Renewable energy facility specification optimization method and system under dual power supply scene

Abstract

The invention provides a renewable energy facility specification optimization method and system in a dual power supply scene, and belongs to the technical field of solar grid-connected dual power supply. Comprising the steps that in a dual-power-supply scene, small base stations in a region are divided into N clusters, each cluster is provided with an energy storage center and K small base stations, and each small base station is provided with a solar cell panel and an energy storage battery; defining an energy sharing model of the energy storage center and a total cost model of renewable energy facilities of each small base station; providing constraint conditions according to parameters of the sharing model and the total cost model, performing minimization calculation on the total cost by using a Markov decision process, and solving an optimal configuration combination of the solar cell panel specification and the energy storage battery specification; according to the method, objective environment data such as local solar irradiance, regional rent and labor cost are utilized, and the scale of renewable energy facilities most suitable for local regional characteristics is explored from the perspective of minimizing the total cost of the facilities on the basis of considering electricity charge loss.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of solar grid-connected dual power supply, and particularly relates to a renewable energy facility specification optimization method and system in a dual power supply scenario equipped with energy storage facilities. BACKGROUND

[0002] With the intensive deployment of base stations, high electricity costs further become a burden for operators. At the same time, the extensive use of traditional power grids also leads to excessive carbon emissions. Renewable energy, as a green energy, is the key to reducing high electricity costs and carbon emissions. However, due to climate factors and high costs, a single renewable energy source is not suitable to replace grid power. Therefore, the dual power supply scheme using the grid and renewable energy is the most promising method at present.

[0003] The dual power supply scenario combines the traditional power grid and the photovoltaic power generation system. In sunny or well-lit conditions, the photovoltaic power generation system provides power, while in bad weather or when the photovoltaic power generation system cannot meet the demand, the traditional power grid will automatically take over power supply. In terms of structure, a storage center is used to share energy with multiple base stations, and the remaining power generated by the base stations is transmitted to the storage center, and the large power grid and the storage center interact.

[0004] At present, photovoltaic power generation technology is very mature and widely used. Under the influence of climate factors, by predicting future photovoltaic output and supplementing grid power purchase, the electricity cost in most areas can be reduced. However, for operators, the size of renewable energy facilities also affects the electricity cost required for base station operation, and the cost of solar panels and energy storage batteries brought by the construction of photovoltaic power stations cannot be ignored. If the size of the renewable energy facility is too small, the base station will have to purchase a large amount of electricity from the grid. If the size of the renewable energy facility is too large, the initial investment cost will be extremely high, and there may be resource waste in the later period. Before the renewable energy facility is laid, the facility specification needs to be selected, and after operation, due to the influence of population growth, economic trends, and climate change in the region, the facility specification still needs to be further optimized to minimize resource waste and save electricity costs. SUMMARY

[0005] Therefore, the present application provides a renewable energy facility specification optimization method and system in a dual power supply scenario, which optimizes the renewable energy facility specification from the facility construction and operation and maintenance perspectives, improves the use rate of green energy, reduces the construction and operation and maintenance costs, uses local solar irradiance, regional rent, labor cost and other objective environmental data, and explores the most suitable renewable energy facility size for the regional characteristics from the perspective of minimizing the total facility cost on the basis of considering electricity loss.

[0006] The technical solution adopted by the embodiment of the present application to solve the technical problems is as follows:

[0007] 1. A method for renewable energy facility specification optimization in a dual power supply scenario, comprising:

[0008] Step S1, in a dual power supply scenario, small base stations in a region are divided into N clusters, each cluster is equipped with one energy storage center and K small base stations, and each small base station is configured with a solar panel and an energy storage battery;

[0009] Step S2, define an energy sharing model of the energy storage center and a total cost model of the renewable energy facility of each small base station, wherein the total cost model is the sum of the purchase cost, replacement cost, installation cost and grid purchase cost of the solar panel and the energy storage battery;

[0010] Step S3, provide constraint conditions according to the parameters of the sharing model and the total cost model, and solve the optimal combination of solar panel specification and energy storage battery specification by minimizing the total cost of the renewable energy facility; wherein the total cost is minimized by using Markov decision process.

[0011] Preferably, the purchase cost in the total cost model of the renewable energy facility comprises:

[0012] A relationship model between the solar panel specification and the solar energy harvesting amount of the small base station is established, and the solar energy collection power P of 1 kW solar panel is PV represented as:

[0013]

[0014] wherein η PV is the efficiency coefficient of the photovoltaic power converter; Rirr is the irradiance level of sunlight; a T is the temperature effect coefficient; T PV represents the average temperature of the solar panel assembly; T amb represents the ambient temperature; T nom represents the nominal operating temperature measured under standard ambient conditions;

[0015] The relationship expression between the energy storage battery specification N B and the theoretical peak value of the remaining energy in the battery is defined as:

[0016]

[0017] wherein δ is a proportional coefficient, representing the proportion of the energy storage battery specification relative to the theoretical peak value of the remaining energy; is the renewable energy generated by the solar panel of the small base station B k in time slot t; and △t p is the time slot in which the harvested solar energy exceeds the consumption of the base station, and △t is a unit time slot.​ is the power consumption of the base station at time slot t; N PV is the number of small base stations B k equipped with solar panels of specification;

[0018] is the purchase cost of solar panels and energy storage batteries C pur is defined as:

[0019] C pur = C PV · N PV + C B · N B

[0020] wherein C PV and C B represent the cost per unit capacity of 1 kW solar panels and energy storage batteries, respectively.

[0021] More preferably, said replacement cost C cha in the total cost model of the renewable energy facility is defined as:

[0022]

[0023] wherein Y represents the estimated service life of solar panels and energy storage batteries; L PV is the service life of solar panels; L B represents the service life of energy storage batteries, the number of cycles of energy storage batteries is defined as:

[0024]

[0025] N D = 7855 · exp(-9.48 · D) + 2508 · exp(-1.605 · D)

[0026]

[0027] wherein N D is the expression of the number of cycles of energy storage batteries corresponding to the depth of discharge D obtained by curve fitting technique; represents the number of cycles of energy storage batteries corresponding to the kth cycle when the depth of discharge is D and the operating temperature is T.

[0028] More preferably, said installation cost C ins in the total cost model of the renewable energy facility is defined as:

[0029] C ins = A PV · N PV · (Y · C a+C b )

[0030] wherein A PV is the footprint of a 1 kW solar panel; Y is the total number of years of operation; C a represents the annual rent per unit of land occupied; C b represents the total cost of construction and installation of the solar panel system per unit of land.

[0031] Preferably, the grid electricity purchase cost C gri in the total cost model of the renewable energy facility is defined as:

[0032]

[0033] wherein G g is the cost of purchasing 1 kWh of grid electricity, is the amount of electricity purchased from the grid by the small base station B k in time slot t, Y is the total number of years of operation; the energy storage center shares electricity to K base stations, is the initial energy of the small base station B k in time slot t.

[0034] Preferably, the total cost model C sum of the renewable energy facility is defined as:

[0035]

[0036] Preferably, the energy storage center energy sharing model includes:

[0037] The remaining amount of electricity k after the energy storage center completes charging and discharging with the small base station B in time slot t is represented as:

[0038]

[0039] wherein, represents the energy sharing relationship between the energy storage center and the base station in time slot t; is the initial energy of the energy storage center in time slot t; is the shared electricity obtained by the small base station B k from the energy storage center in time slot t; when there is a surplus, is the excess energy transmitted by the small base station B k to the energy storage center in time slot t;

[0040] The standby threshold of the energy storage center is represented as:

[0041]

[0042] In the formula, B S This represents the maximum capacity of the energy storage center, and γ is the correlation coefficient.

[0043] Preferably, the minimization of min(C) in step S3 is... sum The constraints are defined as follows:

[0044]

[0045] 0<δ≤1

[0046] In the formula, GCD t PRE represents the grid connection depth of a small base station in time slot t. min PRE represents the lowest possible energy level for the energy storage center. max PRE represents the highest electrical state of the energy storage center. h This is the highest value of the backup threshold. This refers to the remaining energy level of the energy storage center after completing its final charge and discharge cycle in time slot t.

[0047] Preferably, in step S3, a Markov decision process is used to solve for min(C). sum The process includes:

[0048] Min(C) of the total cost sum Rearranging this as a Markov decision process, where:

[0049] State space S t Including N solar cell panels with a capacity of 1kW PV The ratio factor δ of the energy storage battery specifications relative to the theoretical peak remaining energy:

[0050] s t ={N PV (t),δ(t)}

[0051] Action space a t Including the number N of solar cell panels PV increase or decrease and the increase or decrease of the proportionality coefficient δ a δ (t):

[0052]

[0053] Reward space r t Represented as:

[0054]

[0055] In the formula, e is a positive integer, and C sum According to S t Initial value calculation r t, generate a t to update S t ;

[0056] A plurality of combinations of learning rates and discount factors are set, and the solution result obtained by the combination with the highest cumulative reward after convergence is selected as the final output result, so that the best N PV , N B combination is obtained.

[0057] The application provides a renewable energy facility specification optimization system in a dual power supply scenario, which is used to execute the method described above.

[0058] From the above technical solution, the renewable energy facility specification optimization method in the dual power supply scenario provided by the embodiment of the application divides the small base stations in the region into N clusters, each cluster is equipped with one energy storage center and K small base stations, and each small base station is configured with a solar panel and an energy storage battery; an energy storage center energy sharing model and a total cost model of the renewable energy facility of each small base station are defined, the total cost model is the sum of the purchase cost, replacement cost, installation cost and grid power purchase cost of the solar panel and the energy storage battery; constraint conditions are provided according to the parameters of the sharing model and the total cost model, and the optimal solar panel specification and energy storage battery specification configuration combination is solved by minimizing the total cost of the renewable energy facility. The application optimizes the renewable energy facility specification from the facility construction and operation and maintenance point of view, uses local solar irradiance, regional rent, labor cost and other objective environmental data, explores the renewable energy facility scale most suitable for the regional characteristics from the perspective of minimizing the total cost of the facility, and considers the electricity consumption loss. BRIEF DESCRIPTION OF DRAWINGS

[0059] Figure 1 The renewable energy facility specification optimization method in the dual power supply scenario of the application is a flow chart.

[0060] Figure 2 It is a curve diagram of the solar energy harvested by a 1kW solar panel per hour.

[0061] Figure 3 It is a schematic diagram of the cumulative reward change trend corresponding to four learning rates.

[0062] Figure 4 It is a schematic diagram of the cumulative reward change trend corresponding to four discount factors.

[0063] Figure 5 It is a schematic diagram of the convergence comparison effect of the DDPG algorithm and the DQN algorithm. DETAILED DESCRIPTION

[0064] The technical solutions and technical effects of the application will be further described in detail in combination with the drawings of the application.

[0065] Reference Figure 1 As shown in the drawings, the present application provides a renewable energy facility specification optimization method in a dual power supply scenario, which is a specification configuration and optimization scheme in a dual power supply scenario of photovoltaic, energy storage, and large power grid. The purpose is to reduce the dependence on the power grid, maximize the utilization rate of renewable energy, and reduce resource waste. The implementation steps of the method include:

[0066] Step S1, in the dual power supply scenario, small base stations in the region are divided into N clusters, each cluster is equipped with an energy storage center and K small base stations, and each small base station is configured with a solar panel and an energy storage battery;

[0067] Step S2, defining an energy sharing model of the energy storage center and a total cost model of the renewable energy facility of each small base station, the total cost model being the sum of the purchase cost, replacement cost, installation cost and grid purchase cost of the solar panel and the energy storage battery;

[0068] Step S3, providing constraint conditions according to the parameters of the sharing model and the total cost model, and solving the optimal combination of solar panel specification and energy storage battery specification configuration by minimizing the total cost of the renewable energy facility; wherein Markov decision process is used to minimize the total cost.

[0069] In the dual power supply scenario, numerous small base stations are divided into multiple clusters. Each cluster is equipped with an energy storage center and a number of small base stations equipped with solar panels and energy storage batteries. The power grid and the base station are connected to the energy storage center. The power state of the base station is divided into three categories: insufficient supply, balanced supply and demand, and oversupply. Therefore, when configuring the solar panel specification (number of panels) and the energy storage battery specification of the base station (energy storage capacity), the laying cost and the use cost need to be considered. The present application defines each component of the total cost model:

[0070] The purchase cost in the total cost model of the renewable energy facility includes:

[0071] A relationship model between the solar panel specification and the solar energy harvesting amount of the small base station is established, and the solar energy collection power P of 1kW solar panel is expressed as: PV

[0072]

[0073] Wherein, η PV is the efficiency coefficient of the photovoltaic power converter; Rirr is the irradiance level of sunlight; a T is the temperature effect coefficient; T PV represents the average temperature of the solar panel assembly; T amb represents the ambient temperature; T nom ​This indicates the nominal operating temperature of the battery as measured under standard environmental conditions; the solar irradiance level under standard environmental conditions is 800 W / m². 2 The ambient temperature is 20℃ and the wind speed is 1m / s.

[0074] The specifications of small base station energy storage batteries are related to the peak value of the remaining energy in the battery. (Reference) Figure 2 The data shown represents the average hourly solar energy harvest of a 1kW solar panel in a certain location on a certain day. The data comes from an energy meteorological big data platform. Under normal weather conditions, this solar energy harvest trend is representative. Clearly, midday and its vicinity are the times with the most abundant solar energy. Assuming a sufficient number of solar panels, there must exist a continuous period of time Δt... p This ensures that the amount of solar energy harvested always exceeds the energy consumption of the small base station. If the energy storage center within the cluster is fully charged and other energy storage centers have no need for energy sharing, then the remaining energy in the energy storage battery reaches its theoretical peak.

[0075] Assume small base station B k A total of N are equipped PV A solar panel with a capacity of 1kW is used to define the energy storage battery specification N. B The theoretical peak energy of the battery Relational expression:

[0076]

[0077]

[0078] In the formula, δ is the proportionality coefficient, which represents the proportion of the energy storage battery specification to the theoretical peak of the remaining energy. In order to explore the optimal energy storage battery specification, the proportionality coefficient δ (0<δ≤1) is set to represent the proportion of the energy storage battery specification to the theoretical peak of the remaining energy. For small base station B k The amount of renewable energy generated through solar panels in time slot t; Δt p The time slot in which the harvested solar energy exceeds the base station's consumption (since solar energy is at its peak at noon, once it begins to exceed the base station's consumption, it will remain in a state of exceeding it for a continuous period of time. During this period, the energy in the energy storage battery continues to increase), Δt is the unit time slot; N is the power consumption of the base station in time slot t; PV For small base station B k Specifications of the solar panels provided;

[0079] The purchase cost C of solar panels and energy storage batteries pur The expression is defined as:

[0080] C pur =CPV • N PV + C B • N B (6)

[0081] where C PV and C B represent the cost per unit capacity of a 1 kW solar panel and a storage battery, respectively. As time goes on, solar panels and storage batteries can be damaged. Therefore, it is necessary to consider the replacement cost of solar panels and storage batteries in long-term operation. The replacement cost C cha in the total cost model of a renewable energy facility is represented as:

[0082]

[0083] where Y represents the estimated service life of a solar panel and a storage battery; L PV is the service life of a solar panel, which is usually a constant value; and L B represents the service life of a storage battery, which depends on the preset number of charge and discharge cycles and the depth of discharge of the battery in each cycle. The lower the depth of discharge, the longer the service life of the storage battery, i.e., the more charge and discharge cycles the storage battery can tolerate. Since solar energy collection is random and there is little solar radiation at night, the depth of discharge varies greatly in each cycle. The number of cycles N D corresponding to the depth of discharge D is obtained by curve fitting technique. The number of cycles of a storage battery is related to D and represented as:

[0084]

[0085] N D = 7855 · exp(-9.48 · D) + 2508 · exp(-1.605 · D) (9)

[0086]

[0087] where N D is the expression of the number of cycles of a storage battery corresponding to the depth of discharge D obtained by curve fitting technique; represents the number of cycles of a storage battery in the kth cycle when the depth of discharge is D and the operating temperature is T.

[0088] In addition to the equipment cost of solar panels and storage batteries, the site rental, system construction, and facility installation of a solar photovoltaic system also require certain costs. The installation cost depends on the land area occupied by the solar panels, and the installation cost C ins in the total cost model of a renewable energy facility is represented as:

[0089] C ins = A PV · N PV · (Y · C a + C b ) (11)

[0090] where A PV is the land area occupied by 1 kW solar panel; Y is the total number of years of operation; C a represents the annual rent paid for occupying a unit area of land; C b represents the total cost of constructing and installing a unit area of solar panel system and facilities.

[0091] In the dual power supply scenario, if the number of solar panels is insufficient, the energy obtained cannot meet the energy consumption of the small base station, and it may be necessary to purchase electricity from the power grid. To more accurately solve the optimal specifications of solar panels and energy storage batteries, it is necessary to consider the cost of purchasing electricity from the power grid. The cost of purchasing electricity from the power grid C gri is defined as:

[0092]

[0093] where G g is the cost of purchasing 1 kWh of electricity from the power grid, is the amount of electricity purchased from the power grid by the small base station B k in time slot t, Y is the total number of years of operation; the energy storage center shares electricity to K base stations, is the initial energy of the small base station B k in time slot t.

[0094] The total cost model of the renewable energy facility C sum is defined as:

[0095]

[0096] The energy storage center energy sharing model includes:

[0097] The remaining electricity k after the energy storage center completes charging and discharging with the small base station B in time slot t is represented as:

[0098]

[0099] where, represents the energy sharing relationship between the energy storage center and the base station in time slot t; is the initial energy of the energy storage center in time slot t; is the initial energy of the small base station B kshared power obtained from the energy storage center; when there is excess supply, for small base station B k excess energy transmitted to the energy storage center;

[0100] backup threshold of the energy storage center is expressed as:

[0101]

[0102] where B S represents the maximum capacity of the energy storage center, and γ is the correlation coefficient.

[0103] In the dual power supply scenario, the optimal specification solving problem of the solar panel and the energy storage battery is converted into the minimization problem of the total cost of the renewable energy facility. The minimization of step S3 is min(C sum ) and the constraint is defined as:

[0104]

[0105]

[0106] 0 < δ ≤ 1 (18g)

[0107] where GCD t is the grid connection depth of the small base station in time slot t, PRE min is the minimum power state of the energy storage center, PRE max is the maximum power state of the energy storage center, PRE h is the maximum value of the backup threshold, is the remaining power of the energy storage center after completing the last charge and discharge in time slot t.

[0108] where the total cost is established based on the GCD model and the BES model, so the total cost minimization needs to meet all the constraints contained in the two models, i.e. formulas (18a), (18b), (18c), (18d), (18e) and (18f), for limiting the range of the backup threshold and avoiding overcharging / over-discharging of the energy storage center. Formula (18g) ensures that the specification of the small base station energy storage battery will not exceed the theoretical peak value of the remaining energy, thereby avoiding waste of battery capacity.

[0109] Here, the calculation method of the grid connection depth is:

[0110]

[0111] K represents the number of base stations around one energy storage center.

[0112] In dealing with non-convex problems, traditional optimization algorithms tend to transform the problem into a convex one in order to take advantage of the properties of convex problems to achieve more efficient and stable optimization. Although this method can indeed solve non-convex problems, it is often accompanied by complex mathematical transformations and strict condition restrictions, making the optimization process cumbersome. Given the great advantages of DRL algorithms in autonomous learning and decision optimization, using this method to solve non-convex problems has a bright prospect. However, general DRL algorithms such as DQN often fail to produce satisfactory results when faced with problems involving continuous actions. For the DQN algorithm, it is impractical to find the action that maximizes the Q value through exhaustive search. The optimal sizing problem of solar panels and energy storage batteries is a continuous action space, so the DDPG algorithm suitable for continuous action space is chosen as the optimization algorithm.

[0113] To make the total cost minimization problem of renewable energy facilities meet the solution requirements of the DDPG algorithm, the cost function in formula (15) is re-expressed as a Markov Decision Process (MDP). In step S3, the Markov Decision Process is used to solve min(C sum ) process includes:

[0114] The total cost min(C sum ) is re-expressed as a Markov Decision Process, where:

[0115] State space: The state space is used to store the state of the environment, providing guidance for the generation of actions. The state space S t includes the number of 1 kW solar panel blocks N PV and the proportion coefficient δ of the energy storage battery specification relative to the theoretical peak value of the remaining energy:

[0116] s t = {N PV (t), δ(t)} (20)

[0117] Action space: The action space can record the actions performed by the agent. These actions will contribute to the generation of new states. The action space a t includes the increase and decrease of the number of solar panel blocks N PV and the increase and decrease of the proportion coefficient δ a δ (t):

[0118]

[0119] ​Reward space: reward is the feedback of the environment to the action after the action is performed. In the DDPG algorithm, the cumulative total reward will reach the maximum value as the optimization proceeds. However, the optimization objective in formula (15) is to minimize the total cost. Therefore, the reward is set as the inverse of the cost. At the same time, in order to ensure that the algorithm can successfully converge during simulation, it is necessary to make the numerical values of the state space, action space and reward space be in the same order of magnitude. The reward space r t is expressed as:

[0120]

[0121] In the formula, e is a positive integer, which can adjust the order of magnitude of the reward, so that the algorithm converges faster; C sum According to S t The initial value of r t is calculated, and a t is generated t to update S t ;

[0122] Set a combination of multiple learning rates and discount factors, select the solution result of the combination with the highest cumulative reward after convergence as the final output result, and obtain the best N PV , N B combination.

[0123] Further, the renewable energy facility specification optimization system in a dual power supply scenario is used to execute the foregoing method to optimize the calculation of the renewable energy facility specification.

[0124] In order to explore the effect of the method of the application, the proposed DDPG algorithm is simulated and analyzed by Python 3.8.8 and TensorFlow 2.6.0. The simulation will be carried out on a 64-bit Windows 10 system computer equipped with Inter(R) i5-9400 CPU. The simulation is based on the dual power supply strategy recommended in the foregoing, and other simulation parameter settings are shown in Table 1.

[0125] Table 1 DDPG algorithm simulation parameter settings

[0126]

[0127] In order to ensure that the DDPG algorithm can converge as soon as possible, it is necessary to explore the best learning rate and the best discount factor. Figure 3 The cumulative reward change trend corresponding to the four learning rate combinations is shown in Figure 2. Due to the large fluctuation of the cumulative reward of the agent during learning, the adjacent average method is used to simply process the data to highlight the change trend. When the training round number reaches 500 rounds, the cumulative reward corresponding to the four learning rates begins to rise. Among them, α u = 0.001, α Q= 0.0001, the cumulative reward rises fastest and converges earliest. However, the cumulative reward value after convergence is low. To obtain the optimal specifications of solar panels and energy storage batteries, the learning rate combination with the highest cumulative reward after convergence should be selected, i.e., a u = 0.001, a Q = 0.001.

[0128] Figure 4 The cumulative reward change trends corresponding to the four discount factors are shown in FIG. 6. When γ = 0.9, the convergence is slowest and the cumulative reward is lowest. When γ = 0.7, the cumulative reward can converge normally, and the cumulative reward value after convergence is highest. Based on the above analysis of the learning rate and the discount factor, the DDPG algorithm will be compared and analyzed in the following using a u = 0.001, a Q = 0.001, γ = 0.7.

[0129] Comparison and analysis of DDPG algorithm:

[0130] To verify the superiority of the selected DDPG algorithm, the DQN algorithm and GA are used for comparison. Figure 5 FIG. 7 is a comparison chart of the convergence of the DDPG algorithm and the DQN algorithm. The DQN algorithm still has great volatility even after the adjacent average method is used. In addition, the DQN algorithm successfully converges at about 1000 rounds, and the cumulative reward after convergence is significantly lower than that of the DDPG algorithm. This is because the DQN algorithm needs to discretize the action when dealing with continuous action problems, which leads to the neglect of many actions. The DDPG algorithm has superiority when dealing with continuous action problems.

[0131] Table 2 shows the optimal specifications of solar panels and energy storage batteries given by the three algorithms, as well as their corresponding costs. The cost of the specifications given by the genetic algorithm is only 22002.67 yuan higher than that of the DDPG algorithm. However, due to the large value range of the solar panel and energy storage battery specifications, there are countless values in the range of [5, 8] alone, and it is difficult for the genetic algorithm to explore the truly optimal specifications in a limited number of generations. The cost of the specifications given by the DQN algorithm is extremely high, exceeding that of the DDPG algorithm by 543660.88 yuan, and is completely unsuitable for solving the specification optimization problem. In summary, the recommended DDPG algorithm can indeed solve the optimal specifications based on the given parameters and effectively reduce the total cost of renewable energy facilities.

[0132] Table 2 Optimal specifications and corresponding costs given by three algorithms

[0133]

[0134] The application provides a configuration optimization strategy of a base station in a dual power supply scene, a total cost model for a renewable energy facility specification optimization scheme of the base station is constructed by considering various cost factors, renewable energy utilization is maximized, and the dependence on a power grid in the dual power supply is reduced; in addition, an energy storage center shares power and collects excess power with each base station, and the renewable energy utilization rate can be effectively improved.

[0135] The above only discloses preferred embodiments of the application, and of course cannot limit the scope of the rights of the application, and those skilled in the art can understand that all or part of the processes of the above embodiments are implemented, and equivalent changes made according to the claims of the application still belong to the scope covered by the application.

Claims

1. A method for optimizing the specifications of renewable energy facilities in a dual-power supply scenario, characterized in that, Comprising: Step S1, in a dual power supply scenario, small base stations in a region are divided into N clusters, each cluster is equipped with an energy storage center and K small base stations, and each small base station is configured with a solar panel and an energy storage battery; Step S2, defining an energy storage center energy sharing model and a total cost model of the renewable energy facilities of each small base station, the total cost model being the sum of the purchase cost, replacement cost, installation cost and grid purchase cost of the solar panel and the energy storage battery; Step S3, providing constraint conditions according to the parameters of the sharing model and the total cost model, and solving the optimal solar panel specification and energy storage battery specification configuration combination by minimizing the total cost of the renewable energy facilities; wherein Markov decision process is used to minimize the total cost.

2. The method for renewable energy facility sizing in dual power supply scenarios of claim 1, wherein, The definition of the purchase cost in the total cost model of the renewable energy facilities includes: A relationship model between the solar panel specification and the solar energy harvesting amount of a small base station is established, and the solar energy collection power P of a 1 kW solar panel is PV is represented as: wherein η PV is the efficiency coefficient of the photovoltaic power converter; Rirr is the irradiance level of the sunlight; a T is the temperature effect coefficient; T PV represents the average temperature of the solar panel assembly; T amb represents the ambient temperature; T nom represents the nominal operating temperature of the cell measured under standard ambient conditions; Definition of the energy storage battery specification N B in relation to the theoretical peak value of the remaining energy in the battery : In the formula, δ is a proportional coefficient, indicating the proportion of the energy storage battery specification relative to the theoretical peak value of the remaining energy; For small base station B k The renewable energy power generation capacity obtained by the solar panel at time slot t; Δt p For the time slot in which the harvested solar energy exceeds the consumption of the base station, and Δt is a unit time slot; N is the power consumption of the base station at time slot t; PV For small base station B k The specification of the equipped solar panel; The purchase cost C of the solar panels and energy storage batteries pur The expression is defined as: C pur = C PV • N PV + C B • N B where C PV and C B represent the cost per unit capacity of a 1 kW solar panel and energy storage battery, respectively.

3. The method for renewable energy facility sizing in dual power supply scenarios of claim 2, wherein, defining said replacement cost C in a total cost model of said renewable energy facility cha is represented as: Wherein, Y represents the estimated service life of the solar panel and the energy storage battery; L PV is the service life of the solar panel; L B represents the service life of the energy storage battery, and the cycle number of the energy storage battery related to, is represented as: N D = 7855 - exp(-9.48 - D) + 2508 - exp(-1.605 - D) wherein N D is an expression of the number of cycles of the energy storage battery corresponding to the discharge depth D obtained by the curve fitting technique; represents the number of cycles of the energy storage battery corresponding to the kth cycle when the discharge depth is D and the operating temperature is T.

4. The method for renewable energy facility sizing in dual power supply scenarios of claim 3, wherein, defining said installation cost C in a total cost model of said renewable energy facility ins is represented as: C ins = A PV · N PV · (Y · C a + C b ) wherein A PV is the footprint of a 1 kW solar panel; Y is the total number of years of operation; C a represents the annual rent payable per unit area of land; C b represents the total cost of construction and installation of the solar panel system per unit area.

5. The method for renewable energy facility sizing in dual power supply scenarios of claim 4, wherein, defining the grid electricity purchase cost C in a total cost model of the renewable energy facility gri is represented as: where G g the cost required to purchase 1 kWh of grid electricity, is the small base station B k the electricity purchased from the grid in time slot t, Y is the total number of operating years; the energy storage center shares electricity to K base stations, is the initial energy of the small base station B k in time slot t.

6. The method for renewable energy facility sizing in dual power supply scenarios of claim 5, wherein, The total cost model C of the renewable energy facility is defined as: sum is defined as:

7. The method for renewable energy plant sizing in dual-feed scenarios of claim 6, wherein, The energy storage center energy sharing model includes: The energy storage center completes charging and discharging between the small base station B and the small base station C in the time slot t k The remaining power after the energy storage center completes charging and discharging between the small base station B and the small base station C in the time slot t is represented as: wherein represents the energy sharing relationship between the energy storage center and the base station in time slot t; is the initial amount of electricity for the energy storage center in time slot t; is the small base station B k the shared electricity obtained from the energy storage center; when there is excess supply, is the small base station B k the excess energy transmitted to the energy storage center; Energy storage center backup threshold is represented as: In the formula, B S represents the maximum capacity of the energy storage center, and γ is the correlation coefficient.

8. The method for renewable energy facility sizing in dual power supply scenarios of claim 7, wherein, The minimization of step S3 is solved as min(C sum ) and the constraints are defined as: 0 < BES t ≤ 0.75 PRE max • B S , if GCD t = 0 0 < δ ≤ 1 where GCD t is the grid connection depth of the small base station at time slot t, PRE min is the minimum state of charge of the energy storage center, PRE max is the maximum state of charge of the energy storage center, PRE h is the maximum value of the reserve threshold, is the remaining amount of energy of the energy storage center after the last charge and discharge at time slot t.

9. The method for renewable energy facility sizing in dual power supply scenarios of claim 8, wherein, Solving min(C sum ) by Markov Decision Process in step S3 includes: Rewrite the total cost min(C sum ) as a Markov Decision Process, where: State space S t Number of 1 kW solar panel blocks N PV Proportionality factor δ of the remaining energy theoretical peak value to the energy storage battery specification s t = {N PV (t), δ(t)} Action space a t Including the number of solar cell panels N PV Increase and decrease And the increase and decrease of the proportional coefficient δa δ (t): reward space r t is represented as: where e is a positive integer, C sum According to S t Initial value calculation r t , generate a t To update S t ; A plurality of combinations of learning rates and discount factors are set, and a solution result obtained by a combination with the highest cumulative reward after convergence is selected as a final output result, and an optimal N PV is obtained B .

10. A renewable energy facility specification optimization system for a dual power supply scenario, characterized in that, A device for performing the method of any one of claims 1-9. A device for performing the method of any one of claims 1-9.

Images