Power system flow information prediction method, device and computer equipment considering carbon emissions and new energy output uncertainty

Abstract

The present application relates to a method, device and computer equipment for predicting power system flow information that takes into account the uncertainty of carbon emissions and new energy output, and relates to the field of power technology. The method includes: inputting the potential output distribution of the power system into a generative model to obtain the generative output distribution of the power system, and obtaining the joint output distribution of the power system based on the historical output distribution and the generative output distribution of the power system; constructing a sphere that characterizes the uncertainty output boundary of the power system based on the potential output distribution and the joint output distribution; determining the optimal radius of the sphere, and converting the uncertainty output boundary into a deterministic output boundary by approximating the sphere under the optimal radius through a cube; solving the corresponding multi-objective flow information prediction model containing carbon emissions with the deterministic output boundary as a constraint condition, and obtaining the flow information prediction result of the power system. The use of this method can improve the reliability of the power system flow information prediction.

Description

Technical Field

[0001] The present application relates to the field of electric power technology, and in particular to a method, apparatus, computer equipment, computer-readable storage medium, and computer program product for predicting power system flow information taking into account carbon emissions and uncertainty in new energy output. Background Art

[0002] In related technologies, power system planning, such as output scheduling, typically relies on forecasts of power system flow information. However, with the large-scale integration of renewable energy, the uncertainty of renewable energy output has further led to uncertainty in power system output, which in turn reduces the reliability of power system flow information forecasts. Summary of the Invention

[0003] Based on this, it is necessary to address the technical problem of low reliability of the above-mentioned power system flow information prediction and provide a power system flow information prediction method, device and computer equipment that can improve the reliability of the power system flow information prediction and take into account the uncertainty of carbon emissions and new energy output.

[0004] In a first aspect, the present application provides a method for predicting power system flow information taking into account the uncertainty of carbon emissions and renewable energy output, including:

[0005] Inputting a potential output distribution of a power system into a generative model to obtain a generative output distribution of the power system, and obtaining a joint output distribution of the power system based on a historical output distribution of the power system and the generative output distribution;

[0006] Based on the potential output distribution and the combined output distribution, a sphere is constructed for representing an uncertainty output boundary of the power system; the uncertainty output boundary of the power system is used to represent the output uncertainty of the power system, and the output uncertainty of the power system is associated with the uncertainty of the output of renewable energy sources of the power system;

[0007] Determine the optimal radius of the sphere, and convert the uncertain output boundary into a deterministic output boundary by using a cube to approximate the sphere under the optimal radius;

[0008] With the deterministic output boundary as a constraint condition, a power flow information prediction model with system loss, economic cost and carbon emission of the power system as multiple objectives is solved to obtain a power flow information prediction result of the power system.

[0009] In one embodiment, determining the optimal radius of the sphere includes:

[0010] Initializing a hyperparameter upper limit and a hyperparameter lower limit of a hyperparameter of a support distance of the sphere;

[0011] Determining an optimal hyperparameter between the hyperparameter upper limit and the hyperparameter lower limit;

[0012] Obtaining an optimal support distance for the sphere based on the optimal hyperparameter;

[0013] The optimal radius is obtained based on the optimal support distance.

[0014] In one embodiment, determining the optimal hyperparameter between the hyperparameter upper limit and the hyperparameter lower limit includes:

[0015] Determining a hyperparameter median value of the hyperparameter based on the hyperparameter upper limit and the hyperparameter lower limit;

[0016] Determine a first support distance corresponding to the upper limit of the hyperparameter, a second support distance corresponding to the median of the hyperparameter, and a third support distance corresponding to the lower limit of the hyperparameter respectively;

[0017] Based on the numerical comparison result among the first support distance, the second support distance, and the third support distance, updating the hyperparameter upper limit and / or the hyperparameter lower limit to obtain a new hyperparameter upper limit and / or a new hyperparameter lower limit;

[0018] When the hyperparameter difference between the current hyperparameter upper limit and the current hyperparameter lower limit is greater than the preset difference, return to the step of determining the hyperparameter median of the hyperparameter based on the hyperparameter upper limit and the hyperparameter lower limit, until the obtained hyperparameter difference is less than or equal to the preset difference, and determine the corresponding hyperparameter median as the optimal hyperparameter.

[0019] In one embodiment, updating the hyperparameter upper limit and / or the hyperparameter lower limit based on the numerical comparison result between the first support distance, the second support distance, and the third support distance to obtain a new hyperparameter upper limit and / or a new hyperparameter lower limit includes:

[0020] If the second support distance is less than the third support distance, determining the average of the hyperparameter lower limit and the hyperparameter median as the new hyperparameter lower limit, and determining the average of the hyperparameter upper limit and the hyperparameter median as the new hyperparameter upper limit;

[0021] If the second support distance is greater than or equal to the third support distance and less than or equal to the first support distance, the median of the hyperparameter is used as the new upper limit of the hyperparameter;

[0022] If the second support distance is greater than the first support distance, the median of the hyperparameter is used as the new lower limit of the hyperparameter.

[0023] In one embodiment, the step of approximating a sphere with an optimal radius by using a cube includes:

[0024] Initialize the upper limit and lower limit of the side length of the cube;

[0025] According to the optimal radius, the optimal side length of the cube is determined between the upper limit and the lower limit of the side length to approximate a sphere under the optimal radius.

[0026] In one embodiment, determining the optimal side length of the cube between the upper side length limit and the lower side length limit according to the optimal radius includes:

[0027] Determine a median length of the side length based on the side length upper limit and the side length lower limit;

[0028] Determining an acceptable probability of the median side length according to the optimal radius and the median side length;

[0029] Based on the numerical comparison result between the acceptable probability and the preset acceptable probability, updating the side length upper limit or the side length lower limit to obtain a new side length upper limit or a new side length lower limit;

[0030] When the side length difference between the current upper limit of the side length and the current lower limit of the side length is greater than the preset difference, return to the step of determining the median of the side length based on the upper limit of the side length and the lower limit of the side length, until the obtained side length difference is less than or equal to the preset difference, and determine the corresponding median of the side length as the optimal side length.

[0031] In one embodiment, updating the side length upper limit or the side length lower limit based on the numerical comparison result between the acceptable probability and the preset acceptable probability to obtain the new side length upper limit or the new side length lower limit includes:

[0032] If the acceptable probability is greater than the preset acceptable probability, the median of the side length is determined as the new lower limit of the side length;

[0033] If the acceptable probability is less than or equal to the preset acceptable probability, the median of the side length is determined as the new upper limit of the side length.

[0034] In one embodiment, obtaining the combined output distribution of the power system based on the historical output distribution of the power system and the generative output distribution includes:

[0035] Determining a union of the generated output distribution and the historical output distribution;

[0036] The union is determined as the joint output distribution.

[0037] In a second aspect, the present application further provides a power system flow information prediction device that takes into account the uncertainty of carbon emissions and renewable energy output, including:

[0038] a distribution generation module, configured to input a potential output distribution of the power system into a generative model to obtain a generative output distribution of the power system, and obtain a joint output distribution of the power system based on a historical output distribution of the power system and the generative output distribution;

[0039] a sphere construction module for constructing a sphere for representing an uncertain output boundary of the power system based on the potential output distribution and the combined output distribution; the uncertain output boundary of the power system is used to represent the output uncertainty of the power system, and the output uncertainty of the power system is associated with the uncertainty of the output of new energy sources of the power system;

[0040] A boundary conversion module is used to determine the optimal radius of the sphere, and convert the uncertain output boundary into a deterministic output boundary by approximating the sphere under the optimal radius through a cube;

[0041] The information determination module is used to solve the power flow information prediction model with the system loss, economic cost and carbon emission of the power system as multiple objectives with the deterministic output boundary as a constraint condition, and obtain the power flow information prediction result of the power system.

[0042] In a third aspect, the present application further provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0043] Inputting a potential output distribution of a power system into a generative model to obtain a generative output distribution of the power system, and obtaining a joint output distribution of the power system based on a historical output distribution of the power system and the generative output distribution;

[0044] Based on the potential output distribution and the combined output distribution, a sphere is constructed for representing an uncertainty output boundary of the power system; the uncertainty output boundary of the power system is used to represent the output uncertainty of the power system, and the output uncertainty of the power system is associated with the uncertainty of the output of renewable energy sources of the power system;

[0045] Determine the optimal radius of the sphere, and convert the uncertain output boundary into a deterministic output boundary by using a cube to approximate the sphere under the optimal radius;

[0046] With the deterministic output boundary as a constraint condition, a power flow information prediction model with system loss, economic cost and carbon emission of the power system as multiple objectives is solved to obtain a power flow information prediction result of the power system.

[0047] The above-mentioned power system flow information prediction method, device and computer equipment that take into account the uncertainty of carbon emissions and new energy output, first, input the potential output distribution of the power system into the generative model to obtain the generative output distribution of the power system, and obtain the joint output distribution of the power system based on the historical output distribution and the generative output distribution of the power system; then, based on the potential output distribution and the joint output distribution, construct a sphere for characterizing the uncertainty output boundary corresponding to the potential output distribution; the uncertainty output boundary of the power system is used to characterize the output uncertainty of the power system, and the output uncertainty of the power system is associated with the uncertainty of the new energy output of the power system; then, determine the optimal radius of the sphere, and convert the uncertainty output boundary into a deterministic output boundary by approximating the sphere under the optimal radius by cube; finally, with the deterministic output boundary as a constraint condition, solve the flow information prediction model with system loss, economic cost and carbon emissions of the power system as multiple objectives to obtain the flow information prediction result of the power system. In this way, by generating a generative output distribution based on the potential output distribution, a combined output distribution can be obtained by combining the historical output distribution of the power system. Based on the combined output distribution and the potential output distribution, the power system's output distribution expression is enriched, and a sphere is constructed to represent the power system's uncertain output boundary. By solving the radius of the sphere and approximating the sphere with a cube, the uncertain output boundary can be converted into a deterministic output boundary. Based on the deterministic output boundary, a power flow information prediction model with multiple objectives, including system losses, economic costs, and carbon emissions, can be more accurately solved to obtain the power system power flow information prediction results. Based on the above process, the power system power flow information prediction method that takes into account the uncertainty of carbon emissions and renewable energy output can improve the reliability of power system power flow information prediction by considering the uncertainty of carbon emissions and renewable energy output and converting the uncertainty problem into a deterministic problem for solution. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following briefly introduces the drawings required for use in the embodiments of the present application or related technical descriptions. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying any creative work.

[0049] Figure 11 is a flow chart of a method for predicting power system flow information taking into account the uncertainty of carbon emissions and renewable energy output in one embodiment;

[0050] Figure 2 A schematic flow chart of the steps of determining the optimal radius of a sphere in one embodiment;

[0051] Figure 3 1 is a flow chart illustrating a step of determining an optimal hyperparameter between an upper hyperparameter limit and a lower hyperparameter limit in one embodiment;

[0052] Figure 4 1. A flowchart illustrating the steps of updating a hyperparameter upper limit and / or a hyperparameter lower limit based on a numerical comparison result between a first support distance, a second support distance, and a third support distance to obtain a new hyperparameter upper limit and / or a new hyperparameter lower limit in one embodiment;

[0053] Figure 5 A schematic flow chart of steps for approximating a sphere with an optimal radius by using a cube in one embodiment;

[0054] Figure 6 1. A schematic flow chart of a step of determining an optimal side length of a cube between an upper side length limit and a lower side length limit according to an optimal radius in one embodiment;

[0055] Figure 7 1. A flowchart illustrating the steps of updating the upper limit or lower limit of the side length based on the numerical comparison result between the acceptable probability and the preset acceptable probability to obtain a new upper limit or new lower limit of the side length in one embodiment;

[0056] Figure 8 A schematic flow chart of the steps of determining the optimal radius of the sphere and approximating the sphere with the optimal radius by using a cube in one embodiment;

[0057] Figure 9 A flowchart of a power system uncertainty optimization method based on generative artificial intelligence and Wasserstein distance (bulldozer distance) in one embodiment is provided;

[0058] Figure 10 A structural block diagram of a power system power flow information prediction device taking into account carbon emissions and new energy output uncertainty in one embodiment;

[0059] Figure 11 FIG. 1 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0060] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0061] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0062] In one embodiment, Figure 1 As shown, a method for predicting power system flow information that takes into account the uncertainty of carbon emissions and renewable energy output is provided. This embodiment uses the method applied to a server as an example for illustration. It is understandable that the method can also be applied to a terminal, and can also be applied to a system including a server and a terminal, and implemented through the interaction between the server and the terminal; wherein the server can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server providing cloud computing services; the terminal can be, but is not limited to, various personal computers, laptops, smart phones, tablet computers, etc. In this embodiment, the method includes the following steps S102 to S108:

[0063] Step S102 : inputting the potential output distribution of the power system into a generative model to obtain the generative output distribution of the power system, and obtaining the joint output distribution of the power system based on the historical output distribution of the power system and the generative output distribution.

[0064] Among them, the power system includes multiple nodes, which are used to mount units; the units include thermal power units, new energy units and energy storage systems; each node is mounted with at least one type of unit among thermal power units and new energy units; in specific applications, when the node is mounted with a new energy unit, it will also be mounted with an energy storage system.

[0065] The power system output distribution is used to characterize the distribution of the output of each node in the power system. In specific applications, since the output of renewable energy is uncertain, the output of the power system is also uncertain.

[0066] The potential output distribution is an output distribution inferred based on a known empirical distribution, such as the historical output distribution of the power system. In a specific application, there are multiple potential output distributions, forming a potential output distribution set.

[0067] The historical output distribution is the actual output distribution of the power system in the past. In a specific application, there are multiple historical output distributions, which constitute a historical output distribution set.

[0068] A generative model is a machine learning model that learns the joint probability distribution of input data and generates new data similar to the input data. In specific applications, generative models use generative adversarial networks.

[0069] The generative output distribution is the output distribution obtained based on the generative model.

[0070] There are multiple joint output distributions, which constitute a joint output distribution set.

[0071] Specifically, the server infers a potential output distribution based on known historical output distributions. This potential output distribution is then fed into a generative model, such as a generative adversarial network. The generative model learns from the potential output distribution to generate a generated output distribution similar to the potential output distribution. The server then fuses this generated output distribution with the historical output distribution to generate a combined output distribution, enriching the power system's output distribution representation.

[0072] In this embodiment, before inputting the potential output distribution of the power system into the generative model to obtain the generative output distribution of the power system in step S102, and obtaining the combined output distribution of the power system based on the historical output distribution of the power system and the generative output distribution, the server may further construct a power flow information prediction model with multiple objectives, such as system loss, economic cost, and carbon emissions of the power system, as shown in Formula 1:

[0073] (Formula 1)

[0074] in, is the potential output distribution set, taken from the sample space , by random probability definition; is the system loss function of the power system, is the optimal solution of the system loss function; is the economic cost function of the power system, is the optimal solution of the economic cost function; is the carbon emission function of the power system, is the optimal solution of the carbon emission function; 、 、 Obtained by optimizing the single-objective problem; is a parameter vector, which is used to represent the weights of system loss, economic cost and carbon emission.

[0075] in, The variables to be solved in the power flow information prediction model include the active power injected into the nodes of the power system. , node injected reactive power , node phase angle and node voltage amplitude ; Further, the node injects active power Contains the injected active power of each node and the injected reactive power of the node Contains the injected reactive power of each node, node phase angle Contains the phase angle of each node, node voltage amplitude Contains the voltage magnitude at each node.

[0076] Formula 2 shows the constraints of the power flow information prediction model shown in Formula 1:

[0077] (Formula 2)

[0078] Among them, the node set of the power system is recorded as , the node index is recorded as ; For nodes Uncertain output power (because the output of new energy is uncertain); 、 They are node active load and node reactive load respectively (further, node active load Contains active load of each node and reactive load of the node Including reactive load of each node); is a vector of all 1s; is the node conductance matrix, is the node susceptance matrix; is the matrix element product operation; 、 、 Node The coefficients of the quadratic term, linear term and constant term under the economic cost function; 、 are the upper and lower limits of the node phase angle respectively; 、 are the upper and lower limits of the node voltage amplitude respectively; 、 Node The upper and lower limits of active power injected by the node; For nodes Uncertain node carbon emission intensity (because new energy output is uncertain); The preset acceptable probability of violating the uncertain output boundary of the power system.

[0079] in, Indicates the uncertainty output boundary of the power system. In specific applications, the uncertainty output boundary of the power system is used to characterize the output uncertainty of the power system. The output uncertainty of the power system mainly comes from the uncertainty of the output of renewable energy in the power system.

[0080] Furthermore, the uncertainty of a node's carbon emission intensity is related to the carbon emission intensity of the unit mounted on the node and the unit's output power. The carbon emission intensities of thermal power units, new energy units, and energy storage systems are all fixed values, and the output power of thermal power units is deterministic. Therefore, the uncertainty of the carbon emission intensity of an uncertain node mainly comes from the uncertain output power of the new energy units and energy storage systems. The uncertainty of the carbon emission intensity of an uncertain node is shown in Formula 3:

[0081] (Formula 3)

[0082] in, 、 They are the fixed carbon emission intensity of new energy units and the fixed carbon emission intensity of energy storage systems; 、 They are the uncertainty output power of the new energy unit and the uncertainty output power of the energy storage system respectively.

[0083] Furthermore, the uncertainty of the node's output power is expressed as follows:

[0084] (Formula 4)

[0085] Step S104 : constructing a sphere for representing the uncertain output boundary of the power system based on the potential output distribution and the combined output distribution.

[0086] The sphere is used to cover the potential output distribution and the combined output distribution. In a specific application, the sphere is a Wasserstein sphere.

[0087] The sphere has a radius to be solved and a support distance to be solved. The support distance is related to the radius, so the server can solve the radius by solving the support distance. Furthermore, the support distance has a corresponding hyperparameter to be solved, so the server can solve the hyperparameter to solve the support distance.

[0088] Here, constructing a sphere refers to an expression for constructing a sphere.

[0089] Among them, the uncertainty output boundary of the power system is used to characterize the output uncertainty of the power system, and the output uncertainty of the power system is associated with the output uncertainty of the new energy of the power system.

[0090] Specifically, the server constructs an expression of a sphere that can cover the potential output distribution and the combined output distribution; since the sphere covers the potential output distribution and the combined output distribution, the expression of the sphere can be used to characterize the uncertain output boundary of the power system.

[0091] In specific applications, the potential output distribution set is recorded as , the combined output distribution set is , the historical output distribution set is , the generated output distribution set is , the server constructs the Wasserstein sphere as shown in Equation 5 and Equation 6 :

[0092] (Formula 5)

[0093] (Formula 6)

[0094] Among them, the potential output distribution set Taken from the sample space , by random probability definition; is the distribution difference between the potential output distribution set and the joint output distribution set, also known as the Wasserstein sphere radius; is the Wasserstein metric calculation model. Combining Formula 5 and Formula 6, we can see that the Wasserstein distance between the potential output distribution and the joint output distribution is less than or equal to .

[0095] Furthermore, the server can approximate the radius using the following formula 7: :

[0096] (Formula 7)

[0097] in, Wasserstein sphere Support distance; is the number of joint output distributions selected from the joint output distributions; is the confidence interval; Support distance Hyperparameters of For the Generative output distribution; is the mean of the generative output distribution.

[0098] In practical applications, the support distance and hyperparameters It can be solved by the bisection search method (BSM).

[0099] Step S106 , determining the optimal radius of the sphere, and converting the uncertain output boundary into a deterministic output boundary by using a cube to approximate the sphere under the optimal radius.

[0100] Specifically, the server uses a binary search method to determine the optimal hyperparameters and optimal support distance of the sphere, and substitutes them into Formula 7 to obtain the optimal radius of the sphere. Then, the server converts the uncertain output boundary under the Wasserstein sphere into an optimal approximation problem of a cube through covariance variation. As shown in Formula 8, the server converts the uncertain output boundary into a deterministic output boundary by solving the optimal side length of the cube.

[0101] (Formula 8)

[0102] in, is the side length of the cube.

[0103] The constraints of the cube are shown in Equations 9 and 10:

[0104] (Formula 9)

[0105] (Formula 10)

[0106] in, is the scaling ratio; is the acceptable probability of side length under scaling; is the robust adjustment term of the Wasserstein constraint, Ensure that the side length When the time is too small, no meaningless negative value will appear; is the covariance of the generated output distribution.

[0107] Refer to Formula 8 to Formula 10, The uncertainty problem covered by the constructed cube is the uncertainty output boundary of the power system, that is, a The four vertices of a cube can determine the volume of the cube. The vertex positions of the cube are the uncertainty output boundary. Since the cube is transformed by Formula 10, the maximum constraint boundary of the original uncertainty output boundary can be obtained through the inverse transformation of Formula 10 (multiplying by the 1 / 2 power of the covariance plus the mean), thereby converting the uncertainty output boundary into a deterministic output boundary.

[0108] In a specific application, after the server obtains the optimal radius, it substitutes the optimal radius into Formula 9 and Formula 10 to calculate the optimal side length, thereby solving the optimal approximation problem and completing the conversion shown in Formula 8.

[0109] In step S108 , a power flow information prediction model with multiple objectives, including system loss, economic cost, and carbon emission of the power system, is solved with the deterministic output boundary as a constraint condition to obtain a power flow information prediction result of the power system.

[0110] The power flow information prediction results include at least the variables to be solved in formula 1: .

[0111] Specifically, the server substitutes the uncertainty boundary shown in Formula 8 into Formula 2, and under the constraints of Formula 2, solves the power flow information prediction model shown in Formula 1 to obtain the power flow information prediction result of the power system; the server can plan the power system according to the power flow information prediction result, such as output scheduling.

[0112] In specific applications, the server iteratively solves the power flow information prediction model using the interior point method.

[0113] In the above-mentioned power system flow information prediction method that takes into account the uncertainty of carbon emissions and new energy output, first, the server inputs the potential output distribution of the power system into the generative model to obtain the generative output distribution of the power system, and obtains the joint output distribution of the power system based on the historical output distribution and the generative output distribution of the power system; then, the server constructs a sphere based on the potential output distribution and the joint output distribution to characterize the uncertainty output boundary corresponding to the potential output distribution; the uncertainty output boundary of the power system is used to characterize the output uncertainty of the power system, and the output uncertainty of the power system is associated with the uncertainty of the new energy output of the power system; then, the server determines the optimal radius of the sphere, and converts the uncertainty output boundary into a deterministic output boundary by approximating the sphere under the optimal radius by using a cube; finally, the server solves the flow information prediction model with multiple objectives of system loss, economic cost and carbon emissions of the power system under the deterministic output boundary, and obtains the flow information prediction result of the power system. In this way, by generating a generative output distribution based on the potential output distribution, the server can combine the historical output distribution of the power system to obtain a joint output distribution. Based on the joint output distribution and the potential output distribution, the power system's output distribution expression is enriched, and a sphere is constructed to represent the power system's uncertain output boundary. By solving the radius of the sphere and approximating the sphere with a cube, the server can convert the uncertain output boundary into a deterministic output boundary. Based on the deterministic output boundary, the server can more accurately solve the power flow information prediction model with multiple objectives such as system losses, economic costs, and carbon emissions of the power system, and obtain the power flow information prediction results of the power system. Based on the above process, the power system power flow information prediction method that takes into account the uncertainty of carbon emissions and renewable energy output can improve the reliability of power system power flow information prediction by considering the uncertainty of carbon emissions and renewable energy output and converting the uncertainty problem into a deterministic problem for solution.

[0114] In an exemplary embodiment, Figure 2 As shown, in the above step S106, determining the optimal radius of the sphere specifically includes the following steps:

[0115] Step S202 : Initialize the upper limit and lower limit of the hyperparameter of the support distance of the sphere.

[0116] Step S204: Determine the optimal hyperparameter between the upper limit and the lower limit of the hyperparameter.

[0117] Step S206: Obtain the optimal support distance of the sphere based on the optimal hyperparameters.

[0118] Step S208: Obtaining an optimal radius based on the optimal support distance.

[0119] Specifically, the server first initializes the hyperparameters The upper limit of hyperparameters and hyperparameter lower bounds ; Then, the server uses the binary search method to find the upper limit of the hyperparameter and hyperparameter lower bounds Determine the optimal hyperparameters , and the optimal hyperparameters Substitute into formula 7 to obtain the optimal support distance , and then find the optimal radius .

[0120] In this embodiment, the server can optimize the hyperparameters through the binary search method and Formula 7, and can further obtain the optimal support distance and optimal radius through the optimal hyperparameters, thereby determining the radius and support distance of the sphere.

[0121] In an exemplary embodiment, Figure 3 As shown, the above step S204, determining the optimal hyperparameter between the upper limit of the hyperparameter and the lower limit of the hyperparameter, specifically includes the following steps:

[0122] Step S302: Determine a hyperparameter median value of the hyperparameter based on the hyperparameter upper limit and the hyperparameter lower limit.

[0123] Step S304 : Determine the first support distance corresponding to the upper limit of the hyperparameter, the second support distance corresponding to the median of the hyperparameter, and the third support distance corresponding to the lower limit of the hyperparameter.

[0124] Step S306: Based on the numerical comparison results among the first support distance, the second support distance, and the third support distance, update the hyperparameter upper limit and / or the hyperparameter lower limit to obtain a new hyperparameter upper limit and / or a new hyperparameter lower limit.

[0125] Step S308: When the hyperparameter difference between the current hyperparameter upper limit and the current hyperparameter lower limit is greater than the preset difference, return to the step of determining the hyperparameter median of the hyperparameter based on the hyperparameter upper limit and the hyperparameter lower limit, until the obtained hyperparameter difference is less than or equal to the preset difference, and determine the corresponding hyperparameter median as the optimal hyperparameter.

[0126] Specifically, the server first calculates the upper limit of the hyperparameters and hyperparameter lower bounds The average value of the hyperparameter The median value of the hyperparameter .

[0127] The server then sets the hyperparameter upper limit Substitute into formula 7 and the calculated support distance is recorded as the first support distance , similarly, the server will set the median value of the hyperparameter Substitute into formula 7 and record the calculated support distance as the second support distance , the server sets the lower limit of the hyperparameter Substitute into formula 7 and the calculated support distance is recorded as the third support distance .

[0128] Next, the server compares the first support distance , Second support distance , the third support distance The numerical size between the three, and based on the numerical comparison results between the three, update the upper limit of the hyperparameter and hyperparameter lower bounds At least one of them, get a new upper limit for the hyperparameter and / or lower bounds on hyperparameters , to update the optimization range of hyperparameters.

[0129] Then, the server's hyperparameter difference between the current hyperparameter upper limit and the current hyperparameter lower limit is greater than the preset difference In the case of , returns the upper limit of the calculated hyperparameters and hyperparameter lower bounds The average value of the hyperparameter The median value of the hyperparameter Repeat the above steps until the hyperparameter difference is less than or equal to the preset difference , indicating that the hyperparameters have converged at this time, and the server will (That is, the corresponding hyperparameter difference is less than or equal to the preset difference The median value of the hyperparameter corresponding to the upper and lower limits of the hyperparameters when is determined as the optimal hyperparameter .

[0130] It should be noted that the server is based on the first support distance , Second support distance , the third support distance The numerical comparison results among the three can be used to update only the upper limit of the hyperparameter , you can only update the lower limit of the hyperparameter , you can also update the upper limit of hyperparameters at the same time and hyperparameter lower bounds .

[0131] In this embodiment, the server can continuously update the optimization range of the hyperparameters by updating the upper and lower limits of the hyperparameters, so that the optimal hyperparameters can be obtained when the hyperparameters converge.

[0132] In an exemplary embodiment, Figure 4As shown, the above step 306, based on the numerical comparison results between the first support distance, the second support distance, and the third support distance, updates the hyperparameter upper limit and / or the hyperparameter lower limit to obtain a new hyperparameter upper limit and / or a new hyperparameter lower limit, specifically includes the following steps:

[0133] Step S402: If the second support distance is less than the third support distance, the average of the hyperparameter lower limit and the hyperparameter median is determined as the new hyperparameter lower limit, and the average of the hyperparameter upper limit and the hyperparameter median is determined as the new hyperparameter upper limit.

[0134] Step S404: If the second support distance is greater than or equal to the third support distance and less than or equal to the first support distance, the median of the hyperparameter is used as a new upper limit of the hyperparameter.

[0135] Step S406: If the second support distance is greater than the first support distance, the median of the hyperparameter is used as the new lower limit of the hyperparameter.

[0136] Specifically, if the second support distance Less than the third support distance , the server will lower the hyperparameter and the median hyperparameter Average value Determine as the new lower limit of the hyperparameter and set the upper limit of the hyperparameter and the median hyperparameter Average value Determine as the new upper limit of the hyperparameter, that is, the optimization range of the hyperparameter is from Reduce to .

[0137] If the second support distance Greater than or equal to the third support distance , and is less than or equal to the first support distance , the server will set the median value of the hyperparameter As the new upper limit of the hyperparameter, the lower limit of the hyperparameter is not updated at the same time, that is, the optimization range of the hyperparameter is from Reduce to .

[0138] If the second support distance Greater than the first support distance , the server will set the median value of the hyperparameter As the new lower limit of the hyperparameter, the upper limit of the hyperparameter is not updated at the same time, that is, the optimization range of the hyperparameter is from Reduce to .

[0139] In this embodiment, the server can continuously narrow the optimization range of the hyperparameters by updating the upper and lower limits of the hyperparameters, and thus solve the optimal hyperparameters.

[0140] In an exemplary embodiment, Figure 5 As shown, in the above step S106, the sphere with the optimal radius is approximated by a cube, which specifically includes the following steps:

[0141] Step S502 : Initialize the upper limit and lower limit of the side length of the cube.

[0142] Step S504 : determining the optimal side length of the cube between the upper limit and the lower limit of the side length according to the optimal radius, so as to approximate a sphere under the optimal radius.

[0143] Specifically, the server first initializes the edge length The upper limit of the side length and lower limit of side length ; Then, the server uses the binary search method to find the upper limit of the edge length and lower limit of side length Determine the optimal side length between , and the optimal side length Substitute formula 8 into formula 10 to approximate the optimal radius The sphere below.

[0144] In this embodiment, the server can optimize the hyperparameters through the binary search method and Formulas 8 to 10, and can further obtain the optimal support distance and optimal radius through the optimal hyperparameters, thereby determining the radius and support distance of the sphere.

[0145] In an exemplary embodiment, Figure 6 As shown, the above step S504, based on the optimal radius, determines the optimal side length of the cube between the upper limit and the lower limit of the side length, specifically includes the following steps:

[0146] Step S602: Determine the median value of the side length based on the upper limit and the lower limit of the side length.

[0147] Step S604: Determine the acceptability probability of the median side length based on the optimal radius and the median side length.

[0148] Step S606 : Based on the numerical comparison result between the acceptable probability and the preset acceptable probability, the upper limit or the lower limit of the side length is updated to obtain a new upper limit or a new lower limit of the side length.

[0149] Step S608, when the side length difference between the current side length upper limit and the current side length lower limit is greater than the preset difference, return to the step of determining the side length median based on the side length upper limit and the side length lower limit, until the obtained side length difference is less than or equal to the preset difference, and determine the corresponding side length median as the optimal side length.

[0150] Specifically, the server first calculates the upper limit of the edge length and lower limit of side length The average value of the side length The median side length .

[0151] The server then calculates the optimal radius and median side length Substitute into formula 9 and formula 10 to calculate the median side length The acceptable probability .

[0152] Next, the server compares the acceptable probabilities and the preset acceptable probability The numerical value between the two is compared, and based on the numerical comparison result between the two, the upper limit of the side length is updated. or lower limit of side length At least one of them, get a new upper limit on the side length or the new lower limit of the side length , to update the optimal range of edge length.

[0153] Then, the server is at the current edge length limit and the current lower limit of side length The difference in side lengths between the two is greater than the preset difference In the case of , return the upper limit of the calculated side length and lower limit of side length The average value of the side length The median side length Repeat the above steps until the obtained side length difference is less than or equal to the preset difference , indicating that the side length converges at this time, and the server will be the median of the side length at this time (That is, the corresponding side length difference is less than or equal to the preset difference The median value of the side length corresponding to the upper limit and lower limit of the side length when the optimal side length is determined .

[0154] In this embodiment, the server can continuously update the optimization range of the side length by updating the upper and lower limits of the side length, so as to obtain the optimal side length when the side length converges.

[0155] In an exemplary embodiment, Figure 7As shown, the above step S604, based on the numerical comparison result between the acceptable probability and the preset acceptable probability, updates the upper limit or the lower limit of the side length to obtain a new upper limit or a new lower limit of the side length, specifically includes the following steps:

[0156] Step S702: If the acceptable probability is greater than the preset acceptable probability, the median of the side length is determined as a new lower limit of the side length.

[0157] Step S704: If the acceptable probability is less than or equal to the preset acceptable probability, the median of the side length is determined as a new upper limit of the side length.

[0158] Specifically, if the probability is acceptable Greater than the preset acceptable probability , then the server will be the median of the side length Determine as the new lower limit of the edge length, and do not update the upper limit of the hyperparameter. That is, the optimization range of the edge length is from Reduce to .

[0159] If the probability is acceptable Less than or equal to the preset acceptable probability , then the server will be the median of the side length Determine as the new upper limit of the edge length, and do not update the lower limit of the hyperparameter, that is, the optimization range of the edge length is from Reduce to .

[0160] In this embodiment, the server can continuously narrow the optimization range of the side length by updating the upper and lower limits of the side length, and then solve the optimal side length.

[0161] In an exemplary embodiment, Figure 8 As shown, the above step S106, determining the optimal radius of the sphere and approximating the sphere under the optimal radius by using a cube, specifically includes the following steps:

[0162] Initializing hyperparameter upper bounds , hyperparameter lower limit , Upper limit of side length and lower limit of side length .

[0163] Calculating upper limits for hyperparameters and hyperparameter lower bounds The average value of the hyperparameter The median value of the hyperparameter .

[0164] Upper limit of hyperparameters Substitute into formula 7 and the calculated support distance is recorded as the first support distance , similarly, the server will set the median value of the hyperparameter Substitute into formula 7 and record the calculated support distance as the second support distance , the server sets the lower limit of the hyperparameter Substitute into formula 7 and the calculated support distance is recorded as the third support distance .

[0165] If the second support distance Less than the third support distance , the server will lower the hyperparameter and the median hyperparameter Average value Determine as the new lower limit of the hyperparameter and set the upper limit of the hyperparameter and the median hyperparameter Average value Determine as the new upper limit of the hyperparameter, that is, the optimization range of the hyperparameter is from Reduce to If the second support distance Greater than or equal to the third support distance , and is less than or equal to the first support distance , the server will set the median value of the hyperparameter As the new upper limit of the hyperparameter, the lower limit of the hyperparameter is not updated at the same time, that is, the optimization range of the hyperparameter is from Reduce to If the second support distance Greater than the first support distance , the server will set the median value of the hyperparameter As the new lower limit of the hyperparameter, the upper limit of the hyperparameter is not updated at the same time, that is, the optimization range of the hyperparameter is from Reduce to .

[0166] The hyperparameter difference between the current hyperparameter upper limit and the current hyperparameter lower limit is greater than the preset difference In the case of , returns the upper limit of the calculated hyperparameters and hyperparameter lower bounds The average value of the hyperparameter The median value of the hyperparameter until the hyperparameter difference is less than or equal to the preset difference , the median value of the hyperparameter at this time Determined as the optimal hyperparameter , the median value of the hyperparameter at this time Corresponding second support distance Determine the optimal support distance .

[0167] The optimal hyperparameters , optimal support distance and confidence intervals Substitute into formula 7 to calculate the optimal radius .

[0168] Calculate the upper limit of side length and lower limit of side length The average value of the side length The median side length .

[0169] The optimal radius and median side length Substitute into formula 9 and formula 10 to calculate the median side length The acceptable probability .

[0170] If the probability is acceptable Greater than the preset acceptable probability , then the server will be the median of the side length Determine as the new lower limit of the edge length, and do not update the upper limit of the hyperparameter. That is, the optimization range of the edge length is from Reduce to If the probability is acceptable Less than or equal to the preset acceptable probability , then the server will be the median of the side length Determine as the new upper limit of the edge length, and do not update the lower limit of the hyperparameter, that is, the optimization range of the edge length is from Reduce to .

[0171] At the current side length limit and the current lower limit of side length The difference in side lengths between the two is greater than the preset difference In the case of , return the upper limit of the calculated side length and lower limit of side length The average value of the side length The median side length until the obtained side length difference is less than or equal to the preset difference , the median side length at this time Determine the optimal side length .

[0172] At this time, by The uncertainty problem covered by the constructed cube is the uncertainty output boundary of the power system, that is, a The four vertices of a cube can determine the volume of the cube. The vertex positions of the cube are the uncertainty output boundary. Since the cube is transformed by Formula 10, the maximum constraint boundary of the original uncertainty output boundary can be obtained through the inverse transformation of Formula 10 (multiplying by the 1 / 2 power of the covariance plus the mean), thereby converting the uncertainty output boundary into a deterministic output boundary.

[0173] In an exemplary embodiment, in the above step 102, based on the historical output distribution and the generative output distribution of the power system, a joint output distribution of the power system is obtained, which specifically includes the following steps: determining the union of the generative output distribution and the historical output distribution; and determining the union as the joint output distribution.

[0174] Specifically, referring to Formula 5, the server takes the union of the generated output distribution and the historical output distribution, and uses the union as the joint output distribution of the power system.

[0175] In this embodiment, the server enriches the output distribution expression of the power system based on the generated output distribution and the historical output distribution.

[0176] In order to more clearly illustrate the power system flow information prediction method taking into account the uncertainty of carbon emissions and new energy output provided by the embodiment of the present application, the power system flow information prediction method taking into account the uncertainty of carbon emissions and new energy output is specifically described below with a specific embodiment, but it should be understood that the embodiment of the present application is not limited to this. Figure 9 As shown, in one exemplary embodiment, the present application also provides a power system uncertainty optimization method based on generative artificial intelligence and Wasserstein distance, which specifically includes the following steps:

[0177] 1. Taking the system loss, economic cost and carbon emissions of the power system as multiple objectives, a power flow information prediction model including the uncertain output of the power system is constructed.

[0178] Specifically, the server constructs a power flow information prediction model as shown in Formulas 1 to 4.

[0179] 2. Based on generative adversarial networks and Wasserstein distance, a distributed robust optimization model for power systems is constructed.

[0180] Specifically, the server constructs a distributed robust optimization model as shown in Formulas 5 to 7 based on the generative adversarial network and Wasserstein distance, and uses the Wasserstein sphere to describe the uncertainty output boundary in the power flow information prediction model.

[0181] 3. The uncertainty boundary of the distributed robust optimization model is solved by the binary search method, and the uncertainty output boundary in the power flow information prediction model is converted into a deterministic output boundary.

[0182] Specifically, referring to Formulas 8 to 10, the server uses a cube to approximate the Wasserstein sphere, and uses a binary search method to solve the radius and support distance of the Wasserstein sphere, and then obtains the side length of the cube, thereby obtaining the maximum constraint boundary of the original uncertain output boundary and converting the uncertain output boundary into a deterministic output boundary.

[0183] 4. Build a multi-objective time series simulation calculation model and perform iterative solutions to the flow information prediction model.

[0184] In this embodiment, by modeling and solving the distributed robust optimization model containing carbon emissions and renewable energy output uncertainty, it can be used for comprehensive analysis of power planning.

[0185] It should be understood that, although the various steps in the flowcharts involved in the various embodiments described above are displayed in sequence according to the instructions of the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and these steps can be executed in other orders. Moreover, at least a portion of the steps in the flowcharts involved in the various embodiments described above can include multiple steps or multiple stages, and these steps or stages are not necessarily executed and completed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a portion of steps or stages in other steps.

[0186] Based on the same inventive concept, the embodiments of the present application also provide a device for predicting power system flow information taking into account the uncertainty of carbon emissions and renewable energy output, which is used to implement the above-mentioned method for predicting power system flow information taking into account the uncertainty of carbon emissions and renewable energy output. The implementation solution provided by this device is similar to the implementation solution described in the above-mentioned method. Therefore, the specific limitations of one or more embodiments of the device for predicting power system flow information taking into account the uncertainty of carbon emissions and renewable energy output provided below can be found in the above-mentioned limitations of the method for predicting power system flow information taking into account the uncertainty of carbon emissions and renewable energy output, and will not be repeated here.

[0187] In an exemplary embodiment, Figure 10As shown, a power system flow information prediction device taking into account the uncertainty of carbon emissions and new energy output is provided, including: a distribution generation module 1002, a sphere construction module 1004, a boundary conversion module 1006 and an information determination module 1008, wherein:

[0188] The distribution generation module 1002 is used to input the potential output distribution of the power system into the generative model to obtain the generative output distribution of the power system, and obtain the joint output distribution of the power system based on the historical output distribution of the power system and the generative output distribution.

[0189] The sphere construction module 1004 is used to construct a sphere for representing the uncertainty output boundary of the power system based on the potential output distribution and the combined output distribution; the uncertainty output boundary of the power system is used to represent the output uncertainty of the power system, and the output uncertainty of the power system is associated with the uncertainty of the new energy output of the power system.

[0190] The boundary conversion module 1006 is used to determine the optimal radius of the sphere and convert the uncertain output boundary into a deterministic output boundary by approximating the sphere under the optimal radius with a cube.

[0191] The information determination module 1008 is used to solve the power flow information prediction model with the system loss, economic cost and carbon emission of the power system as multiple objectives under the deterministic output boundary, and obtain the power flow information prediction result of the power system.

[0192] In an exemplary embodiment, the boundary conversion module 1006 is further used to initialize the hyperparameter upper limit and hyperparameter lower limit of the hyperparameter of the support distance of the sphere; determine the optimal hyperparameter between the hyperparameter upper limit and the hyperparameter lower limit; obtain the optimal support distance of the sphere based on the optimal hyperparameter; and obtain the optimal radius based on the optimal support distance.

[0193] In an exemplary embodiment, the boundary conversion module 1006 is also used to determine the hyperparameter median of the hyperparameter based on the hyperparameter upper limit and the hyperparameter lower limit; respectively determine the first support distance corresponding to the hyperparameter upper limit, the second support distance corresponding to the hyperparameter median, and the third support distance corresponding to the hyperparameter lower limit; based on the numerical comparison results between the first support distance, the second support distance, and the third support distance, update the hyperparameter upper limit and / or the hyperparameter lower limit to obtain a new hyperparameter upper limit and / or a new hyperparameter lower limit; when the hyperparameter difference between the current hyperparameter upper limit and the current hyperparameter lower limit is greater than the preset difference, return to the step of determining the hyperparameter median of the hyperparameter based on the hyperparameter upper limit and the hyperparameter lower limit, until the obtained hyperparameter difference is less than or equal to the preset difference, and determine the corresponding hyperparameter median as the optimal hyperparameter.

[0194] In an exemplary embodiment, the boundary conversion module 1006 is also used to determine the average of the hyperparameter lower limit and the hyperparameter median as the new hyperparameter lower limit, and the average of the hyperparameter upper limit and the hyperparameter median as the new hyperparameter upper limit if the second support distance is less than the third support distance; if the second support distance is greater than or equal to the third support distance and less than or equal to the first support distance, then the hyperparameter median is used as the new hyperparameter upper limit; if the second support distance is greater than the first support distance, then the hyperparameter median is used as the new hyperparameter lower limit.

[0195] In an exemplary embodiment, the boundary conversion module 1006 is further configured to initialize an upper limit and a lower limit of the side length of the cube; and determine the optimal side length of the cube between the upper limit and the lower limit according to the optimal radius to approximate a sphere under the optimal radius.

[0196] In an exemplary embodiment, the boundary conversion module 1006 is also used to determine the median length of the side length based on the upper limit and lower limit of the side length; determine the acceptable probability of the median length based on the optimal radius and the median length; update the upper limit or lower limit of the side length based on the numerical comparison result between the acceptable probability and the preset acceptable probability to obtain a new upper limit or a new lower limit of the side length; when the side length difference between the current upper limit and the current lower limit is greater than the preset difference, return to the step of determining the median length of the side length based on the upper limit and the lower limit of the side length, until the obtained side length difference is less than or equal to the preset difference, and determine the corresponding median length as the optimal side length.

[0197] In an exemplary embodiment, the boundary conversion module 1006 is further used to determine the median of the side length as the new lower limit of the side length if the acceptable probability is greater than the preset acceptable probability; and to determine the median of the side length as the new upper limit of the side length if the acceptable probability is less than or equal to the preset acceptable probability.

[0198] In an exemplary embodiment, the distribution generation module 1002 is further configured to determine a union of the generated output distribution and the historical output distribution; and determine the union as the joint output distribution.

[0199] Each module in the aforementioned power system power flow information prediction device that takes into account carbon emissions and renewable energy output uncertainty can be implemented in whole or in part through software, hardware, or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in hardware form, or can be stored in a computer device memory in software form, so that the processor can call and execute the corresponding operations of each module.

[0200] In an exemplary embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as shown in FIG. Figure 11As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, abbreviated as I / O) and a communication interface. The processor, memory and input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The database of the computer device is used to store historical output distribution data of the power system. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a method for predicting power system flow information taking into account the uncertainty of carbon emissions and new energy output is implemented.

[0201] Those skilled in the art will understand that Figure 11 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0202] In an exemplary embodiment, a computer device is further provided, including a memory and a processor. The memory stores a computer program, and the processor implements the steps in the above method embodiments when executing the computer program.

[0203] In an exemplary embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0204] In an exemplary embodiment, a computer program product is provided, including a computer program. When the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0205] Those skilled in the art will understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. In particular, any reference to memory, database, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The databases involved in the various embodiments provided herein may include at least one of a relational database and a non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the various embodiments provided herein may be, but are not limited to, general-purpose processors, central processing units (CPUs), graphics processing units (GPUs), digital signal processors (DSPs), programmable logic devices (PLDs), quantum computing-based data processing logic devices, artificial intelligence (AI) processors, and the like.

[0206] The technical features of the above embodiments can be combined arbitrarily. In order to make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0207] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present application. It should be noted that a person of ordinary skill in the art may make various modifications and improvements without departing from the spirit of the present application, and these modifications and improvements fall within the scope of protection of the present application. Therefore, the scope of protection of the present application shall be determined by the appended claims.

Claims

1. A method for predicting power system flow information taking into account the uncertainty of carbon emissions and renewable energy output, characterized in that: The method comprises: Inputting a potential output distribution of a power system into a generative model to obtain a generative output distribution of the power system, and obtaining a joint output distribution of the power system based on a historical output distribution of the power system and the generative output distribution; the generative model includes a generative adversarial network; Based on the potential output distribution and the combined output distribution, a sphere is constructed for representing the uncertainty output boundary of the power system; the uncertainty output boundary of the power system is used to represent the output uncertainty of the power system, and the output uncertainty of the power system is associated with the uncertainty of the renewable energy output of the power system; the sphere is used to cover the potential output distribution and the combined output distribution, and the radius of the sphere is related to the support distance of the sphere, and the support distance has a corresponding hyperparameter; Determining an optimal hyperparameter of the support distance of the sphere, obtaining an optimal support distance of the sphere based on the optimal hyperparameter, and obtaining an optimal radius of the sphere based on the optimal support distance; Converting the uncertain output boundary under the sphere into an optimal approximation problem of a cube, and converting the uncertain output boundary into a deterministic output boundary by approximating the sphere under the optimal radius by the cube; With the deterministic output boundary as a constraint condition, a power flow information prediction model with system loss, economic cost and carbon emission of the power system as multiple objectives is solved to obtain a power flow information prediction result of the power system.

2. The method according to claim 1, characterized in that Determining the optimal hyperparameter of the support distance of the sphere, obtaining the optimal support distance of the sphere based on the optimal hyperparameter, and obtaining the optimal radius of the sphere based on the optimal support distance includes: Initializing a hyperparameter upper limit and a hyperparameter lower limit of a hyperparameter of a support distance of the sphere; Determining an optimal hyperparameter between the hyperparameter upper limit and the hyperparameter lower limit; Obtaining an optimal support distance for the sphere based on the optimal hyperparameter; The optimal radius is obtained based on the optimal support distance.

3. The method according to claim 2, characterized in that Determining the optimal hyperparameter between the hyperparameter upper limit and the hyperparameter lower limit includes: Determining a hyperparameter median value of the hyperparameter based on the hyperparameter upper limit and the hyperparameter lower limit; Determine a first support distance corresponding to the upper limit of the hyperparameter, a second support distance corresponding to the median of the hyperparameter, and a third support distance corresponding to the lower limit of the hyperparameter respectively; Based on the numerical comparison result among the first support distance, the second support distance, and the third support distance, updating the hyperparameter upper limit and / or the hyperparameter lower limit to obtain a new hyperparameter upper limit and / or a new hyperparameter lower limit; When the hyperparameter difference between the current hyperparameter upper limit and the current hyperparameter lower limit is greater than the preset difference, return to the step of determining the hyperparameter median of the hyperparameter based on the hyperparameter upper limit and the hyperparameter lower limit, until the obtained hyperparameter difference is less than or equal to the preset difference, and determine the corresponding hyperparameter median as the optimal hyperparameter.

4. The method according to claim 3, characterized in that The updating of the hyperparameter upper limit and / or the hyperparameter lower limit based on the numerical comparison result among the first support distance, the second support distance, and the third support distance to obtain a new hyperparameter upper limit and / or a new hyperparameter lower limit includes: If the second support distance is less than the third support distance, determining the average of the hyperparameter lower limit and the hyperparameter median as the new hyperparameter lower limit, and determining the average of the hyperparameter upper limit and the hyperparameter median as the new hyperparameter upper limit; If the second support distance is greater than or equal to the third support distance and less than or equal to the first support distance, the median of the hyperparameter is used as the new upper limit of the hyperparameter; If the second support distance is greater than the first support distance, the median of the hyperparameter is used as the new lower limit of the hyperparameter.

5. The method according to claim 1, wherein The step of approximating a sphere with an optimal radius by using a cube includes: Initialize the upper limit and lower limit of the side length of the cube; According to the optimal radius, the optimal side length of the cube is determined between the upper limit and the lower limit of the side length to approximate a sphere under the optimal radius.

6. The method according to claim 5, characterized in that Determining the optimal side length of the cube between the upper side length limit and the lower side length limit according to the optimal radius includes: Determine a median length of the side length based on the side length upper limit and the side length lower limit; Determining an acceptable probability of the median side length according to the optimal radius and the median side length; Based on the numerical comparison result between the acceptable probability and the preset acceptable probability, updating the side length upper limit or the side length lower limit to obtain a new side length upper limit or a new side length lower limit; When the side length difference between the current upper limit of the side length and the current lower limit of the side length is greater than the preset difference, return to the step of determining the median of the side length based on the upper limit of the side length and the lower limit of the side length, until the obtained side length difference is less than or equal to the preset difference, and determine the corresponding median of the side length as the optimal side length.

7. The method according to claim 6, characterized in that The updating of the side length upper limit or the side length lower limit based on the numerical comparison result between the acceptable probability and the preset acceptable probability to obtain a new side length upper limit or a new side length lower limit includes: If the acceptable probability is greater than the preset acceptable probability, the median of the side length is determined as the new lower limit of the side length; If the acceptable probability is less than or equal to the preset acceptable probability, the median of the side length is determined as the new upper limit of the side length.

8. The method according to any one of claims 1 to 7, characterized in that The obtaining, based on the historical output distribution of the power system and the generative output distribution, of the combined output distribution of the power system includes: Determining a union of the generated output distribution and the historical output distribution; The union is determined as the joint output distribution.

9. A power system flow information prediction device taking into account the uncertainty of carbon emissions and new energy output, characterized in that: The device comprises: a distribution generation module, configured to input a potential output distribution of a power system into a generative model to obtain a generative output distribution of the power system, and to obtain a joint output distribution of the power system based on a historical output distribution of the power system and the generative output distribution; the generative model comprising a generative adversarial network; A sphere construction module is configured to construct a sphere for characterizing an uncertainty output boundary of the power system based on the potential output distribution and the combined output distribution; the uncertainty output boundary of the power system is configured to characterize the output uncertainty of the power system, the output uncertainty of the power system being associated with the uncertainty of the output of renewable energy sources of the power system; the sphere is configured to cover the potential output distribution and the combined output distribution, the radius of the sphere being related to a support distance of the sphere, the support distance having a corresponding hyperparameter; a boundary conversion module, configured to determine an optimal hyperparameter for the support distance of the sphere, obtain an optimal support distance for the sphere based on the optimal hyperparameter, and obtain an optimal radius for the sphere based on the optimal support distance; convert the uncertain output boundary under the sphere into an optimal approximation problem of a cube, and convert the uncertain output boundary into a deterministic output boundary by approximating the sphere under the optimal radius with a cube; The information determination module is used to solve the power flow information prediction model with the system loss, economic cost and carbon emission of the power system as multiple objectives with the deterministic output boundary as a constraint condition, and obtain the power flow information prediction result of the power system.

10. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 8 are implemented.

Images