Two-stage optimal dispatching method for distributed power grid based on probabilistic power flow and WOA-PPSO algorithm

Abstract

The application relates to the technical field of optimal scheduling of distributed power grids, and specifically discloses a two-stage optimal scheduling method for distributed power grids based on a probabilistic power flow and a WOA-PPSO algorithm, which comprises the following steps: constructing a probabilistic power flow model, calculating the probability distribution characteristics of node voltage and line flow in the power grid containing wind power, photovoltaic power and load randomness based on the semi-invariant method and Gram-Charlier series expansion. The application introduces the probabilistic power flow modeling into the optimization constraint of the scheduling algorithm, and responds to the uncertainty that may exist in the power grid in real time; by constructing a two-stage scheduling algorithm, the global optimal scheduling plan is generated based on the WOA algorithm in the day-ahead scheduling, and the deviation caused by the uncertainty is quickly responded on the basis of the optimal scheduling plan in the real-time scheduling of the day; and the result of the day-ahead scheduling is taken as the reference starting value of the real-time scheduling of the day, and the projection method is combined with the PSO algorithm to construct the PPSO algorithm, so that the solving speed in the real-time scheduling of the day is maximally improved.

Description

Technical Field

[0001] This invention relates to the field of distributed power grid optimization scheduling technology, and in particular to a two-level optimization scheduling method for distributed power grids based on probabilistic power flow and the WOA-PPSO algorithm. Background Technology

[0002] With the increasing penetration rate of renewable energy, the power system faces increasingly severe challenges in maintaining real-time supply and demand balance. While the electricity spot market, as a core mechanism for resource allocation, can guide power generation and consumption behavior through price signals, its operational effectiveness is highly dependent on the ability of dispatch strategies to cope with uncertainty. Traditional single-time-scale dispatch methods are ill-suited to the random fluctuations of distributed energy sources such as wind and solar power. Therefore, constructing a multi-time-scale collaborative optimized dispatch system has become crucial for ensuring the safe, stable, and economical operation of the electricity spot market.

[0003] To address this issue, the industry generally believes that constructing a two-tiered dispatch system covering both long and short timescales is a crucial way to ensure the safe and economical operation of the electricity spot market. Day-ahead dispatch typically focuses on optimizing the allocation of global resources, aiming to formulate a relatively stable operating plan for the following day; while intraday real-time dispatch emphasizes rapid response to real-time forecast deviations and local disturbances, dynamically adjusting the day-ahead plan. Although the two-tiered dispatch concept has been widely accepted, existing research largely focuses on how to improve system robustness through two-tiered linkage, lacking a scientific and efficient method to handle the dual uncertainties of renewable energy output and load fluctuations.

[0004] Traditional power flow analysis methods, primarily based on deterministic models, cannot accurately assess the operational risks of a system under uncertain conditions. Probabilistic power flow methods, by incorporating statistical characteristics of input variables such as generation and load, can characterize the probability distribution of system state variables, providing a risk assessment basis for scheduling decisions. However, existing probabilistic power flow methods (such as Monte Carlo simulations) often suffer from high computational costs and slow convergence speeds, while point estimation methods struggle to balance computational efficiency and accuracy. How to deeply integrate efficient probabilistic power flow methods with multi-timescale scheduling models and explicitly introduce probabilistic constraints into the optimization model remains a key technical problem that urgently needs to be solved.

[0005] Furthermore, intelligent optimization algorithms have demonstrated significant advantages in solving highly nonlinear, multi-constraint power system dispatch optimization problems. The Whale Algorithm (WOA), with its strong global search capability and fast convergence speed, is suitable for global optimization in the day-ahead dispatch phase. Particle Swarm Optimization (PSO), due to its simplicity and fast convergence speed, is often used for rapid correction in the intraday dispatch phase. However, existing research mostly employs a single intelligent algorithm to solve optimization problems at a single time scale, with few studies combining the advantages of different intelligent algorithms to collaboratively apply them to multi-time scale optimization problems with probabilistic constraints, taking into account the different characteristics of two-level dispatch systems. Summary of the Invention

[0006] This invention aims to at least partially address one of the technical problems in related technologies. Therefore, the purpose of this invention is to propose a two-stage optimal scheduling method for distributed power grids based on probabilistic power flow and the WOA-PPSO algorithm, to solve the problems of existing scheduling strategies for distributed power grids with wind and solar power generation equipment, which suffer from the inability to simultaneously achieve economic optimization and scheduling response speed, as well as poor resistance to fluctuations in power generation equipment.

[0007] To achieve the above objectives, the first aspect of this invention proposes a two-stage optimal scheduling method for distributed power grids based on probabilistic power flow and the WOA-PPSO algorithm, comprising the following steps:

[0008] S1. Construct a probabilistic power flow model, and calculate the probability distribution characteristics of node voltage and line power flow in a power grid containing wind power, photovoltaic and load randomness based on the semi-invariant method and Gram-Charlier series expansion.

[0009] S2. Execute day-ahead scheduling. Based on the Whale Optimization Algorithm (WOA), with the goal of minimizing the total lifecycle operating cost, and combining load forecasting and renewable energy output forecasting, solve to generate the day-ahead baseline scheduling scheme.

[0010] S3. Perform intraday real-time scheduling. Based on the Projected Particle Swarm Optimization (PPSO) algorithm, with the previous day's baseline scheduling scheme as the reference trajectory, and combined with the probabilistic constraints introduced by the probabilistic power flow model described in step S1, dynamically adjust the operation scheme with the goal of minimizing intraday scheduling costs.

[0011] To achieve the above objectives, a second aspect of the present invention proposes a two-stage optimized scheduling system for distributed power grids based on probabilistic power flow and the WOA-PPSO algorithm, comprising:

[0012] The probabilistic power flow modeling module is used to calculate the probabilistic distribution characteristics of node voltage and line power flow in a power grid containing wind power, photovoltaic power and load randomness based on the semi-invariant method and Gram-Charlier series expansion.

[0013] The day-ahead scheduling module is used to generate a day-ahead baseline scheduling scheme based on the Whale Optimization Algorithm (WOA) with the goal of minimizing the total lifecycle operating cost, combined with load forecasting and renewable energy output forecasting.

[0014] The intraday real-time scheduling module is used to dynamically adjust the operation plan based on the Projected Particle Swarm Optimization (PPSO) algorithm, taking the pre-dated baseline scheduling scheme as a reference trajectory and combining the probabilistic constraints introduced by the probabilistic power flow modeling module, with the goal of minimizing intraday scheduling costs.

[0015] To achieve the above objectives, a third aspect of the present invention provides an electronic device, including a memory, a processor, and a computer program stored in the memory. When the computer program is executed by the processor, it implements the above-described two-level optimal scheduling method for distributed power grids based on probabilistic power flow and the WOA-PPSO algorithm.

[0016] The distributed power grid two-level optimization scheduling method based on probabilistic power flow and WOA-PPSO algorithm in this invention introduces probabilistic power flow modeling into the optimization constraints of the scheduling algorithm to respond in real time to uncertainties that may exist in the power grid. By constructing a two-level scheduling algorithm, after generating a globally optimal scheduling plan based on the WOA algorithm in day-ahead scheduling, the method quickly responds to deviations caused by uncertainties in intraday real-time scheduling based on the optimal scheduling plan. Furthermore, this method uses the results of day-ahead scheduling as the reference starting value for intraday real-time scheduling, and combines the projection method with the PSO algorithm to construct the PPSO algorithm, thereby maximizing the solution speed for intraday real-time scheduling.

[0017] In summary, this method, through probabilistic power flow modeling, day-ahead WOA scheduling, and intraday real-time PPSO scheduling, enables the power grid to better ensure its stability and economy in actual operation. It is applicable to scenarios such as economic dispatch of power grids with uncertain states and has broad application prospects and industrial value. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating the two-level optimal scheduling method for distributed power grids based on probabilistic power flow and the WOA-PPSO algorithm provided by this invention.

[0019] Figure 2 This is a schematic diagram illustrating the implementation of the distributed power grid two-level optimized scheduling system based on probabilistic power flow and WOA-PPSO algorithm provided by the present invention.

[0020] Figure 3 This is a schematic diagram of the electronic device provided by the present invention. Detailed Implementation

[0021] Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention.

[0022] The following description, with reference to the accompanying drawings, describes a two-level optimal scheduling method, system, and electronic equipment for distributed power grids based on probabilistic power flow and the WOA-PPSO algorithm, according to embodiments of the present invention.

[0023] Example 1:

[0024] Figure 1 This is a schematic diagram of the overall control of a two-level optimal scheduling method for distributed power grids based on probabilistic power flow and the WOA-PPSO algorithm according to an embodiment of the present invention. The specific framework of this method includes three parts:

[0025] Probabilistic power flow modeling: To achieve quantitative analysis of uncertainties in power systems, this application constructs a probabilistic power flow calculation process based on the semi-invariant method and Gram-Charlier series expansion. This process enables the acquisition of the probabilistic distribution characteristics of node voltage and branch power with relatively low computational complexity, providing a statistical basis for dispatch optimization.

[0026] The day-ahead dispatch plan based on WOA (Load Forecasting and Renewable Energy Output Forecasting) is based on load forecasting and renewable energy output forecasting. It establishes an optimization model that balances economic efficiency and security to rationally allocate the output plans of controllable power sources, energy storage, and grid-purchased electricity. The day-ahead dispatch uses a 1-hour resolution, constructing typical daily curves based on historical load data, and establishing an optimization model by combining wind and solar output forecasts, grid electricity prices, fuel costs, and energy storage parameters. For day-ahead dispatch, accuracy and globality are far more important than solution speed. WOA's advantages in convergence and global search capabilities effectively ensure the robustness of dispatch results under high uncertainty and complex constraints, thus providing a reliable planning baseline for intraday optimization.

[0027] Intraday real-time dispatch based on PPS0: Real-time dispatch fully considers the uncertainties of dispatchable units, combining analysis with the actual system operating status, ultra-short-term expectations, and line probability distribution. Within the operating day, short-term forecasts are performed at a 15-minute resolution. The day-ahead baseline plan serves as a reference trajectory in each rolling window; if the actual load or renewable output deviates from the forecast, an intraday correction model is constructed to reallocate generation, energy storage charging / discharging, and grid power purchase / sale. Information based on probabilistic power flow pre-calculation can be quickly estimated in short-term optimization, avoiding full recalculation in each cycle and improving real-time performance.

[0028] In addition, to ensure that the particle search process remains within the feasible region, this application introduces a projection method to construct the PPSO algorithm based on PSO to actively correct the equality constraints. Furthermore, to achieve two-level optimal scheduling of the power grid, this invention constructs a probabilistic power flow model using a semi-invariant method to reflect potential wind and solar power fluctuations during power grid operation. The specific steps are as follows:

[0029] Step S11: In probabilistic power flow analysis, the uncertainty of node injected power is the main factor affecting the system state probability distribution. Considering that the power injected by each node is independent, a probabilistic power model for the entire network, including PQ nodes, PV nodes, and slack nodes, is constructed.

[0030] By constructing a node probability model based on the node voltage method and ignoring its second-order and higher terms, the corrected equation is obtained as follows:

[0031]

[0032] in: , , This represents the total number of system nodes. These are Jacobian block matrices representing the relationship between node injected power and node phase angle and voltage magnitude, respectively.

[0033] The above block matrices are respectively The specific formula for calculating the diagonal elements is as follows:

[0034] ;

[0035] ;

[0036] ;

[0037] ;

[0038] and For nodes Injected active and reactive power, , and , They are nodes , Voltage amplitude and phase angle;

[0039] , where is the phase difference between the voltages at the two nodes. and These are the elements of the node admittance matrix. The real and imaginary parts; and They are nodes Self-conductivity and self-susceptivity.

[0040] Step S12: Ignore the equivalent capacitance to ground, and assume the transformer turns ratio is... Based on the power flow calculation formula, the corrected equation is obtained as follows:

[0041]

[0042] in: , This represents the total number of system branches (lines); These are the partial derivative matrices of the branch transmission power with respect to the node phase angle and voltage magnitude, respectively.

[0043] In the above partial derivative block matrix, the branches Power to node The specific matrix element expression for taking the partial derivative of the state variables is as follows:

[0044] ;

[0045] ;

[0046] ;

[0047] ;

[0048] ;

[0049] in, , and They represent respectively by Node flow Active power, reactive power, and complex power; , , Then they respectively represent by Node flow Active power, reactive power, and complex power of a node.

[0050] Step S13: Based on the semi-invariant method and Gram-Charlier series expansion, the random variables are standardized: under the assumption of independent load / generation power, the semi-invariant of the node injected power is the sum of the semi-invariants of the same order of its components; it is transferred to the state variables through power flow linearization mapping, and after the semi-invariant-central moment transformation, the probability density function is expanded with a finite order to achieve a high-precision approximation of the true distribution.

[0051] From then on, the day-ahead dispatch strategy based on WOA is based on load forecasting and renewable energy output forecasting, and rationally allocates the day-ahead output plans of controllable power sources, energy storage, and main grid power purchase and sale.

[0052] Step S21: Divide the day into Over several time periods, assuming constant power for each unit, the objective is to minimize the total lifecycle operating cost. The objective function is designed as follows:

[0053]

[0054] in, , , , , Time periods Operating costs for wind turbines, solar power, thermal power generation, electricity purchases from the main grid, and energy storage systems.

[0055] Optionally, operating costs specifically include:

[0056] Wind power cost Functions and photovoltaic costs The function is as follows:

[0057]

[0058]

[0059] in, , These are the numbers of wind power and solar power plants, respectively. , The generation costs are for wind power and solar power, respectively. and The first The unit power generation of a wind and solar power station within a certain time period; For power generation duration;

[0060] The cost of thermal power generation only includes fuel costs and operation and maintenance costs; construction costs are not included in the calculation. The cost function for thermal power generation is:

[0061]

[0062] in, The number of thermal power plants; For the first The power generation cost of a thermal power plant; For the first time period of this thermal power plant The power generation capacity of a thermal power plant;

[0063] The main grid electricity purchase cost function is:

[0064]

[0065] in, , These represent the unit power supplied to the main grid and the corresponding electricity purchase price within a certain time period;

[0066] The operating cost of an energy storage system includes installation, ongoing maintenance, and charging / discharging efficiency losses. Energy storage systems effectively balance electricity supply and demand; their cost function is as follows:

[0067]

[0068] in, The number of energy storage systems; For the first The operating cost per unit of electricity for an energy storage system; For the first The power of a single energy storage system during a given time period; The installation and maintenance costs of the energy storage system are calculated per hour of total usage time. This represents the service life of the energy storage system.

[0069] Step S22: Building upon S21, to achieve the optimal scheduling plan, this application needs to construct inequalities as optimization constraints for the objective function based on actual application requirements. Specifically, this includes:

[0070] 1. Within a certain period, the total power generation of all generating units must at least meet the total load demand:

[0071]

[0072] Among them, the definition The total load of the power grid during a certain period. The overall power grid loss rate;

[0073] The optimal power balance is then obtained as follows:

[0074]

[0075] 2. Scheduling Unit Constraints

[0076] in, , , , , , , , These are the upper and lower limits for wind power, photovoltaic power, thermal power, and power transmission from the main grid.

[0077] 3. Node voltage and branch power constraints

[0078] in, These are the upper and lower limits of the node voltage. This represents the upper limit of the transmission power of the branch.

[0079] 4. Constraints of Energy Storage Systems

[0080] in, , , The first The maximum charging and discharging power of an energy storage system; for The moment, that is, the 1st moment on the end of the time axis The state of charge of an energy storage system; For the previous moment ( (Time) The state of charge of an energy storage system.

[0081] Step S23: Based on the solution objective and solution constraints constructed in S21 and S22, the WOA algorithm is used to solve for the final optimal scheduling plan.

[0082] The current dispatching system uses a 1-hour resolution and constructs typical daily curves based on historical load data. An optimization model is then established by combining wind and solar power output forecasts, grid electricity prices, fuel costs, and energy storage parameters. The model considers the coupling relationship between energy storage and power sources, allowing surplus wind and solar power to be charged first. When the load exceeds the limit, it is supplemented by energy storage discharge or grid power purchase. Simultaneously, cost weights change over time to capture adjustment opportunities of "charging at low prices and using at high prices."

[0083] Due to the nonlinear, strongly constrained, and high-dimensional characteristics of the model, conventional methods are prone to getting trapped in local optima. To improve the globality and accuracy of the solution, this invention uses WOA (World of Origin and Environment) for solving the problem, leveraging its biomimetic mechanism to efficiently optimize in a complex search space and output a baseline scheduling scheme for 24 hours a day.

[0084] For day-ahead scheduling, accuracy and global applicability are far more important than solution speed. The advantages of WOA in convergence and global search capabilities ensure the robustness of scheduling results in the face of high uncertainty and complex constraints, thus providing a reliable planning baseline for intraday optimization. By ensuring that day-ahead optimization results have better feasibility and global applicability, the system can respond flexibly based on this during the intraday phase, avoiding the computational burden and execution risks caused by frequent daily plan refactoring.

[0085] This implementation step is based on the scheduling plan obtained in step S23 above, and takes into account the potential fluctuations in the actual operation of the power grid based on the probabilistic power flow algorithm designed in step S13. Furthermore, the PPS0 algorithm is constructed by combining the projection method with the particle swarm optimization algorithm to improve the response speed and anti-fluctuation capability of intraday real-time scheduling. The specific implementation steps are as follows:

[0086] Step S31: Real-time scheduling fully considers the uncertainties of schedulable units, combining the actual system operating status, ultra-short-term expectations, and line probability distribution for analysis. By adjusting the interruptibility of loads and the operating plans of each scheduling unit, scheduling costs are optimized while meeting safety constraints. The objective function for minimizing scheduling costs is:

[0087]

[0088] in, This refers to the compensation cost for load interruption during the specified period. The penalty charged for the portion of the actual electricity purchased or sold from the main grid that deviates from the electricity volume determined by the day-ahead dispatch when the deviation exceeds the deviation threshold. , , , , These are the system adjustment costs for energy storage, wind power, photovoltaic power, thermal power generation, and main grid power purchase, respectively. , , , , This refers to the power change of the corresponding dispatching unit in the day-ahead dispatching system for energy storage, wind power, photovoltaic power, thermal power generation, main grid power purchase, and power generation within a certain period.

[0089] Step 32: Compared to the cost function of day-ahead scheduling, the cost of intraday real-time scheduling additionally includes:

[0090] 1. Interrupted load compensation fee: Interrupted load refers to the partial load that is disconnected to ensure grid safety when the grid is overloaded or experiences a fault. According to the contract, the compensation fee is regulated using a quadratic function model and is dynamically priced based on the amount of interrupted load.

[0091] The compensation cost is:

[0092]

[0093] in, , The compensation coefficient; The unit power of the interrupted load;

[0094] If the actual electricity purchased by the power system from the main grid deviates from the electricity determined by the day-ahead dispatching mechanism beyond the deviation threshold stipulated in the contract, a penalty will be charged on the excess portion. The penalty is as follows:

[0095]

[0096] in, Main grid power deviation threshold To account for the discrepancy between the actual and planned electricity purchased from the power grid, This refers to the cost of compensating the main grid for each unit of electricity exceeding the threshold.

[0097] Step S32: In addition to the constraints involved in the day-ahead optimization in step S22, since the intraday real-time scheduling needs to take into account the grid fluctuations constrained by the probabilistic power flow in step S1, the following constraints also need to be included:

[0098] 1. Power Flow Calculation Probability Distribution Constraints: To ensure safe line operation, real-time dispatching needs to reserve sufficient safety margins. The probability distribution of node voltage and branch power caused by uncertainties in probabilistic power flow calculations is used to ensure that node voltage and branch power fluctuations do not exceed three standard deviations during dispatching, thus satisfying the constraints.

[0099]

[0100] in, , They are respectively The minimum and maximum voltages of the nodes; , They are Minimum and maximum power of the branch; For nodes Standard deviation of the probability distribution of voltage; branch road Standard deviation of the probability distribution of transmitted power.

[0101] 2. Interruptible load constraints: For the first At the [time]th moment Maximum interruptible load for a node: .

[0102] Step S33: During the daytime operation, short-term forecasts are performed at a 15-minute resolution. The day-ahead baseline plan from Step S2 serves as a reference trajectory in each rolling window; if the actual load or renewable output deviates from the forecast, an intraday correction model is constructed to reallocate generation, energy storage charging / discharging, and grid power purchase / sale. Information based on probabilistic power flow pre-calculation can be quickly estimated in short-term optimization, avoiding full recalculation in each cycle and improving real-time performance.

[0103] For the daily load dispatching problem of combined wind, solar, thermal, and energy storage power generation systems, the optimization decision variable in this application is the power output of various generating units within 24 hours. The model needs to satisfy two strong constraints: power balance constraint and dynamic energy storage constraint. Directly using the standard PSO algorithm would result in a large number of particles in the solution space not satisfying the constraints, making the solution infeasible. To ensure that the particles always remain within the feasible region during the particle search process, this application introduces a projection method to actively correct the equality constraints. Compared with the penalty function method, the projection method can directly project particles that do not satisfy the constraints back into the feasible solution space, increasing the proportion of feasible solutions and enhancing the convergence stability of the algorithm.

[0104] The projection steps in PSO are as follows:

[0105] 1) Construct the power balance residual function

[0106] set up Time of the first The position vectors of the particles are ,in , for Time of the first Processing of each scheduling unit This represents the total number of units participating in the dispatch (including generators, energy storage, wind and solar power, and main grid power purchase interfaces). Define the equation residual function for power balance constraints. for:

[0107]

[0108] in for The power required at any given time The overall power grid loss rate, if This indicates that the current particle does not meet the power balance constraint and needs to be projected and corrected.

[0109] 2) Gradient-based projection correction formula

[0110] To ensure the particles return to the feasible region satisfying power balance with minimal adjustment cost, the gradient projection method is used to correct the particle positions along the normal vector direction of the constraint hyperplane. Since the power balance constraint is a linear equation concerning the output of each unit, its gradient vector is an all-1 vector, i.e. The corrected formula is as follows:

[0111]

[0112] in The particle position vector in the state to be corrected when the constraints are not met. This is the particle position vector after projection correction.

[0113] This embodiment introduces probabilistic power flow modeling into the optimization constraints of the scheduling algorithm to respond in real time to uncertainties that may exist in the power grid. By constructing a two-level scheduling algorithm, a globally optimal scheduling plan is generated based on the WOA algorithm in day-ahead scheduling. In intraday real-time scheduling, the deviation caused by uncertainties is quickly responded to based on the optimal scheduling plan. Furthermore, this method uses the results of day-ahead scheduling as the reference starting value for intraday real-time scheduling. At the same time, the projection method and the PSO algorithm are combined to construct the PPSO algorithm, which maximizes the solution speed in intraday real-time scheduling.

[0114] Example 2:

[0115] like Figure 2 As shown, corresponding to the above method embodiments, this invention also proposes a two-level optimized scheduling system for distributed power grids based on probabilistic power flow and the WOA-PPSO algorithm, comprising:

[0116] The probabilistic power flow modeling module is used to calculate the probabilistic distribution characteristics of node voltage and line power flow in a power grid containing wind power, photovoltaic power and load randomness based on the semi-invariant method and Gram-Charlier series expansion.

[0117] The day-ahead scheduling module is used to generate a day-ahead baseline scheduling scheme based on the Whale Optimization Algorithm (WOA) with the goal of minimizing the total lifecycle operating cost, combined with load forecasting and renewable energy output forecasting.

[0118] The intraday real-time scheduling module is used to dynamically adjust the operation plan based on the Projected Particle Swarm Optimization (PPSO) algorithm, taking the pre-dated baseline scheduling scheme as a reference trajectory and combining the probabilistic constraints introduced by the probabilistic power flow modeling module, with the goal of minimizing intraday scheduling costs.

[0119] The system's workflow is a typical two-stage optimization process: offline modeling, day-ahead planning, and intraday rolling execution. Its core lies in using probabilistic power flow models to quantify uncertainty and guide optimization decisions at different time scales. The specific implementation process includes:

[0120] 1. Initialization and probabilistic power flow modeling:

[0121] After the system starts, the probabilistic power flow modeling module first performs initialization calculations. This module takes into account the power grid topology, line parameters, and the random distribution characteristics of wind power, photovoltaic power, and loads, such as their expected values ​​and variances.

[0122] The module internally uses the semi-invariant method and Gram-Charlier series expansion to calculate and output the probability distribution characteristics of node voltage and line power flow of the entire power grid, such as mean and standard deviation. This model is the theoretical basis for the entire dispatch system to cope with uncertainty.

[0123] 2. Daily dispatch plan:

[0124] The dispatch module has recently started working. It receives grid security operation domain information from the probabilistic power flow modeling module as the basis for constraints, and combines it with load forecast and renewable energy output forecast data for the next 24 hours.

[0125] This module aims to minimize the overall operating cost and employs the Whale Algorithm (WOA), a powerful global search algorithm, to find an economically optimal day-ahead baseline scheduling scheme that covers 24 hours a day. This scheme determines the planned output values ​​of each controllable unit, including thermal power, energy storage, and main grid power purchase plans, for each time period, forming a globally optimal reference trajectory.

[0126] 3. Real-time intraday dispatch:

[0127] On actual operating days, the intraday real-time dispatch module begins rolling execution at a higher time resolution. At the beginning of each dispatch cycle, this module receives real-time load and renewable energy output data, compares it with the day-ahead forecast values, and calculates the power deviation. It uses the planned values ​​at the corresponding time in the day-ahead baseline dispatch scheme as a reference starting point and optimizes with the goal of minimizing intraday dispatch costs.

[0128] During the optimization process, it is strictly constrained by the probability distribution provided by the probabilistic power flow modeling module, such as the limitation that voltage and power flow fluctuations do not exceed 3 times the standard deviation, to ensure that any adjustment will not cause power grid safety risks.

[0129] This module employs the fast-solving Projective Particle Swarm Optimization (PPSO) algorithm to quickly calculate and execute dynamic adjustment commands for the output of each controllable unit. This monitoring-calculation-adjustment process cycles every 15 minutes until the end of the day, thus enabling a rapid and safe response to uncertainties.

[0130] This system, through probabilistic power flow modeling, day-ahead WOA scheduling, and intraday real-time PPSO scheduling methods, enables the power grid to better ensure its stability and economy in actual operation. It is suitable for scenarios such as economic dispatch of power grids with uncertain states and has broad application prospects and industrial value.

[0131] Example 3:

[0132] Corresponding to the above embodiments, the present invention also proposes an electronic device.

[0133] like Figure 3The diagram shows a structural schematic of an electronic device according to the present invention. The electronic device 100 includes a processor 101 and a memory 103. The processor 101 and the memory 103 are connected, for example, via a bus 102. Optionally, the electronic device 100 may further include a transceiver 104. It should be noted that in practical applications, the transceiver 104 is not limited to one unit, and the structure of this electronic device 100 does not constitute a limitation on the embodiments of the present invention.

[0134] Processor 101 may be a CPU, a general-purpose processor, a DSP, an ASIC, an FPGA, or other programmable logic device, transistor logic device, hardware component, or any combination thereof. It may implement or execute the various exemplary logic blocks, modules, and circuits described in connection with this disclosure. Processor 101 may also be a combination that implements computational functions, such as including one or more microprocessor combinations, a combination of a DSP and a microprocessor, etc.

[0135] Bus 102 may include a pathway for transmitting information between the aforementioned components. Bus 102 may be a PCI bus or an EISA bus, etc. Bus 102 may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 3 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0136] The memory 103 stores a computer program corresponding to the two-level optimal scheduling method for distributed power grids based on probabilistic power flow and the WOA-PPSO algorithm in the above embodiments of the present invention. This computer program is executed by the processor 101. The processor 101 executes the computer program stored in the memory 103 to implement the content shown in the aforementioned method embodiments.

[0137] Among them, electronic devices 100 include, but are not limited to: mobile terminals such as laptops and PADs (tablet computers) and fixed terminals such as desktop computers. Figure 3 The electronic device 100 shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments of the present invention.

[0138] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. A two-stage optimal scheduling method for distributed power grids based on probabilistic power flow and the WOA-PPSO algorithm, characterized in that, Includes the following steps: S1. Construct a probabilistic power flow model, and calculate the probability distribution characteristics of node voltage and line power flow in a power grid containing wind power, photovoltaic and load randomness based on the semi-invariant method and Gram-Charlier series expansion. S2. Execute day-ahead scheduling. Based on the Whale Optimization Algorithm (WOA), with the goal of minimizing the total lifecycle operating cost, and combining load forecasting and renewable energy output forecasting, solve to generate the day-ahead baseline scheduling scheme. S3. Perform intraday real-time scheduling. Based on the Projected Particle Swarm Optimization (PPSO) algorithm, with the previous day's baseline scheduling scheme as the reference trajectory, and combined with the probabilistic constraints introduced by the probabilistic power flow model described in step S1, dynamically adjust the operation scheme with the goal of minimizing intraday scheduling costs.

2. The method according to claim 1, characterized in that, The step S1 of constructing the probabilistic power flow model specifically includes: Based on the semi-invariant method, under the assumption of independent load or power generation, the semi-invariant mapping of node injected power is transferred to the state variables. After the semi-invariant-central moment transformation, the probability density function of node voltage and branch power is approximated by Gram-Charlier series expansion.

3. The method according to claim 1, characterized in that, The objective function for minimizing the full-cycle operating cost of day-ahead scheduling in step S2 is... for: ； in, Total number of time periods; The current time period; , , , , They are respectively Operating costs of wind turbines, photovoltaic power generation, thermal power generation, main grid power purchase, and energy storage systems within the specified time period.

4. The method according to claim 3, characterized in that, The operating costs specifically include: Cost of wind and light: ； ； in, , These are the numbers of wind power and solar power plants, respectively. , The generation costs are for wind power and solar power, respectively. and The first The unit power generation of a wind and solar power station within a certain time period; For power generation duration; Cost of thermal power generation: ； in, The number of thermal power plants; For the first The power generation cost of a thermal power plant; For the first time period of this thermal power plant The power generation capacity of a thermal power plant; Main grid electricity purchase cost: ； in, , These represent the unit power supplied to the main grid and the corresponding electricity purchase price within a certain time period; Energy storage system cost: ； in, The number of energy storage systems; For the first The operating cost per unit of electricity for an energy storage system; For the first The power of a single energy storage system during a given time period; The installation and maintenance costs of the energy storage system are calculated per hour of total usage time. This represents the service life of the energy storage system.

5. The method according to claim 1 or 3, characterized in that, In step S2, the execution of day-ahead scheduling must also meet certain constraints, including: Power balance constraints: ； in, The total load of the power grid during a certain period; The overall power grid loss rate; Node voltage and branch power constraints: ； in, Node voltage; For branch transmission power; , These are the upper and lower limits of the node voltage, respectively; This is the upper limit of the branch transmission power; Constraints of energy storage systems: ； in, , respectively the first The maximum charging and discharging power of an energy storage system; for The moment, that is, the 1st moment on the end of the time axis The state of charge of an energy storage system; For the previous moment ( (Time) The state of charge of an energy storage system.

6. The method according to claim 1, characterized in that, The objective function for minimizing the scheduling cost of intraday real-time scheduling in step S3 is... for: ； in, This refers to the compensation cost for load interruption during the specified period. The penalty charged for the portion of the actual electricity purchased or sold from the main grid that deviates from the electricity volume determined by the day-ahead dispatch when the deviation exceeds the deviation threshold. , , , , These are the system adjustment costs for energy storage, wind power, photovoltaic power, thermal power generation, and main grid power purchase, respectively. , , , , This refers to the power change of the corresponding dispatching unit in the day-ahead dispatching system for energy storage, wind power, photovoltaic power, thermal power generation, main grid power purchase, and power generation within a certain period.

7. The method according to claim 6, characterized in that, The compensation cost and default cost in the scheduling cost are specifically defined as follows: Interruption compensation cost: ； in, , The compensation coefficient; The unit power of the interrupted load; Cost of default: ； in, Main grid power deviation threshold To account for the discrepancy between the actual and planned electricity purchased from the power grid, This refers to the cost of compensating the main grid for each unit of electricity exceeding the threshold.

8. The method according to claim 1 or 6, characterized in that, The intraday real-time scheduling in step S3 also needs to meet the probabilistic constraints introduced by the probabilistic power flow model constructed in step S1, ensuring that the fluctuations in node voltage and branch power do not exceed three standard deviations. The constraint conditions are expressed as follows: ； in, , They are respectively The minimum and maximum voltages of the nodes; , They are Minimum and maximum power of the branch; For nodes Standard deviation of the probability distribution of voltage; branch road Standard deviation of the probability distribution of transmitted power.

9. The method according to claim 1, characterized in that, The PPSO algorithm is a particle swarm optimization algorithm that incorporates a projection method. It is used to correct particles that do not satisfy the power balance equation constraint. Its projection steps include: Constructing the projection plane: for the original particle Construct power balance residuals; Projection Correction: If the power balance residual is not zero, the proportional projection correction formula is used to project the particles that do not meet the constraints back into the feasible solution space, thus obtaining the corrected particles. ; Projected particles It is then incorporated into the original PSO algorithm for further solution.

10. The method according to claim 1, characterized in that, In the two-level optimized scheduling: The day-ahead scheduling uses a 1-hour resolution and is solved based on the WOA algorithm to output a baseline scheduling scheme for the entire 24 hours. The intraday real-time scheduling uses a 15-minute resolution, is solved based on the PPSO algorithm, and uses the baseline scheduling scheme as a reference trajectory, making rolling corrections based on real-time prediction information.

Images