Redundant computing power self-adaptive calling method for power system

Abstract

The invention provides a redundancy computing power adaptive calling method for a power system, and relates to the technical field of computing power calling, and the method comprises the steps: constructing a computing power twinning network by an intelligent computing center of an interaction target area; building a power operation and management model, and placing the power operation and management model to each intelligent calculation center; returning the power data set to the target intelligent calculation center, and determining pre-distributed calculation power; determining a calling computing power demand; reading real-time redundant computing power based on the computing power twinning network, and determining a collaborative intelligent computing center based on a calling computing power demand in combination with a matching principle; and performing data differential processing and power dispatching analysis on the basis of the power operation and management model of the target intelligent calculation center and the cooperative intelligent calculation center, and determining a power dispatching operation and management scheme. Therefore, the technical problem is solved, and the objective of intelligent degree of computing power scheduling is achieved.

Description

Technical Field

[0001] The present disclosure relates to the technical field of computing power calling, and in particular to a method for adaptively calling redundant computing power for an electric power system. Background Art

[0002] The redundant computing power of the power system refers to the computing power resources that exist in some devices or nodes that exceed the actual demand when the power system is operating normally. Redundant computing power is usually not fully utilized, resulting in a waste of resources. At present, the existing technology can usually only dispatch redundant computing power according to the preset dispatching strategy, and the ability to perform intelligent dispatch according to the real-time power demand and system status is poor, which may lead to waste of computing power resources or failure to meet the actual power demand. On the other hand, due to the dynamic variability and complexity of the operating state of the power system, the existing technology may not be able to accurately predict and dispatch computing power resources, which may lead to a mismatch between the dispatch of computing power resources and actual demand, thereby affecting the stability and economy of the power system.

[0003] To sum up, the existing technology has technical problems such as the poor ability of the redundant computing power of the power system to intelligently dispatch according to the real-time power demand and system status, resulting in waste of computing power resources or failure to meet the actual power demand. At the same time, the scheduling accuracy of computing power is low, resulting in a mismatch between the scheduling of computing power resources and actual demand. Summary of the invention

[0004] The present disclosure provides a method for adaptively calling redundant computing power for an electric power system, so as to solve the technical problems in the prior art that the redundant computing power of the electric power system has a poor ability to intelligently dispatch according to real-time power demand and system status, resulting in waste of computing power resources or a poor degree of meeting actual power demand, and at the same time, the scheduling accuracy of computing power is low, resulting in a mismatch between the scheduling of computing power resources and actual demand.

[0005] According to a first aspect of the present disclosure, a method for adaptively calling redundant computing power for an electric power system is provided, comprising: interacting with intelligent computing centers in target areas, building a computing power twin network in combination with digital twin technology, wherein a cross-domain data channel is established between each intelligent computing center; building an electric power operation and management model, and delegating the electric power operation and management model to each intelligent computing center, wherein the electric power operation and management model includes a differential conversion unit and an operation and management dispatching unit connected front and back; connecting the electric power system and transmitting the electric power data set back to the target intelligent computing center, determining the pre-allocated computing power in combination with the computing demand, wherein the computing demand is determined based on the operation and management accuracy standard and the computing intensity, and the pre-allocated computing power is determined based on the computing demand of the differential power data; reading the idle computing power of the target intelligent computing center, and verifying it with the pre-allocated computing power to determine the calling computing power demand; based on the computing power twin network, reading the real-time redundant computing power and determining the collaborative intelligent computing center based on the calling computing power demand in combination with the matching principle, wherein the matching principle includes the proximity principle and the integration principle; based on the electric power operation and management models of the target intelligent computing center and the collaborative intelligent computing center, performing data differential processing and electric power dispatching analysis to determine the electric power dispatching and management plan.

[0006] According to a second aspect of the present disclosure, a redundant computing power adaptive calling system for an electric power system is provided, including: a computing power twin network construction module, which is used to interact with the intelligent computing center in the target area, and to build a computing power twin network in combination with the digital twin technology, wherein a cross-domain data channel is established between each intelligent computing center; a power operation management model construction module, which is used to build a power operation management model, and the power operation management model is decentralized to each intelligent computing center, wherein the power operation management model includes a differential conversion unit and an operation management scheduling unit connected front and back; a pre-allocated computing power determination module, which is used to connect to the power system and return the power data set to the target intelligent computing center, and determine the pre-allocated computing power in combination with computing requirements, and the computing requirements are based on the operation management accuracy standard The accuracy and computing intensity are determined, and the pre-allocated computing power is determined based on the computing requirements of differential power data; a computing power demand determination module is called, and the computing power demand determination module is used to read the idle computing power of the target intelligent computing center, and compare it with the pre-allocated computing power to determine the calling computing power demand; a collaborative intelligent computing center determination module is used to read the real-time redundant computing power based on the computing power twin network and determine the collaborative intelligent computing center based on the calling computing power demand in combination with the matching principle, and the matching principle includes the proximity principle and the integration principle; a power dispatching and operation management plan determination module is used to perform data differential processing and power dispatching analysis based on the power operation management model of the target intelligent computing center and the collaborative intelligent computing center to determine the power dispatching and operation management plan.

[0007] One or more technical solutions provided in the present disclosure have at least the following technical effects or advantages: according to the intelligent computing center of the interactive target area adopted in the present disclosure, a computing power twin network is constructed in combination with the digital twin technology, wherein a cross-domain data channel is established between each intelligent computing center; a power operation management model is constructed, and the power operation management model is decentralized to each intelligent computing center, wherein the power operation management model includes a differential conversion unit and an operation management dispatching unit connected to the front and back; the power system is connected and the power data set is transmitted back to the target intelligent computing center, and the pre-allocated computing power is determined in combination with the computing demand, and the computing demand is determined based on the operation management accuracy standard and the computing intensity, and the pre-allocated computing power is determined based on the computing demand of the differential power data; the idle computing power of the target intelligent computing center is read, and the computing power demand is determined by comparing it with the pre-allocated computing power; based on the computing power twin network, the real-time redundant computing power is read and combined The matching principle determines the collaborative intelligent computing center based on the computing power demand, and the matching principle includes the proximity principle and the integration principle; based on the power operation and management model of the target intelligent computing center and the collaborative intelligent computing center, data differential processing and power dispatch analysis are performed to determine the power dispatch and operation management plan, which solves the technical problems in the prior art that the redundant computing power of the power system has poor ability to perform intelligent dispatch according to the real-time power demand and system status, resulting in waste of computing power resources or poor degree of meeting the actual power demand, and at the same time, the scheduling computing power has low accuracy, resulting in the mismatch between the scheduling of computing power resources and the actual demand, and realizes the goal of improving the ability of the redundant computing power of the power system to perform intelligent dispatch according to the real-time power demand and system status and improve the accuracy of the scheduling computing power, and achieves the technical effect of reducing the waste of computing power resources and improving the degree of meeting the actual power demand.

[0008] It should be understood that the contents described in this section are not intended to indicate the key or important features of the embodiments of the present disclosure, nor are they intended to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] In order to more clearly illustrate the technical solutions in the present disclosure or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only exemplary, and for ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0010] Figure 1 A schematic diagram of a flow chart of a method for adaptively calling redundant computing power for a power system provided in an embodiment of the present disclosure;

[0011] Figure 2 A schematic diagram of the structure of a redundant computing power adaptive calling system for an electric power system provided in an embodiment of the present disclosure.

[0012] Explanation of the accompanying drawings: computing power twin network construction module 11, power operation and management model construction module 12, pre-allocated computing power determination module 13, calling computing power demand determination module 14, collaborative intelligent computing center determination module 15, power dispatching and operation management plan determination module 16. DETAILED DESCRIPTION

[0013] The following is a description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0014] Embodiment 1

[0015] The redundant computing power adaptive calling method for the power system provided in the embodiment of the present disclosure is referred to as follows Figure 1 Illustratively, the method comprises:

[0016] The intelligent computing center in the interactive target area combines digital twin technology to build a computing power twin network, in which cross-domain data channels are established between each intelligent computing center;

[0017] Specifically, the target area is the area where computing power is to be called. The target area has an intelligent computing center, and the basic computing power configuration of the intelligent computing center is used to perform digital twin simulation to build a computing power twin network. Among them, there are cross-domain data channels between each intelligent computing center. For example, when the intelligent computing center is a province or city, there is a power transmission channel and a data channel based on the computing power center between the two provinces.

[0018] Build a power operation and management model, and delegate the power operation and management model to each intelligent computing center, wherein the power operation and management model includes a differential conversion unit and an operation and management dispatching unit connected front and back;

[0019] Specifically, a power operation management model is built to perform data differential conversion processing and operation management and dispatch processing before data processing. The power operation management model includes a differential conversion unit and an operation management and dispatch unit connected before and after. Since the transmitted power information is too refined and the amount of data is too large, in order to improve the operation and calculation efficiency, the differential conversion unit is used to perform data differential conversion before data processing to extract overall valuable information. Furthermore, the power operation management model is decentralized to each intelligent computing center to improve data calculation efficiency.

[0020] Connect to the power system and transmit the power data set back to the target intelligent computing center, and determine the pre-allocated computing power based on the computing requirements, the computing requirements are determined based on the operation and management accuracy standards and computing intensity, and the pre-allocated computing power is determined based on the computing requirements of the differential power data;

[0021] Specifically, the power system has a power data set, which connects the power system and transmits the power data set back to the target intelligent computing center for determining the computing power demand. Among them, the target intelligent computing center is a plurality of intelligent computing centers in the target area. The computing demand is obtained based on the operation and management accuracy standard and the computing intensity. Among them, the computing intensity is determined based on the computing time constraint range. The pre-allocated computing power is determined according to the computing demand of the differential power data, that is, it is determined by the respective computing demand in the differential power data. Among them, the differential power data refers to the difference between the power data at adjacent time points or adjacent cycles. By analyzing the differential power data, we can understand the changing trend of the power load, the stability of the power system, the operating status of the power equipment and other information.

[0022] Read the idle computing power of the target intelligent computing center and compare it with the pre-allocated computing power to determine the computing power demand;

[0023] Specifically, the idle computing power of the target intelligent computing center is read, where idle computing power refers to computing power resources that are not fully utilized. In this embodiment, the idle computing power is the basic computing power resources pre-owned by the target intelligent computing center. By comparing the idle computing power with the pre-allocated computing power, the difference between the pre-allocated computing power and the idle computing power is obtained, that is, the computing power requirement that needs to be allocated to the target intelligent computing center is used as the calling computing power requirement.

[0024] Based on the computing power twin network, read the real-time redundant computing power and determine the collaborative intelligent computing center based on the computing power demand in combination with the matching principle, wherein the matching principle includes the proximity principle and the integration principle;

[0025] Specifically, based on the computing power twin network of the target intelligent computing center, the real-time redundant computing power of each intelligent computing center is read and combined with the matching principle, the nearest intelligent computing center among the intelligent computing centers corresponding to the redundant computing power is determined as the collaborative intelligent computing center. Among them, the matching principle includes the proximity principle and the integration principle. The proximity principle is to call the neighboring intelligent computing center, and the integration principle refers to processing based on a small number of calling computing power centers.

[0026] Based on the power operation and management models of the target intelligent computing center and the collaborative intelligent computing center, data differential processing and power dispatch analysis are performed to determine the power dispatch and operation management plan.

[0027] Specifically, based on the power operation and management models of the target intelligent computing center and the collaborative intelligent computing center, data differential processing and power dispatch analysis are performed to determine the power dispatch and operation management plan. Among them, data differential processing refers to eliminating trends and seasonal changes in the data to make the data more stable and easier to analyze and predict. Power dispatch analysis refers to determining the collaborative allocation decision of the collaborative intelligent computing center and determining the computing power decision results.

[0028] Among them, this embodiment can solve the technical problems in the prior art that the redundant computing power of the power system has a poor ability to perform intelligent scheduling according to the real-time power demand and system status, resulting in waste of computing power resources or poor degree of meeting the actual power demand. At the same time, the scheduling computing power has low accuracy, resulting in a mismatch between the scheduling of computing power resources and actual demand. The goal of improving the ability of the redundant computing power of the power system to perform intelligent scheduling according to the real-time power demand and system status and improving the accuracy of scheduling computing power is achieved, and the technical effect of reducing the waste of computing power resources and improving the degree of meeting the actual power demand is achieved.

[0029] The method provided in the embodiment of the present disclosure also includes:

[0030] Reading the basic computing power configuration of the intelligent computing center in the target area, wherein the basic computing power configuration includes installed capacity and system configuration;

[0031] Determining relative distribution characteristics of each computing power center, wherein the relative distribution characteristics are determined based on world coordinates;

[0032] Combining the relative distribution characteristics with the cross-domain data channel, building a network topology structure;

[0033] In combination with the basic computing power configuration, the network topology structure is improved to generate the computing power twin network.

[0034] Specifically, the intelligent computing center refers to an intelligent computing center, which is an infrastructure that integrates computing power, storage, network, and cloud. It provides efficient and secure computing services based on cloud computing, artificial intelligence and other technologies. Read the basic computing power configuration of the intelligent computing center in the target area, where the basic computing power configuration includes the installed capacity and system configuration of the intelligent computing center.

[0035] Furthermore, a world coordinate system of the target area is established, wherein the world coordinate system is an absolute coordinate system of the system, and the coordinates of all points in the target area are determined by the origin of the world coordinate system. Furthermore, the target area has multiple computing centers for calculations through the computing centers. The world coordinates of each computing center in the target area are read, and then the relative distribution characteristics of each computing center in the target area are obtained.

[0036] Furthermore, cross-domain data channels refer to data channels that cross domains. For example, when the computing power centers with relative distribution characteristics are two provinces and cities, there are power transmission channels and data channels based on computing power centers between the two provinces and cities. The computing power centers are connected in pairs according to the cross-domain data channels, and then a network topology structure is built. The network topology structure has relative distribution characteristics and can obtain relative world coordinates.

[0037] Furthermore, the network topology is simulated by digital twins through the basic computing power configuration of the intelligent computing center to obtain a computing power twin network. Among them, building a computing power twin network through digital twin technology can improve the efficiency of computing power call.

[0038] The method provided in the embodiment of the present disclosure also includes:

[0039] Determining the operation management accuracy standard based on the controllable deviation range of the power operation management;

[0040] Determine the data processing time limit constraint based on the operation management accuracy standard;

[0041] Determining a computation intensity based on the differential power data based on a data processing time limit constraint;

[0042] The computing requirement is determined by combining the operation management accuracy standard with the computing intensity.

[0043] Specifically, the controllable deviation interval of power operation management refers to the allowable deviation of power system state quantities, such as voltage, frequency, power, etc., within a certain range during the normal operation of the power system, and the deviation can be restored to the normal range through the control and adjustment of power operation management to ensure the stability and reliability of the power system while taking into account economy and flexibility. According to the controllable deviation interval of power operation management, the operation management accuracy standard is used as the threshold interval.

[0044] Furthermore, based on the operation and management accuracy standard, the data processing time limit constraint is determined. For example, the power system itself is in a fluctuating trend. If processing and decision-making are performed every 15 minutes, the response consistency may not be guaranteed. Therefore, based on the accuracy standard, the data processing time limit is determined. The shorter the time limit, the more accurate the short-term prediction results, the higher the relative computing intensity, and the more computing power is required.

[0045] Furthermore, based on the data processing time limit constraint, the calculation intensity based on the differential power data is determined, wherein when the data processing time limit constraint is shorter, the calculation intensity is higher, and vice versa.

[0046] Furthermore, the computing requirements are determined by combining the operation and management accuracy standards with the computing intensity. The computing is performed according to the operation and management accuracy standards according to the computing intensity. For example, the computing intensity is the computing frequency, such as processing and decision-making every 10 minutes. Obtaining the computing requirements can improve the accuracy of obtaining computing power calls.

[0047] The method provided in the embodiment of the present disclosure also includes:

[0048] Read the computing power call record in the predetermined time zone, and count the call frequency based on the computing power demand, wherein the computing power demand has a computing power tolerance range;

[0049] Determine whether the calling frequency meets the generalization standard, and if so, generate a model optimization instruction;

[0050] As the model optimization instruction is received, the optimization range based on the called computing power demand is determined, and the power operation management model is incrementally learned and updated.

[0051] Specifically, the computing power call records of the predetermined time zone of each intelligent computing center, that is, the calculation records, are read. The call frequency of the computing power demand is obtained from the computing power call records. Among them, the predetermined time zone is a historical time interval, such as within the past 5 days. There is a computing power tolerance range for calling computing power demand. Among them, the computing power tolerance range refers to the error range of calling computing power demand. Statistics are taken for all computing power call records that call computing power demands that are higher or lower than the computing power tolerance range, and then the statistical standards are generalized. It is possible to determine whether it is a situation unique to an individual computing power center through computing power call records that are higher or lower than the computing power tolerance range, or if the call frequency is judged to be low, the model will not be updated, and the call processing of redundant computing power can be performed.

[0052] Further, it is determined whether the calling frequency meets the generalization standard, wherein the generalization standard is obtained by a person skilled in the art through custom settings according to actual conditions, for example, the generalization standard is that the calling frequency is 10 times. If it is determined that the calling frequency meets the generalization standard, a model optimization instruction is generated.

[0053] Furthermore, as the model optimization instruction is received, the optimization range based on the call computing power demand is determined, the call computing power demand is increased or decreased, and the optimized power operation model is incrementally learned and updated to determine the updated call computing power demand. Among them, by obtaining the optimization range of the call computing power demand, the accuracy of the update of the call computing power demand can be improved.

[0054] The method provided in the embodiment of the present disclosure also includes:

[0055] If the calling frequency does not meet the generalization standard, generating a computing power scheduling instruction;

[0056] As the computing power scheduling instruction is received, the computing power twin network is traversed, and the neighborhood redundant computing power reading based on the target intelligent computing center is executed to determine multiple standby scheduling computing powers;

[0057] Based on the computing power demand, traverse the plurality of standby scheduling computing powers for screening and determine a pre-scheduling intelligent computing center;

[0058] If the pre-scheduled intelligent computing center is not empty, one item is randomly selected as the collaborative intelligent computing center.

[0059] Specifically, if the calling frequency does not meet the generalization standard, a computing power scheduling instruction is generated for direct computing power scheduling.

[0060] Furthermore, as the computing power scheduling instructions are received, the computing power twin network corresponding to the target intelligent computing center is accessed in turn, and the redundant computing power of the neighboring intelligent computing centers of the intelligent computing center that generated the computing power scheduling instructions is read within each intelligent computing center in the target area, and the redundant computing power of the neighboring intelligent computing center is determined to be multiple backup scheduling computing powers.

[0061] Furthermore, based on the demand for the computing power required, multiple standby scheduling computing powers are accessed in sequence to screen and obtain the standby scheduling computing power closest to the demand for the computing power required, and the corresponding intelligent computing center is determined as the pre-scheduling intelligent computing center. The standby scheduling computing power closest to the computing power required is obtained by serializing and processing multiple standby scheduling computing powers.

[0062] Furthermore, if the pre-scheduled intelligent computing center is not empty, a smart computing center is randomly selected from multiple pre-scheduled smart computing centers as a collaborative smart computing center. Among them, determining the collaborative smart computing center based on the call computing power demand can improve the efficiency of calling computing power, until the next call, the collaborative smart computing center is extracted for fast call.

[0063] The method provided in the embodiment of the present disclosure also includes:

[0064] If the pre-scheduling intelligent computing center is empty, a collaborative allocation decision is made on the multiple standby scheduling computing powers to determine the computing power decision result;

[0065] If the computing power decision result does not meet the calling computing power requirement, combined with the computing power twin network, a lower-level neighborhood redundant computing power matching analysis based on the target intelligent computing center is performed to determine the collaborative intelligent computing center.

[0066] Specifically, if the pre-scheduled intelligent computing center is empty, it means that there is no intelligent computing center in the pre-scheduled intelligent computing center that is close to the demand for the computing power required, and a collaborative allocation decision is made for multiple standby scheduling computing powers to determine the computing power decision result. Among them, the computing power demand can be evenly divided according to the number of intelligent computing centers corresponding to the standby scheduling computing power, and then a collaborative allocation decision is made.

[0067] Furthermore, if the computing power decision result does not meet the computing power demand, indicating that the computing power of the computing power decision result is less than the computing power required for calling the computing power, then the nearest intelligent computing center is extracted in combination with the computing power twin network, and the intelligent computing center corresponding to the lower-level neighborhood redundant computing power is extracted based on the intelligent computing center corresponding to the neighborhood redundant computing power of the target intelligent computing center except for the intelligent computing center corresponding to the neighborhood redundant computing power, and the second backup scheduling computing power is determined, and the second pre-scheduling intelligent computing center is screened and obtained. If the second pre-scheduling intelligent computing center is not empty, one is randomly selected as the collaborative intelligent computing center. Among them, determining that the pre-scheduling intelligent computing center is empty can improve the efficiency of computing power calling and avoid stagnation when the pre-scheduling intelligent computing center is empty.

[0068] The method provided in the embodiment of the present disclosure also includes:

[0069] Setting data partitioning principles, wherein the data partitioning principles include a time priority principle and a confidentiality priority principle;

[0070] Based on the calling computing power requirement and in combination with the data partitioning principle, the power data set is segmented to determine the local processing data and the calling processing data;

[0071] Based on the collaborative intelligent computing center, identifying a target data channel based on the cross-domain data channel;

[0072] Based on the target data channel, the call processing data is transmitted.

[0073] Specifically, data partitioning principles are set, wherein the data partitioning principles include the timing priority principle and the confidentiality priority principle. The timing priority principle refers to the priority sorting of power dispatching according to different timing requirements. The greater the power demand, the more urgent the demand time, and the higher the power system stability requirement, the higher the corresponding priority, and vice versa. The confidentiality priority principle refers to the priority sorting of confidentiality work for power dispatching according to the confidentiality requirements of power dispatching during the power dispatching process. The higher the confidentiality requirements, the higher the confidentiality priority, and vice versa. For example, confidentiality requirements refer to a large amount of sensitive information and confidential data, such as dispatching instructions, power demand, power generation plans, etc.

[0074] Furthermore, based on the call computing power demand as the standard, the call computing power demand is matched in combination with the data partitioning principle, the power data set is segmented according to the timing priority and confidentiality priority, and the local processing data and the call processing data are determined. Among them, the call computing power demand with a higher priority can be set as the local processing data, and the call computing power demand with a lower priority can be set as the call processing data to improve data processing efficiency.

[0075] Furthermore, the cross-domain data channel of the collaborative intelligent computing center is identified to obtain the target data channel. Furthermore, based on the target data channel, the call processing data is transmitted. Among them, setting the data division principle can improve the efficiency of data processing.

[0076] Embodiment 2

[0077] Based on the same inventive concept as the method for adaptively calling redundant computing power for a power system in the aforementioned embodiment, Figure 2 As shown, the present disclosure also provides a redundant computing power adaptive calling system for a power system, the system comprising:

[0078] A computing power twin network construction module 11, which is used to interact with the intelligent computing center in the target area and build a computing power twin network in combination with the digital twin technology, wherein a cross-domain data channel is established between each intelligent computing center;

[0079] A power operation and management model building module 12, which is used to build a power operation and management model and delegate the power operation and management model to each intelligent computing center, wherein the power operation and management model includes a differential conversion unit and an operation and management dispatching unit connected front and back;

[0080] A pre-allocated computing power determination module 13, which is used to connect to the power system and transmit the power data set back to the target intelligent computing center, and determine the pre-allocated computing power in combination with the computing demand, the computing demand is determined based on the operation management accuracy standard and the computing intensity, and the pre-allocated computing power is determined based on the computing demand of the differential power data;

[0081] Calling a computing power demand determination module 14, wherein the computing power demand determination module 14 is used to read the idle computing power of the target intelligent computing center and compare it with the pre-allocated computing power to determine the computing power demand;

[0082] A collaborative intelligent computing center determination module 15, which is used to read the real-time redundant computing power based on the computing power twin network and determine the collaborative intelligent computing center based on the calling computing power demand in combination with the matching principle, wherein the matching principle includes the proximity principle and the integration principle;

[0083] The power dispatching and operation management plan determination module 16 is used to perform data differential processing and power dispatching analysis based on the power operation management model of the target intelligent computing center and the collaborative intelligent computing center to determine the power dispatching and operation management plan.

[0084] Furthermore, the computing power twin network construction module 11 also includes:

[0085] Reading the basic computing power configuration of the intelligent computing center in the target area, wherein the basic computing power configuration includes installed capacity and system configuration;

[0086] Determining relative distribution characteristics of each computing power center, wherein the relative distribution characteristics are determined based on world coordinates;

[0087] Combining the relative distribution characteristics with the cross-domain data channel, building a network topology structure;

[0088] In combination with the basic computing power configuration, the network topology structure is improved to generate the computing power twin network.

[0089] Furthermore, the pre-allocated computing power determination module 13 also includes:

[0090] Determining the operation management accuracy standard based on the controllable deviation range of the power operation management;

[0091] Determine the data processing time limit constraint based on the operation management accuracy standard;

[0092] Determining a computation intensity based on the differential power data based on a data processing time limit constraint;

[0093] The computing requirement is determined by combining the operation management accuracy standard with the computing intensity.

[0094] Furthermore, the calling computing power requirement determination module 14 further includes:

[0095] Read the computing power call record in the predetermined time zone, and count the call frequency based on the computing power demand, wherein the computing power demand has a computing power tolerance range;

[0096] Determine whether the calling frequency meets the generalization standard, and if so, generate a model optimization instruction;

[0097] As the model optimization instruction is received, the optimization range based on the called computing power demand is determined, and the power operation management model is incrementally learned and updated.

[0098] Furthermore, the calling computing power requirement determination module 14 further includes:

[0099] If the calling frequency does not meet the generalization standard, generating a computing power scheduling instruction;

[0100] As the computing power scheduling instruction is received, the computing power twin network is traversed, and the neighborhood redundant computing power reading based on the target intelligent computing center is executed to determine multiple standby scheduling computing powers;

[0101] Based on the computing power demand, traverse the plurality of standby scheduling computing powers for screening and determine a pre-scheduling intelligent computing center;

[0102] If the pre-scheduled intelligent computing center is not empty, one item is randomly selected as the collaborative intelligent computing center.

[0103] Furthermore, the calling computing power requirement determination module 14 further includes:

[0104] If the pre-scheduling intelligent computing center is empty, a collaborative allocation decision is made on the multiple standby scheduling computing powers to determine the computing power decision result;

[0105] If the computing power decision result does not meet the calling computing power requirement, combined with the computing power twin network, a lower-level neighborhood redundant computing power matching analysis based on the target intelligent computing center is performed to determine the collaborative intelligent computing center.

[0106] Furthermore, the collaborative intelligent computing center determination module 15 also includes:

[0107] Setting data partitioning principles, wherein the data partitioning principles include a time priority principle and a confidentiality priority principle;

[0108] Based on the calling computing power requirement and in combination with the data partitioning principle, the power data set is segmented to determine the local processing data and the calling processing data;

[0109] Based on the collaborative intelligent computing center, identifying a target data channel based on the cross-domain data channel;

[0110] Based on the target data channel, the call processing data is transmitted.

[0111] The specific example of the redundant computing power adaptive calling method for the power system in the aforementioned embodiment 1 is also applicable to the redundant computing power adaptive calling system for the power system in this embodiment. Through the aforementioned detailed description of the redundant computing power adaptive calling method for the power system, those skilled in the art can clearly know the redundant computing power adaptive calling system for the power system in this embodiment, so for the sake of brevity of the specification, it will not be described in detail here. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part description.

[0112] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of this disclosure can be achieved, and this document is not limited here.

[0113] The above specific implementations do not constitute a limitation on the protection scope of the present disclosure. It should be understood by those skilled in the art that various modifications, combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present disclosure should be included in the protection scope of the present disclosure.

Claims

1. A method for adaptively calling redundant computing power for a power system, characterized in that: The method comprises: The intelligent computing center in the interactive target area combines digital twin technology to build a computing power twin network, in which cross-domain data channels are established between each intelligent computing center; Build a power operation and management model, and delegate the power operation and management model to each intelligent computing center, wherein the power operation and management model includes a differential conversion unit and an operation and management dispatching unit connected front and back; Connect to the power system and transmit the power data set back to the target intelligent computing center, and determine the pre-allocated computing power based on the computing requirements, the computing requirements are determined based on the operation and management accuracy standards and computing intensity, and the pre-allocated computing power is determined based on the computing requirements of the differential power data; Read the idle computing power of the target intelligent computing center and compare it with the pre-allocated computing power to determine the computing power demand; Based on the computing power twin network, read the real-time redundant computing power and determine the collaborative intelligent computing center based on the computing power demand in combination with the matching principle, wherein the matching principle includes the proximity principle and the integration principle; Based on the power operation and management models of the target intelligent computing center and the collaborative intelligent computing center, data differential processing and power dispatch analysis are performed to determine the power dispatch and operation management plan.

2. The method according to claim 1, characterized in that The intelligent computing center of the interactive target area combines digital twin technology to build a computing power twin network, including: Reading the basic computing power configuration of the intelligent computing center in the target area, wherein the basic computing power configuration includes installed capacity and system configuration; Determining relative distribution characteristics of each computing power center, wherein the relative distribution characteristics are determined based on world coordinates; Combining the relative distribution characteristics with the cross-domain data channel, building a network topology structure; In combination with the basic computing power configuration, the network topology structure is improved to generate the computing power twin network.

3. The method according to claim 1, characterized in that The computing requirements are determined based on the operation management accuracy standard and computing intensity, including: Determining the operation management accuracy standard based on the controllable deviation range of the power operation management; Determine the data processing time limit constraint based on the operation management accuracy standard; Determining a computation intensity based on the differential power data based on a data processing time limit constraint; The computing requirement is determined by combining the operation management accuracy standard with the computing intensity.

4. The method according to claim 1, characterized in that After determining the computing power requirement, the following steps are included: Read the computing power call record in the predetermined time zone, and count the call frequency based on the computing power demand, wherein the computing power demand has a computing power tolerance range; Determine whether the calling frequency meets the generalization standard, and if so, generate a model optimization instruction; As the model optimization instruction is received, the optimization range based on the called computing power demand is determined, and the power operation management model is incrementally learned and updated.

5. The method according to claim 4, characterized in that The determining of the collaborative intelligent computing center based on the calling computing power requirement includes: If the calling frequency does not meet the generalization standard, generating a computing power scheduling instruction; As the computing power scheduling instruction is received, the computing power twin network is traversed, and the neighborhood redundant computing power reading based on the target intelligent computing center is executed to determine multiple standby scheduling computing powers; Based on the computing power demand, traverse the plurality of standby scheduling computing powers for screening and determine a pre-scheduling intelligent computing center; If the pre-scheduled intelligent computing center is not empty, one item is randomly selected as the collaborative intelligent computing center.

6. The method according to claim 5, characterized in that include: If the pre-scheduling intelligent computing center is empty, a collaborative allocation decision is made on the multiple standby scheduling computing powers to determine the computing power decision result; If the computing power decision result does not meet the calling computing power requirement, combined with the computing power twin network, a lower-level neighborhood redundant computing power matching analysis based on the target intelligent computing center is performed to determine the collaborative intelligent computing center.

7. The method according to claim 1, characterized in that After determining the collaborative intelligent computing center based on the computing power demand, it includes: Setting data partitioning principles, wherein the data partitioning principles include a time priority principle and a confidentiality priority principle; Based on the calling computing power requirement and in combination with the data partitioning principle, the power data set is segmented to determine the local processing data and the calling processing data; Based on the collaborative intelligent computing center, identifying a target data channel based on the cross-domain data channel; Based on the target data channel, the call processing data is transmitted.

8. A redundant computing power adaptive calling system for a power system, characterized in that: The system comprises: A computing power twin network construction module, which is used to interact with the intelligent computing center in the target area and build a computing power twin network in combination with digital twin technology, wherein a cross-domain data channel is established between each intelligent computing center; A power operation and management model building module, which is used to build a power operation and management model and delegate the power operation and management model to each intelligent computing center, wherein the power operation and management model includes a differential conversion unit and an operation and management dispatching unit connected front and back; A pre-allocated computing power determination module, which is used to connect to the power system and transmit the power data set back to the target intelligent computing center, and determine the pre-allocated computing power in combination with the computing demand, the computing demand is determined based on the operation management accuracy standard and the computing intensity, and the pre-allocated computing power is determined based on the computing demand of the differential power data; A computing power demand determination module is called, and the computing power demand determination module is used to read the idle computing power of the target intelligent computing center and compare it with the pre-allocated computing power to determine the computing power demand; A collaborative intelligent computing center determination module, which is used to read the real-time redundant computing power based on the computing power twin network and determine the collaborative intelligent computing center based on the calling computing power demand in combination with a matching principle, wherein the matching principle includes a proximity principle and an integration principle; A power dispatching and operation management plan determination module is used to perform data differential processing and power dispatching analysis based on the power operation management models of the target intelligent computing center and the collaborative intelligent computing center to determine the power dispatching and operation management plan.

Images