1. Research background and significance
In recent years, the progress and development of technology have led to an increasing demand for energy, further exacerbating the contradiction between energy consumption and environmental governance. The progress of the times has prompted the advancement and continuous updating of microgrid systems, which requires accelerating the replacement of basic and aging facilities, reducing power costs, and improving the stability of microgrid systems. The problem of global warming is becoming increasingly serious, and reducing greenhouse gas emissions can help alleviate the severe climate problem. Therefore, it is very necessary to save energy, reduce emissions, and promote the process of renewable energy. Currently, photovoltaic power generation technology is the optimal choice for sustainability and reliability. However, irradiance, temperature, humidity, and other factors are the main factors causing photovoltaic intermittency, and energy storage systems (BESS) are generally used for processing. Figure 1 shows the basic framework of the energy storage microgrid system.
The world is vigorously developing photovoltaic power generation. By the end of 2020, 20 countries have installed more than 1GW of photovoltaic power, with a cumulative global total installed capacity exceeding 760.4GW. Among them, China, the European Union, and the United States rank among the top three in terms of installed capacity, with 48.2GW, 19.6GW, and 19.2GW respectively. China will increase the proportion of non fossil fuels from 15.3% in 2019 to 16% in 2020, with plans to reach 25% by 2030. To compensate for the construction cycle of hydropower stations and the geographical limitations of nuclear power plants, the installed capacity of renewable energy sources such as photovoltaic and wind power will be significantly increased, with an average annual installed capacity of 130GW to 150GW, and is expected to reach around 1800GW to 2000GW by 2030.
Solar energy in China is mainly concentrated in the northwest region with low population density and underdeveloped economy. Due to the relatively backward development and utilization of renewable energy in this region, a large amount of solar energy will be abandoned. Compared to the eastern regions with rapid economic development, high population density, and high electricity consumption, the inability to lay large-scale photovoltaic power stations due to scarce land resources can lead to an imbalance between centralized power generation and high load electricity consumption. Renewable energy sources such as solar energy exhibit significant fluctuations and intermittency under the influence of external climate. Intermittent and fluctuating nature can cause varying degrees of impact on the large power grid, affecting its reliable operation, and directly or indirectly causing economic losses to the country and users. The above issues seriously constrain the development of renewable energy. To ensure the healthy development of green energy, it is necessary to allocate energy storage for the renewable energy system to improve the reliability and safety of the power grid.
Energy storage systems can provide sufficient power for the sustainable development of photovoltaics. Solar energy, energy storage equipment, and DC loads are gradually penetrating into the AC/DC microgrid system and becoming important links in the microgrid system. This will greatly meet the energy consumption in remote areas and the local energy consumption, making the planning and design of the power grid more flexible. Energy storage systems can solve the volatility and intermittency of photovoltaic systems, improve system stability, and enhance the flexible switching of microgrid systems and off grid. However, energy storage technology itself still faces a series of problems such as small scale system, short lifespan, and high cost. Therefore, reasonable optimization of capacity can solve the problems of renewable energy volatility and intermittency, as well as the cost of the system, and is also the key to the future development of microgrids.
At present, the configuration error between photovoltaic installed capacity and energy storage system capacity can lead to energy surplus and load loss. Secondly, high-power load fluctuations and photovoltaic fluctuations can affect the stability of bus voltage. Therefore, this article focuses on addressing the following issues:
1) Optimize the capacity of the renewable energy and energy storage ends of the energy storage microgrid system to reduce system costs, reduce energy surplus rate, and reduce load loss rate.
2) Perform composite frequency division coordination control on DC/DC energy storage converters to effectively suppress power fluctuations at the renewable energy power supply and load ends, and improve the reliability of DC bus voltage.
2. Current research status of energy storage systems
2.1 Current Status of Energy Storage Technology Research
The current energy storage technology is flourishing, mainly including flywheel energy storage, pumped energy storage, battery energy storage, and supercapacitor energy storage. According to its characteristics, it can be classified as physical energy storage and chemical energy storage.
1) Physical energy storage
(1) Pumped storage energy
During the low peak period of electricity consumption, hydropower stations use electricity to pump water upstream of the reservoir. During the high peak period of electricity consumption, they use the potential energy from the upstream to release water downstream for power generation, and use the remaining electricity from the low peak period to fill the gap in electricity consumption for storage. At present, the efficiency can reach between 70% and 85%.
(2) Flywheel energy storage
Flywheels are mostly made of carbon fiber materials. In the energy storage state, excess electrical energy drives the flywheel to store energy. In the non energy storage state, the flywheel drives the load to release energy. Generally, by reducing the weight of the rotor and increasing the angular velocity, the energy storage is increased. Flywheel energy storage, with its high power density and long service life characteristics, is widely used in power distribution systems and industries.
(3) Compressed air energy storage
The method of storing excess electrical energy by pressurizing and sealing the air in specific environments such as caves and gas storage wells for energy storage, and finally releasing compressed air to drive the generator to work to release energy.
2) Chemical energy storage
(1) Lead-acid battery
The positive electrode of lead-acid batteries is PbO2, the negative electrode is lead element, and H2SO4 serves as the electrolyte. This energy storage method has advantages such as stable discharge electromotive force, low environmental corrosion, high efficiency, low cost, and long cycle life. The technology is relatively mature and widely used in the market, but the disadvantage is that it is smaller than energy.
(2) Lithium ion batteries
The abundance of Li+in the positive and negative phases of lithium ions characterizes the different working states of batteries. When the negative polarity of lithium-ion is characterized by excessive lithium, the battery operates in a charged state; On the contrary, the positive polarity of lithium-ion batteries is characterized by excessive lithium, and the battery operates in discharge mode. This battery can be charged and discharged multiple times, with good charging and discharging performance, long service life, and high comprehensive performance, and belongs to a type of secondary battery. However, lithium-ion batteries pose a higher risk.
(3) Sodium sulfur battery
The positive electrode of a sodium sulfur battery is a molten mixture of S and Na, the negative electrode is Na, and the electrolyte is a compound composed of Al2O3 and Na2O. Sodium sulfur batteries have a smaller footprint and a longer lifespan, making them suitable for energy storage in renewable energy generation.
(4) Liquid flow battery
The characteristic of a liquid flow battery is to divide the electrolyte into two parts and react with the positive and negative electrodes respectively. Liquid flow batteries are currently widely used in commercial demonstration stages, with high energy density, long working life, and high safety. However, the power density of liquid flow batteries is low and the cycling efficiency is not high.
(5) Supercapacitor
A supercapacitor consists of a positive electrode, a negative electrode, and a diaphragm. The electrolyte is filled between the two electrodes and the pores, and energy is stored between the two electrodes without consuming energy during energy storage. Therefore, it can be reused tens of thousands of times. Supercapacitors have a high power density, low operation and maintenance costs, and can be recycled from tens of thousands to hundreds of thousands of times. They can also perform well in response to low temperature conditions. However, the current market cost is relatively high.
The main energy storage forms of the energy storage system are summarized in Table 1:
| Energy storage method | Energy density (Wh · L) | Power density (W · L) | Lifetime (years) | Cycle efficiency (%) | Operation and maintenance cost ($/kW/year) | Response time magnitude | Development stage |
| Pumped storage energy | 0.5-1.5 | 0.5-1.5 | 30-60 | 65-85 | 13.3 | ms | Mature |
| Flywheel energy storage | 20-80 | 5000-15000 | 15-20 | 95-98 | 20 | ms | Commercial |
| Compressed air energy storage | 30-60 0.5-2 40 75-95 19 min | 0.5-2 | 40 | 75-95 | 19 | min | R&D – Demonstration/Commercial |
| Lead-acid battery | 50-100 | 10-400 | 5-15 | 63-90 | 50 | ms | Mature |
| Lithium-ion battery | 200-500 | 1500-10000 | 5-20 | 90-97 | 0.46 | ms | Commercial |
| Sodium flow battery | 150-300 | 140-180 | 10-15 | 70-90 | 80 | ms | commercial |
| Liquid flow battery | 16-60 | 0.5-25 | 15-20 | 60-85 | 70 | ms | Demonstration – Commercial |
| Supercapacitor | 10-30 | 100000+ | 20 | 84-95 | 5-6 | ms | Commercial |
From the above table, it can be analyzed that compressed air energy storage has a longer response time, while pumped energy storage and flywheel energy storage are greatly affected by geographical location. Fluid flow batteries are still in the demonstration stage, and lithium-ion sodium sulfur batteries currently have a higher market cost. And lead-acid batteries have mature development and high energy density with low cost, while supercapacitors have high power density and long cycle life, which complement the characteristics of lead-acid batteries. Two types of energy storage devices can be selected as the energy storage part of renewable energy generation, to compensate for the drawbacks that single energy storage cannot balance.
2.2 Research status of non isolated DC/DC converter topology
1) Current research status of non isolated BDC (Bidirectional DC-DC Converter) topology. Currently, non isolated bidirectional DC converter topology structures are mostly obtained by combining Buck circuit, Boost circuit, Cuk circuit, Zeta circuit, Sepic circuit, etc. According to the difference in input and output polarity, it can be divided into homopolarity and antipolarity.
(1) Bidirectional Buck/Boost Converter
The structure of the bidirectional Buck/Boost converter is shown in Figure 2, operating in boost, buck, and alternating operating modes.
The bidirectional Buck/Boost converter is subject to the structural characteristics of the converter, with a small difference in input and output voltage, and is usually used in low-power and small volume applications. And the required switching devices are few, the cost is low, and the control strategy is relatively simple. The inductance current under three different working states: boost mode, buck mode, and alternating working mode is shown in Figure 3.
(2) Bidirectional Buck Boost Converter
The working mode of this converter is similar to that of a bidirectional Buck/Boost converter. The biggest feature is that the circuit topology is symmetrical from left to right, and the input and output voltage polarity is reversed. The absorption and release of energy within the working cycle of an inductor determines the voltage rise and fall working state of the converter. Switch tubes S1 and S2 cannot work simultaneously. The bidirectional Buck Boost converter is suitable for a wide range of voltage regulation and can achieve voltage up and down functions in the same direction. The topology of the bidirectional Buck Boost converter is shown in Figure 4.
(3) Bidirectional Cuk converter
The bidirectional Cuk converter is formed by a Cuk circuit with capacitor C3 as the center and symmetrical left and right, with the same input and output voltage polarity. The topology of the bidirectional Cuk converter is relatively complex, and it uses twice as many switching transistors and nonlinear components as other topologies, which increases the circuit cost. The energy in this topology structure needs to be converted through the capacitor dielectric C3, which is usually suitable for small and medium-sized power applications with low transmission efficiency. The topology of the bidirectional Cuk converter is shown in Figure 5.
(4) Bidirectional Zeta Sepic converter
The bidirectional DC-DC topology composed of bidirectional Zeta and bidirectional Sepic is the same, with the same polarity of input and output, hence it is called a bidirectional Zeta Sepic converter. The transmission characteristics are equivalent to the forward and reverse operating states of Zeta and Sepic converters, and the average value of inductance determines the operating mode of the converter. The bidirectional Zeta Sepic converter has a complex topology and is suitable for small to medium power applications with low transmission efficiency. The topology of the bidirectional Zeta Sepic converter is shown in Figure 6.
The characteristics of the non isolated DC-DC converter topology are shown in Tables 2.
| Advantages | Disadvantages | Usage occasion | |
| Bidirectional Buck/Boot Converter | With fewer switching devices, it is easy to drive and control, and has relatively high conversion efficiency. | It is not possible to change the voltage in both directions, and the inductance only exists on one side, resulting in a large current ripple on the other side. | Suitable for low power, no need for electrical isolation, and unidirectional voltage regulation. |
| Bidirectional Buck Boot Converter | It can undergo one-way voltage up and down conversion, and the drive circuit design is easy. | The polarity of the output voltage is reversed, and due to the lack of inductance on both sides of the input and output, the current ripple on both sides is large. | Suitable for achieving voltage up and down conversion in a short period of time. |
| Bidirectional Cuk converter | The inductance at the input and output terminals can effectively suppress current ripple | The topological structure is complex and requires intermediate energy transmission through capacitive media, resulting in low transmission efficiency. | Suitable for low current ripple on both sides. |
| Bidirectional Zeta Sepic converter | The polarity of the input and output voltage. | The circuit structure is complex and the conversion efficiency is low. | Suitable for high current applications |
The bidirectional Buck/Boost converter has good conversion efficiency, requires fewer electrical components, has a simple topology, and is easy to design for driving. Therefore, this article studies the bidirectional Buck/Boost circuit topology.
2.3 Current Status of Energy Storage Capacity Configuration and Algorithm Research
At present, the power system consists of five links: “generation, transmission, distribution, and storage”. Energy storage technology, as one of the five important links, plays a crucial role in distributed generation, grid connection, transmission, distribution, power supply, retail, end users, and auxiliary services.
The policy and market differences among countries around the world will show different priorities. The United States focuses on the development of photovoltaic power generation, microgrids, and auxiliary services, while Germany focuses on household energy storage systems. Other European countries mainly combine energy storage with photovoltaic systems with relatively abundant resources. In Asian countries such as Japan, energy storage systems are mainly used to solve a series of grid connection problems caused by photovoltaic fluctuations and gaps. Currently, China, a major power system country, involves a wide range of links, Mainly concentrated in large-scale photovoltaic grid connection, microgrid, distributed power generation, and other auxiliary services. With the rapid development of photovoltaic power generation, energy storage systems will play an important role in photovoltaic power generation systems. At present, the cost of energy storage remains high, and reasonable capacity allocation is the core of the performance and efficiency of energy storage systems.
1) Current research status of energy storage capacity configuration:
At present, the configuration of energy storage capacity mainly involves two issues: one is the intermittent problem of photovoltaic power generation systems; The second issue is the cost of energy storage in distribution networks and microgrid systems, as well as the effective utilization of renewable energy. In photovoltaic power generation systems, capacity allocation results are obtained through optimization algorithms with multiple objectives such as improving renewable energy utilization efficiency, suppressing load power fluctuations, and reducing system costs. The energy storage microgrid optimizes system capacity with the objective functions of reducing light abandonment, reducing distributed power fluctuations, improving power quality, and improving grid connection quality.
The current mainstream configuration methods aim to establish mathematical models for the capacity configuration of optical storage systems, with the optimization objectives of minimizing the cost, optimizing the capacity, maximizing profits, achieving high utilization of renewable energy, and achieving low load and power loss.
(1) Application of energy storage in photovoltaic power generation systems
By establishing a multi-objective optimization configuration between capacity and electricity consumption, we can suppress photovoltaic power fluctuations and reduce the impact of photovoltaic grid connection on the power grid. Optimize for maximum efficiency and allocate capacity to address the impact of photovoltaic power fluctuations on system stability. Adopting adaptive genetic algorithm for multi-objective optimization of distributed power generation fluctuations and load defect rates. Considering the load defect rate and energy overflow ratio comprehensively, the optical storage capacity is configured. Application of improving photovoltaic prediction level through capacity configuration of optical storage systems. Capacity allocation based on the lifespan of the energy storage system and the economic operation of the system; Using combination planning and particle swarm optimization algorithm to configure the capacity of the microgrid system for improving abandoned light as the application scenario; Using combination programming combined with particle swarm optimization algorithm to optimize the capacity of photovoltaic distribution networks with the objective of optimizing the economic benefits of energy storage. Considering the best economic performance as the objective function, the capacity configuration of the energy storage system is carried out using the parallel and off grid microgrid system as the application scenario; Taking the utilization rate and revenue of photovoltaic systems as optimization objectives, an improved non inferior sorting genetic algorithm is adopted for multi-objective optimization of energy storage systems; The application scenario is to suppress distributed power fluctuations and power quality, with the configuration goal of minimizing capacity; Target capacity allocation based on electricity purchase cost; Taking the optimal economy of the microgrid system as the objective function, the sampling bat optimization algorithm is used to plan the optimal capacity required for system operation; Based on non inferior sorting genetic algorithm, capacity optimization is carried out with the objective function of green environmental protection and optimal economic cost of power supply. In summary, the capacity configuration of optical storage systems is mostly optimized for economic efficiency, system operation life, and the fluctuation of photovoltaic power supply, without considering the relationship between photovoltaic power generation and load capacity balance. Based on this, this article considers multi-objective optimization of load loss rate and energy surplus rate on the basis of cost, ensuring the economic benefits of the system while increasing the flexibility of system optimization.
2) Common algorithms for capacity optimization:
(1) Ant Colony Algorithm
Ant colony algorithm is inspired by the phenomenon of ants foraging for food. Ants constantly emit odors during their foraging process, which increase in odor concentration as they get closer to the food. The closer you are to the food, the higher the concentration. Make the ants behind follow this scent to find food as quickly as possible.
If the odor concentration accumulates too quickly, an incorrect foraging path will occur, leading to a decrease in the success rate of foraging. However, if the odor accumulates too slowly, it will dissipate, leading to incorrect foraging paths. Equivalent to the process of finding the optimal solution as the number of iterations increases in an algorithm.
(2) Particle Swarm Optimization
The particle swarm optimization algorithm is based on the idea of particle motion. During the optimization process, each particle will share the position between particles, and each particle can move at a corresponding speed based on its own experience. At the same time, the particle’s position and velocity information are updated in real-time, ultimately leading to the correct direction of optimization. This algorithm is suitable for solving planning problems in distributed networks and is easy to solve with fast convergence speed.
(3) Simulated annealing algorithm
Simulated annealing is based on the principle of solid-state annealing and belongs to probabilistic algorithms. The high temperature causes internal particles to become active, generate internal energy, and collide to form irregular motion states. When the temperature decreases, irregular motion reduces the motion of particles inside the material and returns to their original state. The initial temperature of this algorithm is generally set high, and the probability of adaptive transformation occurs as the temperature decreases. By using a certain probability, the global optimal solution of the objective function is found. If the optimization result falls into an incorrect solution set, there is a certain probability that it will jump out of the local solution depending on the parameter settings, and ultimately find the most suitable global solution. This algorithm has a shorter solving time and is suitable for combinatorial optimization solutions.
(4) Genetic algorithm
Genetic algorithm is an optimization algorithm evolved from Darwin’s theory of biological evolution combined with the natural law of the jungle in nature. The initial population is left with species that adapt to the current environment through survival of the fittest. The remaining species reproduce through mating, resulting in gene crossover, mutation, and other processes to produce a new generation. Ultimately, the optimal individual will evolve after several generations. This algorithm is flexible in search and widely used. However, genetic algorithms cannot fully encode constraint problems, have low efficiency, are prone to premature convergence, and the accuracy and complexity of the algorithm cannot be quantitatively analyzed.
The advantages and disadvantages of intelligent algorithms are shown in Table 3.
| Optimization algorithm | Advantages | Disadvantages |
| Simulated annealing algorithm | The algorithm is simple and can be processed through multiple threads | Slow convergence and computation speed |
| Genetic algorithm | Wide range of applications, flexible search, and minimal impact of initial parameters on results | Inaccurate coding, low efficiency, and easy premature convergence |
| Ant colony | Optimization problems for high computational and operational efficiency, suitable path search problems | The optimization results are greatly influenced by the initial parameters and are prone to falling into local optima |
| Particle Swarm Optimization | Easy to implement, easy to adjust parameters, fast convergence, suitable for distributed network planning, and fast computation. | Easy to fall into local optima |
Due to the slow convergence speed and long calculation time of simulated annealing algorithm, while genetic algorithm has low efficiency and is prone to premature phenomena. Ant colony algorithm’s selection of starting positions and roadblocks can have an impact on the optimization results. The particle swarm optimization algorithm is more suitable for solving planning problems in distributed networks, similar to the capacity optimization problem of optical storage systems in this paper. Moreover, this algorithm is easy to implement, converges quickly, and has a short computational time. Therefore, the text selection particle swarm optimization algorithm is used to optimize the capacity of the optical storage system.
2.4 Current Status of Control Strategy Research
A single energy storage system cannot quickly respond to high-frequency or low-frequency signals under high-power load fluctuations and renewable power fluctuations. It can be composed of lead-acid batteries and supercapacitors to quickly respond to high-frequency and low-frequency signals, achieving power distribution between systems. Moreover, a hybrid energy storage system has a longer service life than a single battery energy storage system.
Reported different types of control strategies. This includes fuzzy logic control, control based on adaptive neural fuzzy inference systems, and Model predictive control (MPC). A hybrid energy storage system (HESS) controller based on multi-mode fuzzy logic was designed for photovoltaic systems to extend battery life. However, there is limited description of stability analysis and its practical implementation. A nonlinear control technology has been proposed for strict DC voltage regulation and effective supercapacitor current tracking of electric vehicles equipped with fuel cells and supercapacitors. MPC has been achieved by effectively controlling the current to limit the upper and lower thresholds of the State of Charge (SOC) of the energy storage device, ensuring that there is no overcharging or discharging phenomenon in the energy storage system. However, this method relies on the discrete model and cost function of the system. Using a quadratic programming solution calculator to calculate the optimal control takes a long calculation time. Reference [48] proposes the application of wavelet based frequency decoupling technology in HESS control of electric vehicles. However, the above control strategy requires a significant amount of computational time and resources.
A sliding mode control based on hybrid energy storage system was proposed, and the selection of controller parameters and the determination of supercapacitor size were discussed. The motion of the sliding surface of the system is insensitive to specific types of disturbances and model uncertainties, and the buffeting effect can be observed in practical systems with communication delays, hysteresis, and slow dynamic responses.
A two-state rate limiting control for HESS application in an independent DC power grid system was proposed. This method helps to reduce the operating current of the Battery Energy Storage System (BESS). The proposed method can effectively maintain the state of charge of the battery for a longer period of time. However, excessive charging and discharging rates can affect the service life of energy storage systems, which is not conducive to their long-term use.
Introduced the decentralized control strategy for dynamic power sharing between HESS. This method is based on virtual resistance and capacitance droop coefficient. During power generation and load changes, the droop coefficient essentially acts as a low-pass filter for batteries and a high-pass filter for supercapacitors.
A parameter selection technique based on filter power coupling method for message and supercapacitor size applications in microgrids has been proposed. The application of hybrid energy storage size is introduced in reference [53], and the application of uninterruptible power supply is introduced. Introduced a detailed mathematical analysis of the application of HESS in remote areas.
Introduced the impact of changes in battery internal resistance and voltage on the stability of the converter connected to the battery system. Research has shown that changes in internal parameters during load changes can affect battery response time and DC link voltage recovery time. Due to aging and changes in battery SOC, changes in internal parameters are usually observed. The stability problem of HESS based on cascaded DC-DC converters. Research has shown that changes in battery SOC and capacity of different energy storage types can lead to control stability issues.
In summary, the control strategies of energy storage converters can be divided into two categories: linear and nonlinear based on the relationship between the system’s state variables and output variables. Nonlinear control strategies have certain limitations, as they rely heavily on models and cost functions, consume a large amount of computational resources and time costs, and stability analysis is also complex. Therefore, this article mainly considers using a linear control strategy to analyze the energy storage converter. Currently, for optical storage systems, linear control strategies mainly focus on using virtual resistors or filters to allocate power between supercapacitors and batteries. Based on this, this article uses supercapacitors to compensate for the error signals caused by slow response of battery performance and converter performance, and increases the utilization rate of supercapacitors to indirectly improve the battery’s service life, while considering the impact of load and power fluctuations of photovoltaic power sources on the stability of DC bus voltage, Reduce the peak voltage and recovery time of bus voltage by adding composite control.
3. Summary
Taking the cost, energy excess rate, and load loss rate of the energy storage system as optimization objectives, in order to simplify the analysis, multiple objective functions are fused through weight coefficients. Solve the multi-objective function using particle swarm optimization algorithm to achieve the capacity planning problem of the system. On the basis of capacity configuration, this article studies and analyzes the control strategy of the optical storage system using a bidirectional Buck/Boost energy storage converter as the topology. On the basis of traditional control strategies, a composite frequency division coordinated control strategy is used to improve the speed and stability of the DC bus voltage. Finally, verification was conducted through the semi physical platform RT Box and the energy storage grid connection experimental platform.