Power grid fault analysis method and system based on multi-core task coordination

Abstract

The invention discloses a power grid fault analysis method and system based on multi-core task coordination, and relates to the technical field of power grid fault analysis, and the method comprises the following steps: in a first processing core, obtaining a fault association relationship according to power grid region operation data and environment data; in the second processing core, a power grid fault analysis result is obtained according to the fault incidence relation, the current environment data, the real-time power grid operation data and a fault threshold value; and in the third processing core, dividing the first processing cores according to the types of the transformer substations to obtain division results, obtaining real-time loads of the first processing cores, and executing task scheduling of the first processing cores based on the division results and the real-time loads. The method has the beneficial effects that task scheduling processing is carried out through the third processing core according to the processing data similarity of the first processing core, and the adaptation degree between the scheduling cores is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power grid fault analysis, in particular to a power grid fault analysis method and system based on multi-core task coordination. BACKGROUND

[0002] With the rapid development of smart grids, the scale of power grids continues to expand, and the proportion of distributed energy (such as photovoltaic and wind power) access has significantly increased, resulting in the complexification of power grid operation environment, and the high frequency, diversification and instantaneous characteristics of fault types.

[0003] Traditional power grid fault analysis systems mostly use single-core processors + non-real-time operating systems (such as Linux) architecture, however, single-core processors need to process data collection, feature extraction, fault positioning and other tasks in series, and task accumulation easily occurs under high load, resulting in the reduction of fault analysis efficiency.

[0004] In related technologies using multi-core processors, task scheduling is mostly globally randomly allocated, and task migration is across type cores, which causes task re-entry delay due to poor core adaptability, low migration efficiency, and reduced fault analysis efficiency.

[0005] Patent "RISC-V chip multi-core heterogeneous design method and system for power grid applications", publication number: CN119885987A, publication date: April 25, 2025, specifically discloses including inputting power grid tasks, adaptively exchanging data and scheduling tasks between AI cores and CPU cores according to power grid tasks; performing parallel computing through a special hardware unit and an optimized instruction set according to data exchange and task scheduling requirements; through dynamic power management, real-time monitoring of computing load, and dynamic adjustment of power states of AI cores and CPU cores; through the on-chip network (NOC) communication architecture, multiple parallel communication channels are provided, a network topology structure is introduced inside the chip, and communication and data transmission between processing units are realized. Although this scheme uses multi-core scheduling, it still has the problem of poor adaptability between different processing cores causing task re-entry delay. SUMMARY

[0006] The present application provides a power grid fault analysis method and system based on multi-core task coordination to solve the problem of poor core adaptability causing task re-entry delay in the prior art, which processes data association, global fault judgment and core task scheduling through the first processing core, the second processing core and the third processing core respectively, realizes physical isolation of functional modules, and processes task scheduling according to the processing data similarity of the first processing core through the third processing core, improves the adaptability between scheduling cores, and adapts to the current complex power grid scenario.

[0007] To achieve the above technical purposes, a technical solution provided by the present application is a power grid fault analysis method based on multi-core task coordination, comprising the following steps: dividing the power grid area based on the transformer substation, each power grid area corresponding to a first processing core, all power grid areas corresponding to a second processing core and a third processing core; in the first processing core, obtaining the fault correlation relationship according to the power grid area operation data and the environment data; in the second processing core, obtaining the power grid fault analysis result according to the fault correlation relationship, the current environment data, the real-time power grid operation data and the fault threshold; in the third processing core, obtaining the division result by dividing the first processing core according to the transformer substation type, obtaining the real-time load of each first processing core, and performing first processing core task scheduling based on the division result and the real-time load.

[0008] Further, the obtaining of the fault correlation relationship according to the power grid area operation data and the environment data comprises: obtaining a set of associated operation data corresponding to the fault type according to the power grid area operation data and the environment data of the historical time segment; and constructing the fault correlation relationship corresponding to the fault type based on the environment fluctuation frequency, the environment fluctuation factor and the environment fluctuation intensity according to the set of associated operation data.

[0009] Further, the constructing of the fault correlation relationship corresponding to the fault type based on the environment fluctuation frequency, the environment fluctuation factor and the environment fluctuation intensity according to the set of associated operation data comprises: extracting the environment fluctuation factor in the set of associated operation data, obtaining the environment fluctuation frequency and the environment fluctuation intensity of each environment fluctuation factor according to the associated operation data of the same environment fluctuation factor to obtain a single-factor fluctuation weight; obtaining the multi-factor fluctuation weight of each environment fluctuation factor according to the associated operation data of different environment fluctuation factors and the single-factor fluctuation weight; and constructing the fault correlation relationship corresponding to the fault type based on the single-factor fluctuation weight and the multi-factor fluctuation weight.

[0010] Further, the obtaining of the power grid fault analysis result according to the fault correlation relationship, the current environment data, the real-time power grid operation data and the fault threshold comprises: obtaining the corrected power grid operation data of each area according to the fault correlation relationship, the current environment data and the real-time power grid operation data; and outputting the power grid fault warning of the area if the corrected power grid operation data meets the fault threshold.

[0011] Further, the obtaining of the power grid fault analysis result according to the fault correlation relationship, the current environment data, the real-time power grid operation data and the fault threshold comprises: obtaining the corrected power grid operation data of each area according to the fault correlation relationship, the current environment data and the real-time power grid operation data; and outputting the power grid fault warning of the area if the corrected power grid operation data meets the fault threshold.

[0012] Another technical solution provided in the application is a power grid fault analysis system based on multi-core task coordination, which is used to implement the power grid fault analysis method based on multi-core task coordination, and includes a data acquisition unit configured to acquire power grid region operation data and environment data; a data processing unit including a first processing core, a second processing core, and a third processing core, configured to output a power grid fault analysis result according to the data acquired by the data acquisition unit, and perform task scheduling according to the real-time load of the first processing core.

[0013] Further, the data acquisition unit includes an FPGA data acquisition layer and a PCIe data transmission layer, the FPGA data acquisition layer acquires the power grid region operation data and the environment data and stores them in a memory space, and the first processing core acquires the power grid region operation data and the environment data in the memory space through the PCIe data transmission layer.

[0014] Further, the data processing unit further includes a hard real-time kernel, which controls the interruption of the first processing core, the second processing core, and the third processing core.

[0015] Further, the priority of the second processing core is higher than that of the first processing core, and the priority of the first processing core is higher than that of the third processing core.

[0016] Further, the third processing core acquires load data of the first processing core in real time through a hard real-time kernel interface, and performs task scheduling according to a comparison result of the load data and a preset load threshold.

[0017] The application has the following advantages: 1. The first processing core is responsible for local data association, the second processing core and the third processing core are respectively responsible for global research and scheduling, the physical isolation of the functional modules is realized, the influence of single core failure on the overall system is reduced, the load scheduling of the third processing core can quickly adapt to the fluctuation of the data volume of each region, the processing of task scheduling is performed according to the processing data similarity of the first processing core, the adaptation degree between the scheduling cores is improved, and the current complex power grid scene is adapted.

[0018] 2. The fault association relationship analysis of the corresponding region is performed in each first processing core, the overall power grid fault analysis efficiency is improved through parallel computing, the processing pressure of the single core is reduced, when any first processing core fails, the fault association relationship analysis of the region can be maintained through the task scheduling of the third processing core, the reliability of the fault analysis is improved, the influence of the missing data of a single region on the overall fault analysis is avoided, the accuracy of the power grid fault analysis is improved, and the power grid fault analysis under the current multi-grid connection condition is adapted.

[0019] 3. According to the environmental fluctuation factor of the environmental characteristics fluctuating before and after the fault in the associated operation data set, the influence of the fluctuation frequency of the environmental fluctuation factor and the fluctuation intensity of the environmental fluctuation factor in the associated operation data with the same environmental fluctuation factor on the power grid fault, and the synergistic or antagonistic effect between factors according to the single factor fluctuation weight in the associated operation data with different environmental fluctuation factors, the fault correlation corresponding to the fault type is constructed, so as to quantify the influence of environmental factors on the fault and the correlation between environmental factors, and improve the accuracy of fault tracing. BRIEF DESCRIPTION OF DRAWINGS

[0020] Figure 1 The flowchart of the power grid fault analysis method based on multi-core task coordination of the present application.

[0021] Figure 2 The structural diagram of the power grid fault analysis system based on multi-core task coordination of the present application. DETAILED DESCRIPTION

[0022] In order to make the purpose, technical scheme and advantages of the present application more clear and obvious, the present application will be further described in detail below in combination with the drawings and examples. It should be understood that the specific embodiments described herein are only one of the best embodiments of the present application, which are used to explain the present application and do not limit the protection scope of the present application. All other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0023] As shown in Figure 1 , as an embodiment of the present application, the power grid fault analysis method based on multi-core task coordination includes the following steps: Divide the power grid area based on the transformer substation, each power grid area corresponds to a first processing core, and all power grid areas correspond to a second processing core and a third processing core; In the first processing core, the fault correlation is obtained according to the power grid area operation data and the environmental data; In the second processing core, the power grid fault analysis result is obtained according to the fault correlation, the current environmental data, the real-time power grid operation data and the fault threshold; In the third processing core, the first processing core is divided according to the transformer substation type to obtain the division result, the real-time load of each first processing core is obtained, and the first processing core task scheduling is executed based on the division result and the real-time load.

[0024] The power grid is divided into regions based on substations. Each region has an independent primary processing core responsible for local data processing, enabling parallel computation of environmental correlations. A secondary processing core dynamically updates fault thresholds based on the fault correlations output by the primary processing core and performs real-time analysis of power grid faults. During the computation process of the primary processing core, a third processing core divides the primary processing cores based on substation type and monitors the load of each primary processing core in real time. Task migration between primary processing cores of the same substation type is achieved to avoid single-core overload.

[0025] In this embodiment, the first processing core is responsible for localized data association, while the second and third processing cores are responsible for global analysis and scheduling, respectively. This achieves physical isolation of functional modules, reduces the impact of a single core failure on the overall system, and enables rapid adaptation to fluctuations in data volume in various regions through load scheduling of the third processing core. Task scheduling is performed based on the similarity of the processed data of the first processing core, improving the adaptability between scheduling cores to adapt to the current complex power grid scenario.

[0026] Specifically, obtaining fault correlations based on power grid regional operation data and environmental data includes: Based on historical time-series power grid regional operation data and environmental data, obtain a set of associated operation data corresponding to the fault type; Based on the associated operational data set, fault correlation relationships corresponding to fault types are constructed according to the number of environmental fluctuations, environmental fluctuation factors, and environmental fluctuation intensity.

[0027] Power grid operation data should include at least voltage, current, and power. Environmental data should include at least temperature, humidity, rainfall, wind speed, and icing. By analyzing the frequency and intensity of environmental fluctuations and their impact on power grid operation, fault correlations can be constructed to improve the accuracy of dynamic threshold updates.

[0028] In this embodiment, the fault occurrence sequence is taken as the fault node, and the power grid area operation data and environmental data from the previous fault node to the current fault node are taken as the associated operation data of the fault node. The associated operation data set of the corresponding fault type is obtained by using the associated operation data of all fault nodes with the same fault type.

[0029] Based on the associated operational data set, fault correlation relationships corresponding to fault types are constructed according to the number of environmental fluctuations, environmental fluctuation factors, and environmental fluctuation intensity, including: Extract environmental fluctuation factors from the associated operational data set, and obtain the number of environmental fluctuations and the intensity of environmental fluctuations for each environmental fluctuation factor based on the associated operational data of the same environmental fluctuation factor to obtain the single-factor fluctuation weight. The multi-factor volatility weights of each environmental volatility factor are obtained based on the associated operational data of different environmental volatility factors and the single-factor volatility weights. Fault correlation relationships corresponding to fault types are constructed based on single-factor fluctuation weights and multi-factor fluctuation weights.

[0030] Environmental fluctuation factors are obtained by analyzing environmental characteristics that fluctuate before and after a fault in the associated operational data set. The impact of the fluctuation frequency and intensity of environmental fluctuation factors on power grid faults is determined based on the associated operational data containing the same environmental fluctuation factors. Furthermore, the synergistic or antagonistic effects between factors are obtained based on the single-factor fluctuation weights in associated operational data containing different environmental fluctuation factors. Fault correlation relationships corresponding to fault types are constructed, thereby quantifying the impact of environmental factors on faults and the correlation between environmental factors, improving the accuracy of fault tracing. It is understood that power grid faults are manifested through voltage, current, and power values. Therefore, weight iterative calculations are performed based on the power grid operational data of non-faulty nodes, the power grid operational data of faulty nodes, the initial values ​​of single-factor fluctuation weights, the initial values ​​of multi-factor fluctuation weights, and voltage, current, and power fault thresholds in the associated operational data set. This ensures that non-faulty nodes do not meet the various fault thresholds under single-factor and multi-factor fluctuation weight corrections, while faulty nodes do meet the various fault thresholds under the same corrections. Principal component analysis and neural network learning algorithms can be used to obtain the single-factor and multi-factor fluctuation weights of each environmental fluctuation factor.

[0031] Each first processing core performs fault correlation analysis for its corresponding region. Parallel computing improves the overall efficiency of power grid fault analysis and reduces the processing pressure on a single core. Furthermore, when any first processing core fails, the fault correlation analysis for that region can be maintained through task scheduling of the third processing core, thereby improving the reliability of fault analysis, avoiding the impact of missing data in a single region on the overall fault analysis, improving the accuracy of power grid fault analysis, and adapting to the current multi-grid-connected power grid fault analysis.

[0032] The power grid fault analysis results are obtained based on fault correlation, current environmental data, real-time power grid operation data, and fault thresholds, including: Based on fault correlation, current environmental data, and real-time power grid operation data, corrected power grid operation data for each region is obtained; if corrected power grid operation data meets the fault threshold, a power grid fault warning for that region is output.

[0033] In this embodiment, by incorporating the potential impact of environmental factors on equipment into the operational data, corrected power grid operational data for each region is obtained, making the corrected data closer to the actual fault risk. For example, if the current of a certain line in a certain region is 800A and the current threshold is 1000A, but the current environment is "high temperature 38℃ + humidity 90%", the corrected power grid operational data calculated based on fault correlation is 1067A. This reflects the decrease in line current-carrying capacity caused by high temperature and humidity, reflects the impact of potential faults, and enhances the accuracy and reliability of fault early warning. It is understood that real-time power grid operational data includes real-time power grid operational data for each region, which not only distributes the computational load through the first processing core but also narrows down the investigation area in the event of a potential fault.

[0034] The first processing core is divided according to the substation type, and the division results are obtained. The real-time load of each first processing core is obtained. Based on the division results and the real-time load, the first processing core task scheduling is performed, including: The first processing core obtains the classification results based on the substation type; If the real-time load of a first processing core exceeds a preset load threshold, then the tasks of that first processing core will be scheduled to the other first processing cores in its corresponding partitioning result.

[0035] The first processing core of the same partitioning result has high consistency in task type, data format, and calculation logic. When task migration occurs, there is no need for additional adaptation to dedicated algorithms for different types of substations, improving migration adaptability, reducing load fluctuations within the same group, and enhancing system stability. It can be understood that during task migration, the first processing core with the lowest load is selected as the first processing core corresponding to the highest migration priority.

[0036] like Figure 2 As shown in Embodiment 2 of this application, a power grid fault analysis system based on multi-core task coordination includes: The data acquisition unit is used to acquire regional power grid operation data and environmental data. The data processing unit includes a first processing core, a second processing core, and a third processing core. It is used to output power grid fault analysis results based on the data collected by the data acquisition unit and to perform task scheduling based on the real-time load of the first processing core.

[0037] The data acquisition unit includes an FPGA data acquisition layer and a PCIe data transmission layer. The FPGA data acquisition layer acquires power grid area operation data and environmental data and stores them in memory space. The first processing core obtains power grid area operation data and environmental data in memory space through the PCIe data transmission layer.

[0038] In this embodiment, the FPGA data acquisition layer acquires power grid area operation data through an analog-to-digital converter and environmental data through an optocoupler isolation acquisition circuit. The two types of data are merged into 41 Words with index numbers and stored in 16 dedicated spaces in memory in a cyclic writing manner.

[0039] The data processing unit also includes a hard real-time kernel, which controls the interrupts of the first, second, and third processing cores.

[0040] The number of first processing cores corresponds to the number of power grid regions, with one second and one third processing core each. In other cases, the number of first processing cores can be the number of power grid regions plus one, with one redundant first processing core for load sharing.

[0041] In this embodiment, the hard real-time kernel is the real-time kernel of RT-Linux. The first processing core uses the DMA (Direct Memory Access) mechanism through the PCIe data transfer layer to bypass the Linux kernel's memory management. The hard real-time kernel directly controls the data flow, thereby improving data transfer efficiency, reducing interference from non-real-time kernels, and ensuring low latency and high determinism of the data flow. The interrupt controller is controlled by the real-time kernel, and all interrupts are first intercepted by the real-time kernel. The kernel first performs the relevant interrupt handling and then "passes" the interrupt to the Linux kernel. In this way, no activity of the Linux kernel can cause interrupts to be disabled, thus having no impact on the task scheduling of the real-time kernel, thereby ensuring the hard real-time performance of the kernel.

[0042] In the data processing unit, the priority of the second processing core is higher than that of the first processing core, and the priority of the first processing core is higher than that of the third processing core, thereby ensuring that the fault analysis results are output first.

[0043] In this embodiment, the third processing core collects the load data of the first processing core in real time through a hard real-time kernel interface, and performs task scheduling based on the comparison result of the load data and a preset load threshold. If the load data of the first processing core is greater than the preset load threshold, the remaining first processing cores with the lowest load are retrieved from the substation type partitioning results, and some tasks in the first processing core load data are migrated to the remaining first processing cores with the lowest load. Similarly, when a first processing core fails, the remaining first processing cores with the lowest load are also retrieved from the substation type partitioning results, and the tasks of the failed first processing core are migrated to the remaining first processing cores with the lowest load. The preset load threshold can be 80%.

[0044] In other cases, the first processing cores in the same partition are sorted according to their load. When the load data of any first processing core exceeds a preset load threshold, load migration is performed sequentially. For example, if the load data of processing core A exceeds the preset load threshold, the load data of processing core B is 40%, the load data of processing core C is 20%, and the load data of processing core D is 0%, then the tasks of processing core A are migrated to processing core D first. However, if the load data of processing core B is 70%, the load data of processing core C is 60%, and the load data of processing core D is 50%, and migrating all the tasks of processing core A to processing core D would cause the load data of processing core D to exceed the preset load threshold, then processing core A is migrated to processing core B, processing core C, and then processing core D respectively, so that the load data of the three is balanced.

[0045] By collecting load data from the first processing core in real time through the hard real-time kernel interface, the non-real-time scheduling layer of the Linux kernel is bypassed, ensuring that the third processing core can quickly obtain the load data of the first processing core and improving the timeliness of task scheduling.

[0046] It is understandable that non-real-time tasks such as historical data backup and log data are handled in the ordinary Linux kernel.

[0047] The specific embodiments described above are preferred embodiments of the power grid fault analysis method and system based on multi-core task coordination in this application, and are not intended to limit the specific scope of this application. The scope of this application includes but is not limited to the specific embodiments described above. All equivalent changes made in accordance with the shape and structure of this application are within the protection scope of this application.

Claims

1. A method for power grid fault analysis based on multi-core task coordination, characterized in that: The method comprises the following steps: dividing the power grid area based on the transformer substation, each power grid area corresponding to a first processing core, all power grid areas corresponding to a second processing core and a third processing core; in the first processing core, obtaining the fault correlation relationship according to the power grid area operation data and the environment data; in the second processing core, obtaining the power grid fault analysis result according to the fault correlation relationship, the current environment data, the real-time power grid operation data and the fault threshold; in the third processing core, obtaining the division result according to the transformer substation type division of the first processing core, obtaining the real-time load of each first processing core, and performing the first processing core task scheduling based on the division result and the real-time load.

2. The power grid fault analysis method based on multi-core task coordination according to claim 1, wherein: obtaining the fault correlation relationship according to the power grid area operation data and the environment data comprises: obtaining the associated operation data set corresponding to the fault type according to the power grid area operation data and the environment data of the historical time sequence section; constructing the fault correlation relationship corresponding to the fault type based on the environment fluctuation frequency, the environment fluctuation factor and the environment fluctuation intensity according to the associated operation data set.

3. The power grid fault analysis method based on multi-core task coordination according to claim 2, wherein: constructing the fault correlation relationship corresponding to the fault type based on the environment fluctuation frequency, the environment fluctuation factor and the environment fluctuation intensity according to the associated operation data set comprises: extracting the environment fluctuation factor in the associated operation data set, obtaining the environment fluctuation frequency and the environment fluctuation intensity of each environment fluctuation factor according to the associated operation data of the same environment fluctuation factor to obtain the single-factor fluctuation weight; obtaining the multi-factor fluctuation weight of each environment fluctuation factor according to the associated operation data of different environment fluctuation factors and the single-factor fluctuation weight; constructing the fault correlation relationship corresponding to the fault type based on the single-factor fluctuation weight and the multi-factor fluctuation weight.

4. The power grid fault analysis method based on multi-core task coordination according to claim 1, wherein: obtaining the power grid fault analysis result according to the fault correlation relationship, the current environment data, the real-time power grid operation data and the fault threshold comprises: obtaining the corrected power grid operation data of each area according to the fault correlation relationship, the current environment data and the real-time power grid operation data; if there is corrected power grid operation data meeting the fault threshold, outputting the regional power grid fault warning.

5. The power grid fault analysis method based on multi-core task coordination according to claim 1, wherein: obtaining the division result according to the transformer substation type division of the first processing core, obtaining the real-time load of each first processing core, and performing the first processing core task scheduling based on the division result and the real-time load comprises: obtaining the division result according to the transformer substation type division of the first processing core; if there is a first processing core whose real-time load is greater than the preset load threshold, scheduling the task of the first processing core to the remaining first processing cores in the division result corresponding to the first processing core.

6. A power grid fault analysis system based on multi-core task coordination, for implementing the power grid fault analysis method based on multi-core task coordination as claimed in any one of claims 1 to 5, characterized in that: comprises: a data acquisition unit for obtaining power grid area operation data and environment data; The data processing unit comprises a first processing core, a second processing core and a third processing core, is used for outputting power grid fault analysis results according to data collected by the data collection unit, and performing task scheduling according to real-time load of the first processing core. 7.The power grid fault analysis system based on multi-core task coordination according to claim 6, wherein: The data collection unit comprises an FPGA data collection layer and a PCIe data transmission layer, the FPGA data collection layer collects power grid region operation data and environment data and stores them into a memory space, and the first processing core acquires the power grid region operation data and the environment data in the memory space through the PCIe data transmission layer. 8.The power grid fault analysis system based on multi-core task coordination according to claim 6, wherein: The data processing unit further comprises a hard real-time kernel, and the hard real-time kernel controls interruption of the first processing core, the second processing core and the third processing core. 9.The power grid fault analysis system based on multi-core task coordination according to claim 6, wherein: The priority of the second processing core is higher than that of the first processing core, and the priority of the first processing core is higher than that of the third processing core. 10.The power grid fault analysis system based on multi-core task coordination according to claim 7, wherein: The third processing core collects load data of the first processing core in real time through a hard real-time kernel interface, and performs task scheduling according to a comparison result of the load data and a preset load threshold.

Images