Rail transit station power distribution quality comprehensive evaluation method and system

Abstract

The application provides a rail transit station distribution transformer power quality comprehensive evaluation method and system, comprising: acquiring power quality index data within a given time; constructing a segmented nonlinear single score model according to the preprocessed power quality index data, and obtaining power quality single score according to the single score model; wherein the single score model is constructed based on the historical deviation data of each power quality index collected in the preset sliding window; the initial weight corresponding to each index is determined by the entropy weight method, and the Bayesian algorithm is used to fuse each initial weight to obtain the final weight; based on the final weight and the power quality single score, the comprehensive score result is obtained through the pre-constructed comprehensive score model, and the power quality comprehensive evaluation is carried out according to the comprehensive score result. The application comprehensively measures the power quality from multiple dimensions, and realizes objective and dynamic evaluation of the power quality of the rail transit station distribution transformer.

Description

Technical Field

[0001] This application belongs to the field of power system power quality monitoring and evaluation technology, and in particular relates to a comprehensive evaluation method and system for power quality of distribution transformers in rail transit stations. Background Technology

[0002] With the rapid development of urban rail transit, the electrical loads at its stations (such as elevators, air conditioning, lighting, and signaling systems) exhibit significant nonlinearity, impulsiveness, and imbalance, leading to increasingly prominent power quality problems and posing challenges to the safe, stable, and efficient operation of the power supply system. A scientific and accurate assessment of the power quality at the station's distribution transformer side is a prerequisite for developing effective operation and maintenance strategies and remediation plans.

[0003] Currently, the weighted average method is mostly used for comprehensive evaluation of power quality. However, existing technologies have the following main drawbacks:

[0004] ① The determination of weights is highly subjective: the weights of indicators usually rely on expert experience or the analytic hierarchy process (AHP), which has a great influence from human factors and makes it difficult to guarantee the objectivity and fairness of the evaluation results.

[0005] ② Simplified scoring model: It often uses a linear scoring function, which cannot effectively distinguish between slight fluctuations and severe degradation. The penalty for abnormal situations that exceed the threshold is insufficient or excessive, and it cannot truly reflect the degree of power quality degradation.

[0006] ③ Insufficient data utilization and analysis: Rail transit power quality management systems generally use 15-minute intervals to store data in order to balance storage costs and data value. For this type of data, traditional methods such as wavelet analysis are difficult to apply effectively, resulting in the neglect of the deep value of massive historical data (such as long-term trends and periodic patterns), and the evaluation remains at the "static" or "short-term" level. Summary of the Invention

[0007] In view of this, this application aims to propose a comprehensive evaluation method and system for power quality of distribution transformers in rail transit stations to solve at least one of the above-mentioned problems.

[0008] To achieve the above objectives, the technical solution of this application is implemented as follows:

[0009] Firstly, this application provides a comprehensive evaluation method for the power quality of power distribution transformers in rail transit stations, including:

[0010] Obtain power quality index data within a specified time period;

[0011] A piecewise nonlinear single-item scoring model is constructed based on the preprocessed power quality index data, and a single-item power quality score is obtained based on the single-item scoring model; wherein, the single-item scoring model is constructed based on the historical deviation data of each power quality index collected within a preset sliding window;

[0012] The initial weights for each indicator are determined by the entropy weight method, and the Bayesian algorithm is used to fuse the initial weights to obtain the final weights.

[0013] Based on the final weights and the individual power quality scores, a comprehensive score is obtained through a pre-built comprehensive scoring model, and a comprehensive power quality evaluation is performed based on the comprehensive score results.

[0014] Secondly, based on the same inventive concept, this application also provides a comprehensive evaluation system for power quality of distribution transformers in rail transit stations, including:

[0015] The data acquisition module is configured to acquire power quality index data within a predetermined time period;

[0016] The single-item scoring module is configured to construct a piecewise nonlinear single-item scoring model based on the preprocessed power quality index data, and obtain a single power quality score based on the single-item scoring model; wherein, the single-item scoring model is constructed based on the historical deviation data of each power quality index collected within a preset sliding window;

[0017] The fusion processing module is configured to determine the initial weights corresponding to each indicator using the entropy weight method, and then use the Bayesian algorithm to fuse the initial weights to obtain the final weights.

[0018] The comprehensive scoring module is configured to obtain a comprehensive scoring result based on the final weight and the power quality individual score through a pre-built comprehensive scoring model, and to conduct a comprehensive evaluation of power quality based on the comprehensive scoring result.

[0019] Thirdly, based on the same inventive concept, this application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method described in the first aspect.

[0020] Fourthly, based on the same inventive concept, this application also provides a non-transitory computer-readable storage medium, wherein the non-transitory computer-readable storage medium stores computer instructions for causing the computer to perform the method as described in the first aspect.

[0021] Compared with existing technologies, the comprehensive evaluation method and system for power quality of rail transit station distribution transformers described in this application have the following advantages:

[0022] The comprehensive evaluation method for power quality of distribution transformers in rail transit stations described in this application is based on continuous measured data of distribution transformers in a typical rail transit station. It selects six key indicators: frequency deviation, total harmonic distortion rate of three-phase current, total harmonic distortion rate of three-phase voltage, three-phase voltage imbalance, power factor deviation, and three-phase voltage fluctuation rate. It constructs a multi-dimensional evaluation framework covering steady-state and dynamic characteristics. This method can achieve accurate quantification and dynamic monitoring of power quality in rail transit power distribution systems, providing quantitative basis and decision support for operation and maintenance decisions of rail transit power supply systems. Attached Figure Description

[0023] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings:

[0024] Figure 1 This is a flowchart illustrating a comprehensive evaluation method for power quality in rail transit station distribution transformers, as described in an embodiment of this application.

[0025] Figure 2 This is a timing diagram of the comprehensive power quality scoring described in the embodiments of this application;

[0026] Figure 3 This is a trend analysis chart of power quality scores for May of each year, as described in the embodiments of this application.

[0027] Figure 4 This is a schematic diagram of the structure of a comprehensive evaluation system for power quality of a rail transit station distribution transformer, as described in an embodiment of this application.

[0028] Figure 5 This is a schematic diagram of the hardware structure of the electronic device described in an embodiment of this application. Detailed Implementation

[0029] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with specific embodiments and the accompanying drawings.

[0030] It should be noted that, unless otherwise defined, the technical or scientific terms used in the embodiments of this application should have the ordinary meaning understood by one of ordinary skill in the art to which this application pertains. The terms "first," "second," and similar terms used in the embodiments of this application do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed after the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are only used to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0031] The embodiments of this application are described in detail below with reference to the accompanying drawings.

[0032] Please see Figure 1 As shown in the figure, this embodiment provides a comprehensive evaluation method for power quality of distribution transformers in rail transit stations, which specifically includes the following steps:

[0033] Step S101: Obtain power quality index data within a predetermined time period.

[0034] Specifically, in this embodiment, all power quality index data within one month are acquired, including frequency deviation, three-phase voltage, total harmonic distortion rate of three-phase voltage, total harmonic distortion rate of three-phase current, and total power factor deviation. These data are all taken at 15-minute intervals.

[0035] This step also includes data preprocessing, as follows:

[0036] (1) Data type unification and missing value handling. All numeric columns are forcibly converted to floating-point type, and any value that cannot be converted will be marked as NaN; at 15-minute intervals, all time points are counted to see if there are any missing points, and the missing points are marked as NaN; then, these NaN values ​​are handled using strategies such as mean filling and forward filling.

[0037] (2) Deletion of missing data points. When more than 40% of five consecutive time points are missing, the data at those time points will be deleted.

[0038] To better calculate power quality scores, this embodiment first performs numerical processing on the data in step S101. When the value exceeds 10%, the average value is used instead. Then, the following data is processed to obtain the following six categories of data:

[0039] (1) Frequency deviation: The absolute value of the deviation between the system frequency and the nominal value (50Hz). Calculation method: |actual frequency - 50Hz|.

[0040] (2) Total Harmonic Distortion (THD) of Three-Phase Current: The average value of the harmonic content of the three-phase current.

[0041] ;

[0042] (3) Total Harmonic Distortion (THD) of Three-Phase Voltage: The average value of the harmonic content of three-phase voltage.

[0043] ;

[0044] (4) Three-phase voltage imbalance: the ratio of the maximum phase voltage deviation to the average voltage.

[0045] ;

[0046] In the formula, This represents the average value of the three-phase voltage RMS value, i.e. .

[0047] (5) Power factor deviation: The absolute deviation from the ideal power factor 1: |actual power factor - 1|.

[0048] (6) Three-phase voltage fluctuation rate: The percentage change in the effective voltage value between two adjacent 15-minute time points, reflecting the degree of short-term voltage fluctuation, in percentage (strongly correlated with the sampling interval, which is fixed at 15 minutes here), the main calculation formula is as follows:

[0049] ;

[0050] In the formula, This represents the average three-phase voltage at the current time point (t); This represents the average three-phase voltage at the previous 15-minute time point (t-1).

[0051] Based on an in-depth analysis of the electrical load characteristics of rail transit stations, this embodiment selects six core indicators that can comprehensively reflect key issues of power quality and constructs an evaluation system as shown in Table 1. This system measures power quality from six dimensions: frequency deviation, total harmonic distortion rate of three-phase current, total harmonic distortion rate of three-phase voltage, three-phase voltage imbalance, power factor deviation, and three-phase voltage fluctuation rate, ensuring the comprehensiveness of the assessment.

[0052] Table 1 Power Quality Evaluation System

[0053] Indicator Name illustrate benchmark value Property probability Frequency deviation Absolute value of system frequency deviation from nominal value 50Hz negative Total harmonic distortion of three-phase current Mean harmonic distortion rate of current 0% negative Total harmonic distortion of three-phase voltage Mean voltage harmonic distortion 0% negative Three-phase voltage imbalance Ratio of maximum phase voltage deviation to average voltage 0% negative Power factor deviation Absolute deviation from the ideal power factor of 1 0 negative

[0054] It should be noted that the deviations of the total harmonic distortion (THD) of the three-phase current and the total harmonic distortion (THD) of the three-phase voltage are obtained by subtracting the calculated mean distortion rate from the reference value. Since the reference value is 0 under ideal conditions, the deviations of the three-phase current and the three-phase voltage THD are both the calculated mean distortion rate. The deviation of the three-phase voltage unbalance is obtained by subtracting the ratio of the maximum phase voltage deviation to the average voltage from the reference value. Since the reference value is 0 under ideal conditions, the deviation of the three-phase voltage unbalance is the ratio of the maximum phase voltage deviation to the average voltage.

[0055] Step S102: Construct a piecewise nonlinear single-item scoring model based on the preprocessed power quality index data, and obtain the power quality single-item score based on the single-item scoring model; wherein, the single-item scoring model is constructed based on the historical deviation data of each power quality index collected within a preset sliding window.

[0056] Specifically, in this embodiment, a piecewise nonlinear single-item scoring model is established. This model can effectively distinguish between the two states of "normal fluctuation" and "abnormal deterioration," avoiding the problem of linear scoring being too low or too high in extreme cases. The piecewise function model considers both linear changes within the normal range and nonlinear penalties for abnormal situations. For any index, its deviation value is... The individual scoring model divides the deduction value for each item into two segments:

[0057] ;

[0058] When the deviation is within the allowable threshold Inside( When the deviation exceeds a threshold, a power function is used for smooth deduction to reflect the tolerance for slight deviations. > When this occurs, a logarithmic function is used for non-linear penalty to control the rate of score deduction. Specifically, This is the deduction value for a single item.

[0059] (2) Parameter settings. , , , The numerical values ​​reflect the current power quality operating status. These parameters were determined through calculation, as shown in Table 2.

[0060] Table 2 Power Quality Operation Status Table

[0061]

[0062] (3) Calculation of individual deductions for power quality. Individual deductions are calculated for frequency deviation, total THD of three-phase current, total THD of three-phase voltage, three-phase voltage imbalance, power factor deviation, and three-phase voltage fluctuation rate. calculate.

[0063] (4) Calculation of individual power quality score. The following formula is used to calculate the individual power quality score:

[0064] ;

[0065] In some implementations, the deviation threshold is determined through dynamic threshold self-learning, including:

[0066] Historical deviation data for each power quality indicator are collected within a sliding window to determine the kernel density estimate over discrete time.

[0067] Based on the kernel density estimation, and according to the cumulative distribution function, a new dynamic deviation threshold is determined. Combined with the deviation threshold of the previous time step, the dynamic deviation threshold of the current time step is determined by a dynamic self-learning threshold algorithm.

[0068] Specifically, in traditional methods, the allowable deviation threshold ε parameter in the single-item scoring model is a fixed value (such as the value corresponding to frequency deviation). The total harmonic distortion rate of three-phase current corresponding to The three-phase voltage fluctuation rate corresponding to Furthermore, these values ​​are all empirical and cannot adapt to changes in system operation (such as increased shifts during holidays or new line access; these changes in application scenarios will lead to corresponding changes in power quality data), ultimately resulting in a deviation between the scoring results and the actual results. This embodiment introduces kernel density estimation (KDE) based on a sliding time window to achieve a deviation threshold. Online self-learning and smooth transition. The following are the specific steps to achieve this:

[0069] (1) Calculate historical deviation data based on the collected power quality index data. As shown in Table 3 (the time interval is 15 minutes, and a total of 31×24×4=2976 samples were taken).

[0070] Table 3 Historical Deviation Data for Each Indicator

[0071] Type\Sequence Number (Time Point) 1 2 3 … 2976 Frequency deviation 0.029 0.02 0.03 … 0.003 Total Harmonic Distortion (THD) of Three-Phase Current 0.1573 0.1577 0.1563 … 0.1707 Total Harmonic Distortion (THD) of Three-Phase Voltage 0.0133 0.014 0.1267 … 0.0133 Three-phase voltage imbalance 0.1270 0.141 0.113 … 0.0709 Power factor deviation 0.1 0.1 0.1 … 0.1 Three-phase voltage fluctuation 0 0.1129 0.1833 … 0.0994

[0072] (2) Determine the sliding window: Set a sliding time window with a length of Len (e.g., 7×24×4=672) 15-minute intervals for local analysis of historical data.

[0073] (3) Kernel density estimation (KDE): Within a sliding window, historical deviation data for each power quality indicator (such as frequency deviation) are collected. ,in As an indicator, Given time, determine the kernel density estimate within a discrete time interval.

[0074] The data was processed using the following kernel density estimation formula:

[0075] ;

[0076] In the formula, K(⋅) is the Gaussian kernel function. for Historical deviation data at any given time for time Points to be estimated for power quality indicators For the bandwidth (which can be determined using the Silverman criterion), the Silverman empirical formula can be used:

[0077] ;

[0078] In the formula, The standard deviation of the data. For data volume.

[0079] (4) Calculation of the cumulative distribution function (CDF):

[0080] ;

[0081] (5) Calculation The quantiles are used as the new dynamic threshold. .

[0082] (6) Using the following formula, combined with the dynamic threshold of the previous time step. The dynamic threshold at the current time is calculated. This enables a smooth transition of thresholds and online self-learning.

[0083] ;

[0084] in, (That is, about 10% is updated every week) to ensure that the threshold transitions smoothly within 1-2 weeks.

[0085] Furthermore, in this embodiment, the smoothing penalty coefficient is obtained. The specific methods are as follows;

[0086] (1) Take the sliding time window data of a certain indicator and calculate the proportion of the indicator's normal fluctuation range. The formula used is:

[0087] ;

[0088] in, To meet The number of samples, For a certain number in the active window, This represents the total number of sliding windows.

[0089] (2) Based on the ratio of the national standard limit to the dynamic threshold The formula used is:

[0090] ;

[0091] in, For international reference, the values ​​are shown in Table 4.

[0092] Table 4 International Rooting Table

[0093] name International Roots Frequency deviation 0.2 Total harmonic distortion of three-phase current) 0.1 Total harmonic distortion of three-phase voltage 0.04 Three-phase voltage imbalance 0.02 Power factor deviation 0.1 Three-phase voltage fluctuation 0.02

[0094] (3) Calculate the smoothing penalty coefficient The specific formula is as follows:

[0095] .

[0096] Furthermore, in this embodiment, the nonlinear penalty coefficient The calculation formula.

[0097] (1) Calculate the skewness of the data exceeding the threshold. It reflects the distribution pattern of abnormal data, and the calculation formula is as follows:

[0098] ;

[0099] (2) Calculate the proportion exceeding the threshold :

[0100] ;

[0101] in, To meet The number of samples, For a certain number in the active window, This represents the total number of sliding windows.

[0102] (3) Calculate the nonlinear penalty coefficient The specific formula is as follows:

[0103] .

[0104] Furthermore, in this embodiment, the penalty boundary coefficient The calculation formula.

[0105] (1) Calculate the coefficient of variation (CV) of the index. The calculation formula is as follows:

[0106] ;

[0107] in, The standard deviation of the data. This is the average value.

[0108] (2) Calculate the ratio of the difference between the national standard limit and the dynamic threshold. The calculation formula is:

[0109] ;

[0110] (3) Penalty boundary coefficient The specific formula is as follows:

[0111] .

[0112] In this step of the embodiment, the piecewise nonlinear scoring model can smooth out slight deviations within the normal range and nonlinearly penalize severe degradation exceeding the threshold, thus more realistically reflecting the degree of power quality degradation and avoiding the "one-size-fits-all" problem of linear scoring.

[0113] Step S103: Determine the initial weights corresponding to each indicator using the entropy weight method, and use the Bayesian algorithm to fuse the initial weights to obtain the final weights.

[0114] Specifically, in this embodiment, traditional evaluation methods often rely heavily on expert experience for index weighting, resulting in strong subjectivity. To overcome this drawback, this embodiment employs the entropy weighting method for objective weighting. The entropy weighting method is based on information theory, arguing that the greater the dispersion of data, the more information it provides, and therefore, it should be assigned a higher weight. The specific calculation steps are as follows:

[0115] (1) Data standardization: To eliminate the influence of dimensions, the individual score data are normalized. Since all indicators are negative, the range method is used for positive standardization, and the formula is as follows:

[0116] ;

[0117] in, For the first The sample at the th The original scores on each indicator and These are the maximum and minimum scores for this indicator, respectively. This is the standardized value.

[0118] (2) Calculate the weight of the indicator: Calculate the weight of the first indicator. The indicator in the first The proportion in each sample Its formula is:

[0119] ;

[0120] (3) Calculate the entropy value: Calculate the first... Entropy value of the item index :

[0121] ;

[0122] in, The total number of samples.

[0123] (4) Calculate the weights: Calculate the weights using the weight formula. , among which, the Weight of each indicator Represented as:

[0124] .

[0125] Furthermore, in this embodiment, although the pure entropy weight method is objective, it may cause weight fluctuations when the sample size is insufficient (such as in the early stage of equipment operation). Therefore, a "weak expert rule" is introduced - if a certain indicator is below 60 points for 5 consecutive days, the "weak expert rule" is triggered.

[0126] (1) Detection data, when the rule is triggered, is determined based on the actual situation or expert experience. The value, If the indicator With a score of only around 50 points for 5 consecutive days, you can select... This indicates a need to increase the weighting of this indicator in the overall evaluation.

[0127] (2) Determine the fusion coefficient Its value approaches 1 as the sample size increases. Assuming we are currently in the initial stage of equipment operation and the sample size is small, we select... (If the sample size is sufficient, It will be closer to 1)

[0128] (3) Calculate the final weights using Bayesian fusion:

[0129] ;

[0130] (4) Replace this moment This will be used in the comprehensive scoring calculation process below.

[0131] (5) If a score exceeding 60 is detected, the original weight value is restored. .

[0132] Step S104: Based on the final weight and the power quality individual score, obtain the comprehensive score result through the pre-constructed comprehensive score model, and conduct a comprehensive evaluation of power quality based on the comprehensive score result.

[0133] Specifically, in this embodiment, the comprehensive scoring model uses the cumulative sum of all individual score values ​​multiplied by the corresponding individual indicator weights. The comprehensive score calculation formula at each time point is as follows:

[0134] ;

[0135] At a certain point in time The overall score is .

[0136] By statistically analyzing all data for a certain month, and using the entropy weight method to calculate the weights of each indicator for a comprehensive score, the results show a mean score of 87.2 points and a standard deviation of 2.1. The score range distribution is [85.1, 91.3]. The grade distribution is: Excellent (AAA) 27.2%, Good (AA) 72.0%, Pass (A) 0.8%, with no poor or warning grades. Overall, the power quality is good. Figure 2 As shown.

[0137] In some implementations, further embodiments also perform a comprehensive analysis of power quality based on a comprehensive score, with the specific steps as follows:

[0138] (1) Rating level setting. Based on the comprehensive rating results, the power quality is divided into five levels: ≥90 points is "Excellent (AAA)", 80~89 points is "Good (AA)", 70~79 points is "Qualified (A)", 60~69 points is "Poor (B)", and <60 points is "Poor (warning required)", thereby realizing an intuitive evaluation and warning of the system's operating status.

[0139] (2) Spatiotemporal characteristic analysis of three-dimensional linkage between "indicators, time, and levels". Traditional power quality analysis is mostly limited to "single-indicator time series analysis" or "single-time-point multi-indicator analysis", lacking linkage analysis of "indicator dimension (six core indicators) - time dimension (daily peak / valley, monthly, annual) - level dimension (AAA to warning)", and cannot fully reveal the dynamic evolution law of power quality and the correlation of weak links. The specific content of the three-dimensional linkage spatiotemporal characteristic analysis adopted in this embodiment is as follows:

[0140] 1) Indicator-Level Linkage: By analyzing the frequency of occurrence of six indicators across five levels ("Excellent", "Good", "Qualified", "Poor", and "Very Poor"), "High-Risk Indicators" can be intuitively identified. For example, if the power factor deviation has no "Excellent" level and the "Very Poor (Warning Required)" level occurs 128 times, it can be identified as a core target for governance.

[0141] 2) Time-Indicator Linkage: Combining the daily cycle (morning peak 7:00-9:00, evening peak 18:00-21:00) and peak-valley periods (peak 18:00-21:00 vs. valley 2:00-5:00), the time-specific degradation characteristics of each indicator are analyzed. For example, the three-phase voltage fluctuation rate scores 15-20 points lower during peak hours than during valley hours, revealing the correlation that "peak load exacerbates voltage fluctuations".

[0142] 3) Time-level linkage: By comparing annually (the same month in different years), we can analyze the evolution trend of the comprehensive rating level. For example, the proportion of "excellent" level decreased from 45% in 2022 to 27.2% in 2024, while the proportion of "qualified" level increased from 0.1% to 0.8%, which quantitatively presents the migration pattern of power quality "deteriorating year by year".

[0143] (3) A closed-loop verification mechanism based on "entropy weight - score - grade". In traditional power quality evaluation, weight calculation, individual scoring, and comprehensive grade classification are independent of each other, lacking a "results-backward verification" mechanism, which easily leads to problems such as "weight not matching actual impact" and "score and grade being disconnected" (e.g., a certain indicator has a high weight but the score is always excellent, or the comprehensive score is high but the grade is qualified). This embodiment establishes a closed-loop verification mechanism of "entropy weight calculation - individual scoring - comprehensive grade":

[0144] 1) Weight-score verification: The weight of the index calculated by the entropy weight method needs to match the "fluctuation range" of the single score. For example, the weight of the three-phase voltage fluctuation rate is the highest (0.219), and its standard deviation of the single score is also the largest (12.4). There are also 0-score samples to verify the consistency between the weight and the actual impact of the index.

[0145] 2) Scoring-Level Verification: The overall score must match the "reasonableness" of the level distribution. For example, the average overall score in May 2024 was 87.2 points (in the "good" range), corresponding to a "good" level of 72.0% and an "excellent" level of 27.2%, with no "poor / warning" level, ensuring the logical consistency between the score and the level.

[0146] 3) Level-actual operation verification: The level classification results need to be linked with the station equipment fault records. For example, the time period when the three-phase voltage imbalance "poor (B)" level occurs overlaps with the time period of the station signal system's "instantaneous fault alarm" by 80%, thus verifying the actual reference value of the level classification.

[0147] This closed-loop mechanism can reverse-check the rationality of the evaluation model. If a certain link is mismatched (e.g., the score of a high-weight indicator does not fluctuate), the model parameters are readjusted (e.g., ...). , , This involves using either the weighting calculation logic or other methods to ensure the accuracy and reliability of the entire evaluation system.

[0148] (4) Establish a mapping relationship between rating levels and operation and maintenance response strategies. This embodiment not only provides a rating for power quality, but also establishes a mapping relationship between rating levels and operation and maintenance response strategies:

[0149] Overall score ≥ 90 (AAA): No intervention required, continue monitoring;

[0150] An overall score of 80-89 (AA) suggests optimizing reactive power compensation.

[0151] Overall score <60 (poor): Warning triggered, troubleshooting initiated;

[0152] A score decline for three consecutive days indicates equipment aging and suggests maintenance.

[0153] This embodiment employs this mechanism, which helps to promote the transformation of power quality management from "monitoring" to "decision support".

[0154] (5) Compile a comprehensive power quality assessment report. This step aims to systematically integrate all the aforementioned analysis results to generate a well-structured, clearly concluded, and specifically recommended "Comprehensive Power Quality Assessment Report." This report not only includes data calculation and statistical results but also decision support content based on multi-dimensional in-depth analysis of "indicators-weights-scores-levels-spatiotemporal characteristics-operation and maintenance." It provides operation and maintenance management personnel with a one-stop solution from current status diagnosis to governance measures, realizing a leap from passive monitoring to proactive early warning and precise decision-making in power quality management.

[0155] like Figure 3 As shown, a longitudinal comparative analysis of power quality in May of 2022, 2023, and 2024 reveals the annual evolution trend of the system's operating status. In all years, the power quality peaks in the early morning (6-7 am), gradually declining from morning to afternoon, and rebounding in the evening, which is highly correlated with the daily load curve. Notably, the peak value in 2022 was significantly higher than in the other two years, while the all-weather score in 2024 was lower than in 2022 and 2023, further confirming the annual downward trend in power quality. In summary, this set of figures clearly demonstrates that the power system has undergone a transformation from "high-quality and stable" to "low-quality and volatile" over the past three years, necessitating measures to address the increasingly severe power quality problem.

[0156] Based on continuous measured data of a typical rail transit station power distribution transformer, this embodiment selects six key indicators: frequency deviation, total harmonic distortion rate of three-phase current, total harmonic distortion rate of three-phase voltage, three-phase voltage imbalance, power factor deviation, and three-phase voltage fluctuation rate. It constructs a multi-dimensional evaluation framework covering steady-state and dynamic characteristics. This method can achieve accurate quantification and dynamic monitoring of power quality in rail transit power distribution systems, providing quantitative basis and decision support for the operation and maintenance decisions of rail transit power supply systems.

[0157] It should be noted that the above description describes some embodiments of this application. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in a different order than that shown in the above embodiments and still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0158] Based on the same inventive concept, and corresponding to the methods of any of the above embodiments, the embodiments of this application also provide a comprehensive evaluation system for power quality of distribution transformers in rail transit stations.

[0159] like Figure 4 As shown, the comprehensive evaluation system for power quality of rail transit station distribution transformers includes:

[0160] The data acquisition module is configured to acquire power quality index data within a predetermined time period;

[0161] The single-item scoring module is configured to construct a piecewise nonlinear single-item scoring model based on the preprocessed power quality index data, and obtain a single power quality score based on the single-item scoring model; wherein, the single-item scoring model is constructed based on the historical deviation data of each power quality index collected within a preset sliding window;

[0162] The fusion processing module is configured to determine the initial weights corresponding to each indicator using the entropy weight method, and then use the Bayesian algorithm to fuse the initial weights to obtain the final weights.

[0163] The comprehensive scoring module is configured to obtain a comprehensive scoring result based on the final weight and the power quality individual score, through a pre-built comprehensive scoring model, and to conduct a comprehensive evaluation of power quality based on the comprehensive scoring result.

[0164] For ease of description, the above system is described by dividing it into various modules based on their functions. Of course, in implementing the embodiments of this application, the functions of each module can be implemented in one or more software and / or hardware.

[0165] The system described in the above embodiments is used to implement the corresponding method in any of the foregoing embodiments and has the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0166] Based on the same inventive concept, corresponding to the methods of any of the above embodiments, embodiments of this application also provide an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the methods described in any of the above embodiments.

[0167] Figure 5 This embodiment illustrates a more specific hardware structure of an electronic device, which may include a processor 1010, a memory 1020, an input / output interface 1030, a communication interface 1040, and a bus 1050. The processor 1010, memory 1020, input / output interface 1030, and communication interface 1040 are interconnected internally via the bus 1050.

[0168] The processor 1010 can be implemented using a general-purpose CPU (Central Processing Unit), microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this specification.

[0169] The memory 1020 can be implemented in the form of ROM (Read Only Memory), RAM (Random Access Memory), static storage device, dynamic storage device, etc. The memory 1020 can store the operating system and other applications. When the technical solutions provided in the embodiments of this specification are implemented by software or firmware, the relevant program code is stored in the memory 1020 and is called and executed by the processor 1010.

[0170] The input / output interface 1030 is used to connect input / output modules to realize information input and output. The input / output modules can be configured as components in the device (not shown in the figure) or externally connected to the device to provide corresponding functions. Input devices may include keyboards, mice, touch screens, microphones, various sensors, etc., and output devices may include displays, speakers, vibrators, indicator lights, etc.

[0171] The communication interface 1040 is used to connect a communication module (not shown in the figure) to enable communication between this device and other devices. The communication module can communicate via wired means (such as USB, Ethernet cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.).

[0172] Bus 1050 includes a pathway for transmitting information between various components of the device, such as processor 1010, memory 1020, input / output interface 1030, and communication interface 1040.

[0173] It should be noted that although the above-described device only shows the processor 1010, memory 1020, input / output interface 1030, communication interface 1040, and bus 1050, in specific implementations, the device may also include other components necessary for normal operation. Furthermore, those skilled in the art will understand that the above-described device may only include the components necessary for implementing the embodiments of this specification, and not necessarily all the components shown in the figures.

[0174] The electronic devices described above are used to implement the corresponding methods in any of the foregoing embodiments and have the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0175] Based on the same inventive concept, corresponding to the methods of any of the above embodiments, this application also provides a non-transitory computer-readable storage medium that stores computer instructions for causing the computer to perform the methods described in any of the above embodiments.

[0176] The computer-readable medium of this embodiment includes permanent and non-permanent, removable and non-removable media, and information storage can be implemented by any method or technology. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transfer medium that can be used to store information accessible by a computing device.

[0177] The computer instructions stored in the storage medium of the above embodiments are used to cause the computer to perform the methods described in any of the above embodiments, and have the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0178] Those skilled in the art should understand that the discussion of any of the above embodiments is merely exemplary and is not intended to imply that the scope of this application (including the claims) is limited to these examples; within the framework of this application, the technical features of the above embodiments or different embodiments can also be combined, the steps can be implemented in any order, and there are many other variations of different aspects of the embodiments of this application as described above, which are not provided in the details for the sake of brevity.

[0179] Additionally, to simplify the description and discussion, and to avoid obscuring the embodiments of this application, the well-known power / ground connections to integrated circuit (IC) chips and other components may or may not be shown in the provided drawings. Furthermore, the apparatus may be shown in block diagram form to avoid obscuring the embodiments of this application, and this also takes into account the fact that the details of the implementation of these block diagram apparatuses are highly dependent on the platform on which the embodiments of this application will be implemented (i.e., these details should be fully understood by those skilled in the art). While specific details (e.g., circuits) have been set forth to describe exemplary embodiments of this application, it will be apparent to those skilled in the art that the embodiments of this application can be implemented without these specific details or with variations thereof. Therefore, these descriptions should be considered illustrative rather than restrictive.

[0180] Although this application has been described in conjunction with specific embodiments thereof, many substitutions, modifications, and variations of these embodiments will be apparent to those skilled in the art from the foregoing description. For example, other memory architectures (e.g., dynamic RAM (DRAM)) may be used with the embodiments discussed.

[0181] The embodiments of this application are intended to cover all such substitutions, modifications, and variations that fall within the broad scope of the appended claims. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the embodiments of this application should be included within the protection scope of this application.

Claims

1. A comprehensive evaluation method for power quality of distribution transformers in rail transit stations, characterized in that, include: Obtain power quality index data within a specified time period; A piecewise nonlinear single-item scoring model is constructed based on the preprocessed power quality index data, and a single-item power quality score is obtained based on the single-item scoring model; wherein, the single-item scoring model is constructed based on the historical deviation data of each power quality index collected within a preset sliding window; The initial weights for each indicator are determined by the entropy weight method, and the Bayesian algorithm is used to fuse the initial weights to obtain the final weights. Based on the final weight and the power quality individual score, a comprehensive score result is obtained through a pre-constructed comprehensive score model, and a comprehensive evaluation of power quality is carried out based on the comprehensive score result. The single-item scoring model is divided into two parts, and the specific formula is as follows: ； The formula for calculating the power quality score is as follows: ； in, This indicates the deviation value corresponding to each indicator; This indicates the deduction value for a single item; Indicates the deviation threshold; Indicates the nonlinear penalty coefficient; Indicates the penalty boundary coefficient; Indicates the smoothing penalty coefficient; The deviation threshold is determined through dynamic threshold self-learning, including: Historical deviation data for each power quality indicator are collected within a sliding window to determine the kernel density estimate over discrete time. Based on the kernel density estimation, and according to the cumulative distribution function, a new dynamic deviation threshold is determined. Combined with the deviation threshold of the previous time step, the dynamic deviation threshold of the current time step is determined by the dynamic self-learning threshold algorithm. The dynamic threshold formula is as follows: ； In the formula, , This indicates the new dynamic deviation threshold. This represents the dynamic threshold at the current moment. This represents the dynamic threshold at the previous time step; Obtain the smoothing penalty coefficient The specific methods are as follows; Take the sliding window data of a certain indicator and calculate the percentage of the indicator's normal fluctuation range. The formula used is: ； in, To meet The number of samples, Let be a number in the sliding window. This represents the total number of sliding windows; Based on the ratio of international rooting to dynamic threshold The formula used is: ； in, Rooted in the international community; Smoothing penalty coefficient The specific formula is: ； Obtaining the nonlinear penalty coefficient The specific methods are as follows; Calculate the skewness of data exceeding the threshold It reflects the distribution pattern of abnormal data, and the calculation formula is as follows: ； Calculate the proportion exceeding the threshold : ； in, To meet The number of samples, Let be a number in the sliding window. This represents the total number of sliding windows; Nonlinear penalty coefficient The specific formula is: ； Obtain the penalty boundary coefficient The specific methods are as follows: The coefficient of variation (CV) of the index is calculated using the following formula: ； in, The standard deviation of the data. This is the average value; Calculate the difference ratio between international rooting and dynamic threshold. The calculation formula is: ； Penalty Boundary Coefficient The specific formula is as follows: 。 2. The method according to claim 1, characterized in that: The preprocessing includes data type unification, missing value handling, and indicator deviation handling to obtain the deviation data corresponding to each indicator.

3. The method according to claim 1, characterized in that: The range method is used to normalize the individual score data, and the entropy weight is calculated on the normalized data to obtain the initial weight of each indicator data. Based on the initial weights of each indicator data, the final weights are obtained by fusing them using a Bayesian algorithm. The specific formula for the Bayesian algorithm is as follows: ； In the formula, Indicates the fusion coefficient; Indicates the initial weights; Indicates the final weight; This indicates the weight percentage.

4. The method according to claim 3, characterized in that, The specific formula for the comprehensive scoring model is as follows: ； In the formula, Indicates the amount of data. express Overall score based on the time point.

5. A comprehensive evaluation system for power quality in rail transit station power distribution transformers, characterized in that, include: The data acquisition module is configured to acquire power quality index data within a predetermined time period; The single-item scoring module is configured to construct a piecewise nonlinear single-item scoring model based on the preprocessed power quality index data, and obtain a single power quality score based on the single-item scoring model; wherein, the single-item scoring model is constructed based on the historical deviation data of each power quality index collected within a preset sliding window; The fusion processing module is configured to determine the initial weights corresponding to each indicator using the entropy weight method, and then use the Bayesian algorithm to fuse the initial weights to obtain the final weights. The comprehensive scoring module is configured to obtain a comprehensive scoring result based on the final weight and the power quality individual score through a pre-built comprehensive scoring model, and to conduct a comprehensive evaluation of power quality based on the comprehensive scoring result; The single-item scoring model is divided into two parts, and the specific formula is as follows: ； The formula for calculating the power quality score is as follows: ； in, This indicates the deviation value corresponding to each indicator; This indicates the deduction value for a single item; Indicates the deviation threshold; Indicates the nonlinear penalty coefficient; Indicates the penalty boundary coefficient; Indicates the smoothing penalty coefficient; The deviation threshold is determined through dynamic threshold self-learning, including: Historical deviation data for each power quality indicator are collected within a sliding window to determine the kernel density estimate over discrete time. Based on the kernel density estimation and a new dynamic bias threshold determined according to the cumulative distribution function, combined with the bias threshold from the previous time step, the dynamic bias threshold for the current time step is determined using a dynamic self-learning threshold algorithm; wherein, the dynamic threshold formula is: ； In the formula, , This indicates the new dynamic deviation threshold. This represents the dynamic threshold at the current moment. This represents the dynamic threshold at the previous time step; Obtain the smoothing penalty coefficient The specific methods are as follows; Take the sliding window data of a certain indicator and calculate the percentage of the indicator's normal fluctuation range. The formula used is: ； in, To meet The number of samples, Let be a number in the sliding window. This represents the total number of sliding windows; Based on the ratio of international rooting to dynamic threshold The formula used is: ； in, Rooted in the international community; Smoothing penalty coefficient The specific formula is: ； Obtaining the nonlinear penalty coefficient The specific methods are as follows; Calculate the skewness of data exceeding the threshold It reflects the distribution pattern of abnormal data, and the calculation formula is as follows: ； Calculate the proportion exceeding the threshold : ； in, To meet The number of samples, Let be a number in the sliding window. This represents the total number of sliding windows; Nonlinear penalty coefficient The specific formula is: ； Obtain the penalty boundary coefficient The specific methods are as follows: The coefficient of variation (CV) of the index is calculated using the following formula: ； in, The standard deviation of the data. This is the average value; Calculate the difference ratio between international rooting and dynamic threshold. The calculation formula is: ； Penalty Boundary Coefficient The specific formula is as follows: 。 6. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the program, implements the method as claimed in any one of claims 1-4.

7. A non-transitory computer-readable storage medium, characterized in that, in, The non-transitory computer-readable storage medium stores computer instructions for causing a computer to perform the method described in any one of claims 1-4.

Images