architecture for PV cell systems [120]. In the proposed design, the battery charging and discharging circuits are integrated with the PV power optimizer through a high-frequency AC link, and the converter system is able to regulate the output voltage and use variable switching frequency for maximum power point tracking. The literature [121] proposes a novel power-sharing voltage regulation strategy based on solar PV power microgrids using advanced power management techniques for microgrid systems. This technique improves the stability of the distribution network based on voltage ride-through capability and optimizes the active and reactive currents of the distribution network.

Iranian scholars Fatemeh Jamshidi et al. proposed a backstepping sliding mode control method based on the Liapunov criterion [122]. The proposed method provides faster adaptation to changes in temperature and solar radiation and ensures faster convergence to the MPP. Finally, it is verified that the method has good robustness. Thai scholars Anuwat Chanhome et al. used a control strategy involving the coordination between the central control unit and local control functions of the PV system [123]. An adaptive forbidden search technique was applied to search for the optimal LCF parameters of the PV system and adjusted weekly. The optimization problem is to maximize the true power generation of all PV systems under varying sunlight and three-phase unbalanced loads.

By summarizing the study, it can be seen that the PV voltage control strategy can be mainly divided into the passive control strategy, which regulates the reactive power in the grid through the physical components, thus controlling the voltage level; the active control strategy utilizes advanced control systems and algorithms to dynamically adjust the reactive power in the grid in order to achieve a more accurate voltage control; and the inverter-based control strategy of three kinds. In addition, the different quasi-control locations are mainly divided into three ways, which are in situ, centralized, and distributed, and their specific concepts and advantages and disadvantages are shown in

Table 4. Comparison of advantages and disadvantages of voltage control schemes.

Control Methods	Economics	Robustness	Advantages	Drawbacks
In-situ [125,126]	Excellent	Higher	Fast response time, improved system stability, optimized energy use, reduced transmission losses, and enhanced grid resilience.	High-cost investment, high technical complexity, reliance on accurate measurement and control, potential compatibility issues, and high maintenance requirements.
Centralized [127]	Poor	Lower	Flexibility and stability in site selection and flexibility in operation.	Dependence on long-distance transmission and technical challenges.
Distributed [128]	Average	Very high	Improved energy efficiency, reduced losses, and enhanced system reliability and flexibility.	Voltage control complexity; potential impact on grid stability.

[124].

Directions for future work on voltage control strategy include further optimizing the design of the control system to improve response speed and accuracy, exploring more intelligent and adaptive voltage control strategies to cope with the complex and changing grid environment, and carrying out large-scale empirical studies to verify the effectiveness and feasibility of the strategy.

4. Conclusions

This study reviewed distribution grid consumption strategies containing distributed PV, aiming to provide a comprehensive perspective to understand and address the distribution grid operational challenges that arise with the increased penetration of distributed PV. Through a comprehensive analysis and discussion of the existing literature, this paper draws the following conclusions:

• The incorporation of distributed PV brings significant changes to the distribution grid, including shifts in energy generation patterns, increased complexity in grid operations, and challenges to traditional power system planning and operations management.
• Effective consumption strategies require comprehensive consideration of technical, economic, policy, and social factors. Optimizing the access mode of distributed PV, improving the regulation capability of the distribution grid, adopting advanced control strategies, formulating reasonable policies and incentive mechanisms, and strengthening technical research and development and personnel training are the keys to achieving effective consumption.
• Grid-connected power grids and energy storage systems are currently the main means of consumption. Grid connection can instantly consume a large amount of electricity, but it is necessary to solve the technical problems of grid access and the volatility of grid load. Energy storage systems can improve energy utilization but face cost, lifetime, and maintenance challenges.
• Future development trends indicate that, with advances in materials science and electronics technology, photovoltaic power generation technology will develop in the direction of high efficiency, low cost and long life. The development of smart grid and microgrid technologies will provide more possibilities for the consumption of PV power.
• In order to realize the sustainable development of distributed PV, it is necessary to adopt a comprehensive strategy of consumption and customized design in combination with region-specific environmental, economic, and social conditions. At the same time, the government, industry, and research organizations need to work together to promote policy formulation, technological innovation, and market mechanisms.

In summary, the distribution network consumption strategy with distributed PV is a multidimensional and multidisciplinary research field that requires continuous research and innovation to adapt to the changing technological and social needs. Comprehensive strategies can ensure that distribution grids maintain efficient, stable, and reliable operations while accommodating more distributed PV.