Method for identifying software defects of acquisition terminal based on user power consumption data fluctuation evaluation

Abstract

The application discloses a collection terminal software defect identification method based on user power consumption data fluctuation evaluation, and relates to the technical field of data evaluation.The method comprises the following steps: obtaining a plurality of current collection success rates and optimal periodic success rates of a plurality of to-be-monitored collection terminals; obtaining a group current collection success rate and a group optimal periodic success rate of a plurality of similar terminals; performing self longitudinal comparison and group horizontal comparison to obtain a plurality of longitudinal risk coefficients and a plurality of horizontal risk coefficients; calculating the variation coefficients of each data item collected and obtaining the mean value; obtaining a plurality of environmental confidence degrees; weighting the plurality of longitudinal risk coefficients and the plurality of horizontal risk coefficients to obtain a plurality of terminal abnormal probabilities; clustering a plurality of software versions to obtain a plurality of abnormal versions and a plurality of version influence degrees; and identifying the software defects of the collection terminal.The application solves the technical problem that the existing data evaluation effect is poor, resulting in inaccurate collection terminal software defect identification.

Description

Technical Field

[0001] This invention relates to the field of data evaluation technology, and specifically to a method for identifying defects in data acquisition terminal software based on user electricity consumption data fluctuation evaluation. Background Technology

[0002] With the continuous improvement of the intelligence level of power systems, electricity information acquisition terminals, as key data sensing nodes connecting the user side and the power grid management system, directly affect the power grid's status perception, load analysis, and fault diagnosis capabilities through their operational reliability and data acquisition accuracy. Currently, monitoring the operational status of acquisition terminals and identifying software defects mainly relies on threshold judgment analysis of hard indicators such as communication signal strength. While this method can detect some apparent faults, it is difficult to effectively distinguish between performance degradation caused by the terminal's own software defects and data acquisition anomalies caused by interference from other factors. This can easily lead to misjudgments, resulting in either overly slow or blindly inefficient operation and maintenance responses, and ultimately, inaccurate defect identification results. Summary of the Invention

[0003] This application provides a method for identifying defects in the software of a data acquisition terminal based on the evaluation of fluctuations in user electricity consumption data, which is used to address the technical problem of inaccurate identification of defects in the software of the data acquisition terminal due to poor data evaluation results in the prior art.

[0004] In view of the above problems, this application provides a method for identifying defects in data acquisition terminal software based on user electricity consumption data fluctuation evaluation, the method comprising:

[0005] The system obtains multiple current acquisition success rates for each data item from multiple monitored acquisition terminals, and obtains multiple optimal periodic success rates based on historical data.

[0006] The current collection success rate and the optimal periodic success rate of the group of similar terminals of multiple monitoring collection terminals are obtained, and the multiple monitoring collection terminals are compared vertically on their own and horizontally on the group to obtain multiple vertical risk coefficients and multiple horizontal risk coefficients.

[0007] Calculate the coefficient of variation of each data item collected by multiple monitoring terminals and obtain the mean value to obtain the environmental confidence of multiple monitoring terminals. Perform weighted calculation on multiple longitudinal risk coefficients and multiple horizontal risk coefficients to obtain the anomaly probability of multiple terminals.

[0008] Based on the anomaly probabilities of multiple terminals, the software versions of multiple monitoring terminals are clustered to obtain multiple abnormal versions and the impact of multiple versions, and the software defects of the monitoring terminals are identified.

[0009] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0010] This application proposes a method for identifying software defects in data acquisition terminals based on user electricity consumption data fluctuation evaluation. It constructs a longitudinal benchmark for terminal performance by comprehensively acquiring the current acquisition success rate and historical best-period success rate of each data item of the terminal under monitoring. Simultaneously, it acquires the current acquisition success rate and best-period success rate of a group of similar terminals to establish a horizontal benchmark for group performance. Then, it calculates longitudinal and horizontal risk coefficients to quantify the degree of degradation of the terminal relative to its historical best state and the degree of deviation from the average level of the similar group, respectively. Based on this, it further calculates the coefficient of variation for each data item and calculates the mean to obtain the environmental confidence score, objectively assessing the stability of the terminal's operating environment and effectively distinguishing whether data anomalies originate from internal software defects or external environmental fluctuations. Subsequently, it uses the environmental confidence score to weightedly fuse the longitudinal and horizontal risk coefficients to generate a terminal anomaly probability. This probability value integrates multi-dimensional dynamic information, avoiding the one-sidedness of single-indicator judgment. Finally, it performs cluster analysis on software versions based on the terminal anomaly probability, identifies anomalous versions, and calculates the version impact, achieving a precise correlation between defects and software versions. Compared with traditional methods, the technical solution provided in this application significantly overcomes the limitations of existing technologies that rely on fixed threshold analysis. It achieves the technical effect of comprehensively, accurately and efficiently identifying software defects in the data acquisition terminal under complex operating environments, and realizing intelligent mapping between defects and software versions, thereby improving the reliability, intelligent operation and maintenance level and fault early warning capability of the power grid data acquisition terminal. Attached Figure Description

[0011] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 This is a flowchart illustrating the method for identifying software defects in a data acquisition terminal based on user electricity consumption data fluctuation evaluation, as provided in an embodiment of this application.

[0013] Figure 2 This is a flowchart illustrating the process of obtaining multiple vertical risk coefficients and multiple horizontal risk coefficients in the data acquisition terminal software defect identification method based on user electricity consumption data fluctuation evaluation provided in the embodiments of this application. Detailed Implementation

[0014] This application provides a method for identifying defects in the software of a data acquisition terminal based on the evaluation of fluctuations in user electricity consumption data, which is used to address the technical problem of inaccurate identification of defects in the software of a data acquisition terminal due to poor data evaluation results in the prior art.

[0015] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0016] It should be noted that the terms "comprising" and "having" are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units that are explicitly listed, but may include other steps or modules that are not explicitly listed or that are inherent to these processes, methods, products, or devices.

[0017] Examples, such as Figure 1 As shown, this application provides a method for identifying defects in data acquisition terminal software based on user electricity consumption data fluctuation evaluation, wherein the method includes:

[0018] S10: Obtain multiple current acquisition success rates for each data item of multiple monitoring acquisition terminals, and obtain multiple optimal periodic success rates based on historical data.

[0019] In the daily monitoring of data acquisition terminals, accurately assessing their current data acquisition performance is the first step in identifying potential problems. Existing technologies typically rely on the success rate of data acquisition for each data item within a specific time period as a direct reflection of its operational status. However, this method, which focuses solely on current static indicators, has significant drawbacks. The performance of a data acquisition terminal is not constant; it is affected by various internal factors such as hardware aging, software version updates, and different operating conditions experienced at different historical periods. For example, a seemingly normal current success rate may have already shown a hidden degradation compared to its previous optimal health state, but static comparisons cannot reveal this historical performance decline trend.

[0020] Step S10 in the method provided in this application embodiment includes:

[0021] The actual number of times each data item is collected and the number of times it should be collected are obtained from multiple monitoring and acquisition terminals. Each data item includes total positive active power, total reverse active power, energy indication curve, voltage and current.

[0022] Calculate the ratio of the actual number of times each data item was collected to the number of times it should have been collected, and calculate the average value as the current collection success rate;

[0023] Acquire historical monitoring data from multiple monitoring terminals, extract the success rate of data collection at multiple fixed time periods, and filter to obtain the success rate of multiple optimal periods.

[0024] In this embodiment of the application, multiple current acquisition success rates of each data item of multiple monitoring acquisition terminals are obtained, and multiple optimal periodic success rates are obtained based on historical data.

[0025] Specifically, the actual number of times each data item was collected and the number of times it should have been collected are obtained from multiple monitoring terminals. Each data item includes total forward active power, total reverse active power, energy indication curve, voltage, and current. For example, the power company's electricity consumption information collection system is used as the data source. Detailed collection records for a specified batch of monitoring terminals within the most recent full calendar day are queried and obtained from a database storing massive amounts of terminal-reported data. For each terminal, the "number of times it should have been collected" and the "number of times it was actually successfully collected" for its key data items, including total forward active power, total reverse active power, energy indication curve, voltage, and current, are extracted. The "number of times it should have been collected" is determined by a preset collection task plan, for example, 96 times per data item per day (once every 15 minutes); the "number of times it was actually successfully collected" is calculated based on communication logs and data integrity verification results. This yields two basic values ​​for each terminal and each data item.

[0026] Further, the ratio of the actual number of times each data item was collected to the number of times it should have been collected is calculated, and the average value is taken as the current collection success rate. For a single data item of a single terminal, such as total positive active power, its single-item collection success rate is calculated as: Actual number of collections ÷ Number of collections should have been collected. Assuming that a terminal should have collected 96 times of total positive active power on a given day, and actually successfully collected 92 times, then the current collection success rate for this data item is 92 ÷ 96 = 0.958. The single-item success rate for all five data items of this terminal is calculated sequentially. Then, the arithmetic mean of these five single-item success rates is calculated, and the result is defined as the "current collection success rate" of the terminal. This value is a decimal between 0 and 1, comprehensively reflecting the overall data collection performance of the terminal in the current period.

[0027] Furthermore, historical monitoring data from multiple monitoring terminals is acquired, and the success rates for multiple fixed time periods are extracted. Optimal success rates for each period are then selected. For example, a historical monitoring database is accessed to retrieve the historical sequence of daily success rates for the same terminal over a relatively long period, such as the past year. This historical sequence is then divided into fixed time periods, such as weeks, and the average success rate of all days within each period is calculated, resulting in a series of historical period success rates. Finally, the highest value among these historical period success rates is selected and defined as the terminal's "optimal period success rate." For instance, if a terminal's historically recorded average success rate for week 15 is 0.988, the highest among all weeks, then 0.988 is considered its optimal period success rate. This value represents the best performance level the terminal has achieved under normal historical conditions.

[0028] By acquiring the current acquisition success rate of each data item for each monitored acquisition terminal and simultaneously mining multiple optimal cycle success rates based on historical data, the evaluation method is improved from static to dynamic. This establishes an intrinsic benchmark for subsequent analysis, enabling effective longitudinal comparison of terminals and thus keenly perceiving whether the current performance of a terminal has deviated from its historical best level.

[0029] S20: Obtain the current acquisition success rate and the optimal periodic success rate of the group of the same type of terminals of the multiple acquisition terminals to be monitored, and perform a vertical comparison of the multiple acquisition terminals to be monitored and a horizontal comparison of the group to obtain multiple vertical risk coefficients and multiple horizontal risk coefficients.

[0030] After establishing a historical benchmark for each terminal, accurately quantifying and assessing its current level of anomaly is a pressing technical challenge. While a longitudinal comparison alone can reveal performance degradation, it cannot determine whether this degradation is an isolated, isolated issue or a widespread problem faced by terminals of the same model and batch; the latter is more likely to point to common software defects. Conversely, a horizontal comparison with the group average can identify deviations from the group's performance, but it cannot distinguish whether this deviation is due to the terminal's own continuous deterioration or because its historical performance was inherently inferior to the group average.

[0031] Step S20 in the method provided in this application embodiment includes:

[0032] Obtain the model information and functional information of multiple monitoring and acquisition terminals;

[0033] Based on the model information and the functional information, the data acquisition terminals are filtered to obtain multiple similar terminals of multiple data acquisition terminals to be monitored.

[0034] The average current acquisition success rate of multiple terminals of the same type is obtained as the current acquisition success rate of the group.

[0035] The average of the optimal periodic success rates of multiple terminals of the same type is obtained as the group's optimal periodic success rate;

[0036] Obtain multiple vertical risk coefficients and multiple horizontal risk coefficients, such as Figure 2 As shown, it includes:

[0037] Calculate the ratio of 1 minus the current acquisition success rate and the optimal cycle success rate to obtain the longitudinal risk coefficient;

[0038] Calculate the ratio of 1 minus the current collection success rate and the current collection success rate of the group to obtain the current horizontal risk coefficient;

[0039] Calculate the ratio of 1 minus the optimal cycle success rate and the optimal cycle success rate of the group to obtain the optimal horizontal risk coefficient;

[0040] The mean of the current horizontal risk coefficient and the optimal horizontal risk coefficient is obtained and used as the horizontal risk coefficient.

[0041] Wherein, when the ratio of the current acquisition success rate to the optimal cycle success rate is greater than or equal to 0.95, the vertical risk coefficient is 0.1, and when the horizontal risk coefficient is less than 0.05, the horizontal risk coefficient is 0.1.

[0042] In this embodiment of the application, the current collection success rate and the optimal periodic success rate of the same type of terminals of multiple monitoring collection terminals are obtained, and the multiple monitoring collection terminals are compared vertically on their own and horizontally on the group to obtain multiple vertical risk coefficients and multiple horizontal risk coefficients.

[0043] Specifically, firstly, the model information and functional information of multiple monitoring terminals are obtained. Based on the model information and functional information, the terminals are filtered to obtain multiple similar terminals among the multiple monitoring terminals. For each monitoring terminal, according to model information (e.g., model A) and core functional information (e.g., support for curve data acquisition), filtering conditions are set in the database of all terminals to filter out all other terminals that simultaneously meet the requirements of the same model and consistent core functional configuration, thus obtaining a list of similar terminals for that monitoring terminal. For example, for a monitoring terminal with model A that supports curve data acquisition, its similar terminals are all other terminals of model A that also support curve data acquisition. The process of obtaining similar terminals for all monitoring terminals is repeated.

[0044] Furthermore, the average current acquisition success rate of multiple similar terminals is obtained as the group's current acquisition success rate; the average optimal period success rate of multiple similar terminals is obtained as the group's optimal period success rate. Specifically, the current acquisition success rate and optimal period success rate of each terminal in the same category are extracted. Then, the arithmetic mean of these two success rate sets is calculated. The current acquisition success rates of similar terminals are added together and divided by the number of similar terminals to obtain the group's current acquisition success rate. Furthermore, using the same method, the optimal period success rates of all similar terminals are added together and averaged to obtain the group's optimal period success rate. These two averages represent the general performance level of the acquisition terminals under this model and functional information in the current period, and the general best level reached in historical periods, respectively.

[0045] Further, the ratio of 1 minus the current acquisition success rate to the optimal cycle success rate is calculated to obtain the longitudinal risk coefficient. Specifically, when the ratio of the current acquisition success rate to the optimal cycle success rate is greater than or equal to 0.95, the longitudinal risk coefficient is set to 0.1. Longitudinal risk coefficient = 1 - (current acquisition success rate / optimal cycle success rate). The larger the longitudinal risk coefficient, the more severe the performance degradation compared to its historical best level. If the calculated ratio of the current acquisition success rate to the optimal cycle success rate is greater than or equal to 0.95, it means that the current success rate has reached or exceeded 95% of the historical best level. In this case, the performance degradation is considered very slight. To avoid amplifying minor fluctuations and to consider potential risk impacts, the longitudinal risk coefficient is set to a small fixed value, such as 0.1.

[0046] Further, calculate the ratio of 1 minus the current collection success rate and the current collection success rate of the group to obtain the current horizontal risk coefficient. Calculate the ratio of 1 minus the optimal cycle success rate and the optimal cycle success rate of the group to obtain the optimal horizontal risk coefficient. Calculate the current horizontal risk coefficient as 1 - (current collection success rate / current collection success rate of the group), and the optimal horizontal risk coefficient as 1 - (optimal cycle success rate / optimal cycle success rate of the group).

[0047] Furthermore, the average of the current horizontal risk coefficient and the optimal horizontal risk coefficient is obtained as the horizontal risk coefficient. Horizontal risk coefficient = (current horizontal risk coefficient + optimal horizontal risk coefficient) / 2. The horizontal risk coefficient comprehensively reflects the degree of deviation of the terminal from the group average level in terms of both current performance and historical potential. Specifically, when the horizontal risk coefficient is less than 0.05, it indicates that the difference between the monitored terminal and the group is extremely small; the horizontal risk coefficient is set to 0.1 to prevent potential risks from being overlooked.

[0048] By acquiring success rate data from similar terminal groups and calculating both the longitudinal risk coefficient representing individual performance degradation and the horizontal risk coefficient representing deviation from the group level, a refined and multi-dimensional quantification of terminal anomaly risks was achieved. Risk assessment was expanded from a single dimension to a two-dimensional analysis based on the "self-group" framework. The longitudinal risk coefficient focuses on revealing changes in the terminal's internal state, capturing whether it is worse than before; the horizontal risk coefficient focuses on the terminal's relative position within the same group, revealing whether it differs from its peers. Combining these two allows for the initial differentiation of anomaly patterns: high longitudinal risk accompanied by low horizontal risk points more to internal terminal failures such as hardware damage or individual software errors; while high horizontal risk, especially accompanied by a group-wide longitudinal change trend, may indicate the existence of common problems such as software version or configuration strategy. This dual comparison mechanism enhances the ability to initially assess the nature of anomalies, providing richer and more accurate intermediate risk characteristics for subsequent steps.

[0049] S30: Calculate the coefficient of variation of each data item collected by the multiple monitoring terminals and obtain the mean value to obtain the environmental confidence of the multiple monitoring terminals. Perform weighted calculation on the multiple longitudinal risk coefficients and the multiple horizontal risk coefficients to obtain the abnormal probability of multiple terminals.

[0050] A decline or fluctuation in data acquisition success rate may not necessarily stem from software or hardware defects in the acquisition terminal itself, but is very likely caused by drastic fluctuations in the power supply environment on the user side where the terminal is located. For example, frequent start-ups and shutdowns of user equipment, intermittent production, or poor line contact can cause significant fluctuations or temporary loss of data such as voltage and current. This instability at the data source directly leads to a decrease in acquisition success rate, which is completely consistent with the apparent defects in the terminal software's data processing capabilities.

[0051] Step S30 in the method provided in this application embodiment includes:

[0052] Obtain the coefficient of variation of each data item collected by multiple monitoring terminals, wherein the coefficient of variation is the ratio of the standard deviation to the mean of each data item;

[0053] Calculate 1 minus the mean coefficient of variation of each of the data items, and use this as the environmental confidence level;

[0054] Based on multiple environmental confidence levels, multiple risk weights are obtained for multiple monitoring and data acquisition terminals, wherein the risk weights include vertical risk weights and horizontal risk weights, and the environmental confidence level is proportional to the vertical risk weight.

[0055] Based on multiple risk weights, multiple vertical risk coefficients and multiple horizontal risk coefficients are weighted and calculated to obtain multiple terminal anomaly probabilities.

[0056] In this embodiment, the coefficient of variation of each data item collected by multiple monitoring terminals is calculated and the mean is obtained to acquire the environmental confidence of multiple monitoring terminals. Multiple longitudinal risk coefficients and multiple horizontal risk coefficients are weighted and calculated to acquire the anomaly probability of multiple terminals.

[0057] Specifically, firstly, the coefficient of variation (COP) of each data item collected by multiple monitoring terminals is obtained. The COP is the ratio of the standard deviation to the mean of each data item. For each monitoring terminal, the raw reading sequence of each data item collected in the most recent period is extracted, including total forward active power, total reverse active power, energy indication curve, voltage, and current. For example, for the voltage data item, the voltage values ​​successfully collected by the monitoring terminal in the past 24 hours are extracted to form a numerical list, such as [225.1, 226.3, 220.8, 228.5]. Further, the mean and standard deviation of this sequence are calculated, and the COP is calculated as follows: COP = standard deviation / mean. For example, if the mean of a voltage sequence from a monitoring terminal is 230 volts and the standard deviation is 1.5 volts, then the COP for the voltage item is 1.5 / 230 = 0.0065. The smaller the COP value, the less fluctuation the data item experiences, and the more stable the acquisition environment.

[0058] Furthermore, the mean coefficient of variation of each data item is subtracted from 1 to obtain the environmental confidence score. Environmental confidence score = 1 - mean coefficient of variation. For example, if the mean of the five coefficients of variation for a certain terminal is 0.01, then its environmental confidence score is 1 - 0.01 = 0.99. The environmental confidence score is a value between 0 and 1; the higher the value, the more stable the data environment in which the terminal operates, and the lower the possibility of data acquisition failure due to environmental interference.

[0059] Furthermore, based on multiple environmental confidence levels, multiple risk weights are obtained for multiple monitoring and data acquisition terminals. These risk weights include vertical risk weights and horizontal risk weights, with the environmental confidence level being directly proportional to the vertical risk weight. For example, risk weights are obtained based on environmental confidence levels. A high environmental confidence level indicates a low environmental risk, meaning the assessment based on the terminal itself is more accurate, indicating a higher probability of anomalies caused by a problem within the terminal itself. Therefore, the vertical risk weight is larger, calculated as (0.5 + environmental confidence level) / 2, where 0.5 is a pre-set comprehensive parameter to prevent excessively large vertical risk weights from affecting the results.

[0060] Furthermore, based on multiple risk weights, multiple vertical risk coefficients and multiple horizontal risk coefficients are weighted and calculated to obtain multiple terminal anomaly probabilities. Terminal anomaly probability = (vertical risk coefficient × vertical risk weight) + (horizontal risk coefficient × horizontal risk weight). The obtained terminal anomaly probability comprehensively considers the terminal's own performance changes, as well as the group comparison and environmental stability assessment, making it a more comprehensive and robust quantitative indicator of anomaly risk.

[0061] Data volatility is assessed by calculating the coefficient of variation for each data item, and an environmental confidence score is generated accordingly. This environmental confidence score is then weighted and combined with longitudinal and lateral risk coefficients to obtain the terminal anomaly probability, achieving crucial de-interference and fusion decision-making in the defect identification process. The coefficient of variation reflects the dispersion of the data itself. High volatility indicates an unstable acquisition environment; in this case, even if the acquisition success rate decreases, it should be cautiously attributed to defects in the terminal itself. Conversely, low volatility, i.e., high environmental confidence, indicates a stable data environment, and a decrease in the success rate is more likely due to internal problems within the terminal.

[0062] S40: Based on the anomaly probabilities of multiple terminals, cluster the software versions of multiple monitoring terminals to obtain multiple abnormal versions and the impact of multiple versions, and identify software defects in the monitoring terminals.

[0063] After identifying a high probability of anomalies in one or more terminals, traditional operations and maintenance (O&M) responses often stop at troubleshooting and repairing specific terminals. This approach may be inefficient when dealing with latent or slow-onset problems caused by defects in specific software versions that are prevalent across a large number of terminals. In other words, scattered individual alarms make it difficult for O&M personnel to quickly discern the common patterns behind these anomalies.

[0064] Step S40 in the method provided in this application embodiment includes:

[0065] Based on the multiple terminal anomaly probabilities of the multiple monitoring and acquisition terminals, the software versions of the multiple terminals of the same type are clustered to obtain multiple abnormal versions and an optimal version. The multiple abnormal versions include multiple software versions whose terminal anomaly probability at the cluster center is greater than or equal to the terminal anomaly probability threshold, and the optimal version includes the software version with the minimum terminal anomaly probability at the cluster center.

[0066] The acquisition of the terminal anomaly probability threshold includes:

[0067] The upper quartile of the terminal anomaly probability is obtained as the initial anomaly threshold, and the mean of multiple environmental confidence scores is calculated to correct the initial anomaly threshold, thereby obtaining the terminal anomaly probability threshold.

[0068] Calculate the optimal cycle success rate of multiple abnormal versions and the deviation of the optimal cycle success rate of the optimal version as the version deviation, and calculate the deviation of the mean environmental confidence of multiple abnormal versions and the mean environmental confidence of the optimal version as the environmental deviation.

[0069] The impact of multiple versions is calculated based on the product of multiple version deviations and multiple environmental deviations.

[0070] The impact of the software version of the data acquisition terminal to be monitored is obtained, and combined with the terminal anomaly probability, the software defects of the data acquisition terminal are identified.

[0071] By integrating the terminal anomaly probability, the version impact, the ratio of the actual number of times each data item was collected to the number of times it should have been collected, and the coefficient of variation of each data item, a structured defect identification result list is formed, which serves as the terminal software defect identification result.

[0072] In this embodiment, based on the anomaly probability of multiple terminals, the software versions of multiple monitoring and data acquisition terminals are clustered to obtain multiple abnormal versions and the impact degree of multiple versions, and the software defects of the data acquisition terminals are identified.

[0073] Specifically, based on the anomaly probabilities of multiple monitored terminals, software versions of multiple similar terminals are clustered to obtain multiple anomalous versions and an optimal version. Multiple anomalous versions include software versions whose cluster center terminal anomaly probability is greater than or equal to a terminal anomaly probability threshold, and the optimal version includes the software version with the lowest cluster center terminal anomaly probability. Cluster analysis is performed on terminal groups with different software versions based on terminal anomaly probabilities. The software version is used as the grouping criterion for clustering, and the mean of the terminal anomaly probability for all terminals corresponding to each software version is calculated and used as the feature value of that software version. Then, the KMeans clustering tool is used to cluster the feature values ​​of all software versions, with an example setting of 10 clusters. After clustering, each cluster has a cluster center, which is the mean of the feature values ​​of all versions within that cluster. Software versions contained in clusters whose cluster center value is greater than or equal to the terminal anomaly probability threshold are marked as anomalous versions. Simultaneously, the version with the smallest cluster center value is selected from all software versions and marked as the optimal version.

[0074] The acquisition of the terminal anomaly probability threshold includes: obtaining the upper quartile of the terminal anomaly probability as the initial anomaly threshold, and calculating the mean of multiple environmental confidence levels to correct the initial anomaly threshold, thus obtaining the final terminal anomaly probability threshold. Specifically, the terminal anomaly probability sequence of the terminals to be monitored is obtained and sorted from largest to smallest. The upper quartile of the terminal anomaly probability sequence, i.e., the 75th quartile, is obtained as the initial anomaly threshold. For example, the initial anomaly threshold is calculated to be 0.82. Further, the arithmetic mean of the environmental confidence levels is calculated. Then, a correction formula is used to adjust the initial threshold to consider the mild influence of overall environmental stability on anomaly judgment. The correction formula is: Terminal anomaly probability threshold = Initial anomaly threshold × (1.5 - Mean of environmental confidence level). For example, if the mean of environmental confidence level is 0.85, then the terminal anomaly probability threshold = 0.82 × (1.5 - 0.85) = 0.82 × 0.65 = 0.533. When the overall environmental stability is high, the judgment threshold is lowered to detect anomalies more sensitively; when the overall environmental stability is low, the judgment threshold is raised to avoid false alarms caused by excessive environmental interference.

[0075] Furthermore, the deviation between the optimal cycle success rate of multiple abnormal versions and the optimal cycle success rate of the optimal version is calculated as the version deviation. The deviation between the mean environmental confidence score of multiple abnormal versions and the mean environmental confidence score of the optimal version is also calculated as the environment deviation. Specifically, the average optimal cycle success rate of all terminals under the abnormal version is obtained, as well as the average optimal cycle success rate of all terminals under the optimal version. Version deviation = |Average optimal cycle success rate of abnormal versions - Average optimal cycle success rate of optimal versions| ÷ Average optimal cycle success rate of optimal versions. The version deviation reflects the difference between the abnormal version and the optimal version in terms of historical best performance. Further, the environment deviation is calculated. The average environmental confidence score of all terminals under the abnormal version is obtained, as well as the average environmental confidence score of all terminals under the optimal version. Environment deviation = |Average environmental confidence score of abnormal versions - Average environmental confidence score of optimal versions| ÷ Average environmental confidence score of optimal versions. The environment deviation reflects the difference between the abnormal version and the optimal version in terms of stability in the current operating environment.

[0076] Furthermore, based on the product of multiple version deviations and multiple environment deviations, the impact of multiple versions is calculated. Version impact = Version deviation × Environment deviation.

[0077] Furthermore, the version impact of the software version of the monitored data acquisition terminal is obtained, and combined with the terminal anomaly probability, software defects in the data acquisition terminal are identified. Specifically, for any data acquisition terminal to be monitored, its software version number is first obtained, and then the version impact corresponding to that version number is found. Simultaneously, the terminal itself has a pre-calculated terminal anomaly probability. These two indicators are used together for judgment: if a terminal's terminal anomaly probability exceeds the terminal anomaly probability threshold, and its software version's version impact is also high, then it can be confidently identified that the terminal's anomaly is caused by a potential defect in its software version. This method of combining individual anomaly probabilities with group version impact elevates defect identification from the individual level to the level of common root causes.

[0078] Furthermore, by integrating terminal anomaly probability, version impact, the ratio of actual to expected data collection times for each data item, and the coefficient of variation for each data item, a structured defect identification result list is formed, serving as the terminal software defect identification result. For example, a table is created to generate a record for each terminal identified as having potential software defect risks. Each record contains the following fields: terminal number, software version number, terminal anomaly probability, version impact of its software version, the ratio of actual to expected data collection times for each data item, and the coefficient of variation for each data item. This list constitutes the terminal software defect identification result, systematically presenting the individual terminal anomaly behavior, the group risk of its version, and detailed data evidence, providing comprehensive decision support information for operations and maintenance personnel.

[0079] By clustering software versions based on terminal anomaly probabilities and calculating the version impact of anomalous versions, a closed loop for identifying terminal software defects was achieved. Clustering automatically maps terminal groups with high anomaly probabilities to their corresponding software versions, thus quickly and objectively identifying potential anomalous versions. Furthermore, by calculating the version impact that comprehensively considers the deviation from the success rate and environmental confidence level of the optimal version, not only can anomalous versions be identified, but the severity and scope of their defects can also be quantified. This allows operations and maintenance personnel to prioritize anomalous versions with high impact, formulate unified upgrade or patch strategies, and eliminate potential risks in batches, greatly improving operational efficiency and effectiveness.

[0080] In summary, the embodiments of this application have at least the following technical effects:

[0081] This application proposes a method for identifying software defects in data acquisition terminals based on user electricity consumption data fluctuation evaluation. It constructs a longitudinal benchmark for terminal performance by comprehensively acquiring the current acquisition success rate and historical best-period success rate of each data item of the terminal under monitoring. Simultaneously, it acquires the current acquisition success rate and best-period success rate of a group of similar terminals to establish a horizontal benchmark for group performance. Then, it calculates longitudinal and horizontal risk coefficients to quantify the degree of degradation of the terminal relative to its historical best state and the degree of deviation from the average level of the similar group, respectively. Based on this, it further calculates the coefficient of variation for each data item and calculates the mean to obtain the environmental confidence score, objectively assessing the stability of the terminal's operating environment and effectively distinguishing whether data anomalies originate from internal software defects or external environmental fluctuations. Subsequently, it uses the environmental confidence score to weightedly fuse the longitudinal and horizontal risk coefficients to generate a terminal anomaly probability. This probability value integrates multi-dimensional dynamic information, avoiding the one-sidedness of single-indicator judgment. Finally, it performs cluster analysis on software versions based on the terminal anomaly probability, identifies anomalous versions, and calculates the version impact, achieving a precise correlation between defects and software versions. Compared to traditional methods, the technical solution provided in this application significantly overcomes the limitations of existing technologies that rely on fixed threshold analysis. By tracing historical data longitudinally and referencing groups horizontally, it can keenly capture the slow degradation trend and hidden risks of terminal performance. At the same time, it uses environmental confidence to filter out non-software interference, greatly reducing the possibility of misjudgment. In addition, the version clustering mechanism enables maintenance personnel to quickly locate common problems introduced by specific software versions from a massive number of terminals, achieving more accurate root cause analysis. This achieves the technical effect of comprehensively, accurately, and efficiently identifying software defects in data acquisition terminals in complex operating environments and realizing intelligent mapping between defects and software versions, thereby improving the reliability of power grid data acquisition terminals, the level of intelligent operation and maintenance, and the ability to predict faults.

[0082] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0083] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0084] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and modifications fall within the scope of this application and its equivalents, this application intends to include such modifications and modifications.

Claims

1. A method for identifying software defects in data acquisition terminals based on user electricity consumption fluctuation evaluation, characterized in that, include: The system obtains multiple current acquisition success rates for each data item from multiple monitored acquisition terminals, and obtains multiple optimal periodic success rates based on historical data. The current collection success rate and the optimal periodic success rate of the group of similar terminals of multiple monitoring collection terminals are obtained, and the multiple monitoring collection terminals are compared vertically on their own and horizontally on the group to obtain multiple vertical risk coefficients and multiple horizontal risk coefficients. Calculate the coefficient of variation of each data item collected by multiple monitoring terminals and obtain the mean value to obtain the environmental confidence of multiple monitoring terminals. Perform weighted calculation on multiple longitudinal risk coefficients and multiple horizontal risk coefficients to obtain the anomaly probability of multiple terminals. Based on the anomaly probabilities of multiple terminals, the software versions of multiple monitoring terminals are clustered to obtain multiple anomalous versions and their impact degrees, and software defect identification of the monitoring terminals is performed, including: Based on the multiple terminal anomaly probabilities of the multiple monitoring and acquisition terminals, the software versions of the multiple terminals of the same type are clustered to obtain multiple abnormal versions and an optimal version. The multiple abnormal versions include multiple software versions whose terminal anomaly probability at the cluster center is greater than or equal to the terminal anomaly probability threshold, and the optimal version includes the software version with the minimum terminal anomaly probability at the cluster center. Calculate the optimal cycle success rate of multiple abnormal versions and the deviation of the optimal cycle success rate of the optimal version as the version deviation, and calculate the deviation of the mean environmental confidence of multiple abnormal versions and the mean environmental confidence of the optimal version as the environmental deviation. The impact of multiple versions is calculated based on the product of multiple version deviations and multiple environmental deviations. The impact of the software version of the data acquisition terminal to be monitored is obtained, and combined with the terminal anomaly probability, software defects of the data acquisition terminal are identified.

2. The method for identifying software defects in a data acquisition terminal based on user electricity consumption fluctuation evaluation according to claim 1, characterized in that, Obtain multiple current acquisition success rates for each data item from multiple monitored acquisition terminals, and based on historical data, obtain multiple optimal periodic success rates, including: The actual number of times each data item is collected and the number of times it should be collected are obtained from multiple monitoring and acquisition terminals. Each data item includes total positive active power, total reverse active power, energy indication curve, voltage and current. Calculate the ratio of the actual number of times each data item was collected to the number of times it should have been collected, and calculate the average value as the current collection success rate; Acquire historical monitoring data from multiple monitoring terminals, extract the success rate of data collection at multiple fixed time periods, and filter to obtain the success rate of multiple optimal periods.

3. The method for identifying software defects in a data acquisition terminal based on user electricity consumption fluctuation evaluation according to claim 1, characterized in that, Obtain the current acquisition success rate and the optimal periodic success rate of the group of similar terminals among multiple monitored acquisition terminals, including: Obtain the model information and functional information of multiple monitoring and acquisition terminals; Based on the model information and the functional information, the data acquisition terminals are filtered to obtain multiple similar terminals of multiple data acquisition terminals to be monitored. The average current acquisition success rate of multiple terminals of the same type is obtained as the current acquisition success rate of the group. The average of the optimal periodic success rates of multiple terminals of the same type is obtained as the group's optimal periodic success rate.

4. The method for identifying software defects in a data acquisition terminal based on user electricity consumption fluctuation evaluation according to claim 1, characterized in that, Multiple monitoring and data acquisition terminals are compared longitudinally within themselves and laterally across groups to obtain multiple longitudinal risk coefficients and multiple laterally risk coefficients, including: Calculate the ratio of 1 minus the current acquisition success rate and the optimal cycle success rate to obtain the longitudinal risk coefficient; Calculate the ratio of 1 minus the current collection success rate and the current collection success rate of the group to obtain the current horizontal risk coefficient; Calculate the ratio of 1 minus the optimal cycle success rate and the optimal cycle success rate of the group to obtain the optimal horizontal risk coefficient; The mean of the current horizontal risk coefficient and the optimal horizontal risk coefficient is obtained and used as the horizontal risk coefficient.

5. The method for identifying software defects in a data acquisition terminal based on user electricity consumption fluctuation evaluation according to claim 4, characterized in that, Obtaining multiple vertical risk coefficients and multiple horizontal risk coefficients also includes: When the ratio of the current acquisition success rate to the optimal cycle success rate is greater than or equal to 0.95, the vertical risk coefficient is 0.1; when the horizontal risk coefficient is less than 0.05, the horizontal risk coefficient is 0.

1.

6. The method for identifying software defects in a data acquisition terminal based on user electricity consumption fluctuation evaluation according to claim 1, characterized in that, Calculate the coefficient of variation of each data item collected by the multiple monitoring terminals and obtain the mean value to acquire the environmental confidence level of the multiple monitoring terminals, including: Obtain the coefficient of variation of each data item collected by multiple monitoring terminals, wherein the coefficient of variation is the ratio of the standard deviation to the mean of each data item; Calculate 1 minus the mean coefficient of variation of each of the data items, and use this as the environmental confidence level.

7. The method for identifying software defects in a data acquisition terminal based on user electricity consumption fluctuation evaluation according to claim 1, characterized in that, The multiple vertical risk coefficients and multiple horizontal risk coefficients are weighted and calculated to obtain multiple terminal anomaly probabilities, including: Based on multiple environmental confidence levels, multiple risk weights are obtained for multiple monitoring and data acquisition terminals, wherein the risk weights include vertical risk weights and horizontal risk weights, and the environmental confidence level is proportional to the vertical risk weight. Based on multiple risk weights, multiple vertical risk coefficients and multiple horizontal risk coefficients are weighted and calculated to obtain multiple terminal anomaly probabilities.

8. The method for identifying software defects in a data acquisition terminal based on user electricity consumption fluctuation evaluation according to claim 1, characterized in that, The acquisition of the terminal anomaly probability threshold includes: The upper quartile of the terminal anomaly probability is obtained as the initial anomaly threshold, and the mean of multiple environmental confidence scores is calculated to correct the initial anomaly threshold, thereby obtaining the terminal anomaly probability threshold.

9. The method for identifying software defects in a data acquisition terminal based on user electricity consumption fluctuation evaluation according to claim 1, characterized in that, The identification of defects in the data acquisition terminal software also includes: By integrating the terminal anomaly probability, the version impact, the ratio of the actual number of times each data item was collected to the number of times it should have been collected, and the coefficient of variation of each data item, a structured defect identification result list is formed, which serves as the terminal software defect identification result.

Images