Energy Optimization Strategy for Wind–Solar–Storage Systems with a Storage Battery Configuration

Abstract

With the progressive advancement of the energy transition strategy, wind–solar energy complementary power generation has emerged as a pivotal component in the global transition towards a sustainable, low-carbon energy future. To address the inherent challenges of intermittent renewable energy generation, this paper proposes a comprehensive energy optimization strategy that integrates coordinated wind–solar power dispatch with strategic battery storage capacity allocation. Through the development of a linear programming model for the wind–solar–storage hybrid system, incorporating critical operational constraints including load demand, an optimization solution was implemented using the Artificial Fish Swarm Algorithm (AFSA). This computational approach enabled the determination of an optimal scheme for the coordinated operation of wind, solar, and storage components within the integrated energy system. Based on the case study analysis, the AFSA optimization algorithm achieves a 1.07% reduction in total power generation costs compared to the traditional Simulated Annealing (SA) approach. Comparative analysis reveals that the integrated grid-connected operation mode exhibits superior economic performance over the standalone storage microgrid system. Additionally, we conducted a further analysis of the key factors contributing to the enhancement of economic benefits. The strategy proposed in this paper significantly enhances power supply stability, reduces overall costs and promotes the large-scale application of green energy.

1. Introduction

The ongoing global energy transition has been significantly accelerated by remarkable advancements in renewable energy technologies [1]. In parallel, the dual challenge of optimizing cost-effectiveness while ensuring reliable energy supply has emerged as a critical strategic imperative for the development of sustainable low-carbon economies [2,3]. A dual-phase energy transition strategy has been established by Chinese government, with ambitious targets of achieving a 20% share of non-fossil energy in primary consumption by 2030, followed by carbon neutrality by 2060 [4,5]. To accomplish this objective, the implementation of wind–solar–storage microgrid model becomes particularly crucial, boasting advantages such as environmental friendliness, reduced reliance on fossil fuels, and enhanced utilization efficiency of renewable energy. Nevertheless, this model also faces several challenges, including high initial investment costs, energy losses, and lifespan constraints. According to the latest research data released by the International Renewable Energy Agency (IRENA) [6], global renewable energy capacity additions are projected to experience a substantial surge in 2024, with annual installations anticipated to rise by 25% compared to current levels, surpassing the 700 GW threshold. Notably, solar photovoltaic (PV) technology demonstrates particularly robust momentum within this expansion framework, with its annual newly installed capacity projected to achieve a year-on-year increase of approximately 30%. Accounting for approximately 550 GW of the total additions, solar PV installations are expected to represent nearly 80% of the annual capacity expansion, further solidifying their status as the principal driving force in global clean energy transition initiatives. However, China remains the main driving force behind the increase in installed capacity of renewable energy worldwide. In 2024, the newly added photovoltaic capacity exceeded 340 GW, an increase of 30% compared to 2023, while the newly added wind energy capacity was 80 GW, unchanged from last year. (Figure 1).

Battery energy storage systems have garnered significant research attention due to their crucial role in maintaining grid stability through peak shaving and valley filling operations [7]. These systems effectively mitigate the inherent intermittency of renewable energy generation while enhancing grid flexibility and dispatchability [8]. The key to improving the stability and economic benefits of distributed renewable energy system is not only in optimizing the configuration scheme of energy storage system and ensuring efficient and reasonable dispatching, but also in optimizing the operation mode of power system [9,10]. Although numerous studies have explored the impact of integrating solar and wind energy into power systems, a systematic solution to the grid operation challenges caused by intermittency and volatility has yet to be established [11].

Recently, extensive research has been conducted on the wind–solar–storage microgrid scheduling optimization. Huang et al. developed an energy optimization scheduling model for wind–solar–storage microgrids incorporating comprehensive cost factors with a specific focus on minimizing demand response costs [12]. In a related study, Ma et al. implemented a particle swarm optimization (PSO) algorithm for capacity allocation in wind–solar–storage systems within smart microgrids. However, this solution process was often found to converge to local optima [13]. To enhance the accuracy of search direction and prevent convergence to local optima, Zhang et al. developed an enhanced artificial fish swarm algorithm (AFSA). This improved version incorporates a decay factor and introduces regional search uncertainty, effectively mitigating repetitive search patterns and improving optimization performance [14]. Additionally, Zhang et al. proposed an innovative hybrid optimization approach combining modified genetic algorithms (GAs) with ant colony optimization (ACO) to enhance the reliability of sustainable energy systems by addressing key challenges including renewable energy variability, energy storage limitations, and residential demand fluctuations [15]. Simultaneously, Maklewa Agoundedemba et al. used the combined genetic algorithm (GA) and model predictive control (MPC) to size and optimize the hybrid renewable energy system (PV/Wind/FC/Battery), subject to certain constraints on the power flow and battery state of charge [16]. Furthermore, Parastegari et al. investigated the optimal scheduling problem for both joint operation (JO) and uncoordinated operation (UO) of wind farms and pumped storage power stations, which was systematically verified to enhance the system’s profitability through JO implementation [17].

Although extensive research has been conducted on wind–solar–storage microgrid systems and battery capacity optimization, encompassing diverse technical perspectives, the joint operational mechanisms of microgrid systems remain significantly underexplored in current literature. Meanwhile, the existing fossil fuel-based power generation models are plagued by issues such as environmental pollution, resource depletion, and price volatility [18], whereas independently operated power grid models face challenges like energy wastage and limited anti-interference capabilities. Particularly in the context of China’s comprehensive requirements for energy security assurance and economic performance enhancement [19], several critical research gaps persist. These include the optimization of environmental benefits and the development of advanced operational strategies for storage-integrated microgrid systems, which warrant comprehensive investigation.

This paper explores the optimization of wind–solar–storage configuration schemes. By integrating renewable energy and energy storage technologies for rational configuration and joint operation of multiple power grids, it effectively overcomes the environmental problems of fossil energy generation, while improving the flexibility and anti-interference ability of the power grid. The proposed model incorporates critical system operational constraint condition, particularly focusing on safety and stability requirements. Through comparative performance analysis on the artificial fish swarm algorithm (AFSA) and the conventional Simulated Annealing (SA) approach, an optimal energy storage configuration scheme and energy management strategy are derived, demonstrating significant economic benefits while simultaneously enhancing renewable energy integration and optimizing resource utilization.

2. The Wind–Solar–Storage Microgrid Model

The wind–solar–storage microgrid system structure is illustrated in Figure 2, consisting of a 275 kW wind turbine model, 100 kW photovoltaic model, lithium iron phosphate battery, and user load. When power demand is not fully met, electricity can be obtained from the main grid or supplied by the battery storage system. During periods of excess power, surplus energy is directed to the storage system rather than fed back to the main grid, thereby minimizing grid fluctuations. The symbol list in this model is shown in

Table 1. Basic list of symbols.

Symbols	Meaning
$P_W$	the output power of the fan
$P_R$	the rated power of the fan
$v_c{,v}_R,v_F$	the cut-in, rated, cut-off wind speed
$C_{wind}$	the unit cost of wind power generation
$P_a$	the annual average output power
$C_{solar}$	the unit cost of photovoltaic power generation
$C_p$	the installation cost
$y_p$	lifespan years
$R_{op}{,P}_{loan}{,P}_{intr}$	operating rates, loan ratio and loan interest rate
$H_{fp}$	the equivalent annual power generation hours
$P_{ess,t}$	the charging and discharging power of the energy storage system at time t
${SOC}_t,{SOC}_{t+1}$	the state of charge before and after the battery respectively
$E_{storage}$	the rated capacity of the energy storage system
$C_{bat}$	the life loss cost of lithium iron phosphate battery
$C_{invest}$	the total investment cost of lithium iron phosphate battery

.

2.1. The Wind Turbine Model

The relationship between wind turbine output power and wind speed has been demonstrated to be representable through a piecewise function, as shown in numerous studies [20].

$$P_W=\left\{\begin{array}{c}{ P}_R{ for v}_R<v \le v_F\\ P_R {\displaystyle \frac{v-v_c}{v_R-v_c}} for { v}_c \le v \le v_R \\ 0 otherwise \end{array}\right.$$

(1)

In which $P_W$ is the output power of the fan, $P_R$ is the rated power of the fan, $v_c$ is the cut-in wind speed, $v_R$ is the rated wind speed, $v_F$ is the cut-off wind speed. According to the manufacturer’s data, the fan utilized in this study has a rated power of 275 kW, a cut-in wind speed of 4 m/s, a rated wind speed of 12 m/s, a cut-off wind speed of 20 m/s, and a design life of 20 years.

A unit power generation cost calculation model is established in this paper, based on the equal distribution of total generation costs over the wind farm’s entire expected lifespan [21].

$$C_{wind}={\displaystyle \frac{r{\left(1+r\right)}^{y_w}}{{\left(1+r\right)}^{y_w}-1}}·\left[{\displaystyle \frac{Q}{8760F}}\right]$$

(2)

$$F={\displaystyle \frac{P_a}{{ P}_R}}$$

(3)

In which $C_{wind}$ is the unit cost of wind power generation, $r$ is the annual interest rate of investment loans, $y_w$ refers to the payback time of the investment in the construction of the plant, $Q$ is the unit investment cost of plant construction, $F$ is the capacity factor, $P_a$ is the annual average output power. By consulting the reference [22], the parameters of the cost model for wind power generation are shown in Table 1: The parameters of the economic cost model for wind turbine power generation are provided in

Table 2. Parameters of economic cost model of wind turbine power generation.

Parameter Type	$\mathit{r}$ (%)	${\mathit{y}}_{\mathit{w}}$ (Years)	$\mathit{Q}$ ($/kW)	$\mathit{F}$
numeric value	10	20	1142	0.5

.

2.2. The PV Battery Model

In order to facilitate practical application, the steady-state power output of PV batteries can be simplified as follows, and the formula of PV output power is [23]:

$$P_{PV}=P_{STC}{G}_{AC}\left[1+k\left(T_c-T_{\tau }\right)\right]/G_{STC}$$

(4)

In which $G_{AC}$ is the light intensity, $P_{STC}$ and $G_{STC}$ are the maximum test power and light intensity under standard test conditions respectively, $k$ is the power temperature coefficient, $T_c$ is the working temperature of the battery panel, $T_{\tau }$ is the reference temperature and its value is $25 °\mathrm{C}$. The cost of photovoltaic power generation is primarily determined by installation costs, system efficiency, policy and financing conditions, operational lifespan, and maintenance expenses. Based on these factors, the PV unit generation cost model can be established [22].

$$C_{solar}=C_p({\displaystyle \raisebox{1ex}{$1$}\!\left/ \!\raisebox{-1ex}{$y_p$}\right.}+R_{op}+P_{loan}·P_{intr})/H_{fp}$$

(5)

In which $C_{solar}$ and $C_p$ are the unit cost of photovoltaic power generation and the installation cost, $y_p$ is lifespan years, $R_{op}$ is operating rates, $P_{loan}$ and $P_{intr}$ are loan ratio and loan interest rate, and $H_{fp}$ is the equivalent annual power generation hours. The parameters of the economic cost model for photovoltaic power generation were obtained by consulting reference [22], as shown in

Table 3. Parameters of the economic cost model for PV power generation.

Parameter Type	${\mathit{C}}_{\mathit{p}}$ ($)	${\mathit{y}}_{\mathit{p}}$ (Years)	${\mathit{R}}_{\mathit{o}\mathit{p}}$	${\mathit{P}}_{\mathit{l}\mathit{o}\mathit{a}\mathit{n}}$	${\mathit{P}}_{\mathit{i}\mathit{n}\mathit{t}\mathit{r}}$	${\mathit{H}}_{\mathit{f}\mathit{p}}$ (h)
numeric value	1651	20	2%	70%	7%	1600

.

2.3. The Battery Charging and Discharging Model

2.3.1. Battery State of Charge

The state of charge (SOC), a critical parameter for indicating remaining battery capacity, must be maintained consistent between the start and end of each dispatch cycle to ensure continuous operation. The simplified SOC update formula for the energy storage system can be expressed as [24,25,26]:

$${SOC}_{t+1}={SOC}_t+{\displaystyle \frac{P_{ess,t }· △t}{E_{storage}}}$$

(6)

$${SOC}_0={SOC}_n$$

(7)

In which $P_{ess,t}$ is the charging and discharging power of the energy storage system at time t, ${SOC}_t$ and ${SOC}_{t+1}$ are the state of charge before and after the battery respectively, $E_{storage}$ is the rated capacity of the energy storage system, $△t$ is the time interval of each time period of the energy storage system, $n$ is the last time value of a scheduling cycle of the energy storage system.

2.3.2. Battery Discharge Cost

We compared batteries with other energy storage methods in terms of working principle, energy storage time, cycle life, recycling capacity, and cost by consulting a large number of references, as shown in

Table 4. Performance comparison of various energy storage methods.

	Battery	Pumped-Storage	Supercapacitors	RSOC System
Working principle	electrochemical reaction	Potential energy storage	Electrostatic energy storage	Reversible electrochemistry
Energy storage duration	4–5 h	4–24 h	<30 min	>12 h
cycle life	>3000 times	>50 years	>500,000 times	20,000 h
resources sustainability	Rich in iron and phosphorus	Water resource dependence	Carbon materials are abundant	Ceramic materials are abundant
Recycling capacity	Low	The equipment can be reused	The material can be 100% recycled	Ceramic components can be recycled
Costs	Low	Medium	Average	High

. Overall, batteries have significant advantages [27,28,29,30]. The lithium iron phosphate (LiFePO₄) batteries utilized in this study has its discharge cost determined by its lifecycle, which is influenced by factors such as charge–discharge cycles, depth of discharge, and operating temperature. The equivalent economic loss cost of lithium iron phosphate batteries is expressed as follows [31]:

$$C_j={\displaystyle \frac{C_{invest}}{N_{ESS}}}$$

(8)

$$C_{bat}={\sum }_{j=0}^{N_b}{C}_j$$

(9)

In which $C_{bat}$ and $C_{invest}$ are respectively the life loss cost and total investment cost of lithium iron phosphate battery, $C_j$ is the cost corresponding to the battery charge and discharge depth of $d_j$, $N_b$ and $N_{ESS}$ are the charge and discharge times in a scheduling cycle and the maximum charge and discharge times in the life cycle of the battery. According to the literature [32], LiFePO₄ batteries are employed in this study, with the detailed parameters presented in

Table 5. Relevant parameters of LiFePO₄ battery.

Parameter Type	${\mathit{P}}_{\begin{array}{c}\mathit{c}\mathit{h}\mathit{a}\mathit{r}\mathit{g}\mathit{e},\mathit{t}\mathbf{\left(}\mathit{m}\mathit{a}\mathit{x}\mathbf{\right)}\\ \mathbf{\left(}\mathit{k}\mathit{W}\mathbf{\right)}\end{array}}$	${\mathit{P}}_{\begin{array}{c}\mathit{d}\mathit{i}\mathit{s}\mathit{c}\mathit{h}\mathit{a}\mathit{r}\mathit{g}\mathit{e},\mathit{t}\mathbf{\left(}\mathit{m}\mathit{a}\mathit{x}\mathbf{\right)}\\ \mathbf{\left(}\mathit{k}\mathit{W}\mathbf{\right)}\end{array}}$	${\mathit{S}\mathit{O}\mathit{C}}_{\mathit{m}\mathit{a}\mathit{x}}$	${\mathit{S}\mathit{O}\mathit{C}}_{\mathit{m}\mathit{i}\mathit{n}}$	${\mathsf{\eta }}_{\mathit{c}\mathit{h}\mathit{a}\mathit{r}\mathit{g}\mathit{e}}$	${\mathsf{\eta }}_{\mathit{d}\mathit{i}\mathit{s}\mathit{c}\mathit{h}\mathit{a}\mathit{r}\mathit{g}\mathit{e}}$
numeric value	12	12	10%	90%	95%	95%

:

2.4. The Joint Grid Energy Storage Configuration Model

The joint grid model, as illustrated in Figure 3, is designed to effectively interconnect multiple power grids and energy storage systems, enabling optimal resource allocation and sharing across different regions. This approach has been demonstrated to enhance overall operational efficiency while reducing operational costs of storage facilities, thereby improving economic performance. Additionally, the system’s flexibility is enhanced to better accommodate the stochastic variability of renewable energy generation.

2.5. Operating Costs and Power Purchase Costs

In wind–solar–storage systems, annual equipment replacement costs and operational maintenance expenses are identified as significant components of operational expenditures [33]. At the same time, in the power purchase pricing system, differences between summer and winter are primarily reflected during peak demand periods. On weekdays, peak electricity rates are typically observed during morning and dusk hours, while off-peak electricity rates are typically observed during night hours. Furthermore, weekend rates are generally lower than those on weekdays [34,35]. The power purchase pricing system and peak/off-peak periods are shown in Figure 4.

3. Energy Optimization and Algorithm Solving

3.1. Energy Optimization Strategy

The system energy optimization in this strategy is achieved through a time-segmented dynamic regulation mechanism and the specific workflow is structured as follows: Initial wind–solar–storage power values are collected in real-time and dynamically matched with user load demands for supply-demand analysis. Predefined differentiated energy dispatch strategies are automatically triggered when integrated generation deviates from load requirements. Then, the daily cycle is then divided into 24 equal intervals, during which programmed algorithms are executed to perform dynamic regulation of charge/discharge power, real-time tracking and adjustment of battery state-of-charge (SOC), among other operational tasks. The operational strategies and electricity procurement plans for the energy storage system are ultimately derived, as illustrated in Figure 5.

Initially, the system reads the initial parameters and evaluates whether satisfies the load demand. When the conditions are met, first determine whether the SOC exceeds the upper limit (90%). If exceeded, charging is prohibited. Otherwise, the system evaluates whether the remaining power exceeds the rated charging power of the battery. If exceeded, the battery will be charged at its maximum rated power. On the contrary, if it is below this threshold, the battery will absorb the remaining available power.

When the power generation is insufficient, if the SOC is below 10%, the battery will neither charge nor discharge. If it has not been reached yet, continue to determine whether the remaining capacity meets the demand gap. If satisfied, the battery will discharge at the required power; Otherwise, discharge according to the remaining capacity power. Then update the battery charging status.

3.2. The Linear Programming Model of Wind–Solar Energy Storage Microgrid

To address the linear programming problem in energy management strategies for the wind–solarenergy storage microgrid, an objective function needs to be established for the linear programming model of the wind–solar–storage microgrid.

$$min{C}_{total}=P_W·C_{wind}·T+P_{PV}·C_{solar}·T+C_{bat}·P_{discharge,t}·T+f·P_{grid,t}·T$$

(10)

In which $C_{total}$ is the total operation cost of the microgrid, $f$ is the power purchase price, $P_{discharge,t}$ and $P_{grid,t}$ are the discharge power of the energy storage system at time t and the power purchased from the main grid.

An objective function is formulated to minimize the total operational costs of the microgrid, while simultaneously accounting for critical constraints such as generation capacity limits, charge–discharge power balance, battery state-of-charge limitations, and climbing rate constraint [36].

3.2.1. The Charge–Discharge Power Balance

$$P_W+P_{PV}+P_{ess,t}+P_{grid,t}=P_t+P_{curtailment,t}$$

(11)

In which $P_{curtailment,t}$ is the amount of abandoned wind and solar power of the energy storage system at the moment.

3.2.2. The Generation Capacity Limits

$$-P_{max,t}\ll P_{ess,t}\ll P_{max,t}$$

(12)

$$P_{W,min}\ll P_W\ll P_{W,max}$$

(13)

$$P_{PV,min}\ll P_{PV}\ll P_{PV,max}$$

(14)

In which $P_{max,t}$ is the maximum charging and discharging power of the energy storage at the moment, $P_{W,min}$ and $P_{W,max}$ are the minimum and maximum values of fan output power respectively, $P_{PV,min}$ and $P_{PV,max}$ are the minimum and maximum values of photovoltaic output power.

3.2.3. The Battery State-of-Charge Limitations

$${{SOC}_{min}\le SOC}_t\le {SOC}_{max}$$

(15)

This constraint can prevent overcharging and overdischarging of the battery and increase its service life.

3.2.4. The Climbing Rate Constraint

$$△P_W\le P_{W,control}$$

(16)

$${△P}_{PV}\le P_{PV,control}$$

(17)

In which $△P_W$ and ${△P}_{PV}$ are the difference between the current and previous given values of wind turbine and photovoltaic output, respectively, $P_{W,control}$ and $P_{PV,control}$ are the constraint values for wind turbine and photovoltaic output, respectively.

3.3. The Simulated Annealing Optimization Algorithm (SA)

Given the presence of exceedingly complex constraint condition inherent in this study, we ultimately decided to employ the simulated annealing algorithm The simulated annealing optimization algorithm (SA), a robust global optimization method based on stochastic search, is inspired by the annealing process in solid-state physics [37,38]. This algorithm demonstrates distinctive advantages through the effective escape from local optima to approach global optimal or superior solutions, thereby facilitating the discovery of global or near-global optimal solutions. This algorithm is widely implemented in practical engineering domains including manufacturing scheduling, network optimization, and path planning scenarios. These optimization problems are characterized by multiple local optima, where traditional methods are easily trapped in local solutions, resulting in suboptimal outcomes [39,40,41].

3.4. AFSA Based on Hybrid Mutation Operator and SA

In addition, we introduce another artificial fish swarm optimization algorithm (AFSA), which demonstrates significant advantages in directional search guidance and local optimum avoidance. However, its convergence performance is significantly compromised under two specific operational scenarios. (1) random individual fish movements that disrupt systematic search patterns, and (2) excessive aggregation phenomena at suboptimal solutions. These limitations ultimately lead to reduced optimization accuracy and compromised solution quality [42,43,44]. A genetic algorithm-inspired mutation operation is introduced to enhance algorithm diversity and adaptability in this paper. Through this method, the exploratory capability of AFSA in complex optimization problems is effectively enhanced, thereby improving global optimum identification. The overall process is structured as follows: Initially, a global search is conducted using the AFSA with mutation operators for optimal solution identification in the solution space. Subsequently, the simulated annealing algorithm is applied to perform a localized refinement search on the optimal artificial fish individuals, achieving local optimization [45]. The optimal near-exact extreme value is ultimately achieved through this process. The fundamental workflow of the AFSA based on hybrid mutation operator and SA is illustrated in Figure 6.

4. Example Analysis and Discussion

Three independent park-level wind–solar microgrid systems (Park A, B, C) are analyzed in this study. The only variation between systems is assumed to be in wind turbine and PV cell quantity, and battery energy storage system configurations. The parameters of various components such as wind turbines, PV cells, and batteries are shown in Table 1, Table 2 and Table 3 above.

4.1. Independent Operation and Energy Storage Configuration of Each Park

According to the equipment manual and analysis of numerous case results, independent park-level storage systems employ LiFePO₄ batteries with the following technical and economic parameters: a discharge cost of 110.12 $/kWh, an operational SOC range of 10–90%, charge/discharge efficiency of 95%, and a projected lifespan of 10 years. Subsequently, based on the formula of the wind solar discharge cost model mentioned earlier, the purchase cost of renewable energy was calculated to be 0.0688 USD/kWh for wind energy and 0.0551 USD/kWh for solar energy.

Table 6. Comparison table of basic parameters of each park.

	Microgrid A	Microgrid B	Microgrid C
Operating costs $(\mathrm{\$}$)	5506	6882	6194
Battery configuration scheme	100 kW/150 kWh	100 kW/230 kWh	80 kW/150 kWh
Average cost $(\mathrm{\$}$/kW)	0.1144	0.0983	0.0964

comprehensively presents three key performance metrics for each park: operational costs, optimal battery configurations obtained through SA algorithm, and average unit electricity supply costs.

4.2. Joint Park Operation Energy Storage Configuration

When the aggregate power generation of the three parks matches the total load demand, an integrated joint operation system is established, whose architectural configuration is illustrated in Figure 3. The joint operation maintains consistent renewable energy procurement costs at 0.0688 $/kWh for wind power and 0.0551 $/kWh for solar energy, with the energy storage system configuration remaining identical to individual park operations. The total operational cost is calculated as 18,583 $, while an additional cost of 4,129 $ is required for joint system coordination and inter-park electricity transmission. In addition, other costs are ignored. The optimal battery configuration for the joint park is determined to be 182.48 kW/405 kWh through energy storage optimization performed using the simulated annealing optimization algorithm (SA). Among them, a 0.0925 $/kWh average unit electricity supply cost is achieved for the joint park system, 11.37% lower than the average cost of standalone park operations. Figure 7 provides a comprehensive temporal analysis of energy flow characteristics throughout a 24-h period, illustrating three key parameters for both individual and joint park operations: daily load profiles, electricity procurement from the main grid, and abandoned wind/solar power due to generation overcapacity.

Table 7. Comparison table of corresponding parameters of independent operation and joint operation.

	Microgrid A	Microgrid B	Microgrid C	Joint Microgrids
Total purchased $(\mathrm{k}\mathrm{W}$)	4879.30	2570.00	2699.39	8823.55
Total abandoned $(\mathrm{k}\mathrm{W}$)	1098.65	676.70	1128.02	1578.23
Total cost $(\mathrm{\$}$)	898.54	754.13	753.26	1841.84
Average cost $(\mathrm{\$}$/kW)	0.1144	0.0983	0.0964	0.0925

presents a comprehensive comparison of key operational metrics between standalone and joint operation modes, including: (1) total electricity procurement from the grid, (2) cumulative energy curtailment, (3) aggregate power supply expenditures, and (4) mean unit electricity costs.

From the comprehensive analysis of Figure 7 and Table 5, it can be seen that: The total purchased electricity for independent operation is 10,148.68 kW, and through joint park operation, power resource sharing and optimized scheduling have been achieved. The total purchased electricity is 8823.55 kW, which has been significantly reduced. This change not only highlights the advantages of resource sharing, but also reflects the effectiveness of joint operation in reducing the dependence on the main power grid. By optimizing load balancing and sharing power resources in joint operations, the total abandoned power of the joint park is 1578.23 kW, which significantly reduces power waste and abandonment compared to the total abandoned power of 2903.63 kW of independent parks, thereby improving overall power utilization efficiency. Additionally, internal power demand is effectively balanced by the shared use of energy storage systems across multiple parks or regions. During periods of low electricity prices, energy is stored in the energy storage system and later released during peak price periods, thereby reducing the need to purchase high-priced electricity. This dispatching strategy not only leads to a reduction in enterprise electricity procurement expenditures but also results in decreased production costs, shortened payback period, and enhanced economic efficiency. The joint operation model demonstrates remarkable potential in optimizing energy utilization, particularly through two key aspects: significant minimization of energy waste and substantial improvement in electricity utilization efficiency, establishing it as a promising solution for industrial energy management.

Furthermore, the joint microgrid is optimized through AFSA integrated with mutation operators and SA, resulting in an average power supply cost of 0.0913 $/kWh, which is 1.07% lower than that achieved by the traditional SA. The energy storage configuration for the joint park operation under the two dispatching optimization methods is illustrated in Figure 8. Based on the analysis, the following conclusions can be drawn:

(1) The optimization algorithm outperforms the traditional algorithm in guiding the search trajectory, allowing it to escape local optima and identify superior configuration solutions. (2) Throughout a given day, the optimization algorithm more closely aligns the daily load curve with the power output curve, leading to a significant reduction in total electricity purchased and wasted. In contrast, between 4:00 and 8:00, the curves under the traditional algorithm show a lower degree of coincidence, likely because the algorithm becomes trapped in a local optimum during the search process. (3) Traditional algorithms often charge and discharge batteries between 10:00 and 14:00, which may have a negative impact on their lifespan. In contrast, the optimization algorithm primarily relies on power procurement from the main grid, resulting in a significant reduction in the number of charge–discharge cycles for the joint system’s battery. This helps extend the battery’s lifespan while achieving cost savings. During this period, which coincides with the off-peak electricity pricing, purchasing power from the main grid is more cost-effective compared to obtaining electricity from the battery storage system. (4) The scheduling results of optimization algorithms are superior to traditional algorithms and the SOC curve of the optimization algorithm exhibits narrower fluctuations, with less variation in the depth of battery charge and discharge, which contributes to the deceleration of battery aging. As a consequence, the frequency of replacements and maintenance costs are reduced, ultimately lowering overall production costs and improving the equipment’s economic efficiency and sustainable development.

5. Conclusions

In conclusion, this study establishes a linear programming model for wind–solar–storage integrated microgrid systems addressing the stochastic variability of renewable energy and the coordination capabilities of LiFePO₄ battery storage systems, while comprehensively considering various constraints, including load demand, grid limitations, and battery capacity. The solution is achieved through an enhanced AFSA, which incorporates a genetic algorithm-inspired mutation operator to improve population diversity and adaptive search capabilities. This algorithmic enhancement enables more effective global optimum exploration, ultimately yielding an optimized configuration scheme for coordinated wind–solar–storage system operation. On this basis, this study proposes a joint microgrid energy storage configuration model. A comparative case analysis shows that: (1) The joint operation mode achieves significant economic advantages by reducing the total purchased electricity from 10,148.68 kW in independent operation to 8823.55 kW. (2) The system curtailment is considerably reduced from 2903.63 kW to 1578.23 kW representing a 45.7% reduction in energy waste and a corresponding improvement in energy utilization efficiency. (3) In the joint microgrid, compared with traditional strategies, the optimization strategy proposed in this study reduces the total power generation cost by 1.07%. The proposed strategy offers practical guidance for short-term dispatch operations in wind–solar–storage microgrids while informing future research directions, particularly in further improving the economic optimization scheduling model, considering the impact of factors such as weather changes and labor costs.