A method, apparatus and equipment for energy dispatching of a distributed photovoltaic power generation system

Abstract

This invention provides an energy dispatching method, apparatus, and equipment for a distributed photovoltaic (PV) power generation system, belonging to the field of energy storage technology. It solves the problems of low energy absorption rate and high economic losses in dispatching for distributed PV power generation systems. The method includes: acquiring the PV power generation, energy storage charging / discharging power, and grid interaction power of the distributed PV power generation system during multiple dispatching periods; preprocessing the PV power generation, energy storage charging / discharging power, and grid interaction power to obtain multiple input vectors; and processing the multiple input vectors to obtain energy dispatching data, which is used to control the operating status of the distributed PV power generation system. This solution improves the PV absorption rate, reduces operating costs, and enhances dynamic response capabilities.

Description

Technical Field

[0001] This invention relates to the field of energy storage technology, and in particular to an energy dispatching method, apparatus and equipment for a distributed photovoltaic power generation system. Background Technology

[0002] With the continuous increase in the penetration rate of distributed photovoltaic (PV) power generation systems, their energy dispatch technology faces new technical challenges. Existing dispatch methods are mainly divided into three categories: centralized optimization control, distributed collaborative optimization, and data-driven dispatch. Centralized methods rely on a central controller and employ model predictive control or mixed-integer programming for global optimization, but they suffer from inherent problems such as high communication bandwidth pressure and high risk of single-point failures. While distributed optimization technology improves system reliability, it suffers from slow convergence speed and insufficient dynamic response in practical applications. In recent years, data-driven methods have achieved breakthroughs in the accuracy of PV output prediction, but they are highly dependent on data quality and lack sufficient embedding of physical constraints.

[0003] Existing technologies suffer from four core problems: First, the coupling of uncertainties between the source and load sides intensifies, with traditional model-driven methods experiencing scheduling mismatch rates exceeding 20% ​​under scenarios of fluctuating photovoltaic output and sudden load changes. Second, multi-timescale coordination is lacking, with decoupling biases in the day-ahead, intraday, and real-time optimization objectives, leading to an increase of 1.2-1.5 charge-discharge cycles per day in the energy storage system. Third, the cyber-physical coordination efficiency under distributed architecture is low, with existing consensus algorithms incurring communication overhead more than three times that of traditional methods in a 100-node system. Finally, user-side demand response participation is insufficient, with existing incentive mechanisms resulting in a dispatchable load ratio of less than 35%, severely restricting the actual effectiveness of scheduling strategies. These problems lead to photovoltaic grid integration rates generally below 85% in existing systems, and the levelized cost of energy (LCOE) of scheduling strategies falling short of the theoretical optimum by more than 15%. Summary of the Invention

[0004] This invention provides an energy dispatching method, apparatus, and equipment for a distributed photovoltaic power generation system, which solves the problems of low energy absorption rate and high economic loss in dispatching of distributed photovoltaic power generation systems.

[0005] To solve the above-mentioned technical problems, the technical solution of the present invention is as follows:

[0006] This invention provides an energy dispatching method for a distributed photovoltaic power generation system, comprising:

[0007] Acquire the photovoltaic power generation, energy storage charging and discharging power, and grid interaction power of the distributed photovoltaic power generation system during multiple scheduling periods;

[0008] The photovoltaic power generation, the energy storage charging and discharging power, and the grid interaction power are preprocessed to obtain multiple input vectors;

[0009] The multiple input vectors are processed to obtain energy scheduling data, which is used to control the operating status of the distributed photovoltaic power generation system.

[0010] Optionally, obtaining the photovoltaic power generation, energy storage charging and discharging power, and grid interaction power of the distributed photovoltaic power generation system during multiple scheduling periods includes:

[0011] Based on multiple metering devices in the distributed photovoltaic power generation system, the photovoltaic power generation, energy storage charging and discharging power, and grid interaction power of the distributed photovoltaic power generation system are obtained over multiple natural days. The number of the multiple natural days is a first preset value, the number of scheduling periods included in each natural day is a second preset value, and the duration of each scheduling period is a third preset value.

[0012] Optionally, the photovoltaic power generation, the energy storage charging and discharging power, and the grid interaction power are preprocessed to obtain multiple input vectors, including:

[0013] Anomaly detection processing is performed on the photovoltaic power generation, the energy storage charging and discharging power, and the grid interaction power to obtain sampled data after anomaly detection processing;

[0014] The sampled data after the anomaly detection process is filled with missing values ​​and deleting duplicate data to obtain preprocessed sampled data;

[0015] The preprocessed sampled data is integrated to obtain multiple input vectors.

[0016] Optionally, the multiple input vectors are processed to obtain energy scheduling data, including:

[0017] The multiple input vectors are input into an energy scheduling model, which iteratively processes the input vectors multiple times based on a fitness function to obtain energy scheduling data; wherein, the expression for the fitness function is:

[0018] F(x)=f(x)+αP e (x)+βP i (x)

[0019] Where F(x) is the fitness function, f(x) is the original objective function, α and β are penalty factors, and P e (x) is the equality constraint penalty function, P i (x) is the inequality constraint penalty function.

[0020] Optionally, the inertial weights are updated using a nonlinear dynamic adaptive update method, and the learning factors are updated according to an asynchronous linear law.

[0021] Optionally, the expression for the equality constraint penalty function is:

[0022]

[0023] Among them, g i (x) is the equality constraint function, ε is the tolerance, and i and m are natural numbers;

[0024] The expression for the inequality constraint penalty function is:

[0025]

[0026] Among them, h i (x) is the inequality constraint function, and n is a natural number.

[0027] Optionally, the constraints of the energy dispatch model include one or more of the following: power balance constraints, state of charge constraints, charge and discharge power limit constraints, and grid interaction constraints.

[0028] This invention also provides an energy dispatching device for a distributed photovoltaic power generation system, comprising:

[0029] The acquisition module is used to acquire the photovoltaic power generation, energy storage charging and discharging power, and grid interaction power of the distributed photovoltaic power generation system during multiple scheduling periods;

[0030] The processing module is used to preprocess the photovoltaic power generation, the energy storage charging and discharging power, and the grid interaction power to obtain multiple input vectors; and to process the multiple input vectors to obtain energy dispatch data, which is used to control the operating status of the distributed photovoltaic power generation system.

[0031] This invention also provides a computing device, including: a processor and a memory storing a computer program, wherein the computer program, when run by the processor, executes the above-described method.

[0032] This invention also provides a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the above-described method.

[0033] The technical solution of the present invention has at least the following effects:

[0034] The above-mentioned solution of the present invention obtains the photovoltaic power generation, energy storage charging and discharging power, and grid interaction power of the distributed photovoltaic power generation system in multiple scheduling periods; preprocesses the photovoltaic power generation, energy storage charging and discharging power, and grid interaction power to obtain multiple input vectors; processes the multiple input vectors to obtain energy scheduling data, which is used to control the operating status of the distributed photovoltaic power generation system, thereby improving the photovoltaic absorption rate, reducing operating costs, and enhancing dynamic response capabilities. Attached Figure Description

[0035] Figure 1 This is a flowchart of the energy dispatching method for a distributed photovoltaic power generation system provided in an embodiment of the present invention;

[0036] Figure 2 This is a structural diagram of the distributed photovoltaic power generation system provided in an embodiment of the present invention;

[0037] Figure 3 This is a structural diagram of the energy dispatching device for a distributed photovoltaic power generation system provided in an embodiment of the present invention;

[0038] Figure 4 This is a schematic diagram of the structure of a computing device provided in an embodiment of the present invention. Detailed Implementation

[0039] Exemplary embodiments of the invention will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the invention are shown in the drawings, it should be understood that the invention may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this invention will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art.

[0040] like Figure 1 As shown, an embodiment of the present invention proposes an energy dispatching method for a distributed photovoltaic power generation system, comprising:

[0041] Step 11: Obtain the photovoltaic power generation, energy storage charging and discharging power, and grid interaction power of the distributed photovoltaic power generation system during multiple scheduling periods;

[0042] Step 12: Preprocess the photovoltaic power generation, the energy storage charging and discharging power, and the grid interaction power to obtain multiple input vectors;

[0043] Step 13: Process the multiple input vectors to obtain energy scheduling data, which is used to control the operating status of the distributed photovoltaic power generation system.

[0044] In this embodiment, the entire operating time is first divided into multiple consecutive time periods, such as 15 minutes, 30 minutes, or 1 hour, to achieve a more refined management of the system's energy flow, adapting to changes in factors such as sunlight intensity and electricity demand within different time periods. The system acquires the photovoltaic power generation, energy storage charging / discharging power, and grid interaction power of the distributed photovoltaic power generation system across multiple time periods. Photovoltaic power generation refers to the amount of solar energy converted into electrical energy by the photovoltaic panels within a specific time period, which is affected by various factors such as sunlight intensity, temperature, and the performance of the photovoltaic panels. For example, photovoltaic power generation is higher during the sunny midday hours, while it decreases on cloudy days or in the evening. Energy storage charging / discharging power refers to the power of energy storage devices typically equipped in distributed photovoltaic power generation systems, used to store excess electrical energy or release electrical energy when power generation is insufficient. Energy storage charging / discharging power represents the charging power (absorbing electrical energy from the system) or discharging power (releasing electrical energy to the system) of the energy storage device within a specific time period. For example, when photovoltaic power generation exceeds electricity demand, the energy storage device charges; conversely, when power generation is insufficient, the energy storage device discharges. Grid-interchange power refers to the power exchanged between a distributed photovoltaic (PV) power generation system and the external power grid. When the system's power generation exceeds local electricity demand, the excess power is fed into the grid; when the power generation is insufficient to meet local electricity demand, the system draws power from the grid.

[0045] Because the raw data may contain noise, outliers, or inconsistent data formats, preprocessing is necessary to improve data quality. Preprocessing operations include removing outliers and duplicates, and filling in missing values.

[0046] The preprocessed photovoltaic power generation, energy storage charging and discharging power, and grid interaction power are combined into a vector form in a certain order. For example, each scheduling period corresponds to an input vector, and the elements in the vector are the photovoltaic power generation, energy storage charging and discharging power, and grid interaction power for that period.

[0047] Based on the information in the input vector, analyze the energy flow of the system and predict future energy demand and power generation trends.

[0048] The energy dispatch data obtained after processing the input vector contains information on how to control the operating status of the distributed photovoltaic (PV) power generation system. For example, the energy dispatch data may indicate the charging or discharging power of energy storage devices during specific periods, the output power adjustment strategy of PV panels, and the amount of energy exchanged with the grid. Based on the energy dispatch data, the system can automatically adjust the operating parameters of each component to achieve efficient and stable energy dispatch. For instance, when it is predicted that solar irradiance will decrease in the future, the energy dispatch data may instruct energy storage devices to increase their charging power in advance to store sufficient energy; when electricity demand suddenly increases, the system can adjust the output power of the PV panels or increase the amount of energy obtained from the grid based on the dispatch data.

[0049] This technical solution, through precise energy dispatch, can adjust photovoltaic power generation, energy storage charging and discharging, and grid interaction in real time based on factors such as sunlight intensity and electricity demand, ensuring efficient energy utilization and reducing energy waste throughout the system. Simultaneously, the energy dispatch method can dynamically adjust based on the system's real-time status, effectively addressing uncertainties such as changes in sunlight and fluctuations in electricity demand, ensuring the stable operation of the distributed photovoltaic power generation system. Reasonable energy dispatch can optimize the use of energy storage devices, reducing the number and depth of charge / discharge cycles, extending their lifespan, and lowering equipment maintenance costs. Through intelligent interaction with the grid, electricity can be purchased from the grid during off-peak hours and sold back to the grid during peak hours, reducing electricity costs. This energy dispatch method also helps to better integrate the electricity generated by distributed photovoltaic power generation systems, increasing the proportion of renewable energy absorbed by the local grid, reducing dependence on traditional fossil fuels, and promoting the optimization and sustainable development of the energy structure.

[0050] In an optional embodiment of the present invention, step 11 may include:

[0051] Step 111: Based on multiple metering devices in the distributed photovoltaic power generation system, obtain the photovoltaic power generation, energy storage charging and discharging power, and grid interaction power of the distributed photovoltaic power generation system over multiple natural days. The number of the multiple natural days is a first preset value, the number of scheduling periods included in each natural day is a second preset value, and the duration of each scheduling period is a third preset value.

[0052] In this embodiment, a distributed photovoltaic (PV) power generation system typically consists of multiple PV modules, energy storage devices (such as batteries), and grid connection equipment. These components and devices generate and consume electrical energy at different times, thus requiring real-time monitoring and recording of their power output. Multiple metering devices are installed in the distributed PV power generation system to measure the output power of the PV modules, the charging and discharging power of the energy storage devices, and the power at the grid connection point.

[0053] To obtain more comprehensive data, it is necessary to collect power data from distributed photovoltaic power generation systems over multiple calendar days. The number of calendar days is a first preset value, which can be determined based on actual needs and data analysis requirements. For example, data from a continuous week (7 days) or a month (30 days) can be collected to cover different weather conditions and load variations.

[0054] Each calendar day is divided into multiple scheduling periods, and the number of scheduling periods in each calendar day is a second preset value. The second preset value can be determined based on the system's scheduling needs and the accuracy requirements of data analysis; the duration of each scheduling period is a third preset value, which is determined when dividing the scheduling periods. For example, a day can be divided into 24 periods, each lasting 1 hour; or it can be divided into 96 periods, each lasting 15 minutes.

[0055] During each scheduling period, the metering device collects real-time data on the photovoltaic power generation, energy storage charging and discharging power, and grid interaction power of the distributed photovoltaic power generation system, and transmits the collected data to the data acquisition system. The data acquisition system stores the collected data in a database for subsequent data analysis and processing.

[0056] In an optional embodiment of the present invention, step 12 may include:

[0057] Step 121: Perform anomaly detection processing on the photovoltaic power generation, the energy storage charging and discharging power and the grid interaction power to obtain the sampled data after anomaly detection processing;

[0058] Step 122: Fill missing values ​​and remove duplicate data from the sampled data after anomaly detection processing to obtain preprocessed sampled data;

[0059] Step 123: Integrate the preprocessed sampled data to obtain multiple input vectors.

[0060] In this embodiment, outliers may exist in the originally collected photovoltaic power generation, energy storage charging and discharging power, and grid interaction power data. These outliers may be caused by sensor malfunctions, data transmission errors, extreme weather conditions, etc. The presence of outliers can seriously affect the accuracy of subsequent data analysis and energy dispatch models, therefore, anomaly detection and processing are necessary.

[0061] Anomaly detection methods can employ statistical approaches; for example, calculating the mean and standard deviation of the data, and considering data exceeding a certain multiple of the standard deviation as outliers. For instance, for a dataset, if a data point deviates from the mean by more than three times the standard deviation, it can be considered an outlier. Alternatively, a rule-based approach can be used, where rules are established based on the physical characteristics and operational experience of the distributed photovoltaic (PV) power generation system to determine data anomalies. For example, PV power generation should be zero or close to zero at night; non-zero values ​​may indicate anomalies. Detected outliers can be deleted or replaced with appropriate alternative values. These alternative values ​​can be the average of adjacent data points or values ​​predicted based on historical data and trends. After anomaly detection processing, the sampled data is obtained.

[0062] Data loss may occur during data acquisition and transmission, resulting in missing power data at certain points in time. Missing values ​​affect the integrity and continuity of the data, thereby impacting subsequent analysis and modeling.

[0063] Missing data points can be filled using methods such as mean imputation, where the mean of the sequence containing the missing data point (e.g., a photovoltaic power generation sequence) is used. Alternatively, interpolation methods, such as linear interpolation, can be employed, where the missing value is calculated using a linear equation based on known data points before and after the missing data point. For example, if the photovoltaic power generation at times t1 and t3 is known, and the missing value at time t2 is required, it can be calculated using a linear interpolation formula, expressed as:

[0064]

[0065] Among them, P t2 Let P be the photovoltaic power generation at time t2. t1 Let P be the photovoltaic power generation at time t1. t3 The photovoltaic power generation at time t3;

[0066] During data acquisition, duplicate data collection may occur, resulting in duplicate records. Duplicate data increases redundancy and affects the efficiency and accuracy of data processing. Therefore, it is necessary to check the sampled data after anomaly detection and remove duplicate records. This can be done by comparing key fields such as timestamps and power values ​​to determine if data is duplicated. After filling missing values ​​and removing duplicate data, preprocessed sampled data is obtained.

[0067] Preprocessed sampling data typically consists of dispersed photovoltaic power generation, energy storage charging and discharging power, and grid interaction power data. To input this data into an energy dispatch model for analysis and decision-making, it needs to be integrated into multiple input vectors. Each input vector comprehensively reflects the energy state and characteristics of the system at a specific moment or time period. Integration methods include:

[0068] Step 1231: Align by time. Align the photovoltaic power generation, energy storage charging and discharging power, and grid interaction power data according to the timestamp to ensure that the data at each moment corresponds together.

[0069] Step 1232: Construct a vector. For each moment or time period, combine the corresponding photovoltaic power generation, energy storage charging and discharging power, and grid interaction power values ​​into a vector. For example, if each scheduling time period is taken as a unit, then the i-th input vector can be represented as [P pv (i),P ess (i),P grid (i)], where P pv (i) represents the power generation of the i-th photovoltaic power generation, P ess (i) represents the charging and discharging power of the i-th energy storage unit (charging is positive, discharging is negative), P grid (i) represents the power exchanged between the grid and the grid (selling electricity to the grid is positive, and buying electricity from the grid is negative).

[0070] Step 1233: According to the time sequence, combine the vectors corresponding to each moment or time period to form multiple input vectors. The multiple input vectors can be regarded as a population, wherein the population can be represented as:

[0071] x i =[P pv (i),P ess (i),P grid (i)]

[0072] Where 0≤i≤T, and T is the total number of scheduling periods.

[0073] These input vectors will serve as input to the energy scheduling model for subsequent analysis and decision-making.

[0074] In an optional embodiment of the present invention, step 13 may include:

[0075] Step 131: Input the multiple input vectors into the energy scheduling model. The energy scheduling model iteratively processes the multiple input vectors according to the fitness function to obtain energy scheduling data; wherein, the expression of the fitness function is:

[0076] F(x)=f(x)+αP e(x)+βP i (x)

[0077] Where F(x) is the fitness function, f(x) is the original objective function, α and β are penalty factors, and P e (x) is the equality constraint penalty function, P i (x) is the inequality constraint penalty function.

[0078] In this embodiment, the energy scheduling model is an intelligent model built on mathematical algorithms and logical rules. It can analyze and calculate based on the input system state information (i.e., multiple input vectors) to find the optimal energy scheduling scheme.

[0079] After multiple input vectors are fed into the energy scheduling model, the model performs multiple iterations. The purpose of iteration is to continuously optimize the scheduling scheme, gradually approaching the optimal solution. In each iteration, the model generates a temporary energy scheduling scheme based on the current input vectors and its internal algorithm, and evaluates the scheme using a fitness function. Based on the evaluation results, the model adjusts the input vectors or internal parameters, and then proceeds to the next iteration, until stopping conditions are met (such as reaching the maximum number of iterations, or the fitness function converging).

[0080] The fitness function is a criterion used in the iterative process of an energy scheduling model to evaluate the merits of different scheduling schemes. It calculates a fitness value for each scheduling scheme by comprehensively considering the original objective function and various constraints. A higher fitness value indicates that the scheduling scheme better meets the system's requirements. In this embodiment, the expression for the fitness function is:

[0081] F(x)=f(x)+αP e (x)+βP i (x)

[0082] Where F(x) is the fitness function, f(x) is the original objective function, α and β are penalty factors, and P e (x) is the equality constraint penalty function, P i (x) is the inequality constraint penalty function.

[0083] The original objective function f(x) reflects the core objective of energy dispatch in a distributed photovoltaic power generation system. In this embodiment, the original objective function f(x) minimizes the system's operating costs, including equipment depreciation costs and maintenance costs. The expression for the original objective function is:

[0084]

[0085] Among them, C grid (i) represents the electricity price for time period i, C essλ is the energy storage charging and discharging loss cost coefficient, and λ is the curtailment penalty coefficient. The maximum photovoltaic output is for time period i.

[0086] The velocity v of the i-th input vector (also called particle i) in the t-th iteration i (t) and position x i The updated formula for (t) is:

[0087] v i (t+1)=ωv i (t)+c1r1(t)[p i (t)-x i (t)]+c2r1(t)[p g (t)-x i (t)]

[0088] x i (t+1)=x i (t)+v i (y+1), i = 1, 2, ..., T

[0089] In the formula, ω is the inertia weight; C1 and c2 are learning factors, representing the particle's own cognition and social cognition, respectively; r1 and r2 are independent random variables between [0,1] that follow a uniform distribution; p i (t) represents the historical optimal position of particle i; p g (t) represents the optimal position of the particles in the entire population.

[0090] Because particles in the population may initially cluster in the vicinity of a local optimum, the mutual influence between particles quickly stabilizes the search process there, losing particle diversity and becoming trapped in a local optimum, leading to significant errors in the results. The improvement strategy involves two aspects: adjusting the parameter update method and adding penalty functions to handle constraints.

[0091] (1) Improvement of inertia weight and learning factor

[0092] The inertia weight ω and learning factors c1 and c2 are key parameters of the model. Their impact on the model's optimization ability is as follows: increasing the value of ω can improve the algorithm's global search ability, while decreasing it can enhance the local search ability. c1 and c2 represent the weights of individual experience and group experience in the population search process, respectively. Therefore, to improve the algorithm's global search ability, c2 should be increased and c1 decreased. Conversely, if it is necessary to improve the algorithm's local search ability, c1 should be increased and c2 decreased.

[0093] The inertia weight ω is updated using a nonlinear dynamic adaptive update method, while the learning factors c1 and c2 are updated according to an asynchronous linear law. In the early stage of optimization iteration, the values ​​of the inertia weight ω and the learning factor c1 are relatively large, while the value of c2 is relatively small. In the later stage of iteration, the opposite is true. This approach helps to speed up the search in the early stage of iteration, prevent the model from converging to a local optimum too early, and improve the accuracy of optimization in the later stage of iteration, thereby improving the accuracy of the model.

[0094]

[0095] In the formula, ω max ω min ω represents the upper and lower limits of the inertia weight; f is the fitness function value of a particle; f avg f is the average value of the current population fitness function; max This represents the maximum value of the current population fitness function.

[0096]

[0097] In the formula, c 1s c 1e Here are the initial and final values ​​of the learning factor c1; c 2s c 2e Here are the initial and final values ​​of the learning factor c2; t iter t is the current iteration number; iter,max This represents the maximum number of iterations for the algorithm.

[0098] (2) Add penalty functions to handle constraints

[0099] Because energy scheduling models contain other complex equality and inequality constraints, such as electrical balance constraints, thermal balance constraints, and energy storage constraints, simply restricting the position of particles in space is insufficient to achieve optimal solutions. Therefore, a penalty function method is added to the model, transforming equality constraints into penalty terms in the fitness function. In this way, when a particle violates a constraint, its fitness value (i.e., its fitness function value) is penalized, thus guiding the particle to move towards a region that satisfies the constraints.

[0100] For complex equality constraints, they are all transformed into g. i Let (x) = 0, where i = 1, 2, ..., m, and m be the number of equality constraints. Then, it is transformed into inequality constraints g. i (x)≤ε and g i (x)≥-ε, where ε is a very small positive number, i.e., tolerance, used to handle the accuracy issues in numerical computation. The penalty term can be defined as:

[0101]

[0102] In the formula, max(0,|g i (x)|-ε) ensures that only when |g i The penalty term is non-zero only when (x)|>ε, that is, when the constraint is violated.

[0103] For inequality constraints, transform them all into h. i If (x)≤0, the penalty term can be defined as:

[0104]

[0105] In the formula, max(0,h) j (x) ensures that only when h j The penalty is non-zero only when (x)>0, that is, when the constraint is violated.

[0106] Integrating the two penalty terms mentioned above into the original objective function f(x) yields a new objective function F(x):

[0107] F(x)=f(x)+αP e (x)+βP i (x)

[0108] In one optional embodiment of the present invention, the constraints of the energy dispatch model include one or more of the following: power balance constraints, state of charge constraints, charge and discharge power limit constraints, and grid interaction constraints.

[0109] In this embodiment, the energy dispatch method for distributed photovoltaic (PV) power generation systems aims to efficiently and rationally allocate and utilize the electricity generated by the PV system through intelligent algorithms and strategies, thereby meeting user load demands, optimizing system performance, and reducing operating costs. Specifically, the power balance constraint requires that the total output power of the distributed PV power generation system (including the power directly output by the PV panels and the power released by the energy storage system) must be equal in real time to the total power demand of the user load plus the power interacting with the grid. Power balance is the foundation for stable system operation, helping to prevent voltage fluctuations and frequency deviations, ensuring that the system can provide sufficient electricity to meet user power needs at any time, and reducing excess or insufficient electricity through precise supply and demand matching, thus improving energy utilization efficiency. The power balance constraint can be expressed as:

[0110] P pv (t)+P ess (t)+P grid (t)=P load (t)

[0111] In the formula, P pv (t) represents the photovoltaic power generation at time t, P ess (t) represents the energy storage charging and discharging power at time t, P grid(t) represents the grid interaction power at time t, P load (t) represents the load power demand at time t.

[0112] State of Charge (SOC) constraint refers to the requirement that the electrical charge level of an energy storage system must be maintained within a certain range. Typically, upper and lower limits for SOC are set to prevent damage to energy storage devices due to overcharging or over-discharging, thereby extending their lifespan. By rationally controlling the charging and discharging power, the charging and discharging process of the energy storage system can be optimized, improving charging and discharging efficiency and preventing accelerated aging of energy storage devices due to frequent high-power charging and discharging, thus extending their lifespan. The state of charge constraint can be expressed as:

[0113] (1) State of charge range: 20% ≤ SOC(t) ≤ 95%;

[0114] (2) Charging and discharging power limitations:

[0115] (3) SOC dynamic equation:

[0116]

[0117] In the formula, SOC(t) represents the state of charge at time t. E represents the maximum charge / discharge power. ess Let η be the energy storage capacity, and η be the charge / discharge efficiency.

[0118] Grid interaction constraints refer to the limitations on the power exchange between distributed photovoltaic (PV) power generation systems and the power grid. These limitations are necessary to meet grid dispatch requirements and electricity market rules. By restricting the power exchange, the impact of distributed PV systems on the grid or their stable operation can be avoided. This ensures that the operation of distributed PV systems complies with electricity market trading rules and dispatch requirements. Reasonable control of the power exchange can achieve harmonious interaction between distributed PV systems and the grid, improving energy efficiency and the absorption capacity of renewable energy. Grid interaction constraints can be expressed as:

[0119]

[0120] In the formula, This represents the maximum power exchanged between the power grid and the grid.

[0121] like Figure 2 As shown, a specific embodiment of the energy dispatching method for a distributed photovoltaic power generation system provided by this invention is as follows:

[0122] Step 1: Based on multiple metering devices in the distributed photovoltaic power generation system, obtain the photovoltaic power generation P of the distributed photovoltaic power generation system over 30 consecutive natural days. pv Energy storage charging and discharging power Pess The power exchange with the power grid is divided into 24 scheduling periods per day, each of which lasts for 1 hour.

[0123] Step 2: Perform anomaly detection processing on the photovoltaic power generation, the energy storage charging and discharging power, and the grid interaction power to obtain anomaly-detected sampling data; perform missing value filling and duplicate data deletion processing on the anomaly-detected sampling data to obtain preprocessed sampling data; integrate the preprocessed sampling data to obtain 30 input vectors, where the i-th input vector can be represented as: [P pv (i),P ess (i),P grid (i)], where P pv (i) represents the power generation of the i-th photovoltaic power generation, P ess (i) represents the charging and discharging power of the i-th energy storage unit (charging is positive, discharging is negative), P grid (i) represents the power exchanged with the grid at the i-th grid (selling electricity to the grid is positive, buying electricity from the grid is negative);

[0124] The 30 input vectors can be considered as a population, wherein the population can be represented as:

[0125] x i =[P pv (i),P ess (i),P grid (i)]

[0126] Where 0≤i≤30.

[0127] Step 3, construct the energy scheduling model, where the fitness function is:

[0128] F(x)=f(x)+αP e (x)+βP i (x)

[0129] Where f(x) is the original objective function, and:

[0130]

[0131] Equality constraint penalty function P e (x) is,

[0132]

[0133] Inequality constraint penalty function P i (x) is

[0134]

[0135] The velocity v of the i-th input vector (i.e., particle i) in the t-th iteration i (t) and position x i The updated formula for (t) is:

[0136] v i (t+1)=ωv i (t)+c1r1(t)[p i (t)-x i (t)]+c2r2(t)[p g (t)-x i (t)]

[0137] x i (t+1)=x i (t)+v i (t+1), i = 1, 2, ..., 30

[0138] The inertia weight ω is updated using a nonlinear dynamic adaptive update method, and the learning factors c1 and c2 are updated according to an asynchronous linear law, expressed as:

[0139]

[0140]

[0141] Step 4, transfer population x i An input energy scheduling model is used, which iterates through the multiple input vectors multiple times according to a fitness function, with the number of iterations being 200; for each particle, according to [P pv (i),P ess (i),P grid (i)], substitute into the fitness function to calculate the total cost; if the global optimal fitness changes by less than 1 for 20 consecutive generations, terminate early; and output the optimal scheduling scheme, i.e., including [P pv (t),P ess (t),P grid The time series of [(t)] and the fitness function value, i.e., the minimum total cost, are analyzed. Finally, constraint verification is performed to check whether the SOC trajectory and power balance are satisfied.

[0142] Step 5: Convert the optimal scheduling scheme into control commands to control the operation of photovoltaic power generation equipment and energy storage equipment in the distributed photovoltaic power generation system.

[0143] like Figure 3 As shown, embodiments of the present invention also propose an energy dispatching device 30 for a distributed photovoltaic power generation system, comprising:

[0144] The acquisition module 31 is used to acquire the photovoltaic power generation, energy storage charging and discharging power and grid interaction power of the distributed photovoltaic power generation system during multiple scheduling periods;

[0145] The processing module 32 is used to preprocess the photovoltaic power generation, the energy storage charging and discharging power and the grid interaction power to obtain multiple input vectors; and to process the multiple input vectors to obtain energy dispatch data, which is used to control the operating status of the distributed photovoltaic power generation system.

[0146] Optionally, module 31 is specifically used for:

[0147] Based on multiple metering devices in the distributed photovoltaic power generation system, the photovoltaic power generation, energy storage charging and discharging power, and grid interaction power of the distributed photovoltaic power generation system are obtained over multiple natural days. The number of the multiple natural days is a first preset value, the number of scheduling periods included in each natural day is a second preset value, and the duration of each scheduling period is a third preset value.

[0148] Optionally, processing module 32 is specifically used for:

[0149] Anomaly detection processing is performed on the photovoltaic power generation, the energy storage charging and discharging power, and the grid interaction power to obtain sampled data after anomaly detection processing;

[0150] The sampled data after the anomaly detection process is filled with missing values ​​and deleting duplicate data to obtain preprocessed sampled data;

[0151] The preprocessed sampled data is integrated to obtain multiple input vectors.

[0152] Optionally, the processing module 32 is also specifically used for:

[0153] The multiple input vectors are input into an energy scheduling model, which iteratively processes the input vectors multiple times based on a fitness function to obtain energy scheduling data; wherein, the expression for the fitness function is:

[0154] F(x)=f(x)+αP e (x)+βP i (x)

[0155] Where F(x) is the fitness function, f(x) is the original objective function, α and β are penalty factors, and P e (x) is the equality constraint penalty function, P i (x) is the inequality constraint penalty function.

[0156] Optionally, the inertial weights are updated using a nonlinear dynamic adaptive update method, and the learning factors are updated according to an asynchronous linear law.

[0157] Optionally, the expression for the equality constraint penalty function is:

[0158]

[0159] Among them, g i (x) is the equality constraint function, ε is the tolerance, and i and m are natural numbers;

[0160] The expression for the inequality constraint penalty function is:

[0161]

[0162] Among them, h i (x) is the inequality constraint function, and n is a natural number.

[0163] Optionally, the constraints of the energy dispatch model include one or more of the following: power balance constraints, state of charge constraints, charge and discharge power limit constraints, and grid interaction constraints.

[0164] It should be noted that this device is a device corresponding to the above method. All implementation methods in the above method embodiments are applicable to this embodiment and can achieve the same technical effect.

[0165] like Figure 4 As shown, this embodiment of the invention also provides a computing device 40, including a processor 41, a memory 42, and a program or instructions stored in the memory 42 and executable on the processor 41. When the program or instructions are executed by the processor 41, they implement the various processes of the energy dispatching method embodiment of the distributed photovoltaic power generation system described above, and achieve the same technical effect. To avoid repetition, they will not be described again here. It should be noted that the computing device in this embodiment of the invention includes the aforementioned mobile electronic devices and non-mobile electronic devices.

[0166] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0167] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0168] In the embodiments provided by this invention, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0169] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0170] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0171] The data is stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, ROM, RAM, magnetic disks, or optical disks.

[0172] Furthermore, it should be noted that in the apparatus and method of the present invention, it is obvious that the components or steps can be decomposed and / or recombined. These decompositions and / or recombinations should be considered equivalent solutions of the present invention. Moreover, the steps performing the above-described series of processes can naturally be executed in the order described, but are not necessarily required to be executed in chronological order; some steps can be executed in parallel or independently of each other. Those skilled in the art will understand that all or any step or component of the method and apparatus of the present invention can be implemented in any computing device (including processors, storage media, etc.) or network of computing devices, in hardware, firmware, software, or a combination thereof. This is something that those skilled in the art can achieve by using their basic programming skills after reading the description of the present invention.

[0173] Therefore, the object of the present invention can also be achieved by running a program or a set of programs on any computing device. The computing device can be a known general-purpose device. Therefore, the object of the present invention can also be achieved simply by providing a program product containing program code for implementing the method or apparatus. That is, such a program product also constitutes the present invention, and the storage medium storing such a program product also constitutes the present invention. Obviously, the storage medium can be any known storage medium or any storage medium developed in the future. It should also be noted that in the apparatus and method of the present invention, it is obvious that the components or steps can be decomposed and / or recombined. These decompositions and / or recombinations should be considered equivalent to the present invention. Furthermore, the steps for performing the above series of processes can naturally be performed in the order described, but are not necessarily required to be performed in chronological order. Some steps can be performed in parallel or independently of each other.

[0174] The above are preferred embodiments of the present invention. It should be noted that, for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. An energy dispatching method for a distributed photovoltaic power generation system, characterized in that, include: Acquire the photovoltaic power generation, energy storage charging and discharging power, and grid interaction power of the distributed photovoltaic power generation system during multiple scheduling periods; The photovoltaic power generation, the energy storage charging and discharging power, and the grid interaction power are preprocessed to obtain multiple input vectors; The multiple input vectors are processed to obtain energy scheduling data, which is used to control the operating status of the distributed photovoltaic power generation system. The acquisition of photovoltaic power generation, energy storage charging and discharging power, and grid interaction power of the distributed photovoltaic power generation system during multiple scheduling periods includes: Based on multiple metering devices in the distributed photovoltaic power generation system, the photovoltaic power generation, energy storage charging and discharging power and grid interaction power of the distributed photovoltaic power generation system are obtained over multiple natural days. The number of the multiple natural days is a first preset value, the number of scheduling periods included in each natural day is a second preset value, and the duration of each scheduling period is a third preset value. The photovoltaic power generation, energy storage charging and discharging power, and grid interaction power are preprocessed to obtain multiple input vectors, including: Anomaly detection processing is performed on the photovoltaic power generation, the energy storage charging and discharging power, and the grid interaction power to obtain sampled data after anomaly detection processing; The sampled data after the anomaly detection process is filled with missing values ​​and deleting duplicate data to obtain preprocessed sampled data; The preprocessed sampled data is integrated to obtain multiple input vectors; The energy scheduling data is obtained by processing the multiple input vectors, including: The multiple input vectors are input into an energy scheduling model, which iteratively processes the input vectors multiple times based on a fitness function to obtain energy scheduling data; wherein, the expression for the fitness function is: in, F ( x ) is the fitness function. f ( x ) is the original objective function. α and β As a penalty factor, P e ( x ) is the equality constraint penalty function. P i ( x ) is the penalty function for inequality constraints; Wherein, the original objective function The expression is: in, For time period i Electricity price, The energy storage charging and discharging loss cost coefficient. This is the penalty coefficient for discarded light. For time period i The maximum output of photovoltaic power, T The total number of scheduling periods. P grid ( i ) is the first i Power grid interaction power during each time period P ess ( i ) is the first i Energy storage charging and discharging power during each time period, P pv ( i ) is the first i Photovoltaic power generation capacity during each time period.

2. The energy dispatching method for a distributed photovoltaic power generation system according to claim 1, characterized in that, The inertial weights are updated using a nonlinear dynamic adaptive update method, and the learning factors are updated according to an asynchronous linear law.

3. The energy dispatching method for a distributed photovoltaic power generation system according to claim 1, characterized in that, The expression for the equality constraint penalty function is: in, g i ( x ) is the equality constraint function. ε For tolerance, i , m It is a natural number; The expression for the inequality constraint penalty function is: in, h i ( x ) is the inequality constraint function. n It is a natural number.

4. The energy dispatching method for a distributed photovoltaic power generation system according to claim 1, characterized in that, The constraints of the energy dispatch model include one or more of the following: power balance constraints, state of charge constraints, charge and discharge power limit constraints, and grid interaction constraints.

5. An energy dispatching device for a distributed photovoltaic power generation system, characterized in that, include: The acquisition module is used to acquire the photovoltaic power generation, energy storage charging and discharging power, and grid interaction power of the distributed photovoltaic power generation system during multiple scheduling periods; The processing module is used to preprocess the photovoltaic power generation, the energy storage charging and discharging power, and the grid interaction power to obtain multiple input vectors; and to process the multiple input vectors to obtain energy dispatch data, which is used to control the operating status of the distributed photovoltaic power generation system. The acquisition of photovoltaic power generation, energy storage charging and discharging power, and grid interaction power of the distributed photovoltaic power generation system during multiple scheduling periods includes: Based on multiple metering devices in the distributed photovoltaic power generation system, the photovoltaic power generation, energy storage charging and discharging power and grid interaction power of the distributed photovoltaic power generation system are obtained over multiple natural days. The number of the multiple natural days is a first preset value, the number of scheduling periods included in each natural day is a second preset value, and the duration of each scheduling period is a third preset value. The photovoltaic power generation, energy storage charging and discharging power, and grid interaction power are preprocessed to obtain multiple input vectors, including: Anomaly detection processing is performed on the photovoltaic power generation, the energy storage charging and discharging power, and the grid interaction power to obtain sampled data after anomaly detection processing; The sampled data after the anomaly detection process is filled with missing values ​​and deleting duplicate data to obtain preprocessed sampled data; The preprocessed sampled data is integrated to obtain multiple input vectors; The energy scheduling data is obtained by processing the multiple input vectors, including: The multiple input vectors are input into an energy scheduling model, which iteratively processes the input vectors multiple times based on a fitness function to obtain energy scheduling data; wherein, the expression for the fitness function is: in, F ( x ) is the fitness function. f ( x ) is the original objective function. α and β As a penalty factor, P e ( x ) is the equality constraint penalty function. P i ( x ) is the penalty function for inequality constraints; Wherein, the original objective function The expression is: in, For time period i Electricity price, The energy storage charging and discharging loss cost coefficient. This is the penalty coefficient for discarded light. For time period i The maximum output of photovoltaic power, T The total number of scheduling periods. P grid ( i ) is the first i Power grid interaction power during each time period P ess ( i ) is the first i Energy storage charging and discharging power during each time period, P pv ( i ) is the first i Photovoltaic power generation capacity during each time period.

6. A computing device, characterized in that, include: A processor, a memory storing a computer program, wherein the computer program, when executed by the processor, performs the method as described in any one of claims 1 to 4.

7. A computer-readable storage medium, characterized in that, The system stores instructions that, when executed on a computer, cause the computer to perform the method as described in any one of claims 1 to 4.

Images