Source network load storage coordinated regulation and control method and device

Abstract

The invention provides a source network load storage coordinated regulation and control method and device, and the method comprises the steps: inputting a predicted value of first power data in a current time window into a source network load storage regulation and control model, and obtaining an original power regulation and control scheme, outputted by the source network load storage regulation and control model, of a target power system in the current time window; and when the current moment is the starting moment of the current time window, starting to execute the original power regulation and control scheme, regulating and controlling the active power and reactive power of the target power system, and when the corrected power regulation and control scheme output by the source network load storage regulation and control model is obtained, starting to execute the original power regulation and control scheme. And replacing the power regulation and control scheme which is being executed with the corrected power regulation and control scheme. According to the source-grid-load-storage coordinated regulation and control method and device provided by the invention, more accurate source-grid-load-storage coordinated scheduling can be carried out on a power system comprising a large number of distributed power supplies, a high-flexibility energy storage device and a flexible load.

Description

Technical Field

[0001] The present invention relates to the field of electric power technology, and in particular to a source-grid-load-storage coordinated control method and device. Background Art

[0002] With the decreasing total reserves of non-renewable energy such as oil, coal and natural gas, as well as a series of environmental problems caused by the extensive use of fossil energy, renewable clean energy has gradually become a key direction for the transformation of the global energy system.

[0003] With the large-scale access of highly random distributed power sources and highly flexible energy storage devices to the distribution network, and the increase of flexible loads on the load side, the control objectives and constraints for regulating the new power system composed of sources, grids, loads and storage have become more diverse. The control objectives include not only the overall comprehensive economy of the above-mentioned new power system, but also the low-carbon economy and operational safety of the above-mentioned new power system. The constraints include not only the grid flow constraints and basic safety constraints in the above-mentioned new power system, but also various constraints that need to be met for the coordinated operation of sources, grids, loads and storage in the above-mentioned new power system.

[0004] However, based on the traditional source-grid-load-storage control method in the relevant technology, it is difficult to accurately coordinate the source-grid-load-storage for a power system including a large number of distributed power sources, highly flexible energy storage devices, and flexible loads. Therefore, how to more accurately coordinate the source-grid-load-storage for a power system including a large number of distributed power sources, highly flexible energy storage devices, and flexible loads is a technical problem that needs to be solved urgently in this field. Summary of the invention

[0005] The present invention provides a source-grid-load-storage coordinated control method and device, which are used to solve the defect in the prior art that it is difficult to accurately perform source-grid-load-storage coordinated scheduling for an electric power system including a large number of distributed power sources, highly flexible energy storage devices and flexible loads, and to achieve more accurate source-grid-load-storage coordinated scheduling for an electric power system including a large number of distributed power sources, highly flexible energy storage devices and flexible loads.

[0006] The present invention provides a source-grid-load-storage coordinated control method, comprising:

[0007] Inputting the predicted value of the first power data in the current time window into the source-grid-load-storage regulation model, and obtaining the original power regulation scheme of the target power system in the current time window output by the source-grid-load-storage regulation model;

[0008] When the current moment is the starting moment of the current time window, the original power control scheme is started to be executed to control the active power and reactive power of the target power system, and when the revised power control scheme output by the source-grid-load-storage control model is obtained, the power control scheme being executed is replaced with the revised power control scheme;

[0009] Wherein, the source-grid-load-storage control model is used to correct the power control scheme of active power in the power control scheme being executed based on the actual value of the second power data at the current moment and the predicted value of the first power data in the current time window when the current moment is the first node moment in the current time window; the source-grid-load-storage control model is also used to correct the power control scheme of reactive power in the power control scheme being executed based on the actual value of the second power data at the current moment and the actual value of the electrical quantity data of the target power system when the current moment is the second node moment in the current time window and it is determined that there is an operational safety risk in the target power system at the current moment;

[0010] The first power data includes the power generation power of the distributed power source nodes and the load power of each load node in the target power system; the second power data includes the power data of each node in the target power system; the first time interval between any two adjacent first node moments in the current time window is the first time interval; the second time interval between any two adjacent second node moments in the current time window is the second time interval; the first time interval is greater than the second time interval; the active power in the target power system is a continuous variable, and the reactive power in the target power system is a discrete variable.

[0011] According to a source-grid-load-storage coordinated control method provided by the present invention, a source-grid-load-storage control model includes: a day-ahead scheduling layer, an intra-day rolling layer, and a real-time optimization layer;

[0012] The day-ahead scheduling layer is used to obtain and output the power control scheme based on the predicted value of the first power data in the current time window, according to the target constraint and the first control target, using the improved particle swarm algorithm;

[0013] The intra-day rolling layer is used to correct the power control scheme of active power in the power control scheme being executed based on the actual value of the second power data at the current moment and the predicted value of the first power data in the current time window, according to the target constraint and the first control target, when the current moment is the first node moment in the current time window, and then output the corrected power control scheme;

[0014] The real-time optimization layer is used to correct the power control scheme of reactive power in the power control scheme being executed based on the actual value of the second power data at the current moment and the actual value of the electrical quantity data of the target power system, according to the target constraint and the second control target, using the improved particle swarm algorithm, when the current moment is the second node moment within the current time window and it is determined that the target power system at the current moment has an operation safety risk, and then output the corrected power control scheme;

[0015] Among them, the first control target includes minimizing the comprehensive economic cost of the target power system; the second control target includes minimizing the operating safety risk of the target power system; the constraints include power balance equation constraints, source-grid-load-storage operation constraints, and reactive equipment control constraints.

[0016] According to a source-grid-load-storage coordinated control method provided by the present invention, in the improved particle swarm algorithm, when the global optimal solution of the particle population is not updated for consecutive target number of iterations, the position vector of each particle in this iteration is determined based on the adversarial extreme value distance between the position vector of each particle in the previous iteration and the position vectors of any two particles in the particle population in the previous iteration; the adversarial extreme value distance is determined based on the cosine similarity coefficient between the position vectors of the any two particles and the extreme value distance of the position vector of each particle in the previous iteration.

[0017] According to a source-grid-load-storage coordinated control method provided by the present invention, in the improved particle swarm algorithm, the position vector of any particle in the particle swarm in this iteration is determined based on the corrected position vector of any particle in the previous iteration; the corrected position vector of any particle in the previous iteration is obtained by correcting the position vector of any particle in the previous iteration based on the equal distance; the equal distance is determined based on the upper and lower limits of the position of any particle.

[0018] According to a source-grid-load-storage coordinated control method provided by the present invention, the inertia weight in the improved particle swarm algorithm is negatively correlated with the number of iterations, and the inertia weight corresponding to any iteration is calculated based on the number of iterations of the any iteration and the predefined maximum number of iterations, the maximum and minimum values ​​of the inertia weight, and the adjustment coefficient.

[0019] According to a source-grid-load-storage coordinated control method provided by the present invention, the real-time optimization layer is specifically used to determine that there is an operation safety risk in the target power system at the current moment when the current moment is the second node moment within the current time window and it is determined that at least one of the voltage over-limit evaluation index, the grid voltage distortion rate evaluation index and the voltage safety margin evaluation index of the target power system at the current moment exceeds the corresponding threshold range.

[0020] According to a source-grid-load-storage coordinated control method provided by the present invention, before inputting the predicted value of the first power data in the current time window into the source-grid-load-storage control model and obtaining the original power control scheme of the target power system in the current time window output by the source-grid-load-storage control model, the method further includes:

[0021] Obtaining an actual value of the first power data in the previous time window of the current time window;

[0022] Inputting the actual value of the first power data in the previous time window into a power data prediction model, and obtaining a predicted value of the first power data in the current time window output by the power data prediction model;

[0023] The power data prediction model is constructed based on a BP neural network and is trained based on actual values ​​of the first power data in multiple continuous historical time windows.

[0024] The present invention also provides a source-grid-load-storage coordinated control device, comprising:

[0025] A scheme prediction module is used to input the predicted value of the first power data in the current time window into the source-grid-load-storage regulation model to obtain the original power regulation scheme of the target power system in the current time window output by the source-grid-load-storage regulation model;

[0026] The dynamic optimization module starts to execute the original power control scheme when the current moment is the starting moment of the current time window, controls the active power and reactive power of the target power system, and replaces the power control scheme being executed with the revised power control scheme when the revised power control scheme output by the source-grid-load-storage control model is obtained;

[0027] Wherein, the source-grid-load-storage control model is used to correct the power control scheme of active power in the power control scheme being executed based on the actual value of the second power data at the current moment and the predicted value of the first power data in the current time window when the current moment is the first node moment in the current time window; the source-grid-load-storage control model is also used to correct the power control scheme of reactive power in the power control scheme being executed based on the actual value of the second power data at the current moment and the actual value of the electrical quantity data of the target power system when the current moment is the second node moment in the current time window and it is determined that there is an operational safety risk in the target power system at the current moment;

[0028] The first power data includes the power generation power of the distributed power source nodes and the load power of each load node in the target power system; the second power data includes the power data of each node in the target power system; the first time interval between any two adjacent first node moments in the current time window is the first time interval; the second time interval between any two adjacent second node moments in the current time window is the second time interval; the first time interval is greater than the second time interval; the active power in the target power system is a continuous variable, and the reactive power in the target power system is a discrete variable.

[0029] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the source-grid-load-storage coordinated control method as described above is implemented.

[0030] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the source-grid-load-storage coordinated control methods described above.

[0031] The present invention also provides a computer program product, including a computer program, which, when executed by a processor, implements any of the above-mentioned source-grid-load-storage coordinated control methods.

[0032] The source-grid-load-storage coordinated control method and device provided by the present invention inputs the predicted value of the first power data in the current time window into the source-grid-load-storage control model, obtains the original power control scheme of the target power system in the current time window output by the source-grid-load-storage control model, and then starts to execute the original power control scheme when the current moment is the starting moment of the current time window to control the active power and reactive power of the target power system. When the revised power control scheme output by the source-grid-load-storage control model is obtained, the power control scheme being executed is replaced with the revised power control scheme. The source-grid-load-storage control model is used to correct the power control scheme being executed based on the actual value of the second power data at the current moment and the predicted value of the first power data in the current time window when the current moment is the first node moment in the current time window. The power control scheme for active power in the executed power control scheme, the source-grid-load-storage control model is also used to correct the power control scheme for reactive power in the power control scheme being executed based on the actual value of the second power data at the current moment and the actual value of the electrical quantity data of the target power system when the current moment is the second node moment within the current time window and it is determined that there is an operational safety risk in the target power system at the current moment. This can more accurately coordinate source-grid-load-storage scheduling for power systems including a large number of distributed power sources, highly flexible energy storage devices and flexible loads, and control the active power and reactive power of the above-mentioned power systems from three time scales. In the control, the decoupling of active power and reactive power is achieved, the amount of calculation is reduced, the calculation efficiency is improved, the operational risk of the above-mentioned power system can be reduced, and the comprehensive cost of the above-mentioned power system can be reduced. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0034] Figure 1 It is a flow chart of the source-grid-load-storage coordinated control method provided by the present invention;

[0035] Figure 2 It is a schematic diagram of a source-grid-load-storage control model in the source-grid-load-storage coordinated control method provided by the present invention;

[0036] Figure 3 It is a topological diagram of IEEE33 nodes;

[0037] Figure 4 It is a schematic diagram comparing the iterative convergence of the improved particle swarm algorithm and the traditional particle swarm algorithm in the source-grid-load-storage method provided by the present invention;

[0038] Figure 5 It is a schematic diagram of a power control scheme for active power in the source-grid-load-storage coordinated control method provided by the present invention;

[0039] Figure 6 It is a schematic diagram of a power control scheme for reactive power in the source-grid-load-storage coordinated control method provided by the present invention;

[0040] Figure 7 It is a structural schematic diagram of the source-grid-load-storage coordinated control device provided by the present invention;

[0041] Figure 8 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0042] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0043] In the description of the invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0044] In the description of the present application, the terms "first", "second", etc. are used to distinguish similar objects, and are not used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable under appropriate circumstances, so that the embodiments of the present application can be implemented in an order other than those illustrated or described here, and the objects distinguished by "first", "second", etc. are usually a class, and the number of objects is not limited. For example, the first object can be one or more. In addition, in the description of the present application, "and / or" represents at least one of the connected objects, and the character " / " generally represents that the front and back associated objects are in an "or" relationship.

[0045] It should be noted that in the field of power technology, source, grid, load and storage is a concept involving the balance of power supply and demand and energy storage. The source in source, grid, load and storage refers to the power generation source in the power system, including thermal power, hydropower, wind power, solar power and other power generation methods. The above power generation sources provide the electric energy required by the power system; the network in source, grid, load and storage refers to the transmission network and distribution network in the power system, which are responsible for transmitting the electric energy generated by the power generation source to the users and ensuring the stability and reliability of the power supply; the load in source, grid, load and storage refers to the load in the power system, that is, the user's electricity demand. The size of the load will change over time. The power system needs to adjust the power generation according to the change of load to maintain the balance of supply and demand; the storage in source, grid, load and storage refers to the energy storage device in the power system, which is used to store excess electric energy in the power system so that it can be released for power supply when needed.

[0046] The installed capacity of distributed power sources such as solar photovoltaic (PV) and wind turbine (WT) is continuing to grow at a high rate, and the proportion of distributed power sources connected to the distribution network is also increasing, which has a significant impact on the power flow, power quality, and dispatching control of the distribution network.

[0047] With the large-scale access of highly random distributed power sources and highly flexible energy storage devices to the distribution network, and the increase of flexible loads on the load side, the objects for regulating the new power system composed of source, grid, load and storage no longer include only traditional reactive power control equipment. The grid side also has added highly random distributed power sources and highly flexible energy storage devices, and the load side also has added flexible and controllable flexible loads.

[0048] The control objectives and constraints for regulating the above-mentioned new power systems are also more diverse. The control objectives include not only the overall comprehensive economy of the above-mentioned new power systems, but also the low-carbon economy and operational safety of the above-mentioned new power systems. The constraints include not only the grid flow constraints and basic safety constraints in the above-mentioned new power systems, but also the various constraints required for the coordinated operation of sources, grids, loads and storage in the above-mentioned new power systems. The time granularity for regulating the above-mentioned new power systems is also finer, and can be determined based on multiple dimensions such as equipment control characteristics, control time requirements, and control target requirements. Therefore, accurate coordinated scheduling of sources, grids, loads and storage, including a large number of distributed power sources, highly flexible energy storage devices and flexible loads, is of great significance for improving the economy and operational safety of the above-mentioned new power systems.

[0049] The coordinated dispatch of sources, grids, loads and storage is a power-coupled dispatch, because the regulation of active power in sources, grids, loads and storage mainly includes coordinating the power generation power and the power load power. The purpose is to improve the absorption rate of distributed power sources and increase the benefits of time-of-use electricity price users, thereby ensuring economic optimization while meeting the energy needs of users.

[0050] The regulation of reactive power in sources, grids, loads and storage not only includes regulating voltage quality, but also changes the power grid trend, affects network losses, and affects the consumption of distributed power sources, which will ultimately affect users' energy demand and benefits.

[0051] Therefore, the active power and reactive power in the source, grid, load and storage are coupled and affect each other. However, the amount of calculation required for comprehensive regulation of the active power and reactive power in the source, grid, load and storage is large. It is possible to consider reasonably decoupling the active power and reactive power in the source, grid, load and storage and then regulating them independently.

[0052] The regulation of reactive equipment (such as transformer taps, capacitor banks, etc.) in the source-grid-load-storage system is a discrete regulation of one gear or one group, and the frequency of daily regulation is relatively low, so the time scale of regulation is relatively long. However, the regulation of photovoltaic or wind power inverters, energy storage devices, and flexible loads in the source-grid-load-storage system is smooth and continuous, so the time scale of regulation is relatively short. Therefore, the regulation characteristics and time scale of power equipment in the source-grid-load-storage system provide a path for the decoupling of active power and reactive power.

[0053] In addition, due to the high randomness of distributed power sources, the prediction of distributed power sources often has errors, and the scheduling plan needs to be revised based on the real-time collected power data. Therefore, it is necessary to propose a multi-time scale control model for source-grid-load-storage that includes a large number of distributed power sources, highly flexible energy storage devices, and flexible loads. The model also needs to consider the control characteristics and computational efficiency of power equipment, decouple and optimize active power and reactive power, and propose an optimization solution method suitable for mixed variables for the solution algorithm.

[0054] The traditional source-grid-load-storage control method uses a coordinated approach of two time scales, day-ahead prediction and intraday control, to dispatch the source-grid-load-storage. The day-ahead goal is to minimize the comprehensive costs such as energy purchase cost, load response cost, network loss, carbon emissions, and operating cost, and obtain the optimal power of wind, light, electricity, gas, and energy storage. During the day, based on the day-ahead dispatch results, ultra-short load forecasts for photovoltaic and wind power are carried out, with the goal of minimizing the power fluctuation of the power system, and feedback is provided to correct the day-ahead regulation plan.

[0055] However, the above traditional source-grid-load-storage control method does not consider the power coupling control nature, and fine-tunes the control according to the characteristics of the equipment. In response to this problem, the relevant technology further establishes an optimization model with three time scales: day-ahead scheduling, intraday rolling and real-time feedback. The wind and solar output and load power data of 24 hours are predicted on the day-ahead, and the active power and reactive power coupling control are considered. On the day-ahead, the wind, solar, storage and gas turbines and reactive power regulation equipment are used as variables, with the goal of maximizing benefits; the power data of wind power and photovoltaic power are corrected every 4 hours during the day, and the day-ahead plan is corrected in a rolling manner; the real-time feedback stage uses 5 minutes as a model to calculate the predicted deviation value of wind and solar output and load power data. If the deviation value is too large, it is fed back to the intraday layer and the intraday optimization calculation is performed again. This method establishes three time scales, and the day-ahead regulation performs the coupling regulation of active and reactive power. However, the above traditional entire model only takes economic benefits as the goal, does not consider carbon emissions, and does not consider the safety of the power grid. The day-ahead time scale is 4 hours, which is relatively long and the control precision is not enough.

[0056] In response to the above problems, the relevant technologies further optimize the 24-hour active power and reactive power schemes with the comprehensive economic benefits as the goal in the day; for the problem of large amount of rolling adjustment calculations in 1h, the method of optimizing the division of time periods within the day is used to replace the optimization results of a longer time with the optimization results of the time scale, simplify the calculation process, and improve the efficiency of the algorithm; at the same time, the real-time stage uses 1min as the scale to monitor the voltage limit of each load node in real time, and adjust the photovoltaic and energy storage power when the voltage exceeds the limit. This method improves the calculation efficiency, and the optimization and regulation takes into account both economic performance and voltage safety risks, but this method does not consider the carbon reduction and emission reduction goals; when the voltage exceeds the limit during the day, the photovoltaic and energy storage power are only used as variables, and the demand response load is not considered, and the adjustment of reactive equipment for special reactive optimization is not considered; only the voltage limit is considered, and other safety risk conditions such as voltage fluctuations and voltage safety margins are not considered. Therefore, there is still room for further optimization and improvement on the above issues.

[0057] Figure 1 Schematic diagram of the flow chart of the source-grid-load-storage coordinated control method provided by the present invention. Figure 1 The source-grid-load-storage coordinated control method of the present invention is described. Figure 1 As shown, the method includes: step 101, inputting the predicted value of the first power data in the current time window into the source-grid-load-storage regulation model, and obtaining the original power regulation plan of the target power system in the current time window output by the source-grid-load-storage regulation model.

[0058] It should be noted that the executor of the embodiment of the present invention is a source-grid-load-storage coordinated control device.

[0059] Optionally, the source-grid-load-storage coordinated control device in the embodiment of the present invention can be deployed in a remote server.

[0060] Specifically, the target power system is the control object of the source-grid-load-storage coordinated control method provided by the present invention, and the target power system includes a large number of highly random distributed power sources, highly flexible energy storage devices, and flexible loads. Based on the source-grid-load-storage coordinated control method provided by the present invention, the target power system including a large number of distributed power sources, highly flexible energy storage devices, and flexible loads can be more accurately dispatched in a source-grid-load-storage coordinated manner.

[0061] It is understandable that the target power system in the embodiment of the present invention can be determined according to actual needs. For example, the power system of a province, a city or a region can be determined as the target power system according to actual needs. The target power system is not specifically limited in the embodiment of the present invention.

[0062] It should be noted that in the power system, a node refers to a connection point or interface in the power system, and is a basic unit for power system analysis and calculation. The target power system in the embodiment of the present invention may include non-flexible load nodes, flexible load nodes, distributed power generation nodes, and energy storage nodes.

[0063] Among them, non-flexible load nodes refer to nodes connected to non-flexible loads; flexible load nodes refer to nodes connected to flexible loads; distributed power supply nodes refer to nodes connected to distributed power supply equipment; and energy storage nodes refer to nodes connected to energy storage devices.

[0064] In the embodiment of the present invention, the power generation power of the distributed power source node in the target power system and the load power of each load node in the target power system may be determined as the first power data.

[0065] It should be noted that distributed energy resources (DER) refers to small-scale, renewable and non-renewable energy generation devices that are dispersed on the user side. Unlike traditional centralized energy systems, distributed power sources can meet energy needs through local and personal energy production and use, which helps to improve energy security, reliability and reduce energy transmission losses. Distributed power sources can include various types of equipment, such as solar photovoltaic power generation equipment (Photovoltaic, PV), wind power generation equipment (Wind Turbine, WT), small gas turbine power generation equipment, fuel cells, micro hydropower generation equipment, etc. Distributed power sources can be installed in end-user buildings, factories, residential areas or communities to meet local energy needs. However, the power generation of distributed power nodes such as solar photovoltaic power generation equipment and wind power generation equipment is not stable, and energy storage devices (Energy Storage System, ESS) need to be configured to buffer the adverse effects caused by the unstable output power of distributed power sources.

[0066] Therefore, the target power system in the embodiment of the present invention includes not only distributed power supply nodes but also energy storage nodes.

[0067] In order to facilitate the understanding of the source-grid-load-storage coordinated control method provided by the present invention, the source-grid-load-storage coordinated control method provided by the present invention is described below by taking the distributed power source in the target power system as a solar photovoltaic power generation device as an example. Accordingly, the distributed power source node in the target power system in the embodiment of the present invention is a photovoltaic node connected to the solar photovoltaic power generation device.

[0068] In an embodiment of the present invention, the length of the time window and the starting time of the time window can be determined based on prior knowledge and / or actual conditions. In an embodiment of the present invention, the length of the time window can be determined as 24 hours, and the starting time of the time window can be zero o'clock.

[0069] It should be noted that before inputting the predicted value of the first power data in the current time window into the source-grid-load-storage control model to obtain the original power control scheme of the target power system in the current time window output by the source-grid-load-storage control model, the first power data in the current time window can be predicted in a variety of ways to obtain the predicted value of the first power data in the current time window. For example, in an embodiment of the present invention, the first power data in the current time window can be predicted based on deep learning technology, mathematical statistics, and numerical calculation to obtain the predicted value of the first power data in the current time window.

[0070] As an optional embodiment, the predicted value of the first power data in the current time window is input into the source-grid-load-storage control model, and before the original power control scheme of the target power system in the current time window output by the source-grid-load-storage control model is obtained, the method also includes: obtaining the actual value of the first power data in the previous time window of the current time window.

[0071] It should be noted that in the embodiments of the present invention, data collection devices such as smart meters and grid-edge smart collection devices can be used to obtain the operating data of the target power system, including the output power of solar photovoltaic power generation equipment in the target power system, the charging and discharging power of energy storage devices, flexible loads, basic loads, transformer taps, capacitor banks and other electrical quantity data and adjustment information. The above operating data can be transmitted to a remote server by means of data communication, and then stored in a local database as historical data.

[0072] In the embodiment of the present invention, the actual value of the first power data in the previous time window of the current time window can be obtained by means of data query.

[0073] Inputting the actual value of the first power data in the previous time window into the power data prediction model, and obtaining the predicted value of the first power data in the current time window output by the power data prediction model;

[0074] The power data prediction model is constructed based on a BP neural network and is trained based on actual values ​​of the first power data in multiple continuous historical time windows.

[0075] Specifically, after obtaining the actual value of the first power data in the previous time window of the current time window, the actual value of the first power data in the previous time window can be input into the power data prediction model to obtain the predicted value of the first power data in the current time window output by the power data prediction model.

[0076] It should be noted that in the embodiment of the present invention, an initial model can be constructed based on a BP neural network, and the initial model can be trained based on actual values ​​of the first power data in multiple continuous historical time windows to obtain the power data prediction model.

[0077] It should be noted that, in the embodiment of the present invention, the actual value of the first power data in a plurality of continuous historical time windows can be obtained by means of data query.

[0078] When the above-mentioned initial model is trained based on the actual value of the first power data in multiple continuous historical time windows, any two adjacent historical time windows in the above-mentioned multiple continuous historical time windows can be taken as a historical time window pair, and then the actual value of the first power data in the historical time window with an earlier time sequence in any historical time window pair can be taken as sample data, and the actual value of the first power data in the historical time window with a later time sequence in the above-mentioned historical time window pair can be taken as sample label to train the above-mentioned initial model to obtain the above-mentioned power data prediction model.

[0079] After obtaining the predicted value of the first power data in the current time window, the predicted value of the first power data in the current time window may be input into the source-grid-load-storage regulation model.

[0080] The source-grid-load-storage control model can obtain the original power control scheme of the target power system in the current time window based on the first power data in the current time window through various methods such as deep learning technology and numerical calculation. In the embodiment of the present invention, the specific method for the source-grid-load-storage control model to obtain the original power control scheme of the target power system in the current time window based on the first power data in the current time window is not limited.

[0081] It should be noted that the active power and reactive power in the target power system can be regulated based on the above original power regulation scheme.

[0082] Accordingly, the control object of the above original power control scheme may include a first control device for smoothly controlling the active power in the target power system and a second control device for non-smoothly controlling the reactive power in the target power system.

[0083] Among them, the first control device may include an inverter for controlling the power generation power of the photovoltaic node and the charging and discharging power of the energy storage node, and an intelligent switch group for controlling the flexible load power of the flexible load node; the second control device may include a control device for adjusting the transformer tap and the capacitor group.

[0084] Step 102: when the current time is the start time of the current time window, start to execute the original power control plan, control the active power and reactive power of the target power system, and when the revised power control plan output by the source-grid-load-storage control model is obtained, replace the power control plan being executed with the revised power control plan;

[0085] Among them, the first power data includes the power generation power of the distributed power source node and the load power of each load node in the target power system; the source-grid-load-storage control model is used to correct the power control scheme of active power in the power control scheme being executed based on the actual value of the second power data at the current time and the predicted value of the first power data in the current time window when the current moment is the first node moment in the current time window. The source-grid-load-storage control model is also used to correct the power control scheme of reactive power in the power control scheme being executed when the current moment is the second node moment in the current time window and it is determined that there is an operation safety risk in the target power system at the current moment, based on the actual value of the second power data at the current time and the actual value of the electrical quantity data of the target power system; the second power data includes the power data of each node in the target power system; the first time interval between any two adjacent first node moments in the current time window is the first time interval; the second time interval between any two adjacent second node moments in the current time window is the second time interval; the first time interval is greater than the second time interval; the active power in the target power system is a continuous variable, and the reactive power in the target power system is a discrete variable.

[0086] It should be noted that, in the embodiment of the present invention, the power data of each node in the target power system can be determined as the second power data. The power data of each node in the target power system may include but is not limited to the power generation power of the photovoltaic node, the power generation power of the generator node, the non-flexible load power of the non-flexible load node, the flexible load power of the flexible load node, the charging and discharging power of the energy storage node, etc.; the generator node refers to the node connected to the generator.

[0087] It should be noted that since there is a deviation between the predicted value of the first power data in the current time window and the actual value of the first power data in the current time window, and the power generation power of the photovoltaic node is random, the source-grid-load-storage control model in the embodiment of the present invention will correct the power control scheme being executed according to the real-time operation data of the target power system in the current time window.

[0088] Moreover, if the source-grid-load-storage control model optimizes the coupling of active power and reactive power in the target power system at the same time, the variable dimension will be increased, which significantly increases the amount of calculation. In addition, the control of active power is usually smooth control with a high control frequency, while the control of reactive power is usually non-smooth regulation with a low control frequency. Therefore, in order to reduce the amount of calculation and in combination with the control characteristics of the control equipment, the source-grid-load-storage control model in the embodiment of the present invention performs decoupling optimization on active power and reactive power.

[0089] Specifically, after obtaining the original power control plan of the target power system within the current time window output by the source-grid-load-storage control model, if the current moment is the starting moment of the current time window, the above-mentioned original power control plan is executed to control the first control device and the second control device in the target power system, thereby starting power control of the active power and reactive power in the target power system.

[0090] When the revised power control scheme output by the source-grid-load-storage control model is obtained for the first time, the power control scheme being executed is exactly the original power control scheme mentioned above. Therefore, the original power control scheme being executed can be replaced by the revised power control scheme mentioned above.

[0091] When the revised power control scheme output by the source-grid-load-storage control model is not obtained for the first time, the power control scheme being executed can be replaced with the revised power control scheme.

[0092] It should be noted that the first duration and the second duration in the embodiment of the present invention can be determined based on prior knowledge and / or actual conditions. The specific values ​​of the first duration and the second duration in the embodiment of the present invention are not limited.

[0093] Optionally, the first duration may be set to 1 hour; the second duration may be set to 5 minutes.

[0094] As an optional embodiment, the source-grid-load-storage regulation model includes: a day-ahead scheduling layer, an intra-day rolling layer, and a real-time optimization layer;

[0095] The day-ahead scheduling layer is used to obtain and output a power control scheme based on the predicted value of the first power data in the current time window, according to the target constraint and the first control target, using the improved particle swarm algorithm;

[0096] The intraday rolling layer is used to correct the power control scheme of active power in the power control scheme being executed based on the actual value of the second power data at the current moment and the predicted value of the first power data in the current time window, according to the target constraint and the first control target, when the current moment is the first node moment in the current time window, and then output the corrected power control scheme;

[0097] The real-time optimization layer is used to correct the power control scheme of reactive power in the power control scheme being executed based on the actual value of the second power data at the current moment and the actual value of the electrical quantity data of the target power system, according to the target constraint and the second control target, using the improved particle swarm algorithm, when the current moment is the second node moment within the current time window and it is determined that the target power system at the current moment has an operation safety risk, and then output the corrected power control scheme;

[0098] Among them, the first control target includes minimizing the comprehensive economic cost of the target power system; the second control target includes minimizing the operating safety risk of the target power system; the constraints include power balance equation constraints, source-grid-load-storage operation constraints, and reactive equipment control constraints.

[0099] Figure 2 Schematic diagram of the source-grid-load-storage control model in the source-grid-load-storage coordinated control method provided by the present invention. Figure 2 As shown, the source-grid-load-storage control model in the embodiment of the present invention includes power optimization models of three time scales: day-ahead scheduling layer, intra-day rolling layer and real-time optimization layer, which can perform power decoupling control on the active power and reactive power in the target power system according to the time scale and power regulation characteristics.

[0100] It should be noted that the day-ahead dispatching layer in the embodiment of the present invention takes the minimum comprehensive economic cost of the target power system as the goal, takes into account the time-of-use electricity price factor, so that the power "peak shaving and valley filling" benefit of the target power system can be maximized within the current time window, takes into account the low-carbon benefit, so that the low-carbon operation of the target power system can be promoted, and takes into account the abandoned light factor, so that the photovoltaic absorption rate of the target power system can be improved. Therefore, in the embodiment of the present invention, the minimum comprehensive economic cost of the target power system is determined as the first control target.

[0101] Specifically, in the embodiment of the present invention, the comprehensive economic cost of the target power system may include energy purchase cost, equivalent carbon emission processing cost, network loss cost, equipment regulation cost, and curtailment penalty cost.

[0102] The objective function corresponding to the day-ahead scheduling layer can be expressed by the following formula:

[0103] F Day =f energy-f carbon +f ops +f punish (1)

[0104] In the formula, F Day represents the comprehensive economic cost of the target power system; f energy represents the energy purchase cost of the target power system; f carbon represents the equivalent carbon reduction benefit of the target power system; f ops represents the equipment control cost of the target power system; f punish Represents the penalty cost of abandoned solar power in the target power system.

[0105] It should be noted that the unit of cost in the embodiments of the present invention is Yuan.

[0106] The target power system's energy purchase cost f energy It can be calculated by the following formula:

[0107]

[0108] Where P load (t,n) represents the non-flexible load power of the nth non-flexible load node in the target power system at time t, in kW; N represents the total number of non-flexible load nodes in the target power system; P flex (t,m) represents the flexible load power of the mth flexible load node in the target power system at time t, in kW; M represents the total number of flexible load nodes in the target power system; P loss (t) represents the network loss power of the target power system at time t, in kW; P ESS (t,l) represents the charge and discharge power of the lth energy storage node in the target power system at time t, where a positive value represents the charging state and a negative value represents the discharging state, and the unit is kW; L represents the total number of energy storage nodes in the target power system; P PV (t,k) represents the output power of the kth photovoltaic node in the target power system at time t, in kW; K represents the total number of photovoltaic nodes in the target power system; C price (t) represents the time-of-use electricity price, in yuan / kwh.

[0109] Equivalent carbon reduction benefit of the target power system carbon Calculated by the following formula:

[0110]

[0111] In the formula, represents the power generation power of the lth node in the target power system at time t, in kW. When the lth node is a storage node, the power generation power is 0 kW;ESS represents the energy storage charging and discharging efficiency of the target power system. In the embodiment of the present invention, ESS The value of is 0.8; 1 , 2 Indicates the unit coal consumption of the energy storage node in the target power system when it is generating electricity or charging instead of the thermal power unit. In the embodiment of the present invention, 1 , 2 Take 0.3kg / kWh; C carbon Indicates the carbon emission trading price. In the embodiment of the present invention, C carbon Take 0.07 yuan / kg.

[0112] Equipment control cost of target power system f ops It can be calculated by the following formula:

[0113] f ops =C OLTC ·N T +C CB ·N C +f flex +f ESS +f PV (4)

[0114] in:

[0115]

[0116] In the formula, C OLTC represents the unit adjustment cost of the transformer tap in the target power system, in RMB / time. In the embodiment of the present invention, C OLTC The value of is 10 yuan / time; N T represents the number of transformer tap adjustments in the target power system; C CB represents the unit adjustment cost of the capacitor bank in the target power system, in RMB / time. In the embodiment of the present invention, C CB The value of is 2 yuan / time; f flex 、f ESS 、f PV They represent the control costs of flexible load nodes, energy storage nodes and photovoltaic nodes in the target power system, respectively, in yuan; ΔP flex (t,m),ΔP ESS (t,l),ΔP PV (t, k) represents the power change of the flexible load, energy storage and photovoltaic power generation at the corresponding node at time t, in kW; C flex , C ESS , C PV They are respectively expressed as the adjustment cost per unit flexible load, per unit energy storage, and per unit photovoltaic power, in units of RMB / kWh. In the embodiment of the present invention, C flex , CESS , C PV The values ​​are 0.032 yuan / kWh, 0.04 yuan / kWh, and 0.07 yuan / kWh respectively.

[0117] The penalty cost of abandoned solar power in the target power system is f punish It can be calculated by the following formula:

[0118]

[0119] Where P PV_rated (t,k) represents the maximum power that can be generated by the kth photovoltaic node in the target power system at time t, in kW; P PV_rated (t,k)-P PV (t,k) represents the abandoned power of the kth PV node at time t; η PV represents the light abandonment penalty coefficient, in the embodiment of the present invention, η PV The value of is 3.

[0120] It should be noted that the intraday rolling layer is used to correct the power control scheme of active power in the power control scheme being executed in the current time window, and keep the power control scheme of reactive power in the power control scheme being executed unchanged. It can eliminate the prediction error and the volatility of the power generation power of the photovoltaic node, so as to keep the comprehensive economic cost of the target power system in a better state at the same time as reducing the amount of calculation.

[0121] It should be noted that the day-ahead dispatching layer in the embodiment of the present invention takes the minimum comprehensive economic cost of the target power system as the goal, and the objective function corresponding to the intraday rolling layer can be expressed by the following formula:

[0122] F Hour =f energy -f carbon +f′ ops +f punish (7)

[0123] in:

[0124] f′ ops =f flex +f ESS +f PV (8)

[0125] In the formula, F Hour represents the intraday rolling regulation cost of the target power system when the current moment is the first node moment in the current time window, in yuan; f′ opsIt represents the intraday rolling control cost of the target power system after the power control plan for active power in the power control plan being executed in the current time window is corrected and updated, when the current moment is the first node moment in the current time window, in yuan.

[0126] It should be noted that the real-time optimization layer is used to establish indicators of voltage over-limit, voltage distortion and overall voltage safety margin of each node. When the current moment is the second node moment in the current time window and it is determined that there is an operational safety risk in the target power system at the current moment, the optimization of reactive power in the power control scheme being executed is triggered.

[0127] The real-time optimization layer aims to minimize the operational safety risk of the target power system. Therefore, in the embodiment of the present invention, the lowest operational safety risk of the target power system is determined as the second control target.

[0128] In the embodiment of the present invention, the operation safety risk index of the target power system can be used to describe the operation safety risk of the target power system.

[0129] The objective function corresponding to the real-time optimization layer can be expressed by the following formula:

[0130]

[0131] in:

[0132]

[0133]

[0134]

[0135] Where U optimize It represents the operation safety risk index of the target power system. When the voltage over-limit evaluation index, grid voltage distortion rate evaluation index and voltage safety margin evaluation index of the target power system are within the limit, the value of the operation safety risk index of the target power system is the smallest. The smaller the value of the operation safety risk index of the target power system, the lower the operation safety risk of the target power system and the better the voltage quality.

[0136] U punish1 (t), U punish2 (t), U punish3 (t) are penalty functions for voltage over-limit evaluation index, grid voltage distortion rate evaluation index, and voltage safety margin evaluation index, respectively, which are used to impose hard constraints on reactive power in the target power system. margin The negative sign before (t) is taken to solve the target minimum direction, which is convenient for subsequent optimization calculations.

[0137] The voltage over-limit evaluation index of the target power system can be calculated by the following formula:

[0138]

[0139] Where U over (t,n) represents the voltage over-limit evaluation index of the nth node in the target power system at time t. When the voltage of the nth node in the target power system at time t exceeds the voltage upper and lower limits, U over (t,n) takes the value of 1. When the voltage of the nth node in the target power system at time t does not exceed the upper and lower voltage limits, U over (t,n) takes the value of 0; U n ′(t) represents the voltage per unit value of the nth node in the target power system at time t; U max and U min are the upper and lower limits of voltage respectively. In the embodiment of the present invention, U max and U min The values ​​of are 1.05 and 0.95 respectively.

[0140] The grid voltage distortion rate evaluation index of the target power system can be calculated by the following formula:

[0141]

[0142] Where U dis (t+1,n) represents the grid voltage distortion rate evaluation index of the nth node in the target power system at time t+1, which is defined in U dis When the value of (t+1,n) is greater than 1, it indicates that there is a serious voltage fluctuation at the nth node in the target power system at time t+1.

[0143] The voltage safety margin evaluation index of the target power system can be calculated by the following formula:

[0144]

[0145] Where U margin (t) represents the voltage safety margin evaluation index of the target power system at time t, U margin When the value of (t) is positive, U margin The closer (t) is to 0, the smaller the voltage safety margin of the target power system at time t is, that is, the voltage of the target power system at time t is closer to the voltage upper limit or voltage lower limit, and the greater the risk of voltage exceeding the limit in the target power system; U margin When the value of (t) is negative, the voltage of the target power system at time t has exceeded the limit, and U marginThe smaller the value of (t), the greater the voltage limit of the target power system at time t.

[0146] As an optional embodiment, the real-time optimization layer is used to determine that there is an operational safety risk in the target power system at the current moment when the current moment is the second node moment within the current time window and at least one of the voltage over-limit evaluation index, the grid voltage distortion rate evaluation index and the voltage safety margin evaluation index of the target power system at the current moment exceeds the corresponding threshold range.

[0147] Specifically, the threshold range corresponding to the voltage over-limit evaluation index in the embodiment of the present invention is 0.

[0148] Correspondingly, when the current time is time t and the current time is the second node time in the current time window, if the voltage over-limit evaluation index U of the nth node in the target power system at time t is over If the value of (t,n) is 1, it can be determined that there is an operation safety risk in the target power system at the current moment. The nth node in the target power system can represent any node in the target power system.

[0149] In the embodiment of the present invention, the threshold value range corresponding to the grid voltage distortion rate evaluation index is not greater than 1.

[0150] Correspondingly, when the current time is t+1 and the current time is the second node time in the current time window, if the grid voltage distortion rate evaluation index U of the nth node in the target power system at t+1 is dis If the value of (t+1,n) is greater than 1, it can be determined that there is an operation safety risk in the target power system at the current moment. The nth node in the target power system can represent any node in the target power system.

[0151] In the embodiment of the present invention, the threshold range corresponding to the voltage safety margin evaluation index is not less than 0.02.

[0152] When the current time is time t and the current time is the second node time in the current time window, if the voltage safety margin evaluation index U of the target power system at time t is margin If (t) is less than 0.02, it can be determined that there is an operational safety risk in the target power system at the current moment.

[0153] It should be noted that the target constraint in the embodiment of the present invention may be predefined based on prior knowledge and / or actual conditions. The power balance equation constraint may be expressed by the following formula:

[0154]

[0155] Where P G_nand Q G_n represents the active power and reactive power of the generator node in the target power system; P Load_n and Q Load_n represents the active power and reactive power of the load node in the target power system; M represents the total number of nodes in the target power system; U n is the voltage of the nth node in the target power system; G nm , B nm ,θ nm is the conductance, susceptance, and voltage phase difference between the nth node and the mth node in the target power system.

[0156] The source-grid-load-storage operation constraints can be expressed by the following formula:

[0157]

[0158] Where P ESS_max (l) represents the rated charge and discharge power of the lth energy storage node in the target power system, in kW; SOC(t,l) represents the charge capacity of the lth energy storage node in the target power system, in kWh; P PV_rated (k), P PV_setting (k), P flex_max (m) represent the maximum power generation capacity of the kth PV node, the installed power of the kth PV node and the maximum power of the mth flexible load node in the target power system respectively.

[0159] The reactive equipment control constraint can be expressed by the following formula:

[0160]

[0161] Where N T_max 、N C_max Respectively represent the daily maximum adjustment times of transformer taps and capacitor banks in the target power system. In the embodiment of the present invention, N T_max 、N C_max The value of is 10 times; N O_max 、N O_min They represent the maximum and minimum gear numbers that can be adjusted by the transformer in the target power system. In the embodiment of the present invention, N O_max 、N O_min The values ​​of are 4 and -4, which means the maximum upward adjustment is 4 gears, the minimum downward adjustment is 4 gears, and the adjustment amount for each gear is ±1.25%; N CB_max Indicates the maximum number of capacitors in the target power system. In the embodiment of the present invention, N CB_max The value of is 4 groups.

[0162] As an optional embodiment, in the improved particle swarm algorithm, the position vector of any particle in the particle swarm in this iteration is determined based on the corrected position vector of any particle in the previous iteration; the corrected position vector of any particle in the previous iteration is obtained by correcting the position vector of any particle in the previous iteration based on the equal distance; the equal distance is determined based on the upper and lower limits of the position of any particle.

[0163] Specifically, since the active power in the target power system is a continuous variable and the reactive power is a discrete variable, if based on the traditional intelligent algorithm, there is a defect that the discrete variables are rounded at each iterative update, which makes the optimal solution easily fall into the local optimal solution, resulting in the original power control scheme or the revised power control scheme output by the source-grid-load-storage control model having unsatisfactory control effect.

[0164] Therefore, in the embodiment of the present invention, on the basis of the traditional particle swarm optimization (PSO), an iterative link of counteracting the extreme value distance is introduced, and an adaptive inertia weight parameter is designed. The state of the optimal solution is tracked in real time during the evolution process and parameter optimization link, and the iterative strategy is updated in time to ensure that the particle swarm algorithm jumps out of the local optimal solution.

[0165] If the total number of particles in the traditional particle swarm algorithm is I, i is the particle identifier, the number of iterations is s, and the position vector of particle i in the sth iteration is The velocity sequence of particle i is J represents the total number of variables, then the position and velocity of particles in the particle population in the traditional particle swarm algorithm can be calculated by the following formula:

[0166]

[0167] In the formula, X i (s) and V i (s) represent the position vector and velocity sequence of particle i in the sth iteration respectively; w represents the inertia weight of the particle swarm algorithm; β 1 and β 2 is a random coefficient, with a value range of [0,1]; X P_i and X G_i They represent the individual optimal solution and the global optimal solution respectively. The individual optimal solution is the optimal solution of particle i to the sth iteration, while the global optimal solution is the optimal solution of the particle population to the sth iteration; α 1 and α 2 is the acceleration factor, and this paper takes the value as 1.49.

[0168] For continuous variables, the above traditional particle swarm algorithm can smoothly adjust the position and speed of particles, but the above traditional particle swarm algorithm is not suitable for discrete variables. It not only requires sustainable rounding operations, but also avoids falling into the local optimal solution after rounding. In addition, the iteratively updated X in formula (19) i (s+1), V i The probability that (s+1) is not an integer is high, so it is necessary to update X i (s+1), V i (s+1) is rounded.

[0169] Integer variable rounding operation, if the rounding direct rounding truncation method is adopted, the values ​​at the upper and lower limits of the boundary will be directly truncated, and the probability of taking an integer is not equal to other values. Therefore, in the embodiment of the present invention, the equal distance probability value method is adopted, and the upper and lower limits of the particle position are set to be X i_max and X i_min .

[0170]

[0171] X′ i (s+1)=X i_min +ROUND[ΔX i ·|X i (s+1)-X i_min |] (21)

[0172] Where, X i (s+1) represents the position vector of particle i in the s+1th iteration; X′ i (s+1) represents the position vector of particle i in the s+1th iteration after correction; ΔX i Indicates equal distance; ROUND[·] indicates rounding to the nearest integer.

[0173] In the embodiment of the present invention, the distance of the coordinate axis can be re-divided by the equal distance probability value method, and the adjacent coordinate difference is divided from 1 to ΔX i , so that the two boundary values ​​have the same probability value.

[0174] As an optional embodiment, the inertia weight in the improved particle swarm algorithm is negatively correlated with the number of iterations, and the inertia weight corresponding to any iteration is calculated based on the number of iterations of any iteration and the predefined maximum number of iterations, the maximum and minimum values ​​of the inertia weight and the adjustment coefficient.

[0175] Specifically, the inertia weight w in the particle swarm algorithm is directly related to the global and local optimization capabilities. When the number of iterations is small, the larger the inertia weight w, the stronger the global optimization capability and the weaker the local optimization capability. On the contrary, when the number of iterations is large, the smaller the inertia weight w, the weaker the global optimization capability and the stronger the local optimization capability.

[0176] Therefore, in the embodiment of the present invention, an adaptive function of the inertia weight w changing with the number of iterations s is designed according to the number of iterations s, so as to ensure that the improved particle swarm algorithm has a larger inertia weight w when the number of iterations is small, thereby increasing the global optimization ability. In the case of a large number of iterations in the late iteration, the smaller the inertia weight w, the greater the local optimization ability. The adaptive function of the inertia weight w changing with the number of iterations s can be expressed by the following formula:

[0177]

[0178] Where w(s) represents the inertia weight corresponding to the sth iteration; w max 、w min Respectively represent the maximum and minimum values ​​of the inertia weight. In the embodiment of the present invention, w max 、w min Take 0.9 and 0.5 respectively; η o represents the adjustment coefficient, and its range is [1,30]. In the present invention, the value is 3; s max Indicates the maximum number of iterations.

[0179] As an optional embodiment, in the improved particle swarm algorithm, when the global optimal solution of the particle population is not updated for consecutive target number of iterations, the position vector of each particle in this iteration is determined based on the position vector of each particle in the previous iteration and the adversarial extreme value distance between the position vectors of any two particles in the particle population in the previous iteration; the adversarial extreme value distance is determined based on the cosine similarity coefficient between the position vectors of any two particles and the extreme value distance of the position vector of each particle in the previous iteration.

[0180] Specifically, in order to prevent the particle swarm algorithm from falling into a local optimal solution and enhance the ability of the particle swarm algorithm to jump out of the local optimal solution, an embodiment of the present invention proposes a population evolution strategy that resists extreme distance.

[0181] First, the embodiment of the present invention sets that if the continuous S gen If there is no global optimal solution update, the particle population is considered to have entered the local optimal solution state, that is, S gen is the target number. At this time, it is necessary to enhance the ability of the algorithm to jump out of the local optimal solution, perform global optimization multiple times, and activate the link of updating the counteracting extreme value distance.

[0182] For any two particles i' and i'' in the particle population, the position vectors of particles i' and i'' in the sth iteration are respectively The position vector of particle i in the s+1th iteration can be updated based on the following formula:

[0183] X i (s+1)=X i (s)+D resist [X i′ (s),X i″ (s)] (23)

[0184]

[0185] in:

[0186]

[0187]

[0188] In the formula, j represents any variable; D resist [] represents the antagonistic extreme distance between the position vectors of two particles; D[X i (s)] represents the extreme distance of the position vector of particle i, that is, the variable distance of each variable in particle i; COS[X i′ (s),X i″ (s)] represents the position vector X i′ (s) and X i″ The cosine similarity coefficient between (s), the position vector X i′ (s) and X i″ The closer (s) is, the better COS[X i′ (s),X i″ The closer (s)] is to 1, the position vector X i′ (s) and X i″ The greater the difference in (s), the greater the COS[X i′ (s),X i″ (s)] is closer to 0; randsrc represents a function that randomly takes ±1.

[0189] Because when using traditional particle swarm algorithm to solve mixed variables, the traditional particle swarm algorithm is easy to fall into local solutions. There is no optimal solution updated between particles for many consecutive times. The traditional particle swarm algorithm is easy to move toward the current global "optimal solution", thus falling into the local solution and finding it difficult to escape.

[0190] At this time, each particle has a high similarity in the variable sequence, that is, the closer the solution set is to the "optimal solution", the higher the similarity of the particles is, and the more difficult it is to jump out of the local optimal solution. Therefore, it is necessary to reversely change this local aggregation situation. The present invention proposes an update strategy based on the confrontation extreme value distance. First, two different particles are randomly calculated and their confrontation extreme value distance is calculated. When the two vectors are closer, the cosine similarity is higher, and the value is closer to 1. At this time, the confrontation extreme value distance D obtained is resist The larger the value of (), the The greater the change, the stronger the ability to jump out of the optimal solution. On the contrary, if the similarity of the two selected particles is lower, it means that the difference between the particles is large, and there is no need for a large particle change. The smaller the cosine similarity value, the stronger the ability to resist the extreme value distance D. resist The smaller the value of (), the smaller the value. At this time, not only the positions of all particles change, but also the initial velocity sequence V of all particles i (s) Initialize again.

[0191] The embodiment of the present invention inputs the predicted value of the first power data in the current time window into the source-grid-load-storage control model, obtains the original power control scheme of the target power system in the current time window output by the source-grid-load-storage control model, and then starts to execute the original power control scheme when the current moment is the starting moment of the current time window to control the active power and reactive power of the target power system. When the revised power control scheme output by the source-grid-load-storage control model is obtained, the power control scheme being executed is replaced with the revised power control scheme. The source-grid-load-storage control model is used to correct the power control scheme being executed based on the actual value of the second power data at the current moment and the predicted value of the first power data in the current time window when the current moment is the first node moment in the current time window. The power control scheme for active power in the scheme, the source-grid-load-storage control model is also used to correct the power control scheme for reactive power in the power control scheme being executed based on the actual value of the second power data at the current moment and the actual value of the electrical quantity data of the target power system when the current moment is the second node moment within the current time window and it is determined that there is an operational safety risk in the target power system at the current moment. This can more accurately coordinate source-grid-load-storage scheduling for power systems including a large number of distributed power sources, highly flexible energy storage devices and flexible loads, and control the active power and reactive power of the above-mentioned power systems from three time scales. In the control, the decoupling of active power and reactive power is achieved, the amount of calculation is reduced, the calculation efficiency is improved, the operational risk of the above-mentioned power system can be reduced, and the comprehensive cost of the above-mentioned power system can be reduced.

[0192] In order to facilitate the understanding of the source-grid-load-storage coordinated regulation method provided by the present invention, the source-grid-load-storage coordinated regulation provided by the present invention is explained and verified through an example as follows.

[0193] In this example, the IEEE33 node calculation example is used to verify the method provided by the present invention. The voltage registration of the example is 12.66kV, the system reference capacity is 10MVA, the initial voltage reference value of node 1 is 1, the adjustment ratio of the transformer is ±4 gears, each gear can be adjusted by 1.25%, the photovoltaic nodes are located at nodes 17 and 29, and the rated installed capacity is 400kW; the energy storage is located at node 14, the energy storage capacity is 200kWh, the initial capacity is 40kWh, and the maximum rated charging and discharging power is 50kW. The nodes 18 and 22 of the power grid are equipped with capacitor group devices, and each node is equipped with a 4×20kvar reactive compensation device. The initial gear of the capacitor is 0, and the initial gear of the transformer tap is 0; the flexible electric load is located at node 25, and the flexible load power range is 0 to 50kW.

[0194] Figure 3 The topology diagram of IEEE33 nodes is shown in Figure 1. Figure 3 shown.

[0195] In this example, the off-peak period of the power grid is 0:00-7:00h, and the electricity price is 0.56 yuan / kWh; the normal period is 12:00-16:00, 22:00-23:00, and the electricity price is 1.12 yuan / kWh; the peak period is 8:00-11:00, 17:00-21:00, and the electricity price is 1.56 yuan / kWh.

[0196] In this example, the particle swarm algorithm has a particle population size of 50, a maximum number of iterations of 200, and a target number S. gen For 20 times.

[0197] In this example, an initial model can be constructed based on the BP neural network, and the above initial model can be trained based on the actual values ​​of the first power data in multiple continuous historical time windows to obtain a power data prediction model. Then, the actual value of the first power data in the previous day can be input into the above power data prediction model to obtain the predicted value of the first power data on the day output by the above power data prediction model.

[0198] Specifically, it includes: collecting the first power data of the target power system for 30 days in the same month of the previous year, taking the first power data of each day as a group of samples, for a total of 30 groups of samples; then taking the first power data of each day as a sample, and taking the first power data of the next day of each day as a sample label, training the BP neural network model, and after 29 times of training, obtaining the above-mentioned power data prediction model; inputting the actual value of the first power data of the previous day into the above-mentioned power data prediction model, and obtaining the predicted value of the first power data of the day output by the above-mentioned power data prediction model.

[0199] Based on the predicted value of the first power data of the day and the source-grid-load-storage regulation model, the active power and reactive power in the target power system of the day are regulated.

[0200] Firstly, the effectiveness of the improved particle swarm algorithm is verified. Figure 4 The improved particle swarm algorithm in the source-grid-load-storage method provided by the present invention is compared with the iterative convergence of the traditional particle swarm algorithm. The improved particle swarm algorithm proposed by the present invention is called ARED-PSO (Adaptive resist extreme distance particle swarm optimization), and the iterative convergence process of the traditional particle swarm algorithm PSO, the particle swarm algorithm RED-PSO (Resistextreme distance particle swarm optimization) algorithm and the adaptive particle swarm algorithm APSO (Adaptive particle swarm optimization) and the improved particle swarm algorithm proposed by the present invention is ARED-PSO. Figure 4 shown.

[0201] from Figure 4 It can be seen that the traditional particle swarm algorithm PSO has slow iteration convergence when solving mixed variables. It slowly enters the convergence stage after 180 iterations, and the convergence accuracy is not ideal.

[0202] When the APSO algorithm is used and the adaptive link is added, the inertia weight w takes a larger value in the initial stage of the solution to increase the global convergence strength. In the convergence stage, the inertia weight w takes a smaller value to increase the local search capability. Therefore, the optimization effect in the early stage of the convergence curve is significant. The algorithm converges in 72 times and the convergence efficiency is improved.

[0203] The REDPSO algorithm is adopted, and the re-iteration link of the anti-extreme distance is added. When the iteration enters the gentle convergence stage, the particles will be updated again. For example, Figure 4 At the dotted “circle” in the figure, the iterative calculation of the adversarial extreme value distance is performed again to further improve the convergence accuracy of the algorithm.

[0204] By adopting the improved particle swarm algorithm ARED-PSO proposed in the present invention and adding adaptive weight and anti-extreme distance evolution links, the algorithm has the highest convergence accuracy and the most ideal result. At the same time, the algorithm has a high convergence efficiency and basically enters a gentle convergence period at the 91st time.

[0205] Therefore, the improved particle swarm algorithm ARED-PSO proposed in the present invention can better solve the problem of high-dimensional mixed variables. It can not only effectively process discrete variables, but also effectively improve its efficiency and accuracy compared with other particle swarm algorithms.

[0206] The source-grid-load-storage method provided by the present invention performs intraday optimization on the basis of the original power regulation scheme, and the results of various indicators are shown in Table 1.

[0207] Table 1 Comparison of economic costs of intraday rolling optimization

[0208] Objective function value Day-ahead scheduling Active power intraday rolling Comprehensive control cost (yuan) 70181.05 69739.36 Energy purchase cost (yuan) 68338.28 68061.33 Equivalent carbon reduction benefits (yuan) 73.71 78.28 Comprehensive operation and maintenance cost (yuan) 1276.09 1140.97 Abandoned light penalty cost (yuan) 640.39 615.34

[0209] From the data in Table 1, it can be seen that the comprehensive cost of intraday rolling optimization scheduling is 69739.36 yuan, of which the main proportion is the energy purchase cost, which is 68051.33 yuan, followed by the comprehensive operation and maintenance cost of 1140.97 yuan. Compared with the day-ahead scheduling plan, its comprehensive regulation cost is reduced by 135.12 yuan, and the reduced comprehensive cost is mainly reflected in the reduction of operation and maintenance costs. Because the intraday rolling optimization is based on the day-ahead scheduling plan, and the real photovoltaic and load data are updated every 1 hour, it is more accurate than the day-ahead scheduling plan based on prediction. Therefore, it can adjust the active power of photovoltaic, energy storage, and flexible load in real time, reduce the power regulation deviation caused by prediction deviation, and thus reduce the operation and maintenance cost. In addition, through the rolling adjustment of the present invention, the uncertainty of photovoltaic can be weakened, the penalty cost of abandoning light can be reduced, and the absorption rate of photovoltaic can be improved. Through the real-time adjustment of the light-storage linkage, the photovoltaic absorption rate can be improved, and the equivalent carbon reduction benefits can be improved.

[0210] The present invention performs real-time voltage safety monitoring on the basis of the rolling active power optimization scheme every 1 hour during the day. When the safety risk index is found to be abnormal, reactive power optimization is performed in real time. The comparison results of the real-time reactive power optimization are shown in Table 2.

[0211] Table 2 Comparison of voltage safety risks of real-time monitoring of reactive power optimization

[0212] Objective function value Intraday rolling Real-time reactive power optimization Voltage over-limit time ratio (%) 29.17 4.16 Average voltage distortion rate (%) 10.34 4.73 Average voltage safety margin (%) 2.31 3.19

[0213] As can be seen from Table 2, on the basis of intraday rolling optimization of active power, the present invention obtains power data and electrical quantity data of each node in the entire network every 5 minutes, establishes three indicators of voltage over-limit evaluation index, power grid voltage distortion rate evaluation index and voltage safety margin evaluation index and corresponding threshold ranges, and starts reactive power optimization when it is determined based on the above three indicators that the target power system has an operation safety risk, optimizes the power control scheme of reactive power in the power control scheme being executed, and improves the voltage safety performance of the target power system. The proportion of voltage over-limit time is reduced from 29.17% to 4.16%, the average voltage distortion rate is reduced from 10.34% to 4.73%, and the voltage safety margin is increased from 2.31% to 3.19%. The voltage quality is effectively improved, and the voltage safety performance is effectively improved.

[0214] By generating the original power control plan a day ago, performing intraday rolling optimization of active power and real-time optimization of reactive power, coordinating the three time scales, and performing hierarchical optimization according to the nature of the equipment and whether the variables are discrete, more accurate source-grid-load-storage coordinated scheduling can be achieved for the target power system that includes a large number of distributed power sources, highly flexible energy storage devices, and flexible loads.

[0215] Figure 5 Schematic diagram of the power control scheme of active power in the source-grid-load-storage coordinated control method provided by the present invention. The power control scheme of the active power of the target power system on the same day is as follows: Figure 5 shown.

[0216] Depend on Figure 5 It can be seen that PV 1 and PV 2 have significant synchronization, and are basically at their maximum output at noon, with basically similar power per hour. Energy storage has the typical characteristics of "peak shaving and valley filling". At night, when the load and electricity price are low, it is basically in a charging state. At the peak of electricity prices in the morning and evening, it is in a discharging state. At noon, when the photovoltaic power is fully generated, it is in a charging state, assisting the photovoltaic power to further absorb. The power regulation of flexible loads also has the typical characteristics of "peak shaving and valley filling". When the electricity consumption and electricity price are low, the load is at a large level, and the electricity load will be reduced during peak hours, which plays an indirect peak regulation role.

[0217] Figure 6 Schematic diagram of the reactive power control scheme in the source-grid-load-storage coordinated control method provided by the present invention. Figure 6 shown.

[0218] Depend on Figure 6It can be seen that in the power control scheme of the active power of the standard power system on that day, the transformer tap position is mostly at -2 gear, followed by -1 gear, and only once was it adjusted to 2 gears, and the overall gear adjustment range is not large. Because the transformer tap adjustment cost is relatively high, it is only adjusted 6 times throughout the day. The adjustment cost of the capacitor group is relatively low, so the adjustment is more frequent, 12 and 13 times throughout the day, respectively, and the adjustment range of the group is not large, basically each adjustment is in 1 to 2 groups.

[0219] The source-grid-load-storage control model provided by the present invention has the following key technologies and innovations: First, in terms of hierarchical structure, three time scales are established: 24-hour scheduling optimization, 1-hour rolling optimization within the day, and voltage monitoring and reactive power optimization every 5 minutes. The BP neural network predicts 24-hour photovoltaic and load data, and can formulate control plans in advance; by collecting electrical quantity data every 1 hour, updating the real-time data of photovoltaic and load, and making rolling corrections to the day-ahead plan, the uncertainty of photovoltaic is eliminated and the timeliness of the control plan is improved; the real-time voltage monitoring and reactive power optimization every 5 minutes monitors voltage safety and reduces operational risks at a shorter real-time level. The three time scales cooperate and correct each other to improve the operation effect of the algorithm;

[0220] Secondly, in terms of power decoupling optimization, not only is the three layers divided according to the time scale, but the power is also decoupled before optimization. According to the regulation characteristics of controllable power equipment, photovoltaic, energy storage, flexible loads and other power equipment that can be smoothly adjusted and frequently adjusted are divided into the intra-day rolling layer to optimize active power. Transformer taps, capacitor banks and other non-smooth equipment and equipment that are relatively infrequently adjusted are divided into the real-time layer for conditional reactive power optimization. On the one hand, this decoupling optimization method reduces the dimensions of optimization variables in the intra-day layer and the real-time layer, which is conducive to greatly reducing the amount of model calculation and improving the efficiency of algorithm solution; on the other hand, the intra-day layer is mainly based on the fine-tuning of active power, mainly adjusting the active power flow and dynamically correcting the economic cost, while the reactive regulation of the real-time layer directly affects the voltage. The reactive optimization can more directly and effectively improve the voltage quality, making the optimization of the two layers more targeted.

[0221] Thirdly, the goal of the source-grid-load-storage control model is to consider the comprehensive economic cost at the day-ahead dispatch layer and the intraday rolling layer. It not only takes into account the energy purchase cost, network loss cost, and equipment control cost, but also adds the equivalent carbon reduction benefit and the penalty cost for abandoned light. While ensuring economic benefits, it further promotes carbon emission reduction and photovoltaic absorption, which is in line with the overall development strategy goal of energy conservation and carbon reduction of the power grid.

[0222] Finally, in the real-time optimization layer, three indicators, namely, voltage over-limit evaluation index, grid voltage distortion rate evaluation index and voltage safety margin evaluation index and their corresponding threshold ranges, are proposed. A corresponding mathematical model and threshold range are established for each indicator. Real-time monitoring is carried out based on the three indicators to ensure the grid voltage quality while ensuring the safe operation level of the grid.

[0223] The present invention proposes an improved particle swarm algorithm for solving mixed variables. The improved particle swarm algorithm has the following technical key points and innovations: first, in view of the situation where the probability of integer variables after rounding is unequal, a rounding method based on equal distance probability is proposed to re-divide the distance of the coordinate axis, improve the shortcomings of the conventional rounding and truncation method, and make the probability of each discrete integer variable being selected equal;

[0224] Secondly, an adaptive inertia weight was designed. According to the characteristics of population optimization iteration in the evolutionary algorithm, the inertia weight was dynamically improved according to the number of iterations, and the optimization focus of the algorithm was dynamically adjusted. The global optimization ability of the algorithm was strengthened in the early stage, and the local optimization ability of the algorithm was strengthened in the later stage, thereby improving the computational efficiency of the algorithm.

[0225] Finally, in view of the fact that the particle swarm algorithm is prone to fall into the local optimal solution when iterating to optimize the mixed variables, the present invention improves the PSO algorithm and proposes a starting threshold. If there is no optimal solution update for a certain continuous scale, it is determined to be trapped in the local solution. Then the update link of the counteracting extreme value distance proposed by the present invention is started, and all particles in the population are updated again, thereby increasing the ability to jump out of the local solution and improving the calculation accuracy of the algorithm.

[0226] The source-grid-load-storage coordinated control method provided by the present invention has the following advantages: first, the three time scales of day-ahead scheduling, intra-day rolling, and real-time optimization are established to cooperate with each other, which can not only issue control plans in advance to guide the operation of the power grid in advance, but also strategically correct the plans in real time on the same day, reduce the deviations caused by factors such as predicted values ​​and photovoltaic randomness, improve the optimization results, and real-time voltage monitoring and reactive power optimization ensure the quality of power grid voltage and the safety of power grid operation;

[0227] Secondly, a multi-time scale hierarchical optimization model is established. On the basis of the 24h control plan determined the day before, the intraday rolling layer and the real-time optimization layer are decoupled and controlled according to the nature and frequency of the control equipment. The continuous and relatively frequently adjusted active power control equipment is adjusted intraday, and calculation and adjustment are performed every 1 hour. The non-continuous and relatively infrequently adjusted reactive power equipment is monitored in real time, and then calculated and adjusted according to the operating conditions. The active and reactive power are divided into intradays, which reduces the amount of calculation for intraday rolling optimization. Moreover, the continuous and non-continuous variables are calculated separately, which simplifies the calculation process of the algorithm. This optimization model not only takes into account the comprehensive economic benefits, but also takes into account the energy-saving and carbon-reduction benefits and the grid voltage safety risks, and the optimization objectives are relatively comprehensive.

[0228] Finally, the optimization algorithm of the present invention is applicable to the field of rapid solution of mixed variables. The equidistant probability rounding further ensures the probability of each integer variable being taken, and provides a new rounding reference method. The designed adaptive inertia weight and anti-extreme distance evolution solution link can not only dynamically adjust the global and local optimization capabilities of the algorithm and improve the efficiency of the algorithm, but also improve the problem that the mixed variable solution is prone to fall into the local optimum and improve the accuracy of the algorithm. The algorithm has good scalability and provides effective support for the solution of mixed variables.

[0229] Figure 7 This is a schematic diagram of the structure of the source-grid-load-storage coordinated control device provided by the present invention. Figure 7 The source-grid-load-storage coordinated control device provided by the present invention is described. The source-grid-load-storage coordinated control device described below and the source-grid-load-storage coordinated control method provided by the present invention described above can be referred to each other. Figure 7 As shown, the device includes: a solution prediction module 701 and a dynamic optimization module.

[0230] A scheme prediction module 701 is used to input the predicted value of the first power data in the current time window into the source-grid-load-storage regulation model to obtain the original power regulation scheme of the target power system in the current time window output by the source-grid-load-storage regulation model;

[0231] The dynamic optimization module 702 starts to execute the original power control scheme when the current time is the starting time of the current time window, controls the active power and reactive power of the target power system, and replaces the power control scheme being executed with the revised power control scheme when the revised power control scheme output by the source-grid-load-storage control model is obtained;

[0232] Among them, the source-grid-load-storage control model is used to correct the power control scheme of active power in the power control scheme being executed based on the actual value of the second power data at the current moment and the predicted value of the first power data in the current time window when the current moment is the first node moment in the current time window. The source-grid-load-storage control model is also used to correct the power control scheme of reactive power in the power control scheme being executed based on the actual value of the second power data at the current moment and the actual value of the electrical quantity data of the target power system when the current moment is the second node moment in the current time window and it is determined that there is an operational safety risk in the target power system at the current moment;

[0233] The first power data includes the power generation power of the distributed power nodes and the load power of each load node in the target power system; the second power data includes the power data of each node in the target power system; the first time duration is the interval between any two adjacent first node moments in the current time window; the second time duration is the interval between any two adjacent second node moments in the current time window; the first time duration is greater than the second time duration; the active power in the target power system is a continuous variable, and the reactive power in the target power system is a discrete variable.

[0234] The source-grid-load-storage collaborative control device in the embodiment of the present invention inputs the predicted value of the first power data in the current time window into the source-grid-load-storage control model, obtains the original power control scheme of the target power system in the current time window output by the source-grid-load-storage control model, and then starts to execute the original power control scheme when the current moment is the starting moment of the current time window to control the active power and reactive power of the target power system. When the revised power control scheme output by the source-grid-load-storage control model is obtained, the power control scheme being executed is replaced with the revised power control scheme. The source-grid-load-storage control model is used to correct the power control scheme being executed based on the actual value of the second power data at the current moment and the predicted value of the first power data in the current time window when the current moment is the first node moment in the current time window. The power control scheme for active power in the power control scheme being executed, the source-grid-load-storage control model is also used to correct the power control scheme for reactive power in the power control scheme being executed based on the actual value of the second power data at the current moment and the actual value of the electrical quantity data of the target power system when the current moment is the second node moment in the current time window and it is determined that there is an operational safety risk in the target power system at the current moment. This can more accurately coordinate source-grid-load-storage scheduling for power systems including a large number of distributed power sources, highly flexible energy storage devices and flexible loads, and control the active power and reactive power of the above-mentioned power systems from three time scales. In the control, the decoupling of active power and reactive power is achieved, the amount of calculation is reduced, the calculation efficiency is improved, the operational risk of the above-mentioned power system can be reduced, and the comprehensive cost of the above-mentioned power system can be reduced.

[0235] Figure 8 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 8 As shown, the electronic device may include: a processor (processor) 810, a communication interface (Communications Interface) 820, a memory (memory) 830 and a communication bus 840, wherein the processor 810, the communication interface 820, and the memory 830 communicate with each other through the communication bus 840. The processor 810 can call the logic instructions in the memory 830 to execute the source-grid-load-storage coordinated control method, which includes: inputting the predicted value of the first power data in the current time window into the source-grid-load-storage control model, and obtaining the original power control plan of the target power system in the current time window output by the source-grid-load-storage control model; when the current moment is the starting moment of the current time window, starting to execute the original power control plan, regulating the active power and reactive power of the target power system, and when the revised power control plan output by the source-grid-load-storage control model is obtained, replacing the power control plan being executed with the revised power control plan; wherein the source-grid-load-storage control model is used to correct the power control plan being executed based on the actual value of the second power data at the current moment and the predicted value of the first power data in the current time window when the current moment is the first node moment in the current time window. The power control scheme for active power in the power control scheme, the source-grid-load-storage control model is also used to correct the power control scheme for reactive power in the power control scheme being executed based on the actual value of the second power data at the current moment and the actual value of the electrical quantity data of the target power system when the current moment is the second node moment in the current time window and it is determined that there is an operation safety risk in the target power system at the current moment; the first power data includes the power generation power of the distributed power source node and the load power of each load node in the target power system; the second power data includes the power data of each node in the target power system; the first time interval between any two adjacent first node moments in the current time window is the first time interval; the second time interval between any two adjacent second node moments in the current time window is the second time interval; the first time interval is greater than the second time interval; the active power in the target power system is a continuous variable, and the reactive power in the target power system is a discrete variable.

[0236] In addition, the logic instructions in the above-mentioned memory 830 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0237] On the other hand, the present invention also provides a computer program product, which includes a computer program, which can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the source-grid-load-storage coordinated control method provided by the above methods, which method includes: inputting the predicted value of the first power data in the current time window into the source-grid-load-storage control model, and obtaining the original power control scheme of the target power system in the current time window output by the source-grid-load-storage control model; when the current moment is the starting moment of the current time window, starting to execute the original power control scheme, regulating the active power and reactive power of the target power system, and when the revised power control scheme output by the source-grid-load-storage control model is obtained, replacing the power control scheme being executed with the revised power control scheme; wherein the source-grid-load-storage control model is used to, when the current moment is the first node moment in the current time window, based on the second power control scheme at the current moment. The power control scheme of active power in the power control scheme being executed is corrected based on the actual value of the rate data and the predicted value of the first power data in the current time window. The source-grid-load-storage control model is also used to correct the power control scheme of reactive power in the power control scheme being executed when the current moment is the second node moment in the current time window and it is determined that there is an operation safety risk in the target power system at the current moment, based on the actual value of the second power data at the current moment and the actual value of the electrical quantity data of the target power system; the first power data includes the generated power of the distributed power source node and the load power of each load node in the target power system; the second power data includes the power data of each node in the target power system; the first time interval between any two adjacent first node moments in the current time window is the first time interval; the second time interval between any two adjacent second node moments in the current time window is the second time interval; the first time interval is greater than the second time interval; the active power in the target power system is a continuous variable, and the reactive power in the target power system is a discrete variable.

[0238] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the source-grid-load-storage coordinated control method provided by the above-mentioned methods, the method comprising: inputting the predicted value of the first power data in the current time window into the source-grid-load-storage control model, and obtaining the original power control scheme of the target power system in the current time window output by the source-grid-load-storage control model; when the current moment is the starting moment of the current time window, starting to execute the original power control scheme, regulating the active power and reactive power of the target power system, and when the revised power control scheme output by the source-grid-load-storage control model is obtained, replacing the power control scheme being executed with the revised power control scheme; wherein the source-grid-load-storage control model is used to, when the current moment is the first node moment in the current time window, based on the actual value of the second power data at the current moment and the current time window The predicted value of the first power data in the mouth is used to correct the power control scheme of active power in the power control scheme being executed. The source-grid-load-storage control model is also used to correct the power control scheme of reactive power in the power control scheme being executed based on the actual value of the second power data at the current moment and the actual value of the electrical quantity data of the target power system when the current moment is the second node moment in the current time window and it is determined that there is an operation safety risk in the target power system at the current moment; the first power data includes the power generation power of the distributed power source node and the load power of each load node in the target power system; the second power data includes the power data of each node in the target power system; the first time interval between any two adjacent first node moments in the current time window is the first time interval; the second time interval between any two adjacent second node moments in the current time window is the second time interval; the first time interval is greater than the second time interval; the active power in the target power system is a continuous variable, and the reactive power in the target power system is a discrete variable.

[0239] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0240] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0241] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A source-grid-load-storage coordinated control method, characterized in that: include: Inputting the predicted value of the first power data in the current time window into the source-grid-load-storage regulation model, and obtaining the original power regulation scheme of the target power system in the current time window output by the source-grid-load-storage regulation model; When the current moment is the starting moment of the current time window, the original power control scheme is started to be executed to control the active power and reactive power of the target power system, and when the revised power control scheme output by the source-grid-load-storage control model is obtained, the power control scheme being executed is replaced with the revised power control scheme; Wherein, the source-grid-load-storage control model is used to correct the power control scheme of active power in the power control scheme being executed based on the actual value of the second power data at the current moment and the predicted value of the first power data in the current time window when the current moment is the first node moment in the current time window; the source-grid-load-storage control model is also used to correct the power control scheme of reactive power in the power control scheme being executed based on the actual value of the second power data at the current moment and the actual value of the electrical quantity data of the target power system when the current moment is the second node moment in the current time window and it is determined that there is an operational safety risk in the target power system at the current moment; The first power data includes the power generation power of the distributed power source nodes and the load power of each load node in the target power system; the second power data includes the power data of each node in the target power system; the first time interval between any two adjacent first node moments in the current time window is the first time interval; the second time interval between any two adjacent second node moments in the current time window is the second time interval; the first time interval is greater than the second time interval; the active power in the target power system is a continuous variable, and the reactive power in the target power system is a discrete variable.

2. The source-grid-load-storage coordinated control method according to claim 1 is characterized in that: The source-grid-load-storage control model includes: day-ahead scheduling layer, intraday rolling layer and real-time optimization layer; The day-ahead scheduling layer is used to obtain and output the power control scheme based on the predicted value of the first power data in the current time window, according to the target constraint and the first control target, using the improved particle swarm algorithm; The intra-day rolling layer is used to correct the power control scheme of active power in the power control scheme being executed based on the actual value of the second power data at the current moment and the predicted value of the first power data in the current time window, according to the target constraint and the first control target, when the current moment is the first node moment in the current time window, and then output the corrected power control scheme; The real-time optimization layer is used to correct the power control scheme of reactive power in the power control scheme being executed based on the actual value of the second power data at the current moment and the actual value of the electrical quantity data of the target power system, according to the target constraint and the second control target, using the improved particle swarm algorithm, when the current moment is the second node moment within the current time window and it is determined that the target power system at the current moment has an operation safety risk, and then output the corrected power control scheme; Among them, the first control target includes minimizing the comprehensive economic cost of the target power system; the second control target includes minimizing the operating safety risk of the target power system; the constraints include power balance equation constraints, source-grid-load-storage operation constraints, and reactive equipment control constraints.

3. The source-grid-load-storage coordinated control method according to claim 2 is characterized in that: In the improved particle swarm algorithm, when the global optimal solution of the particle population is not updated for consecutive target number of iterations, the position vector of each particle in the current iteration is determined based on the position vector of each particle in the previous iteration and the adversarial extreme value distance between the position vectors of any two particles in the particle population in the previous iteration; the adversarial extreme value distance is determined based on the cosine similarity coefficient between the position vectors of the any two particles and the extreme value distance of the position vector of each particle in the previous iteration.

4. The source-grid-load-storage coordinated control method according to claim 2 is characterized in that: In the improved particle swarm algorithm, the position vector of any particle in the particle population in this iteration is determined based on the corrected position vector of any particle in the previous iteration; the corrected position vector of any particle in the previous iteration is obtained by correcting the position vector of any particle in the previous iteration based on the equal distance; the equal distance is determined based on the upper limit and lower limit of the position of any particle.

5. The source-grid-load-storage coordinated control method according to claim 2 is characterized in that: The inertia weight in the improved particle swarm algorithm is negatively correlated with the number of iterations, and the inertia weight corresponding to any iteration is calculated based on the number of iterations of the any iteration and the predefined maximum number of iterations, the maximum and minimum values ​​of the inertia weight and the adjustment coefficient.

6. The source-grid-load-storage coordinated control method according to claim 2 is characterized in that: The real-time optimization layer is specifically used to determine that there is an operational safety risk in the target power system at the current moment when the current moment is the second node moment within the current time window and at least one of the voltage over-limit evaluation index, the grid voltage distortion rate evaluation index and the voltage safety margin evaluation index of the target power system at the current moment exceeds the corresponding threshold range.

7. The source-grid-load-storage coordinated control method according to any one of claims 1 to 6, characterized in that: Before inputting the predicted value of the first power data in the current time window into the source-grid-load-storage regulation model and obtaining the original power regulation scheme of the target power system in the current time window output by the source-grid-load-storage regulation model, the method further includes: Obtaining an actual value of the first power data in the previous time window of the current time window; Inputting the actual value of the first power data in the previous time window into a power data prediction model, and obtaining a predicted value of the first power data in the current time window output by the power data prediction model; The power data prediction model is constructed based on a BP neural network and is trained based on actual values ​​of the first power data in multiple continuous historical time windows.

8. A source-grid-load-storage coordinated control device, characterized in that: include: A scheme prediction module is used to input the predicted value of the first power data in the current time window into the source-grid-load-storage regulation model to obtain the original power regulation scheme of the target power system in the current time window output by the source-grid-load-storage regulation model; The dynamic optimization module starts to execute the original power control scheme when the current moment is the starting moment of the current time window, controls the active power and reactive power of the target power system, and replaces the power control scheme being executed with the revised power control scheme when the revised power control scheme output by the source-grid-load-storage control model is obtained; Wherein, the source-grid-load-storage control model is used to correct the power control scheme of active power in the power control scheme being executed based on the actual value of the second power data at the current moment and the predicted value of the first power data in the current time window when the current moment is the first node moment in the current time window; the source-grid-load-storage control model is also used to correct the power control scheme of reactive power in the power control scheme being executed based on the actual value of the second power data at the current moment and the actual value of the electrical quantity data of the target power system when the current moment is the second node moment in the current time window and it is determined that there is an operational safety risk in the target power system at the current moment; The first power data includes the power generation power of the distributed power source nodes and the load power of each load node in the target power system; the second power data includes the power data of each node in the target power system; the first time interval between any two adjacent first node moments in the current time window is the first time interval; the second time interval between any two adjacent second node moments in the current time window is the second time interval; the first time interval is greater than the second time interval; the active power in the target power system is a continuous variable, and the reactive power in the target power system is a discrete variable.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the source-grid-load-storage coordinated control method as described in any one of claims 1 to 7 is implemented.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the source-grid-load-storage coordinated control method as described in any one of claims 1 to 7 is implemented.

Images