Lightning protection grounding resistance data processing method and system for smart park

Abstract

The invention relates to the field of data processing, in particular to a lightning protection grounding resistance data processing method and system for a smart park, and the method comprises the steps: obtaining the resistance of each measurement point in each day according to a preset sampling frequency, and constructing a resistance time sequence of each day; for any two measuring points, calculating a correction correlation coefficient between resistance time sequences of the two measuring points on the same day; constructing a measuring point relation graph by taking each measuring point as a node and taking the corrected correlation coefficient as an edge weight; performing clustering analysis based on a graph structure on the measuring point relation graph, dividing all the measuring points into a plurality of categories, and calculating health evaluation values of the measuring points in each category; and evaluating the long-term health state of the corresponding measuring point according to the health evaluation value. According to the invention, a complete analysis chain from data to relation to insight is formed, and an automatic and intelligent solution is provided for long-term health state monitoring of hidden projects such as a smart park grounding grid.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of data processing. More particularly, the present application relates to a lightning protection grounding resistance data processing method and system for a smart park. BACKGROUND

[0002] The lightning protection safety of a smart park highly depends on the reliability of the grounding system. The grounding resistance value is a key indicator for evaluating the grounding performance, and its size directly affects the discharge capacity of lightning current. In actual operation, the grounding resistance value is affected by many factors such as soil humidity, temperature, material aging, etc., and presents the characteristics of continuous dynamic change.

[0003] At present, distributed grounding resistance measuring points are generally deployed in smart parks to realize online monitoring of grounding resistance. However, when analyzing these massive time series data to evaluate the health status of the resistance, the existing methods have obvious deficiencies. The traditional periodic detection or simple threshold alarm method cannot effectively mine the deep information contained in the multi-measuring point data. In particular, for how to accurately identify from the numerous resistance measuring points in the park those groups that have common change trends due to being in the same soil, moisture and other environments. This leads to inaccurate evaluation of the health status of the lightning protection grounding network of the park. SUMMARY

[0004] To solve the technical problems that the existing technology cannot effectively mine the deep information contained in the multi-measuring point data, leading to inaccurate evaluation of the health status of the lightning protection grounding network of the park, the present application provides solutions in the following aspects.

[0005] In a first aspect, a lightning protection grounding resistance data processing method for a smart park comprises: obtaining the resistance of each measuring point for each day according to a preset sampling frequency, and constructing a resistance time series for each day; for any two measuring points, calculating the corrected correlation coefficient between the resistance time series of the same day of the two measuring points; the corrected correlation coefficient is calculated by introducing a weight related to the resistance value difference change of the two measuring points at adjacent time points; constructing a measuring point relationship graph with each measuring point as a node and the corrected correlation coefficient as an edge weight; performing graph structure-based clustering analysis on the measuring point relationship graph, dividing all measuring points into multiple categories, and calculating the health evaluation value of the measuring points in each category; evaluating the long-term health status of the corresponding measuring point according to the health evaluation value.

[0006] Preferably, the weight acquisition comprises: calculating the absolute value of the difference between the resistance values of the two measuring points at the same time of the same day as the first element; for each time, calculating the difference between the first element of the current time and the previous time as the second element of the current time; calculating the average of all second elements of the day; calculating the exponential function value of the absolute value of the difference between the second element of the current time and the average of all second elements of the day as the stability factor; calculating the exponential function value of the absolute value of the second element of the current time as the synchronization factor; multiplying the stability factor and the synchronization factor as the weight between the two measuring points at the current time.

[0007] Preferably, the obtaining of the corrected correlation coefficient comprises: respectively calculating the weighted average of the two measuring points on the same day; calculating the weighted covariance and the weighted standard deviation of the two measuring points according to the weight and the weighted average, and calculating the weighted Pearson correlation coefficient between the two measuring points according to the weighted covariance and the weighted standard deviation, and taking the weighted Pearson correlation coefficient as the corrected correlation coefficient.

[0008] Preferably, the cluster analysis adopts a spectral clustering algorithm.

[0009] Preferably, the obtaining of the health evaluation value comprises: selecting any one of the categories as the target category based on the results of the cluster analysis of the current day; tracing back and obtaining the category containing the most target category elements in the results of the cluster analysis of each previous day, and constructing an evolution process set of the target category in chronological order; selecting a reference measuring point of the target category according to the appearance probability of each measuring point in the evolution process set; for any measuring point in the target category, calculating the health evaluation value of the measuring point according to the appearance probability of the measuring point in the evolution process set and the corrected correlation coefficient between the measuring point and the reference measuring point.

[0010] Preferably, the obtaining of the reference measuring point of the target category comprises: taking the measuring point corresponding to the maximum appearance probability of all measuring points in the target category in the evolution process set as the reference measuring point of the target category.

[0011] Preferably, the product of the appearance probability of the measuring point in the evolution process set and the corrected correlation coefficient between the measuring point and the reference measuring point is taken as the health evaluation value of the measuring point.

[0012] Preferably, the evaluation of the long-term health status of the corresponding measuring point according to the health evaluation value comprises: When the health evaluation value is less than a preset safety threshold, it is determined that the grounding resistance health condition of the corresponding measuring point is abnormal, and an alarm information is automatically generated.

[0013] In a second aspect, a lightning protection grounding resistance data processing system for a smart park comprises a processor and a memory, and the memory stores computer program instructions which, when executed by the processor, implement any of the lightning protection grounding resistance data processing methods for the smart park.

[0014] The present application has the following advantages: The present application introduces weight calculation correction correlation coefficient based on resistance time sequence dynamic change characteristics, constructs a measuring point relationship graph and uses spectral clustering for intelligent division, combines probability analysis of historical clustering evolution and benchmark measuring point correlation evaluation, and realizes accurate mining of deep collaborative relationship and stability between multiple measuring points of the park lightning protection grounding network. This method significantly improves the early identification ability of the hidden abnormal state, effectively overcomes the misjudgment and omission problem caused by traditional single measuring point evaluation or simple correlation coefficient analysis, thereby greatly improving the accuracy and reliability of the grounding resistance health state evaluation, and providing intelligent decision support for preventive maintenance and risk warning of the park lightning protection system. BRIEF DESCRIPTION OF DRAWINGS

[0015] Figure 1 is a method flowchart of steps S1-S4 in a lightning protection grounding resistance data processing method for a smart park according to an embodiment of the present application.

[0016] Figure 2 is a structural block diagram of a lightning protection grounding resistance data processing system for a smart park according to an embodiment of the present application. DETAILED DESCRIPTION

[0017] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments of the present application.

[0018] The monitoring object of the present application is each measuring point, i.e., lightning protection grounding resistance, in a smart park. For ease of description, the measuring point and the resistance value of the measuring point are described below.

[0019] Referring to Figure 1 A lightning protection grounding resistance data processing method for a smart park comprises steps S1-S4, and specifically as follows: S1: Obtain the resistance of each measuring point in each day according to a preset sampling frequency, and construct a resistance time sequence for each day.

[0020] The resistance values ​​at the measurement points are significantly affected by environmental factors, and the resistance values ​​collected under different environments will vary. To ensure strict alignment of the timestamps of the data from each measurement point, a high-precision clock synchronization protocol can be used, and a consistency verification mechanism can be added when the acquisition command is issued. The acquisition terminal at each measurement point starts sampling after receiving unified time synchronization, thereby eliminating time deviations caused by start-up delays or clock drift and ensuring the comparability of data at the same moment.

[0021] In one embodiment, a fixed data collection period of one day is set, with resistance values ​​collected in real time at multiple measuring points within the smart park, in one-hour increments. Over the course of a day, each measuring point will generate a resistance time series ordered chronologically.

[0022] In addition, to ensure the quality of the collected data, all the collected data were subjected to routine preprocessing operations such as filling missing values, removing outliers, and smoothing the data.

[0023] S2: For any two measuring points, calculate the corrected correlation coefficient between the resistance timing of these two measuring points on the same day.

[0024] When multiple measuring points are in the same or similar environment, their resistance values ​​will show similar trends, such as rising at the same time, falling at the same time, or remaining stable at the same time.

[0025] The traditional Pearson correlation coefficient measures a strictly linear relationship, requiring that the changes in two sequences be perfectly proportional (i.e., satisfying a linear relationship). However, in real-world environments: two measuring points in the same area may have different resistivity values ​​due to uneven soil moisture distribution, and environmental factors may have a non-linear effect on resistivity. These factors cause the traditional Pearson correlation coefficient to underestimate the correlation between measuring points with consistent trends.

[0026] Therefore, by introducing weights based on local trend stability, the traditional Pearson correlation coefficient is improved, making it better reflect the consistency in the changing trends of two resistance time series data (even if there is no strict linear relationship between them). The specific calculation steps are as follows: For any two measuring points, firstly, calculate the absolute value of the resistance difference between the two measuring points at the same time on the same day, and use it as the first element. Thus, there is a first element corresponding to a certain time. Calculate the difference between the first element at the current time and the first element at the previous time, and use it as the second element at the current time.

[0027] If the second element is close to 0, it means that the difference in resistance values ​​between the two measuring points remains stable at adjacent moments, that is, they rise / fall synchronously or remain unchanged synchronously; if the second element is large (positive or negative), it means that the direction or magnitude of the change in resistance values ​​between the two measuring points is inconsistent.

[0028] Then, the mean value of all second elements of the day is calculated, the exponential function value of the absolute value of the difference between the second element of the current time and the mean value of all second elements of the day is calculated as a stability factor; the exponential function value of the absolute value of the second element of the current time is calculated as a synchronism factor.

[0029] The product of the stability factor and the synchronism factor calculated above is taken as the weight between the two measuring points at the current time.

[0030] Exemplarily, the weight between the two measuring points at the current time is expressed by a relational expression as follows: In the formula, is the weight between the two measuring points at the current time, is the second element, is the mean value of all second elements of the day, is the exponential function with the natural number e as the base.

[0031] Among them, is the stability factor, which punishes the time point deviating from the average change amount, and if is close to , the weight increases, indicating stable change, and if is far from , the weight decreases, indicating abnormal fluctuation; and even if the change of a certain time point is large, as long as this change pattern is consistent throughout the day, i.e. is close to , a higher weight can still be obtained.

[0032] is the synchronism factor, which punishes the time point with large change amount, and when the resistance difference between the two measuring points changes very small, i.e. is small, it indicates that the two are highly consistent in pace, and regardless of the absolute value, a high weight is given; if is large, the weight decreases, indicating inconsistent trend.

[0033] The weight above is the largest when the change amount is small and stable, and the smallest when the change amount is large or unstable. The time point with high weight represents that the change trend of the two resistance values is highly consistent in this period.

[0034] Finally, the weight calculated above is used to calculate the corrected correlation coefficient between the resistance time series of the two measuring points, i.e. the weighted mean values of the two measuring points on the same day are calculated respectively, the weighted covariance and weighted standard deviation of the two measuring points are calculated according to the weight and the weighted mean value, the weighted Pearson correlation coefficient between the two measuring points is calculated according to the weighted covariance and weighted variance, and the weighted Pearson correlation coefficient is taken as the corrected correlation coefficient. ​

[0035] The above operation is actually a weighted Pearson correlation coefficient, The trend-consistent moment is given a higher weight. Through this weighting method, the calculation focus of the correlation coefficient is directed to the period of "trend consistency and stability". For example, two measuring points may have synchronous resistance decline due to synchronous temperature rise during the day (trend-consistent period), and temporary divergence due to local interference at night (trend-inconsistent period). The traditional method will significantly lower the correlation coefficient due to the divergence at night, while the present method reduces the weight of the night data, so that the final result can better reflect the essential correlation of the two under the influence of the dominant environmental factor (temperature), thereby more accurately distinguishing between "influenced by the common environment" and "independent changes" of the measuring point pair.

[0036] For example, the above modified correlation coefficient is expressed as a relationship: In the formula, is the modified correlation coefficient between the resistance time series of the two measuring points on the same day, is the total number of moments of the day, is the current moment the weight between the two measuring points, is the current moment the resistance value of the measuring point , is the current moment the resistance value of the measuring point , is the weighted mean of the resistance value of the measuring point at the current moment , is the weighted mean of the resistance value of the measuring point at the current moment . The numerator is the weighted covariance, and the denominator is the product of the weighted standard deviations of the two measuring points.

[0037] Since the above weighted Pearson correlation coefficient is an extension of the standard Pearson correlation coefficient, it will not be described in detail here.

[0038] Further, according to the above operation, the modified correlation coefficient between the resistance time series of any two measuring points on any day can be obtained.

[0039] S3: Construct a measuring point relationship graph with each measuring point as a node and the modified correlation coefficient as an edge weight.

[0040] After obtaining the modified correlation coefficient between the resistance time series of any two measuring points on any day, the discrete measuring point resistance values can be converted into a connected graph by taking the measuring points as nodes and the modified correlation coefficient as edge weights. This structure can intuitively show the network relationship between the measuring points, for example: Strongly connected edges: Indicate two measurement points influenced by common factors (e.g., soil stratification, lightning strikes), possibly forming parallel or series paths; Weakly connected edges: May reflect independent faults or local environmental differences.

[0041] Thus, the measurement point relationship graph of each day is obtained.

[0042] S4: Perform graph structure-based clustering analysis on the measurement point relationship graph, divide all measurement points into multiple categories, and calculate the health evaluation value of the measurement points in each category; according to the health evaluation value, the long-term health state of the corresponding measurement point is evaluated.

[0043] In the application scenario of the present application, the change of resistance value with time may present a complex pattern, and the resistance change pattern under different environments may not be convex, and the spectral clustering algorithm can process such non-convex clustering.

[0044] Spectral clustering obtains a low-dimensional representation of data through the eigenvectors of the Laplacian matrix of a graph, which considers the similarity between all nodes in the graph (i.e., global information), not just local information. This helps to discover the potential global structure in the data.

[0045] In this measurement point relationship graph: Community structure: Through clustering analysis, a group of strongly connected nodes may correspond to areas with similar soil composition, humidity, or underground structure in the park, and the electrical characteristics of the grounding grid in these areas tend to converge.

[0046] Key node identification: Measurement points with a large number of strongly connected edges (high-degree), which may be located in sensitive areas of environmental factors or pivotal positions of electrical networks, their state changes will have a wide impact on surrounding measurement points, and are the focus of monitoring.

[0047] Weak edges and bridges: Weakly connected edges connecting different strong communities, which may reflect weak coupling between different geological areas or electrical circuits, and sudden increases in their weights may indicate the spread of anomalies or the emergence of new common interference sources.

[0048] The measurement point relationship graph generated each day is a "snapshot" of the environmental and electrical state of that day. By comparing the graph structures of consecutive days, the evolution of edge weights, the formation and dissipation of communities can be observed, and the spatiotemporal propagation path of environmental factors (such as seasonal changes in humidity) on the overall grounding network of the park can be dynamically tracked.

[0049] Further, spectral clustering algorithm analysis is performed on the measurement point relationship graph of each day to obtain the corresponding spectral clustering result of each day.

[0050] But the spectrum clustering result of each day will change dynamically, because the environmental conditions and resistance values are changing, and it is necessary to find the measuring points that exist stably in each cluster category for a long time as the representative measuring points of the category, namely the reference measuring points.

[0051] In one embodiment, on the basis of the spectrum clustering result of the current day, any category is selected as the target category, and the category containing the most elements of the target category in the spectrum clustering result of each previous day is obtained by tracing back, and the obtained categories are constructed in time sequence to obtain the evolution process set of the target category. For any measuring point in the target category, the appearance probability of the measuring point in the evolution process set is calculated.

[0052] Tracing back and constructing the evolution process set of the target category is essentially a filtering in the time dimension. The high appearance probability of a measuring point in the evolution process set not only represents that it belongs to the target category for a long time, but also means that its resistance change mode is associated with the core environmental driving factor of the target category for a long time, which is a sign of its "reliable identity".

[0053] Further, the measuring point with the maximum appearance probability is selected as the reference measuring point of the target category, which represents the measuring point that is stably in the environment corresponding to the target category for a long time.

[0054] Using the calculation method of the modified correlation coefficient of S2 above, the modified correlation coefficient of the reference measuring point and other measuring points in the target category on the same day is further calculated.

[0055] Finally, the product of the appearance probability of other measuring points in the target category in the evolution process set and the modified correlation coefficient of the measuring point and the reference measuring point is taken as the health evaluation value of the measuring point.

[0056] By coupling the two dimensions of time stability and state coordination, when the appearance probability is low, it means that the category to which the measuring point belongs is unstable, and it may be in the transition zone or frequently interfered by independent factors, and its own state is worth attention, and when the modified correlation coefficient with the reference measuring point is low, it means that even if it is in the same category, its change trend is not synchronized with the reference measuring point of the category, which may indicate local fault or performance degradation.

[0057] The smaller the product is, the more likely it is that the measuring point has deviated from its long-term stable mode.

[0058] Finally, according to the preset safety threshold, when the calculated health evaluation value is less than the preset safety threshold, it is determined that the grounding resistance health condition of the corresponding measuring point is abnormal, and an alarm information is automatically generated, which contains the measuring point ID, abnormal time, evaluation value, and can be associated with the category where it is located, the current state of the reference measuring point, and the comparison of the health trend of the same period in history. This helps the operation and maintenance personnel to quickly locate whether the problem is local (single measuring point failure), category (measuring point group abnormality in a certain environment) or systemic (reference measuring point itself failure or environment change).

[0059] The safety threshold can be set as: The average health degree when the resistance is replaced in the past is counted, the safety threshold is set as the difference of the average health degree minus 0.1, and the 0.1 can reserve a safety margin of 10%.

[0060] The method of the present application guarantees data quality through synchronous acquisition and preprocessing, depicts the dynamic correlation between measuring points through the introduction of trend consistency weight correction correlation coefficient, inspects the system structure from the network perspective through the construction of the clustering measuring point relationship diagram, and finally realizes precise abnormal early warning through the fusion of time stability and state cooperativity health evaluation value. The whole process forms a complete analysis chain from data to relationship to insight, and provides an automatic and intelligent solution for long-term health state monitoring of hidden projects such as grounding grids in smart parks.

[0061] The present application also provides a lightning protection grounding resistance data processing system for a smart park. As shown in Figure 2 The system includes a processor and a memory, and the memory stores computer program instructions, which realize the lightning protection grounding resistance data processing method for a smart park according to the first aspect of the present application when executed by the processor.

[0062] The system also includes a communication bus and a communication interface and other components familiar to those skilled in the art, the settings and functions of which are known in the art, so they will not be described here.

[0063] It should be noted that for those skilled in the art, without departing from the concept of the present application, several modifications and improvements can be made, which are within the scope of the present application. Therefore, the protection scope of the present application patent should be subject to the appended claims.

Claims

1. A method for processing lightning protection grounding resistance data of a smart park, characterized in that, The method comprises the following steps: Obtaining the resistance value of each measuring point in each day according to a preset sampling frequency, and constructing a resistance time sequence of each day; For any two measuring points, calculating the correction correlation coefficient between the resistance time sequences of the same day of the two measuring points; The correction correlation coefficient is calculated by introducing a weight related to the difference value change of the resistance values of the two measuring points at adjacent time points; Taking each measuring point as a node and the correction correlation coefficient as an edge weight, a measuring point relationship graph is constructed; Performing graph structure-based clustering analysis on the measuring point relationship graph, dividing all measuring points into multiple categories, and calculating the health evaluation value of the measuring points in each category; According to the health evaluation value, the long-term health state of the corresponding measuring point is evaluated. 2.The lightning protection grounding resistance data processing method of a smart park according to claim 1, wherein, The weight acquisition comprises the following steps: Calculating the absolute value of the resistance difference value of the same time in the same day of the two measuring points as a first element; for each time, calculating the difference value of the first element between the current time and the previous time as a second element of the current time; calculating the mean value of all second elements of the day; Calculating the exponential function value of the absolute value of the difference between the second element of the current time and the mean value of all second elements of the day as a stability factor; Calculating the exponential function value of the absolute value of the second element of the current time as a synchronization factor; The product of the stability factor and the synchronization factor is taken as the weight between the two measuring points at the current time. 3.The lightning protection grounding resistance data processing method of a smart park according to claim 2, wherein, The correction correlation coefficient acquisition comprises the following steps: Respectively calculating the weighted mean value of the two measuring points in the same day; calculating the weighted covariance and weighted standard deviation of the two measuring points according to the weight and the weighted mean value, and calculating the weighted Pearson correlation coefficient between the two measuring points according to the weighted covariance and the weighted standard deviation, and taking the weighted Pearson correlation coefficient as the correction correlation coefficient.

4. The lightning protection grounding resistance data processing method for a smart park according to claim 1, characterized in that, The clustering analysis adopts a spectral clustering algorithm.

5. The lightning protection grounding resistance data processing method for a smart park according to claim 1, characterized in that, The health evaluation value acquisition comprises the following steps: Selecting any category in the clustering analysis result of the current day as a target category based on the clustering analysis result of the current day; Tracing back and obtaining the category containing the most target category elements in the clustering analysis result of each previous day, and constructing an evolution process set of the target category in chronological order; Selecting a reference measuring point of the target category according to the appearance probability of each measuring point in the target category in the evolution process set; For any measuring point in the target category, the health evaluation value of the measuring point is calculated according to the appearance probability of the measuring point in the evolution process set and the correction correlation coefficient between the measuring point and the reference measuring point.

6. The lightning protection grounding resistance data processing method of the smart park according to claim 5, characterized in that, The reference measuring point of the target category is obtained by: Taking the measuring point corresponding to the maximum appearance probability of all measuring points in the target category in the evolution process set as the reference measuring point of the target category.

7. The lightning protection grounding resistance data processing method of the smart park according to claim 6, characterized in that, The product of the appearance probability of the measuring point in the evolution process set and the correction correlation coefficient between the measuring point and the reference measuring point is taken as the health evaluation value of the measuring point. 8.The data processing method for lightning protection grounding resistance of a smart park according to claim 1, wherein, The long-term health state of the corresponding measuring point is evaluated according to the health evaluation value, which comprises the following steps: When the health evaluation value is less than a preset safety threshold, it is determined that the grounding resistance health condition of the corresponding measuring point is abnormal, and an alarm information is automatically generated.

9. A lightning protection grounding resistance data processing system for a smart park, characterized in that, The method comprises the following steps: A processor and a memory, the memory storing computer program instructions which, when executed by the processor, implement the lightning protection grounding resistance data processing method of the smart park according to any one of claims 1-8.

Images