A new energy centralized control branch current anomaly detection method, device and medium

Abstract

The application relates to the technical field of new energy centralized control center safety detection, and discloses a new energy centralized control branch current anomaly detection method, device and medium. The application can effectively identify current changes caused by environmental factors by taking the environmental conditions at the current time as an index to obtain matching historical data from historical current data and calculate the similarity. The application can accurately judge the degree of deviation from the normal level by taking the average value of the current of the same group of branches as a reference to calculate the deviation degree and realize horizontal space comparison. The application can determine the abnormal duration and extract the morphological characteristics at the same time. Finally, the deviation degree, the abnormal duration and the morphological characteristics are respectively quantitatively mapped and weightedly fused, and whether the branch current is abnormal is determined according to the fusion result. The application effectively solves the problems in the prior art that the centralized control center relies on manual screen patrol for branch current anomaly detection, resulting in low efficiency, high false alarm rate of fixed threshold alarm, lack of horizontal comparison and root cause positioning ability.

Description

Technical Field

[0001] This invention relates to the field of safety detection technology for new energy centralized control centers, specifically to a method, device, and medium for detecting abnormal current in a new energy centralized control branch. Background Technology

[0002] As the scale of new energy bases continues to expand, the number of wind turbines and photovoltaic inverters under the jurisdiction of the central control center often reaches hundreds or even thousands, involving tens of thousands of branch current measurement points. In daily operation, due to factors such as equipment aging, loose wiring, component obstruction, and combiner box failure, abnormal branch currents occur frequently, such as branch current deviation, sudden current drop, and current interruption, which seriously affect power generation efficiency and equipment safety.

[0003] Currently, the central control center primarily relies on manual screen inspection and threshold alarms to monitor branch currents. For manual screen inspection, maintenance personnel must navigate through tree-like menus such as "station-inverter" or "wind turbine-branch," manually reviewing the current curves of thousands of branches. This is time-consuming, labor-intensive, and prone to missed detections due to visual fatigue. Regarding threshold alarms, the normal current baseline varies across different branches due to differences in micro-topography and dust obstruction. Using a uniform, fixed threshold, such as triggering an alarm when the current is below 20% of the average, easily leads to numerous false alarms, especially during periods of low irradiance in the early morning or late evening, or when there is localized cloud cover, resulting in an alarm storm.

[0004] Furthermore, existing systems only display a single value. When a branch current is abnormal, they cannot automatically correlate it with other branches in the same combiner box, adjacent inverters, or meteorological data. Maintenance personnel find it difficult to quickly distinguish whether the problem is due to environmental factors (such as cloud cover), a single branch fault, or a combiner box busbar fault. Additionally, the characteristics of abnormal wind turbine phase current and abnormal photovoltaic string current differ, and traditional systems lack a unified one-click inspection and intelligent judgment mechanism. Summary of the Invention

[0005] This invention provides a method, device, and medium for detecting abnormal branch current in new energy centralized control systems, in order to solve the problems in the prior art where the detection of abnormal branch current in centralized control centers relies on manual screen inspection, resulting in low efficiency, high false alarm rate of fixed threshold alarms, and lack of horizontal comparison and root cause localization capabilities.

[0006] In a first aspect, the present invention provides a method for detecting abnormal current in a new energy centralized control branch, comprising: acquiring current current data, historical current data, and current current data of branches in the same group as the branch to be detected, wherein branches in the same group refer to branches located in the same convergence unit as the branch to be detected; using the current environmental conditions as an index, acquiring matching historical data from the historical current data of the branch to be detected, and calculating the similarity between the current current data of the branch to be detected and the matching historical data; if the similarity is lower than a preset threshold, using the current current data of the branches in the same group as a benchmark, calculating the deviation of the current current data of the branch to be detected, determining the duration of the abnormality, and extracting morphological features; performing quantization mapping and weighted fusion on the deviation, the duration of the abnormality, and the morphological features, and determining whether the current data of the branch to be detected is abnormal based on the weighted fusion result.

[0007] This invention first acquires the current current data of the branch to be tested and its group branches, as well as the historical current data of the branch to be tested. By using the current environmental conditions as an index to obtain matching historical data from the historical current data and calculating the similarity, it can effectively identify current changes caused by environmental factors. If the similarity is lower than a preset threshold, it indicates that it is not affected by environmental factors. Then, the deviation is calculated based on the average current of the group branches, realizing horizontal spatial comparison and accurately judging the degree of deviation from the normal level. At the same time, the duration of the abnormality is determined and morphological features are extracted, providing a reliable data foundation for fault root cause analysis. Finally, the deviation, abnormal duration, and morphological features are quantified, mapped, and weighted and fused to form a multi-dimensional quantitative index that can comprehensively reflect the degree of deviation, duration, and fault type, and determine whether the branch current is abnormal. This invention forms a complete multi-dimensional detection system, effectively solving multiple technical problems such as low efficiency of traditional manual screen inspection, high false alarm rate of fixed threshold alarms, and lack of horizontal comparison and root cause location capabilities. It realizes automated and accurate detection and attribution diagnosis of branch current anomalies, significantly improving the inspection efficiency and intelligent operation and maintenance level of the central control center.

[0008] This invention can be applied to batch anomaly detection of multiple branches under the jurisdiction of a new energy central control center. After a one-click trigger, all branches of the entire station can be scanned in parallel. Maintenance personnel can directly obtain anomaly detection results without manually reviewing a large number of current curves. This reduces the inspection time of tens of thousands of measurement points in the entire station from several hours to seconds, significantly improving the efficiency of the central control center's screen inspection and the accuracy of anomaly diagnosis, and providing reliable technical support for the refined operation and maintenance of new energy power plants.

[0009] In one optional implementation, using the current environmental conditions as an index, matching historical data is obtained from the historical current data of the branch to be tested. This includes using historical current data from the historical current data of the branch to be tested that is the same as or similar to the current environmental conditions as the matching historical data. In this implementation, by using the current environmental conditions as an index to obtain matching historical data under the same or similar conditions from the historical current data, the similarity calculation is performed under the same or similar environmental conditions. This avoids the misjudgment of normal current fluctuations caused by changes in environmental factors such as irradiance and wind speed as equipment malfunctions. Environmental interference is eliminated at the source, significantly reducing the false alarm rate and effectively solving the problem of traditional fixed threshold alarms generating a large number of false alarms in the early morning and late evening when irradiance is low or when there is local cloud cover.

[0010] In one optional implementation, if the similarity is not lower than a preset threshold, it is determined to be due to environmental factors, and abnormal alarms are suppressed. In this implementation, when the similarity is not lower than the preset threshold, it is determined to be due to environmental factors and abnormal alarms are suppressed. By introducing an intelligent environmental factor recognition mechanism into the detection process, current fluctuations caused by environmental changes such as cloud cover and gusts are effectively distinguished from actual equipment faults. This avoids invalid alarms for normal environmental fluctuations, significantly reduces the false alarm rate, and eliminates the drawback of traditional threshold alarms that are prone to alarm storms when the environment changes. This allows maintenance personnel to focus on actual equipment anomalies, improving the accuracy and efficiency of fault handling.

[0011] In one optional implementation, the deviation of the current current data of the branch to be detected is calculated based on the current current data of the branches in the same group. This includes: using the average value of the current current data of the branches in the same group as a reference value; and calculating the degree of deviation between the current current data of the branch to be detected and the reference value to obtain the deviation. In this implementation, by using the average value of the current current data of the branches in the same group as a reference to calculate the deviation, the characteristics of each branch in the same converging unit being under the same environmental conditions and enduring the same irradiance and temperature are fully utilized. This allows the reference value to be adjusted in real time with environmental changes, avoiding false alarms caused by differences in the micro-topography and dust obstruction of different branches in traditional fixed thresholds. It can accurately identify faulty branches that truly deviate from the normal level of the same group in environmental fluctuations, significantly improving the accuracy and reliability of anomaly detection.

[0012] In one optional implementation, determining the abnormal duration includes: starting timing when the deviation exceeds a preset deviation threshold, stopping timing when the deviation returns to within the preset deviation threshold, and using the timing result as the abnormal duration. In this implementation, by using exceeding the preset deviation threshold as the starting condition and returning to within the threshold as the ending condition to determine the abnormal duration, the duration is quantified into an objective indicator. This provides a reliable data foundation for subsequent weighted fusion, avoids misjudgments caused by instantaneous fluctuations or occasional interference, and makes fault determination more accurate and reliable.

[0013] In one optional implementation, morphological features are extracted, including: extracting the curve morphological features of the current current data of the branch to be detected, wherein the curve morphological features include at least one of step drop, sawtooth fluctuation, and persistently low current. In this implementation, by extracting the curve morphological features of the current current data of the branch to be detected and specifying the morphological features into quantifiable and identifiable types such as step drop, sawtooth fluctuation, and persistently low current, automatic classification and identification of abnormal current patterns are achieved. This enables fault detection to no longer rely on a single numerical judgment, but to identify the fault type from the waveform morphology level. It can effectively distinguish different fault causes such as sudden disconnection, poor contact, and dust obstruction, providing maintenance personnel with accurate fault attribution basis. This avoids the limitation of traditional threshold alarms that only indicate anomalies but cannot locate the root cause, significantly improving the depth of anomaly diagnosis and the accuracy of maintenance decisions.

[0014] In one optional implementation, the deviation, abnormal duration, and morphological features are quantized, mapped, and then weighted and fused. This includes: using a Sigmoid function to perform dimensionless mapping on the deviation to obtain a deviation probability index; using an exponential function to perform dimensionless mapping on the abnormal duration to obtain a duration probability index; performing confidence-weighted mapping based on the type of morphological features to obtain a morphological probability index; and then weighting and fusing the deviation probability index, duration probability index, and morphological probability index. In this implementation, by quantizing and mapping the deviation, abnormal duration, and morphological features using the Sigmoid function, exponential function, and confidence-weighted methods respectively, and then weighting and fusing them, multidimensional features of different dimensions and types are uniformly mapped to the probability index space. This solves the problem that traditional single-threshold judgment cannot integrate multidimensional information, enabling the final comprehensive fault probability to fully reflect the joint contribution of multiple factors such as deviation degree, duration, and morphological type to branch current anomalies, thus achieving accurate quantitative assessment of abnormal states. Meanwhile, the Sigmoid function and the exponential function can map the changes in deviation and duration to the probability interval and reflect their saturation effect. Confidence weighting can assign differentiated weights according to the degree of certainty of the fault mechanism corresponding to different morphological features. The three work together to achieve scientific and reasonable feature fusion, which significantly improves the accuracy and reliability of anomaly detection.

[0015] In one optional implementation, determining whether the current data of the branch to be detected is abnormal based on the weighted fusion result includes: obtaining the comprehensive fault probability of the branch to be detected based on the weighted fusion result; if the comprehensive fault probability is greater than or equal to a first threshold, it is determined to be a severe anomaly; if the comprehensive fault probability is greater than or equal to a second threshold and less than the first threshold, it is determined to be a general anomaly; if the comprehensive fault probability is greater than or equal to a third threshold and less than the second threshold, it is determined to be a minor anomaly. This implementation obtains the comprehensive fault probability based on the weighted fusion result and compares it with the first, second, and third thresholds to classify the detection result into three levels: severe anomaly, general anomaly, and minor anomaly. This achieves refined classification and discrimination of the severity of anomalies, enabling maintenance personnel to quickly formulate differentiated handling strategies based on the anomaly level.

[0016] Secondly, the present invention provides a new energy centralized control branch current anomaly detection device, comprising: a data acquisition module, used to acquire the current current data, historical current data, and current current data of branches in the same group as the branch to be detected; a similarity calculation module, used to acquire matching historical data from the historical current data of the branch to be detected using the current environmental conditions as an index, and calculate the similarity between the current current data of the branch to be detected and the matching historical data; an anomaly index extraction module, used to calculate the deviation of the current current data of the branch to be detected based on the current current data of the branches in the same group if the similarity is lower than a preset threshold, determine the duration of the anomaly, and extract morphological features; and an anomaly determination module, used to perform quantized mapping and weighted fusion of the deviation, the duration of the anomaly, and the morphological features, and determine whether the current data of the branch to be detected is abnormal based on the weighted fusion result.

[0017] Thirdly, the present invention provides a computer-readable storage medium storing computer instructions, which are used to cause a computer to execute the new energy centralized control branch current anomaly detection method of the first aspect or any corresponding embodiment described above. Attached Figure Description

[0018] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0019] Figure 1 This is a schematic diagram of the first process of the new energy centralized control branch current abnormality detection method according to an embodiment of the present invention; Figure 2 This is a schematic diagram of the second process of the new energy centralized control branch current abnormality detection method according to an embodiment of the present invention; Figure 3 This is a structural block diagram of a new energy centralized control branch current abnormality detection device according to an embodiment of the present invention; Figure 4 This is a schematic diagram of the hardware structure of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] It is understood that before using the technical solutions disclosed in the various embodiments of the present invention, users should be informed of the types, scope of use, and usage scenarios of the personal information involved in the present invention and their authorization should be obtained in accordance with relevant laws and regulations through appropriate means.

[0022] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0023] As the scale of new energy bases continues to expand, the number of wind turbines and photovoltaic inverters under the jurisdiction of the central control center often reaches hundreds or thousands, involving tens of thousands of branch current measurement points. Affected by factors such as equipment aging, loose wiring, component obstruction, and combiner box failure, abnormal branch currents occur frequently, seriously affecting power generation efficiency and equipment safety. Currently, the central control center mainly relies on manual screen inspection and threshold alarms to monitor branch currents. Maintenance personnel need to manually browse through thousands of branch current curves by clicking through the tree menu, which is time-consuming, labor-intensive, and prone to missed detections due to visual fatigue. As for the fixed threshold alarm method, the normal current baseline varies due to differences in the micro-topography and dust obstruction of different branches. A uniform threshold is prone to generating a large number of false alarms when the irradiance is low in the morning and evening or when there is local cloud cover, which can trigger an alarm storm. At the same time, the existing system only displays a single value. When the branch current is abnormal, it cannot automatically correlate with other branches in the same combiner box, adjacent inverters, or meteorological data. Maintenance personnel have difficulty quickly distinguishing whether it is an environmental factor, a single branch fault, or a combiner box busbar fault. In addition, the characteristics of abnormal wind turbine phase current and abnormal photovoltaic string current are different. Traditional systems lack a unified one-click inspection and intelligent judgment mechanism. Based on this, the present invention provides a method, device and medium for detecting abnormal branch current in new energy centralized control, in order to solve the problems of low efficiency, high false alarm rate of fixed threshold alarm and lack of horizontal comparison and root cause location capabilities in the existing technology where the centralized control center relies on manual screen inspection for abnormal branch current detection.

[0024] According to an embodiment of the present invention, a method for detecting abnormal current in a new energy centralized control branch is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0025] This embodiment provides a method for detecting abnormal current in a new energy centralized control branch. Figure 1 This is a flowchart of a new energy centralized control branch current anomaly detection method according to an embodiment of the present invention, such as... Figure 1 As shown, the process includes the following steps: Step S101: Obtain the current current data, historical current data, and current current data of the branch to be tested, as well as the current current data of the branches in the same group. The branches in the same group refer to the branches that are in the same converging unit as the branch to be tested.

[0026] This step is used to read real-time data from the data acquisition and monitoring system of the central control center, as well as historical stored data from the historical database. This can be achieved through data interface calls or database queries.

[0027] The branch to be tested refers to any branch within the jurisdiction of the new energy control center that needs to be tested for abnormal current. For photovoltaic power generation scenarios, the branch refers to the photovoltaic string branch, and for wind power generation scenarios, the branch refers to the wind turbine phase current branch.

[0028] Current current data refers to real-time current data collected at the current moment or within the current time period, reflecting the actual current value of the branch under the current operating state.

[0029] Historical current data refers to current data stored within a time period prior to the current moment, and is stored in the historical database of the central control center. The purpose of obtaining the historical current data of the branch to be tested is to provide a reference benchmark for longitudinal comparison in subsequent steps.

[0030] Both current data and historical current data are time-series data, that is, data sequences with time as the horizontal axis and current value as the vertical axis. They are used to characterize the changing trend of branch current over time, providing a data foundation for subsequent similarity calculation, deviation calculation and morphological feature extraction.

[0031] A branch in the same group refers to a branch that is located in the same aggregation unit as the branch under test. In a photovoltaic power generation scenario, the aggregation unit is a combiner box, meaning that all photovoltaic string branches under the same combiner box are branches in the same group. In a wind power generation scenario, it is a converter, meaning that the arm branches under the same converter are branches in the same group. The purpose of setting up branches in the same group is that branches under the same aggregation unit operate under the same environmental and electrical conditions, and their current output has theoretical similarity, thus serving as a horizontal comparison reference for the branch under test.

[0032] It should be noted that the data acquired in step S101 is the foundation for all subsequent calculations and analyses, and its accuracy and completeness directly affect the reliability and precision of the entire detection method. Furthermore, the acquisition operation in step S101 can be automatically executed by the data acquisition module of the central control center, or it can be executed in response to a one-click inspection command triggered by maintenance personnel, thereby achieving batch acquisition of all station branch data.

[0033] Step S102: Using the current environmental conditions as an index, obtain matching historical data from the historical current data of the branch to be detected, and calculate the similarity between the current current data of the branch to be detected and the matching historical data.

[0034] The environmental conditions in this step refer to the external environmental parameters that affect the branch current output. For photovoltaic power generation scenarios, environmental conditions include irradiance, and for wind power generation scenarios, environmental conditions include wind speed. When irradiance or wind speed changes, the branch current will fluctuate accordingly. This fluctuation is a normal physical phenomenon and not a equipment failure.

[0035] The environmental conditions at the current moment refer to the environmental parameter values ​​collected at the current moment or within the current time period, which serve as the index for retrieving historical data.

[0036] Using the current environmental conditions as an index means using the current environmental conditions as the retrieval basis to search for historical records with the same or similar environmental conditions in the historical database. The purpose is to ensure that the matching historical data and the current current data are under the same environmental conditions, thereby avoiding the misjudgment of current fluctuations caused by differences in environmental factors as abnormal.

[0037] Historical matching data refers to the current data retrieved from the historical current data of the branch under test that is consistent with or similar to the current environmental conditions. This data serves as a reference for the normal operating status of the branch under test under the same environmental conditions.

[0038] Similarity is used to quantify the degree of consistency between the current current and historical normal currents under the same environmental conditions. The purpose of this step is to identify normal current fluctuations caused by environmental changes through environmental condition indexing and similarity calculation before subsequent fault diagnosis, thereby distinguishing them from actual equipment faults in subsequent steps and reducing the false alarm rate.

[0039] It should be noted that the current current data and the matching historical data in step S102 are both time series data. The similarity calculation is essentially a quantitative evaluation of the degree of matching between the two sets of time series data in terms of waveform morphology. It can be implemented using time series similarity algorithms such as dynamic time warping. The specific algorithm does not constitute a limitation on this step.

[0040] Step S103: If the similarity is lower than the preset threshold, the deviation of the current current data of the branch to be detected is calculated based on the current current data of the branch in the same group, the duration of the abnormality is determined, and the morphological features are extracted.

[0041] In this step, a preset threshold is used to determine whether the current current change is caused by environmental factors. If the similarity is lower than the threshold, it indicates that the waveform of the current current data differs significantly from that of the historical data, meaning the current current change is not caused by environmental factors but may be due to a fault in the branch itself. In this case, the subsequent fault detection process needs to be initiated. Conversely, if the similarity is not lower than the threshold, it indicates that the current current change is highly correlated with environmental factors and is a normal fluctuation, requiring no fault detection. This judgment step effectively separates environmental factors from equipment faults.

[0042] This step uses the current current data of the same group of branches as the reference standard for horizontal comparison. Since the branches in the same group are located in the same converging unit and are subjected to the same environmental and electrical conditions, their current data are comparable. By using the concentration trend of the current in the same group of branches as the benchmark, the influence of environmental factors on the current of a single branch can be eliminated.

[0043] Deviation refers to the degree of deviation of the current data of the branch under test from the current reference value of the same group of branches. It is used to quantify the current difference between the branch under test and other branches in the same group. The greater the deviation, the more significant the deviation of the branch from the same group of branches.

[0044] The abnormal duration refers to the length of time that the current of the branch under test deviates from the normal range relative to the branches in the same group. It is used to reflect the duration of the abnormal state. Incorporating the duration into the subsequent fusion calculation can avoid occasional misjudgments caused by instantaneous fluctuations.

[0045] Morphological features refer to the waveform shape characteristics of the current data of the branch under test as it changes over time. Different types of faults will result in different current waveform shapes. For example, a sudden wire break corresponds to a step drop, poor contact corresponds to a sawtooth fluctuation, and dust obstruction corresponds to a persistently low current. The purpose of extracting morphological features is to provide a classification basis for subsequent fault attribution analysis, so that anomaly detection can not only determine whether there is an anomaly in the branch, but also identify the specific type of anomaly.

[0046] It should be noted that in step S103, the calculation of deviation, the determination of abnormal duration, and the extraction of morphological features are all operations performed in parallel or sequentially based on the same group of branch references. They respectively characterize the branch current from three dimensions: amplitude, time, and waveform, providing a data basis for the multi-dimensional feature fusion in step S104.

[0047] Step S104: After quantizing and mapping the deviation, abnormal duration and morphological characteristics, the data are weighted and fused. Based on the weighted fusion result, it is determined whether the current data of the branch to be detected is abnormal.

[0048] Since deviation, abnormal duration, and morphological features belong to different data dimensions, with deviation being a dimensionless relative value, abnormal duration being a time-based measure, and morphological features being a categorical variable, the three have different dimensions and cannot be directly integrated for calculation. Therefore, it is necessary to unify the three to the same measurement scale through quantization mapping.

[0049] The weighted and fused comprehensive quantitative value is then used as the basis for judgment. By comparing it with the preset judgment conditions, it is determined whether the current data of the branch to be detected is in an abnormal state.

[0050] The purpose of step S104 is to integrate the features extracted from the three different dimensions in step S103 by unifying their dimensions, and form a comprehensive judgment index that fully reflects the branch current state, thereby achieving accurate quantitative assessment of abnormal branch current states.

[0051] It should be noted that there are various ways to implement quantization mapping and weighted fusion. For example, quantization mapping can be done using function mapping, lookup table mapping, or confidence weighting mapping, while weighted fusion can be done using linear weighting, nonlinear weighting, or enhanced weighting. This step only limits the overall concept of weighted fusion after quantization mapping, without specifying the specific mapping function or weight configuration. The specific implementation method can be flexibly selected according to the actual application scenario, and all of them can achieve the technical effect of comprehensive judgment after unifying the dimensions of multi-dimensional features.

[0052] It is worth noting that when it is necessary to perform batch anomaly detection on multiple branches under the jurisdiction of the New Energy Central Control Center, each branch is treated as the branch to be detected, and the above steps S101 to S104 are executed to obtain the anomaly detection results of each branch.

[0053] The new energy centralized control branch current anomaly detection method provided in this embodiment first acquires the current data of the branch to be detected and its group of branches, as well as the historical current data of the branch to be detected. By using the current environmental conditions as an index to obtain matching historical data from the historical current data and calculating the similarity, it can effectively identify current changes caused by environmental factors. If the similarity is lower than a preset threshold, it indicates that it is not affected by environmental factors. Then, the deviation is calculated based on the average current of the group of branches, realizing horizontal spatial comparison and accurately judging the degree of deviation from the normal level. At the same time, the duration of the anomaly is determined and morphological features are extracted, providing a reliable data foundation for fault root cause analysis. Finally, the deviation, the duration of the anomaly, and the morphological features are quantified, mapped, and weighted and fused to form a multi-dimensional quantitative index that can comprehensively reflect the degree of deviation, duration, and fault type, and determine whether the branch current is abnormal based on this index. This invention forms a complete multi-dimensional detection system, which effectively solves multiple technical problems such as low efficiency of traditional manual screen inspection, high false alarm rate of fixed threshold alarm, and lack of horizontal comparison and root cause location capabilities. It realizes automated and accurate detection and attribution diagnosis of branch current anomalies, and significantly improves the inspection efficiency and intelligent operation and maintenance level of the central control center.

[0054] This embodiment provides another method for detecting abnormal current in the centralized control branch of a new energy source. Figure 2 This is a flowchart of a new energy centralized control branch current anomaly detection method according to an embodiment of the present invention, such as... Figure 2 As shown, the process includes the following steps: Step S201: Obtain the current current data, historical current data, and current current data of the branch to be tested, as well as the current current data of the branches in the same group. The branches in the same group refer to the branches that are in the same converging unit as the branch to be tested.

[0055] In this invention, the aggregation unit refers to an electrical node that aggregates the currents from multiple branches and outputs the current. In a photovoltaic power generation scenario, the aggregation unit is a combiner box, and the branches in the same group are the photovoltaic string branches under the same combiner box; in a wind power generation scenario, the aggregation unit is a converter, and the branches in the same group are the bridge arm branches under the same converter.

[0056] Step S202: Using the current environmental conditions as an index, obtain matching historical data from the historical current data of the branch to be detected, and calculate the similarity between the current current data of the branch to be detected and the matching historical data.

[0057] In this step, a longitudinal time comparison mechanism is used to retrieve matching historical data from the historical current data of the branch under test, using the current environmental conditions as an index. The similarity between the current current data and the matching historical data is then calculated, enabling self-calibration between the branch under test and its own historical health status. When the similarity is not lower than a preset threshold, it indicates that the current current change is highly correlated with environmental factors, and the issue is determined to be influenced by environmental factors, thus suppressing abnormal alarms. When the similarity is lower than the preset threshold, the subsequent fault detection process begins.

[0058] Specifically, step S202 above includes: Step S2021: Use historical current data from the branch to be tested that are the same as or similar to the current environmental conditions as the historical current data.

[0059] In one optional implementation, meteorological data from one hour prior to the current time is used as an index to retrieve branch current data under similar meteorological conditions from a historical database, which is then used as matching historical data. The aforementioned "similar meteorological conditions" refer to irradiance deviation ≤ 5% and wind speed deviation ≤ 5%. Using meteorological data from one hour prior as an index ensures that the retrieved historical data has a high environmental similarity to the current state, thereby improving the accuracy of similarity calculation.

[0060] Step S2022: Calculate the similarity between the current current data of the branch to be detected and the historical matching data.

[0061] After calculating the similarity in step S202, the relationship between the similarity and the preset threshold is determined. If the similarity is not lower than the preset threshold, it is determined that the problem is due to environmental factors, and the abnormal alarm is suppressed; otherwise, steps S203 and S204 are executed.

[0062] In one optional implementation, if the similarity is not lower than a preset threshold, the following verification steps are performed before suppressing the abnormal alarm: When multiple inverter branch currents in the same area are detected to decrease simultaneously and proportionally, real-time irradiance or wind speed data from the local meteorological station are automatically imported. If the correlation coefficient is greater than 0.8, it is determined to be caused by environmental factors such as cloud cover or gusts, and the alarm is automatically suppressed and marked as "environmental impact," without outputting an abnormal equipment alarm. This implementation, through multi-source data correlation analysis, can effectively identify the phenomenon of simultaneous decrease caused by collective environmental factors, avoiding a large number of false alarms.

[0063] Step S203: If the similarity is lower than the preset threshold, the deviation of the current current data of the branch to be detected is calculated based on the current current data of the branch in the same group, the duration of the abnormality is determined, and the morphological features are extracted.

[0064] Specifically, step S203 above includes: Step S2031: Using the current current data of the same group of branches as a benchmark, calculate the deviation of the current current data of the branch to be detected.

[0065] In one optional implementation, step S2031 includes: Step a1: Use the average value of the current current data of the same branch as the reference value.

[0066] In one optional implementation, for photovoltaic power generation scenarios, the benchmark value is the average value after removing the maximum and minimum current values ​​within the same combiner box; for wind power generation scenarios, the benchmark value is the average value of the current in each phase of the same converter. By removing extreme values ​​and taking the average, interference caused by extreme values ​​of a single branch on the benchmark value can be avoided, making the deviation calculation more accurate and reliable.

[0067] Step a2: Calculate the degree of deviation between the current current data of the branch to be tested and the reference value, and obtain the deviation degree.

[0068] It should be noted that the lateral spatial comparison mechanism used in this step extracts the real-time average current of each branch under the same aggregation unit as the benchmark value, calculates the deviation of the current of the branch to be detected from the benchmark value, and if a single branch deviates significantly from other branches in the same group, it is marked as having a suspected fault, thus achieving mutual calibration within the same source. This mechanism utilizes the characteristic that each branch within the same aggregation unit is under the same environmental and electrical conditions, which can effectively eliminate the interference of environmental factors on the judgment of anomalies in a single branch.

[0069] Step S2032: Determine the duration of the anomaly based on the deviation.

[0070] In one optional implementation, the method for determining the abnormal duration based on the deviation is as follows: when the deviation exceeds a preset deviation threshold, timing begins; when the deviation returns to within the preset deviation threshold, timing stops; and the timing result is taken as the abnormal duration.

[0071] Step S2033: Extract the morphological features of the current current data of the branch to be detected.

[0072] In one alternative implementation, the morphological feature in this embodiment refers to the curve morphological feature, specifically including at least one of step drop, sawtooth fluctuation and low persistence.

[0073] It should be noted that the morphological recognition extraction mechanism used in this step extracts the micro-features of the current current curve of the branch to be detected. The micro-features include at least one of step drop, sawtooth fluctuation and persistent low current, which correspond to different fault types such as sudden wire breakage, poor contact or arcing, dust obstruction or component attenuation, and form a feature vector to realize automatic attribution and identification of abnormal types.

[0074] In one alternative implementation, persistently low current means that the current data of the branch under test is consistently lower than the reference value of the current data of the same group of branches, corresponding to cumulative faults such as dust obstruction or component degradation.

[0075] Step S204: After quantizing and mapping the deviation, abnormal duration and morphological characteristics, the data are weighted and fused. Based on the weighted fusion result, it is determined whether the current data of the branch to be detected is abnormal.

[0076] In this embodiment, before weighted fusion after quantization mapping, a fault probability calculation formula is pre-constructed. The core idea of ​​this formula is as follows: for continuous numerical deviation and abnormal duration, the Sigmoid function and exponential function are used to nonlinearly map them to the [0,1] probability interval, which not only eliminates the influence of dimensions, but also conforms to the physical objective law that "the more obvious the feature, the more the probability tends to saturate"; for categorical morphological feature vectors, a morphological confidence weighting method is introduced, and the three morphologies of step drop, sawtooth fluctuation and persistent low are assigned different confidence weights according to the certainty of their corresponding physical faults, and combined with the morphological recognition probability, they are transformed into scalar indicators in the [0,1] interval.

[0077] In the fusion calculation, considering the coupled enhancement effect between spatial deviation and temporal duration in electrical faults—that is, the greater the deviation and the longer the duration, the more exponentially (but not nonlinearly) the fault certainty increases—this invention does not employ simple linear weighting. Instead, it constructs an enhanced weighted model that includes the interactive product of deviation probability and duration probability, as well as the interactive product of morphological probability and duration probability, to scientifically and accurately calculate the comprehensive fault probability. The following details the specific implementation of this fusion calculation and hierarchical determination in steps S2041 and S2042.

[0078] Specifically, step S204 above includes: Step S2041: After quantizing and mapping the deviation, abnormal duration, and morphological features, the data are weighted and fused.

[0079] In one optional implementation, step S2041 includes: Step b1: Use the Sigmoid function to perform dimensionless mapping on the deviation to obtain the deviation probability index.

[0080] In step b1 above, the Sigmoid function is used to perform a dimensionless mapping of the deviation. Specifically, let the current of the branch to be detected be I, and the average value of the reference in the same group be I. base Then the original deviation D raw =|II base | / I baseThe deviation probability index SD is obtained by using the sigmoid function to perform dimensionless mapping to the [0,1] interval: ; Wherein, SD is the deviation probability index; k1 is the curve steepness coefficient (controlling the sensitivity of deviation to probability growth); and D0 is the baseline deviation threshold (e.g., set to 0.2, meaning that the probability reaches 0.5 when the deviation is 20%).

[0081] Since the greater the deviation, the more obvious the fault, but when the deviation exceeds a certain threshold, its contribution to the fault probability will tend to saturate. For example, deviations of 50% and 80% are both highly indicative of faults. Therefore, using the Sigmoid function can conform to the physical objective law that "the more obvious the feature, the more the probability tends to saturate".

[0082] Step b2 involves using an exponential function to perform a dimensionless mapping of the abnormal duration to obtain a duration probability index.

[0083] In step b2 above, an exponential function is used to perform a dimensionless mapping of the anomaly duration, resulting in the duration probability index ST. Specifically, let the anomaly duration be T. raw The unit is minutes. The longer the time, the higher the probability of failure, but there is also a saturation effect. Therefore, the offset exponential function is used to perform dimensionless mapping to the [0,1] interval. Where k2 is the time decay coefficient. The longer the time, the higher the probability of failure, but there is also a saturation effect. Therefore, using an exponential function can reflect the diminishing marginal contribution of duration to the failure probability. It should be noted that k2 can be adjusted according to different anomaly types: for step drops, k2 takes a larger value, and the failure can be confirmed in a short time; for low persistence, k2 takes a smaller value, and confirmation takes a longer time.

[0084] Step b3: Perform confidence weighting mapping based on the type of morphological features to obtain morphological probability index.

[0085] In step b3 above, confidence levels are weighted and mapped according to the type of morphological features. Specifically, morphological features are discrete categorical variables and cannot be directly substituted into the formula for calculation. Therefore, based on the severity and physical mechanism of the fault corresponding to different morphological features, a morphological confidence vector M=[C step C saw C cont The system outputs the matching probability P for three patterns: step drop, sawtooth fluctuation, and low persistence, through pattern recognition. step P saw and P cont And P step +P saw +Pcont =1.

[0086] Among them, the step-type drop (step) corresponds to a sudden disconnection or short circuit, with a clear fault mechanism and a confidence level (C) in the assigned form. step =0.95; Sawtooth-shaped fluctuations correspond to poor contact or arcing; the fault mechanism is relatively clear, and the confidence level of the assigned shape is C. saw =0.75; Low persistence (cont): corresponds to dust obstruction or attenuation, possibly affected by environmental interference; assigned a morphological confidence level (C). cont =0.45.

[0087] The comprehensive probability index SM=P for calculating morphological features step ·C step +P saw ·C saw +P cont ·C cont This scalarizes the morphological vector into a value between [0,1].

[0088] Step b4 involves weighted fusion of the deviation probability index, duration probability index, and morphology probability index.

[0089] In step b4 above, the deviation probability index SD, the duration probability index ST, and the comprehensive probability index SM of morphological features are weighted and fused. Specifically, conventional linear weighting is prone to missing detection when the deviation is low but the morphological features are extremely poor, such as a slight step drop with a very short duration. Therefore, this invention adopts an enhanced weighting model based on spatiotemporal coupling, using deviation as the base probability, morphological features as a correction factor, and duration as a confirmation enhancement factor to calculate the comprehensive failure probability P. fault : P fault =α·SD+β·SM+γ·(SD·ST)+δ·(SM·ST), where α is the deviation weight, β is the morphological weight, γ is the coupling weight between deviation and time, representing the physical fact that "the greater the deviation and the longer the duration, the higher the probability of failure"; δ is the coupling weight between morphological features and time, representing that "the longer the adverse morphology lasts, the higher the certainty"; and α+β+γ+δ=1.

[0090] In one embodiment, α=0.4, β=0.3, γ=0.2, and δ=0.1.

[0091] Step S2042: Determine whether the current data of the branch to be detected is abnormal based on the weighted fusion result.

[0092] In an optional implementation, the abnormality of the branch current data to be detected is determined solely based on the weighted fusion result. In this case, step S2042 includes: Step c1: Obtain the comprehensive fault probability of the branch to be detected based on the weighted fusion result.

[0093] The weighted fusion result above is the comprehensive fault probability, which is used to characterize the likelihood of a fault occurring in the branch to be detected.

[0094] Step c2: If the overall failure probability is greater than or equal to the first threshold, it is determined to be a serious anomaly.

[0095] Step c3: If the overall failure probability is greater than or equal to the second threshold and less than the first threshold, it is determined to be a general anomaly.

[0096] Step c4: If the overall failure probability is greater than or equal to the third threshold and less than the second threshold, it is determined to be a minor anomaly.

[0097] In another optional implementation, the current data of the branch to be detected is determined to be abnormal based on the weighted fusion result and the principal component of the morphological features. In this case, step S2042 includes: Step d1: Obtain the comprehensive fault probability of the branch to be detected based on the weighted fusion result; Step d2: If the overall failure probability is greater than or equal to the first threshold and the principal component of the morphological feature is a step drop, then it is judged as a serious anomaly. Step d3: If the overall failure probability is greater than or equal to the second threshold and less than the first threshold, and the principal component of the morphological feature is sawtooth fluctuation, then it is judged as a general anomaly. Step d4: If the overall failure probability is greater than or equal to the third threshold and less than the second threshold, and the principal component of the morphological feature is persistently low, then it is judged as a minor anomaly.

[0098] For example, the method for determining the principal components of morphological features is as follows: Compare P step P saw and P cont The size of the shape is used to determine the principal component of the morphological feature, with the highest matching probability. If P step At its highest, the main component of the morphological feature is a step-like drop; If P saw At its highest, the principal component of the morphological feature is a sawtooth-shaped fluctuation; If P cont If the highest value is found, then the principal component of the morphological features is persistently low.

[0099] Then, based on the calculated overall failure probability P fault For each ∈[0,1], a hierarchical determination is performed: If P fault If the value is ≥0.8 and the principal component of the morphological feature is a step drop, it is judged as a serious fault, corresponding to a sudden interruption or short circuit fault, which requires immediate on-site handling. If 0.5≤P fault If the value is less than 0.8 and the principal component of the morphological characteristic is sawtooth fluctuation, it is judged as a general abnormality, corresponding to poor contact or arcing fault, and needs to be inspected at a later date. If 0.2≤P fault If the value is less than 0.5 and the principal component of the morphological feature is consistently low, it is considered a minor anomaly, corresponding to dust obstruction or component attenuation fault, and should be recorded and monitored. If P fault If the value is less than 0.2, it is considered to be in a normal state and no alarm will be output.

[0100] In an optional implementation, the method provided in this embodiment can be triggered by a one-click trigger button on the "Intelligent Screen Patrol" feature set on the central control center monitoring interface. When maintenance personnel click this button, the system automatically calculates the branch data of all wind turbines and inverters in the entire station in parallel in the background, and each branch is treated as the branch to be detected, executing the above steps S201 to S204. This one-click trigger mechanism allows maintenance personnel to trigger the detection of abnormal branch currents throughout the entire station without having to manually browse through the data of each branch by clicking through the tree menu.

[0101] In one optional implementation, after the anomaly determination of all branches to be tested is completed in step S204, the monitoring screen in the central control center automatically drills down from the global map to the topology map of the abnormal devices, hides the normal branches in grayscale, and highlights the abnormal branches in red, orange, and yellow according to their severity. The current curve, comparison curve of the same group, and meteorological correlation curve of the abnormal branch are automatically displayed. Red corresponds to severe anomalies, orange to general anomalies, and yellow to minor anomalies. Maintenance personnel can click on any highlighted branch to view its multi-dimensional comparison curve and detailed determination information.

[0102] In one optional implementation, after the abnormal topology perspective display is completed, the system automatically generates an inspection report, which includes a list of abnormal devices, the type of abnormality, and an estimated power loss. The inspection report is then pushed to the mobile maintenance APP in the form of a work order. Maintenance personnel can view the inspection results and perform corresponding on-site handling or maintenance tasks through their mobile terminals.

[0103] In one optional implementation, after all the branches to be tested have been anomaly identified, the central control center's monitoring screen displays the results using a "grayscale hiding + color highlighting" mechanism: normal branch nodes are hidden in grayscale, only abnormal branch nodes are retained, and they are marked with different colors according to the anomaly level. Severe anomalies are displayed with flashing red, general anomalies with solid orange, and minor anomalies with yellow. Maintenance personnel can intuitively view the distribution of anomalies across the entire station through the monitoring screen, and clicking on any highlighted branch will allow them to view the branch's current curve, group comparison curve, and meteorological correlation curve.

[0104] The method for detecting abnormal current in the centralized control branch of new energy sources provided in this embodiment has the following beneficial effects: This invention is the first to propose an intelligent patrol screen interaction mode of "one-click triggering - full-site parallel scanning - abnormal topology perspective" in the central control center, which transforms the traditional "human searching for data" into "data finding people". Maintenance personnel can trigger automatic inspection of all branches in the entire station with a single click, without having to navigate through a tree-like menu, greatly improving inspection efficiency. At the same time, this invention is the first to combine DTW historical similarity with source spatial comparison and apply it to the identification of abnormal current in wind and solar branches. Through the collaborative mechanism of longitudinal time comparison and horizontal spatial comparison, it achieves multi-dimensional and accurate identification of branch current anomalies.

[0105] This invention addresses the pain point of high false alarm rates at new energy power stations by innovatively introducing real-time regional meteorological data as a condition for preventing false alarms. By comparing similarity with a preset threshold and verifying it in conjunction with regional meteorological data, it effectively isolates normal fluctuations caused by environmental factors such as cloud cover, solving the problem of alarm storms easily triggered by traditional fixed threshold alarms when the environment changes. At the same time, this invention, based on an attribution and classification mechanism of current morphology characteristics, extracts three morphological characteristics—step drop, sawtooth fluctuation, and persistently low current—and assigns confidence weights to them, achieving a leap from "detecting anomalies" to "diagnosing the cause," providing maintenance personnel with clear fault types and handling directions.

[0106] This invention significantly reduces the workload of centralized control personnel in patrolling the screens, shortening the inspection time for tens of thousands of measurement points across the entire station from several hours to seconds, effectively avoiding missed detections due to human visual fatigue. At the same time, through precise three-level classification and work order push, it avoids ineffective on-site inspections, enabling maintenance personnel to rationally allocate resources and formulate differentiated handling strategies based on the anomaly level. This invention has extremely high practical value for improving the level of refined operation and maintenance and the power recovery rate of new energy power stations.

[0107] This embodiment also provides a new energy centralized control branch current abnormality detection device, which is used to implement the above embodiments and preferred embodiments, and will not be repeated as already described. As used below, the term "module" can be a combination of software and / or hardware that implements a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0108] This embodiment provides a new energy centralized control branch current abnormality detection device, such as... Figure 3 As shown, it includes: The data acquisition module 301 is used to acquire the current current data, historical current data and current current data of the branch to be tested as well as the current current data of the branch in the same group. The branch in the same group refers to the branch that is in the same converging unit as the branch to be tested. The similarity calculation module 302 is used to obtain matching historical data from the historical current data of the branch to be detected, using the current environmental conditions as an index, and calculate the similarity between the current current data of the branch to be detected and the matching historical data. The abnormal index extraction module 303 is used to calculate the deviation of the current current data of the branch to be detected based on the current current data of the same group of branches if the similarity is lower than the preset threshold, determine the duration of the abnormality, and extract the morphological features. The anomaly determination module 304 is used to perform quantization mapping and weighted fusion of deviation, anomaly duration and morphological characteristics, and determine whether the current data of the branch to be detected is abnormal based on the weighted fusion result.

[0109] In some alternative implementations, the similarity calculation module 302 includes: The historical data matching unit is used to select historical current data of the branch to be tested that are the same as or similar to the environmental conditions at the current moment as the matching historical data. The similarity calculation unit is used to calculate the similarity between the current current data of the branch to be detected and the matching historical data.

[0110] In some optional implementations, the anomaly indicator extraction module 303 includes: The deviation calculation unit is used to calculate the deviation of the current current data of the branch to be detected based on the current current data of the same group of branches; An abnormal duration determination unit is used to determine the abnormal duration based on the deviation. The morphological feature extraction unit is used to extract the morphological features of the current current data of the branch to be detected.

[0111] In some optional implementations, the anomaly determination module 304 includes: Anomaly index fusion unit is used to quantitatively map and weight fusion the deviation, anomaly duration, and morphological features; The anomaly detection unit is used to determine whether the current data of the branch to be detected is abnormal based on the weighted fusion result.

[0112] The new energy centralized control branch current anomaly detection device provided in this embodiment of the invention can execute the new energy centralized control branch current anomaly detection method provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the method. Further functional descriptions of the above modules and units are the same as in the corresponding embodiments described above, and will not be repeated here.

[0113] Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention.

[0114] The following is a detailed reference. Figure 4 This diagram illustrates a structural schematic suitable for implementing an electronic device according to embodiments of the present invention. The electronic device may include a processor (e.g., a central processing unit, graphics processor, etc.) 401, which can perform various appropriate actions and processes according to a program stored in read-only memory (ROM) 402 or a program loaded from memory 408 into random access memory (RAM) 403. The RAM 403 also stores various programs and data required for the operation of the electronic device. The processor 401, ROM 402, and RAM 403 are interconnected via a bus 404. An input / output (I / O) interface 405 is also connected to the bus 404.

[0115] Typically, the following devices can be connected to I / O interface 405: input devices 406 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 407 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; memory devices 408 including, for example, magnetic tapes, hard disks, etc.; and communication devices 409. Communication device 409 allows electronic devices to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 4 Electronic devices with various devices are shown, but it should be understood that it is not required to implement or have all of the devices shown, and more or fewer devices may be implemented or have instead.

[0116] In particular, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device 409, or installed from a memory 408, or installed from a ROM 402. When the computer program is executed by the processor 401, it performs the functions defined in the new energy centralized control branch current anomaly detection method of the present invention.

[0117] Figure 4 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of use of the embodiments of the present invention.

[0118] This invention also provides a computer-readable storage medium. The methods described above according to embodiments of the invention can be implemented in hardware or firmware, or implemented as computer code that can be recorded on a storage medium, or implemented as computer code downloaded via a network and originally stored on a remote storage medium or a non-transitory machine-readable storage medium and then stored on a local storage medium. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code. When the software or computer code is accessed and executed by the computer, processor, or hardware, the new energy centralized control branch current anomaly detection method shown in the above embodiments is implemented.

[0119] A portion of this invention can be applied as a computer program product, such as computer program instructions, which, when executed by a computer, can invoke or provide the methods and / or technical solutions according to the invention through the operation of the computer. Those skilled in the art will understand that the forms in which computer program instructions exist in a computer-readable medium include, but are not limited to, source files, executable files, installation package files, etc. Correspondingly, the ways in which computer program instructions are executed by a computer include, but are not limited to: the computer directly executing the instructions, or the computer compiling the instructions and then executing the corresponding compiled program, or the computer reading and executing the instructions, or the computer reading and installing the instructions and then executing the corresponding installed program. Here, the computer-readable medium can be any available computer-readable storage medium or communication medium accessible to a computer.

[0120] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A method for detecting abnormal current in a new energy centralized control branch circuit, characterized in that, The method includes: Acquire the current current data, historical current data, and current current data of the branch to be tested, as well as the current current data of the branches in the same group. The branches in the same group refer to the branches that are in the same converging unit as the branch to be tested. Using the current environmental conditions as an index, matching historical data is obtained from the historical current data of the branch to be detected, and the similarity between the current current data of the branch to be detected and the matching historical data is calculated. If the similarity is lower than a preset threshold, the deviation of the current current data of the branch to be detected is calculated based on the current current data of the branch in the same group, the duration of the abnormality is determined, and the morphological features are extracted. The deviation, the duration of the anomaly, and the morphological features are quantized, mapped, and then weighted and fused. The result of the weighted fusion is used to determine whether the current data of the branch to be detected is abnormal.

2. The method for detecting abnormal current in a new energy centralized control branch according to claim 1, characterized in that, The step of obtaining matching historical data from the historical current data of the branch to be detected, using the current environmental conditions as an index, includes: The historical current data of the branch to be tested that is the same as or similar to the environmental conditions at the current moment is used as the matching historical data.

3. The method for detecting abnormal current in a new energy centralized control branch according to claim 1, characterized in that, If the similarity is not lower than the preset threshold, it is determined to be due to environmental factors, and the abnormal alarm is suppressed.

4. The method for detecting abnormal current in a new energy centralized control branch according to claim 1, characterized in that, The step of calculating the deviation of the current current data of the branch to be detected based on the current current data of the same group of branches includes: The average value of the current current data of the branches in the same group is used as the reference value; The deviation degree is obtained by calculating the degree of deviation between the current current data of the branch to be detected and the reference value.

5. The method for detecting abnormal current in a new energy centralized control branch according to claim 1, characterized in that, The determination of the duration of the anomaly includes: Timing starts when the deviation exceeds a preset deviation threshold and stops when the deviation returns to within the preset deviation threshold. The timing result is taken as the abnormal duration.

6. The method for detecting abnormal current in a new energy centralized control branch according to claim 1, characterized in that, The extracted morphological features include: Extract the curve shape characteristics of the current current data of the branch to be detected. The curve shape characteristics include at least one of step drop, sawtooth fluctuation and persistent low current.

7. The method for detecting abnormal current in a new energy centralized control branch according to claim 1, characterized in that, The weighted fusion of the deviation, the duration of the anomaly, and the morphological features after quantization mapping includes: The deviation is mapped dimensionlessly using the Sigmoid function to obtain a deviation probability index. An exponential function is used to perform a dimensionless mapping on the abnormal duration to obtain a duration probability index. Based on the type of the morphological features, a confidence weighting mapping is performed to obtain a morphological probability index; The deviation probability index, the duration probability index, and the morphology probability index are weighted and fused.

8. The method for detecting abnormal current in a new energy centralized control branch according to claim 1, characterized in that, The step of determining whether the current data of the branch to be detected is abnormal based on the weighted fusion result includes: The overall fault probability of the branch to be detected is obtained based on the weighted fusion result; If the overall failure probability is greater than or equal to the first threshold, it is determined to be a serious anomaly; If the overall failure probability is greater than or equal to the second threshold and less than the first threshold, it is determined to be a general anomaly; If the overall failure probability is greater than or equal to the third threshold and less than the second threshold, it is determined to be a minor anomaly.

9. A new energy centralized control branch current abnormality detection device, characterized in that, The device includes: The data acquisition module is used to acquire the current current data, historical current data and current current data of the branch to be tested as well as the current current data of the branches in the same group. The branches in the same group refer to the branches that are in the same converging unit as the branch to be tested. The similarity calculation module is used to obtain matching historical data from the historical current data of the branch to be detected, using the current environmental conditions as an index, and calculate the similarity between the current current data of the branch to be detected and the matching historical data. An anomaly index extraction module is used to calculate the deviation of the current current data of the branch to be detected based on the current current data of the same group of branches if the similarity is lower than a preset threshold, determine the duration of the anomaly, and extract morphological features. The anomaly determination module is used to perform quantization mapping and weighted fusion of the deviation, the anomaly duration and the morphological features, and determine whether the current data of the branch to be detected is abnormal based on the weighted fusion result.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing the computer to execute the new energy centralized control branch current anomaly detection method according to any one of claims 1 to 8.

Images