Data intelligent analysis system and method for power cable grounding resistance

Abstract

The invention discloses a data intelligent analysis system and method for power cable grounding resistance, and relates to the technical field of grounding resistance analysis, and the system comprises a time sequence data generation module, a feature data group analysis module, a mathematical model establishment module and an adjustment strategy distribution module. The time sequence data generation module is used for generating continuous time sequence data from the data collected according to the historical record according to a preset time interval; the feature data group analysis module is used for performing feature extraction on the data in all the sets to generate feature data groups; the mathematical model establishing module is used for establishing a mathematical model for measuring the grounding resistance and each characteristic data in the characteristic data group; and the adjustment strategy distribution module is used for performing type judgment on the feature influence factors in the target data set and executing different adjustment strategies based on different types.

Description

Technical Field

[0001] This invention relates to the field of grounding resistance analysis technology, specifically to a data intelligent analysis system and method for the grounding resistance of power cables. Background Technology

[0002] Accurate measurement of grounding resistance is crucial for ensuring the safe and stable operation of high-voltage power cables. While existing methods and devices for measuring power cable grounding resistance based on data curve fitting have addressed online measurement to some extent, they still have several shortcomings. Firstly, traditional methods often employ fixed measurement parameters, such as fixed sampling frequencies and capacitor pre-charge voltages, which cannot be dynamically adjusted according to the actual conditions during power cable operation. However, the operating environment of power cables is complex and variable; environmental temperature, humidity, and electromagnetic interference constantly affect the accuracy of grounding resistance measurement. When environmental interference is strong, a fixed sampling frequency may not be able to acquire sufficient data to accurately calculate the grounding resistance, leading to increased measurement errors. Secondly, the grounding resistance itself changes with time and cable operating conditions. When grounding resistance changes rapidly, existing measurement parameters are difficult to adapt to these changes, also reducing measurement accuracy. Furthermore, existing power cable grounding resistance measurement technologies only adjust sampling parameters for a single factor (such as electromagnetic interference), failing to address the error amplification problem caused by the coupling of multiple factors. For example, increased environmental humidity may exacerbate moisture absorption in the cable insulation layer, and when accompanied by electromagnetic interference, traditional fixed sampling frequencies cannot account for both types of error sources, resulting in a significant decrease in measurement accuracy. Furthermore, the lack of a quantitative assessment mechanism for the degree of influence of each factor can easily lead to delayed or over-adjusted adjustment strategies, affecting measurement efficiency. Therefore, developing an intelligent analysis system and method for power cable grounding resistance data that can adaptively adjust measurement parameters is of great practical significance. Summary of the Invention

[0003] The purpose of this invention is to provide a data intelligent analysis system and method for the grounding resistance of power cables, so as to solve the problems raised in the prior art.

[0004] To achieve the above objectives, the present invention provides the following technical solution: a data intelligent analysis method for the grounding resistance of power cables, the method comprising: Step S1: Mark the power cable monitoring area based on the function of voltage sensors and environmental monitoring equipment; extract the capacitance voltage and environmental data in the power cable grounding circuit collected under the normal working condition of the historical grounding resistance in the monitoring area; generate continuous time series data from the historical data collected according to the preset time interval; Step S2: Divide the time series data evenly into several equal parts according to the same time interval; store the data within each time interval into the same set; and extract features from the data in all sets to generate feature data groups; Step S3: Based on the feature data set, establish a mathematical model of the measured grounding resistance and each feature data in the feature data set; Step S4: Based on the established mathematical model, extract the data set when the grounding resistance response is abnormal and use it as the target data set; determine the type of the characteristic influencing factors in the target data set, and implement different adjustment strategies based on different types.

[0005] Furthermore, step S2 includes the following specific steps: Step S21: The time series data includes the voltage value V(t) of the capacitor collected by the voltage sensor during the second-order oscillation process, the ambient temperature W(t) and ambient humidity H(t) recorded by the temperature and humidity sensor, and the electromagnetic interference intensity Iem(t) recorded by the electromagnetic interference sensor; each set contains the voltage value V(t), ambient temperature W(t), ambient humidity H(t), and electromagnetic interference intensity Iem(t). Step S22: Extract features from the voltage values ​​in each set, calculate the rate of change of capacitor voltage dV(t) / dt, where t represents the time interval length recorded in each set; and calculate the standard deviation q of voltage fluctuation within the time interval length t based on the voltage value V(t). V ; The ambient temperature W(t) and ambient humidity H(t) recorded within the time interval are acquired, and the response times when the temperature and humidity change are marked. Based on the time interval between each response time and the initial monitoring time of the set, a temperature parameter response period Z1 and a humidity parameter response period Z2 are generated for each temperature data and ambient humidity data. The temperature parameter response periods and ambient temperature are combined to form a temperature data group [Z1, W(t)], and the humidity parameter response periods and ambient humidity are combined to form a humidity data group [Z2, H(t)]. The temperature function relationship w is constructed for each, w = K. W *Z1+ε1, humidity function relationship h, h=K H *Z2+ε2, where ε1 represents the error term coefficient corresponding to the temperature function, and ε2 represents the error term coefficient corresponding to the humidity function; K W The slope of temperature change over time, K H This represents the slope of humidity change over time; the corresponding slope K is calculated by substituting the temperature data set into the temperature function formula. W As feature data, the humidity data set is substituted into the humidity function to calculate the corresponding slope K. H As feature data; Step S23: Calculate the rate of change of capacitor voltage, standard deviation, slope of temperature change over time, slope of humidity change over time, and mean electromagnetic interference intensity I. em0 This constitutes the characteristic data groups for each corresponding set.

[0006] Furthermore, step S3 includes the following specific steps: Step S31: Extract the maximum grounding resistance Rmax recorded in the corresponding time period of each set, construct the maximum grounding resistance and the feature data group in the corresponding set into a correlation data pair, and generate the correlation data pair corresponding to each set; Step S32: Establish the resistance regression model R. R = β0 + β1 * [dV(t) / dt] + β2 * q V +β3*K W +β4*K H +β5*I em0 +ε0; where β0, β1, β2, β3, β4, and β5 represent the regression coefficients of the corresponding independent variables; ε0 is the corresponding error term; Step S33: Substitute all the corresponding associated data pairs into the set, calculate the corresponding regression coefficients and error terms, and construct the complete resistance regression model R.

[0007] This application establishes a mathematical model between measurement parameters, grounding resistance, and environmental factors, which can effectively and data-drivenly distinguish the factors that may affect grounding resistance during the measurement process, as well as the resistance prediction estimate when the factors change dynamically, thereby greatly improving the precision monitoring of grounding resistance.

[0008] Furthermore, the data set extracted when the grounding resistance response is abnormal is designated as the target data set; the type determination of the characteristic influencing factors in the target data set includes the following specific steps: Step S41: Obtain the response event applied after the model is established. The response event refers to the event that triggers the response mechanism when the input data obtained during the target time period is substituted into the model for calculation, and the predicted grounding resistance value R is obtained, the difference between the predicted grounding resistance value R and the actual measured grounding resistance R0 is calculated, and the difference is greater than or equal to the difference threshold. A difference greater than the difference threshold indicates that the measurement error may increase. Step S42: Extract the input data recorded in the response event as the target data set; judge each feature data in the target data set, the judgment process is as follows: Obtain the input capacitor voltage change rate and the corresponding set change rate threshold. If the input capacitor voltage change rate is greater than the change rate threshold, mark the capacitor voltage change rate as an influencing factor of the target feature; otherwise, do not mark it. Obtain the standard deviation q of the input. Vand the corresponding standard deviation threshold q th , in q V >q th At that time, the standard deviation was labeled as an influencing factor of the target feature; The system acquires the input temperature and humidity data, and calculates the corresponding real-time temperature change slope and humidity change slope for the corresponding monitoring period. If the difference between the temperature change slope and the preset normal temperature change slope is greater than the first slope difference threshold, the temperature data is marked as a target feature influencing factor. If the difference between the humidity change slope and the preset normal humidity change rate is greater than the second slope difference threshold, the humidity data is marked as a target feature influencing factor. Obtain the input electromagnetic interference intensity Iem0 and the set electromagnetic interference intensity threshold Ith. If Iem0>Ith, then mark the electromagnetic interference intensity as a target feature influencing factor. Step S43: Obtain the number M of feature data marked as target feature influencing factors in the same response event. When M=1, the output is a Class I response event, and the target feature influencing factors corresponding to the event are Class I influencing factors. When M>1, the output is a Class II response event, and the target feature influencing factors in the Class II response event are Class II influencing factors.

[0009] Furthermore, the implementation of different adjustment strategies based on different types includes the following steps: For each type of influencing factor in each response event, an independent response regulation strategy is implemented; Priority analysis was performed on the two types of influencing factors in each response event. The specific analysis process is as follows: Calculate the standard value X for each of the two types of influencing factors. norm , X norm =(X real -X min ) / (X max -X min ), where X real In response to the real-time monitoring value corresponding to the event, X min X max These represent the minimum and maximum values ​​of the two types of influencing factors within the normal working range; Obtain the historical record of the i-th type II influencing factor as the number Di of type II influencing factors in the response event, using the formula: R i =D i / [(D1+D2+...+D n )] Calculate the weight coefficient R of the i-th type II response factor in the corresponding response event. i , i∈[1,n]; D1+D2+...+D nThis represents the summation of the D values ​​for the first to nth binary influencing factor records in the response event, where n represents the total number of binary influencing factors in the corresponding response event; Based on the weighting coefficients and corresponding standard values, calculate the comprehensive impact quantification value Qi of the i-th type II influencing factor, Qi=R i *X norm,i ;X norm,i This represents the standard value of the i-th type II influencing factor; For the same response event, each of the two types of influencing factors is sorted in descending order of comprehensive impact quantification value to generate a priority sequence; The strategies corresponding to each of the two types of influencing factors are responded to sequentially based on the priority sequence.

[0010] This application can dynamically adjust the weights of different influencing factors based on real-time dynamic monitoring data, making the analyzed execution strategy priority accurate and adaptive; adapting to the differences in working conditions in different seasons and regions; and achieving "prioritizing the adjustment of high-impact factors" by quantifying the comprehensive impact value of each factor and dynamically sorting them, thus avoiding the blindness of traditional empirical methods. For example, when humidity and electromagnetic interference exceed the standard at the same time, the system can automatically identify that humidity has a more significant impact and prioritize increasing high-frequency sampling to capture subtle changes in the voltage curve caused by moisture.

[0011] A data intelligent analysis system for the grounding resistance of power cables, the system comprising a time series data generation module, a feature data group analysis module, a mathematical model establishment module, and a regulation strategy allocation module; The time series data generation module is used to generate continuous time series data from historical data collected at preset time intervals. The feature data set analysis module is used to extract features from the data in all sets to generate feature data sets. The mathematical model building module is used to build a mathematical model for measuring grounding resistance and each characteristic data in the characteristic data group; The adjustment strategy allocation module is used to determine the type of the feature influencing factors in the target data set and to execute different adjustment strategies based on different types.

[0012] Furthermore, the mathematical model building module includes a data pair generation unit and a resistance regression model building unit; The associated data pair generation unit is used to extract the maximum grounding resistance recorded in the corresponding time period of each set, construct an associated data pair with the maximum grounding resistance and the feature data group in the corresponding set, and generate an associated data pair for each set. The resistance regression model building unit is used to substitute all the corresponding associated data pairs into the set, calculate the corresponding regression coefficients and error terms, and construct a complete resistance regression model.

[0013] Furthermore, the adjustment strategy allocation module includes a response event determination unit and a feature data differentiation unit; The response event determination unit is used to obtain the response event applied after the model is established. The response event refers to the event that triggers the response mechanism when the input data obtained within the target time period is substituted into the model for calculation, the predicted grounding resistance value is obtained, the difference between the predicted grounding resistance value and the actual measured grounding resistance is calculated, and the difference is greater than or equal to the difference threshold. The feature data differentiation unit is used to extract the input data recorded in the response event as the target data set; to judge each feature data in the target data set, and to output feature data of different types.

[0014] Furthermore, the adjustment strategy allocation module also includes a priority analysis unit and a sorting response unit; The priority analysis unit is used to perform priority analysis on the two types of influencing factors in each response event; The sorting response unit is used to sort the two types of influencing factors in the same response event in descending order based on the comprehensive influence quantification value to generate a priority sequence. The strategies corresponding to each of the two types of influencing factors are responded to sequentially based on the priority sequence.

[0015] Compared with the prior art, the beneficial effects of the present invention are: 1. This invention acquires real-time data on the grounding circuit of power cables and environmental data, performs feature extraction and modeling analysis, and can dynamically adjust measurement parameters according to actual conditions. When electromagnetic interference is significant, the sampling frequency is increased to ensure sufficient data for accurate calculations; when grounding resistance changes drastically, the capacitor pre-charge voltage is adjusted to optimize the measurement process, thereby effectively reducing measurement errors, improving the accuracy of grounding resistance measurement, and providing more reliable data support for the safe operation of power cables.

[0016] 2. This invention achieves "prioritizing the adjustment of high-impact factors" by quantifying the comprehensive impact value of each factor and dynamically sorting them, thus avoiding the blindness of traditional empirical methods; it optimizes factor weights in real time through correlation analysis of historical data to adapt to the differences in working conditions in different seasons and regions; at the same time, it reduces the number of invalid sampling frequency adjustments and extends the service life of equipment. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the structure of a data intelligent analysis method for the grounding resistance of power cables according to the present invention. Detailed Implementation

[0018] Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Example: Figure 1 As shown, this invention provides a data intelligent analysis method for the grounding resistance of power cables, the method comprising: Step S1: Mark the power cable monitoring area based on the function of voltage sensors and environmental monitoring equipment; extract the capacitance voltage and environmental data in the power cable grounding circuit collected under the normal working condition of the historical grounding resistance in the monitoring area; generate continuous time series data from the historical data collected according to the preset time interval; Step S2: Divide the time series data evenly into several equal parts according to the same time interval; store the data within each time interval into the same set; and extract features from the data in all sets to generate feature data groups; Step S3: Based on the feature data set, establish a mathematical model of the measured grounding resistance and each feature data in the feature data set; Step S4: Based on the established mathematical model, extract the data set when the grounding resistance response is abnormal and use it as the target data set; determine the type of the characteristic influencing factors in the target data set, and implement different adjustment strategies based on different types.

[0020] Step S2 includes the following specific steps: Step S21: The time series data includes the voltage value V(t) of the capacitor collected by the voltage sensor during the second-order oscillation process, the ambient temperature W(t) and ambient humidity H(t) recorded by the temperature and humidity sensor, and the electromagnetic interference intensity Iem(t) recorded by the electromagnetic interference sensor; each set contains the voltage value V(t), ambient temperature W(t), ambient humidity H(t), and electromagnetic interference intensity Iem(t). Step S22: Extract features from the voltage values ​​in each set, calculate the rate of change of capacitor voltage dV(t) / dt, where t represents the time interval length recorded in each set; and calculate the standard deviation q of voltage fluctuation within the time interval length t based on the voltage value V(t). V ; The ambient temperature W(t) and ambient humidity H(t) recorded within the time interval are acquired, and the response times when the temperature and humidity change are marked. Based on the time interval between each response time and the initial monitoring time of the set, a temperature parameter response period Z1 and a humidity parameter response period Z2 are generated for each temperature data and ambient humidity data. The temperature parameter response periods and ambient temperature are combined to form a temperature data group [Z1, W(t)], and the humidity parameter response periods and ambient humidity are combined to form a humidity data group [Z2, H(t)]. The temperature function relationship w is constructed for each, w = K. W *Z1+ε1, humidity function relationship h, h=K H *Z2+ε2, where ε1 represents the error term coefficient corresponding to the temperature function, and ε2 represents the error term coefficient corresponding to the humidity function; K W The slope of temperature change over time, K H This represents the slope of humidity change over time; the corresponding slope K is calculated by substituting the temperature data set into the temperature function formula. W As feature data, the humidity data set is substituted into the humidity function to calculate the corresponding slope K. H As feature data; Step S23: Calculate the rate of change of capacitor voltage, standard deviation, slope of temperature change over time, slope of humidity change over time, and mean electromagnetic interference intensity I. em0 This constitutes the feature data groups corresponding to each set.

[0021] Step S3 includes the following specific steps: Step S31: Extract the maximum grounding resistance Rmax recorded in the corresponding time period of each set, construct the maximum grounding resistance and the feature data group in the corresponding set into a correlation data pair, and generate the correlation data pair corresponding to each set; Step S32: Establish the resistance regression model R. R = β0 + β1 * [dV(t) / dt] + β2 * q V +β3*K W +β4*K H +β5*I em0 +ε0; where β0, β1, β2, β3, β4, and β5 represent the regression coefficients of the corresponding independent variables; ε0 is the corresponding error term; Step S33: Substitute all the corresponding associated data pairs into the set, calculate the corresponding regression coefficients and error terms, and construct the complete resistance regression model R.

[0022] In this application, the frequency of historical measurement of grounding resistance is a fixed frequency under normal operating conditions of the power cable, meaning that the sampling frequency is the same in each time interval corresponding to each set.

[0023] This application establishes a mathematical model between measurement parameters, grounding resistance, and environmental factors, which can effectively and data-drivenly distinguish the factors that may affect grounding resistance during the measurement process, as well as the resistance prediction estimate when the factors change dynamically, thereby greatly improving the precision monitoring of grounding resistance.

[0024] The data set extracted when the grounding resistance response is abnormal is the target data set; the characteristic influencing factors in the target data set are classified, including the following specific steps: Step S41: Obtain the response event applied after the model is established. The response event refers to the event that triggers the response mechanism when the input data obtained during the target time period is substituted into the model for calculation, and the predicted grounding resistance value R is obtained, the difference between the predicted grounding resistance value R and the actual measured grounding resistance R0 is calculated, and the difference is greater than or equal to the difference threshold. A difference greater than the difference threshold indicates that the measurement error may increase. Step S42: Extract the input data recorded in the response event as the target data set; judge each feature data in the target data set, the judgment process is as follows: Obtain the input capacitor voltage change rate and the corresponding set change rate threshold. If the input capacitor voltage change rate is greater than the change rate threshold, mark the capacitor voltage change rate as an influencing factor of the target feature; otherwise, do not mark it. Obtain the standard deviation q of the input. V and the corresponding standard deviation threshold q th , in q V >q th At that time, the standard deviation was labeled as an influencing factor of the target feature; The system acquires the input temperature and humidity data, and calculates the corresponding real-time temperature change slope and humidity change slope for the corresponding monitoring period. If the difference between the temperature change slope and the preset normal temperature change slope is greater than the first slope difference threshold, the temperature data is marked as a target feature influencing factor. If the difference between the humidity change slope and the preset normal humidity change rate is greater than the second slope difference threshold, the humidity data is marked as a target feature influencing factor. Obtain the input electromagnetic interference intensity Iem0 and the set electromagnetic interference intensity threshold Ith. If Iem0>Ith, then mark the electromagnetic interference intensity as a target feature influencing factor. Step S43: Obtain the number M of feature data marked as target feature influencing factors in the same response event. When M=1, the output is a Class I response event, and the target feature influencing factors corresponding to the event are Class I influencing factors. When M>1, the output is a Class II response event, and the target feature influencing factors in the Class II response event are Class II influencing factors.

[0025] In this application, the case of M=0 is not considered. M=0 indicates that although the grounding resistance measurement is abnormal, it is not caused by the above input data.

[0026] The implementation of different adjustment strategies based on different types includes the following steps: For each type of influencing factor in each response event, an independent response regulation strategy is implemented; As shown in the example: when one type of influencing factor is the intensity of electromagnetic interference, the model determines that the current measurement environment has a large interference. At this time, the system automatically increases the sampling frequency from the original n times per second to m times per second; m>n; When one of the influencing factors is the rate of change of grounding resistance, the system automatically adjusts the pre-charge voltage of the capacitor to optimize the second-order oscillation process and make the measurement results more accurate. When one type of influencing factor is ambient humidity or temperature, the system automatically increases the sampling frequency from n samples per second to m samples per second. When one type of influencing factor is the standard deviation of voltage fluctuation, if qV is greater than 1.5 times the mean of the standard deviation under normal operating conditions, the measurement time interval should be shortened and the number of measurement points within the voltage fluctuation cycle should be increased to obtain voltage change information more accurately. Priority analysis was performed on the two types of influencing factors in each response event. The specific analysis process is as follows: Calculate the standard value X for each of the two types of influencing factors. norm , X norm =(X real -X min ) / (X max -X min ), where X real In response to the real-time monitoring value corresponding to the event, X min X max These represent the minimum and maximum values ​​of the two types of influencing factors within the normal working range; Obtain the historical record of the i-th type II influencing factor as the number Di of type II influencing factors in the response event, using the formula: R i =D i / [(D1+D2+...+D n )] Calculate the weight coefficient R of the i-th type II response factor in the corresponding response event. i , i∈[1,n]; D1+D2+...+D n This represents the summation of the D values ​​for the first to nth binary influencing factor records in the response event, where n represents the total number of binary influencing factors in the corresponding response event; Based on the weighting coefficients and corresponding standard values, calculate the comprehensive impact quantification value Qi of the i-th type II influencing factor, Qi=R i *Xnorm,i ;X norm,i This represents the standard value of the i-th type II influencing factor; For the same response event, each of the two types of influencing factors is sorted in descending order of comprehensive impact quantification value to generate a priority sequence; The strategies corresponding to each of the two types of influencing factors are responded to sequentially based on the priority sequence.

[0027] As shown in the example: If the analysis of a certain response event includes two types of influencing factors, namely electromagnetic interference, ambient humidity, and voltage fluctuation; If the normal range of electromagnetic interference intensity is (0–100 dB), and the real-time value is 150 dB, then the standardized value is: X norm =(X real -X min ) / (X max -X min ) = (150-100) / 100 = 0.5; The comprehensive impact values ​​were calculated to be 0.15, 0.3, and 0.2 respectively. At this point, the corresponding priority sequence is: ambient humidity > voltage fluctuation > electromagnetic interference; The algorithm prioritizes increasing the sampling frequency for humidity, then handles electromagnetic interference, and finally optimizes the voltage fluctuation filtering algorithm.

[0028] This application can dynamically adjust the weights of different influencing factors based on real-time dynamic monitoring data, making the analyzed execution strategy priority accurate and adaptive; adapting to the differences in working conditions in different seasons and regions; and achieving "prioritizing the adjustment of high-impact factors" by quantifying the comprehensive impact value of each factor and dynamically sorting them, thus avoiding the blindness of traditional empirical methods. For example, when humidity and electromagnetic interference exceed the standard at the same time, the system can automatically identify that humidity has a more significant impact and prioritize increasing high-frequency sampling to capture subtle changes in the voltage curve caused by moisture.

[0029] A data intelligent analysis system for the grounding resistance of power cables, the system comprising a time series data generation module, a feature data group analysis module, a mathematical model establishment module, and a regulation strategy allocation module; The time series data generation module is used to generate continuous time series data from historical data collected at preset time intervals. The feature data set analysis module is used to extract features from the data in all sets to generate feature data sets. The mathematical model building module is used to build a mathematical model for measuring grounding resistance and each characteristic data in the characteristic data group; The adjustment strategy allocation module is used to determine the type of the feature influencing factors in the target data set and to execute different adjustment strategies based on different types.

[0030] The mathematical model building module includes a correlation data pair generation unit and a resistance regression model building unit. The associated data pair generation unit is used to extract the maximum grounding resistance recorded in the corresponding time period of each set, construct an associated data pair with the maximum grounding resistance and the feature data group in the corresponding set, and generate an associated data pair for each set. The resistance regression model building unit is used to substitute all the corresponding associated data pairs into the set, calculate the corresponding regression coefficients and error terms, and construct a complete resistance regression model.

[0031] The adjustment strategy allocation module includes a response event determination unit and a feature data differentiation unit; The response event determination unit is used to obtain the response event applied after the model is established. The response event refers to the event that triggers the response mechanism when the input data obtained within the target time period is substituted into the model for calculation, the predicted grounding resistance value is obtained, the difference between the predicted grounding resistance value and the actual measured grounding resistance is calculated, and the difference is greater than or equal to the difference threshold. The feature data differentiation unit is used to extract the input data recorded in the response event as the target data set; to judge each feature data in the target data set, and to output feature data of different types.

[0032] The adjustment strategy allocation module also includes a priority analysis unit and a sorting response unit; The priority analysis unit is used to perform priority analysis on the two types of influencing factors in each response event; The sorting response unit is used to sort the two types of influencing factors in the same response event in descending order based on the comprehensive influence quantification value to generate a priority sequence. The strategies corresponding to each of the two types of influencing factors are responded to sequentially based on the priority sequence.

[0033] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for data intelligent analysis of power cable grounding resistance, characterized by: The method comprises: Step S1: Marking the power cable monitoring area based on the voltage sensor and the environmental monitoring device; extracting the capacitive voltage in the power cable grounding loop and the environmental data collected under the normal working state of the historical record of the monitoring area; and generating continuous time series data according to the data collected by the historical record in a preset time interval; Step S2: uniformly dividing the time series data into several equal parts according to the same interval period; storing the data in each interval period into the same set; and extracting the feature data group from the data in all sets; Step S3: based on the feature data group, a mathematical model of measuring the grounding resistance and each feature data in the feature data group is established; Step S4: based on the established mathematical model, the data set when the measurement grounding resistance responds abnormally is extracted as the target data set; the type of the feature influencing factor in the target data set is judged, and different adjustment strategies are executed based on different types.

2. A data intelligent analysis method for power cable grounding resistance according to claim 1, characterized in that: The step S2 comprises the following specific steps: Step S21: the time series data comprises the voltage value V(t) of the capacitive voltage collected by the voltage sensor in the second-order oscillation process, the environmental temperature W(t) recorded by the temperature and humidity sensor, the environmental humidity H(t) and the electromagnetic interference intensity Iem(t) recorded by the electromagnetic interference sensor; the voltage value V(t), the environmental temperature W(t), the environmental humidity H(t) and the electromagnetic interference intensity Iem(t) are contained in each set; Step S22: feature extraction is performed on the voltage values in each set, the rate of change of the capacitor voltage dV(t) / dt is calculated, t represents the length of the time interval recorded in each set; and the standard deviation q of the voltage fluctuation within the time interval length t is calculated based on the voltage value V(t) V ; The recorded ambient temperature W(t) and ambient humidity H(t) in the time interval length are obtained, the response moments of the temperature and humidity generating numerical changes are marked, the temperature parameter response period Z1 corresponding to each temperature data and the humidity parameter response period Z2 corresponding to the ambient humidity data are generated based on the time interval of each response moment and the initial monitoring moment of the set, the temperature parameter response period and the ambient temperature constitute a temperature data group [Z1, W(t)], the humidity parameter response period and the ambient humidity constitute a humidity data group [Z2, H(t)], and a temperature function relationship w, w=K W *Z1+ε1, a humidity function relationship h, h=K H *Z2+ε2, wherein ε1 represents an error term coefficient corresponding to the temperature function relationship, and ε2 represents an error term coefficient corresponding to the humidity function relationship; K W represents a change slope of the temperature with time, and K H represents a change slope of the humidity with time; the temperature data group is substituted into the temperature function relationship to calculate the corresponding change slope K W as characteristic data, the humidity data group is substituted into the humidity function relationship to calculate the corresponding change slope K H as characteristic data. Step S23: Calculate the rate of change of the capacitor voltage, the standard deviation, the slope of the change in temperature over time, the slope of the change in humidity over time, and the mean value of the electromagnetic interference intensity I em0 The feature data sets corresponding to each set are constituted.

3. A data intelligent analysis method for power cable grounding resistance according to claim 2, characterized in that: The step S3 comprises the following specific steps: Step S31: extracting the maximum value Rmax of the grounding resistance recorded in each set corresponding period, and constructing the associated data pair of the maximum value of the grounding resistance and the feature data group in the corresponding set to generate the associated data pair corresponding to each set; Step S32: establishing a resistance regression model R, R = β0 + β1 * [dV(t) / dt] + β2 * q V + β3 * K W + β4 * K H + β5 * I em0 + ε0; wherein β0, β1, β2, β3, β4, β5 represent regression coefficients corresponding to the independent variables, respectively; and ε0 is the error term corresponding. Step S33: substituting the associated data pairs corresponding to all sets to calculate the corresponding regression coefficients and error terms to construct the complete resistance regression model R.

4. A data intelligent analysis method for power cable grounding resistance according to claim 3, characterized in that: The data set when the measurement grounding resistance responds abnormally is extracted as the target data set; the type of the feature influencing factor in the target data set is judged, which comprises the following specific steps: Step S41: obtaining the response event applied after the model is established, the response event refers to obtaining the input data in the target period, substituting the input data into the model to calculate the predicted grounding resistance value R, calculating the difference between the predicted grounding resistance value R and the actual measured grounding resistance R0, and triggering the response mechanism when the difference is greater than or equal to the difference threshold value; Step S42: extracting the input data recorded in the response event as the target data set; judging each feature data in the target data set, and the judgment process is as follows: obtaining the input capacitive voltage change rate and the corresponding set change rate threshold value, when the input capacitive voltage change rate is greater than the change rate threshold value, marking the capacitive voltage change rate as the target feature influencing factor; otherwise, not marked; standard deviation q of the input is acquired V and a standard deviation threshold q corresponding to the setting th , when q V > q th , the standard deviation is marked as a target feature influencing factor; Obtaining input temperature data and humidity data, and corresponding monitoring period, calculating corresponding real-time temperature change slope and humidity change slope; if there is a temperature change slope and a preset normal temperature change slope difference greater than a first slope difference threshold, the temperature data is marked as a target feature influencing factor, and if there is a humidity change slope and a preset normal humidity change rate difference greater than a second slope difference threshold, the humidity data is marked as a target feature influencing factor; Obtaining input electromagnetic interference intensity Iem0 and setting electromagnetic interference intensity threshold Ith, if Iem0>Ith, marking the electromagnetic interference intensity as a target feature influencing factor; Step S43: obtaining the number M of feature data marked as target feature influencing factors in the same response event, when M=1, outputting a type of response event, and the target feature influencing factor of the event being a type of influencing factor; when M>1, outputting a second type of response event, and the target feature influencing factors in the second type of response event being a second type of influencing factor.

5. A method for data intelligent analysis of power cable grounding resistance according to claim 4, characterized in that: The different adjustment strategies based on different types include the following steps: Independently performing an adjustment strategy for responding to each type of influencing factor in each response event; Performing priority analysis on each second type of influencing factor in each response event, and the specific analysis process is as follows: calculating the standard value X of each of the two types of influence factors norm , X norm =(X real -X min ) / (X max -X min ), wherein X real is a real-time monitoring value corresponding to the response event, X min and X max are minimum and maximum values of the second type of influence factor in the normal working range; The i-th second-type influence factor historical record is marked as the number of times Di of the second-type influence factor in the response event, and the formula R i =D i / [(D1+D2+...+D n )], is used to calculate the weight coefficient R i of the i-th second-type response factor in the corresponding response event, i∈[1,n]; D1+D2+...+D n represents the sum of the D values of the first to n-th second-type influence factor records in the response event, and n represents the total number of second-type influence factors in the corresponding response event. Based on the weight coefficient and the corresponding standard value, the comprehensive influence quantitative value Qi of the i-th secondary influence factor is calculated, Qi=R i X norm,i ; X norm,i represents the standard value of the i-th secondary influence factor; Ordering each second type of influencing factor in the same response event based on the comprehensive influence quantitative value from large to small to generate a priority sequence; Based on the priority sequence, the strategies corresponding to each second type of influencing factor are responded to in turn.

6. A data intelligent analysis system for power cable earth resistors, such as a data intelligent analysis method for power cable earth resistors according to any one of claims 1 - 5, characterized by The system includes a time series data generation module, a feature data set analysis module, a mathematical model establishment module, and an adjustment strategy distribution module; The time series data generation module is used to generate continuous time series data according to a preset time interval based on the data collected by the historical record; The feature data set analysis module is used to extract features from all data sets to generate feature data sets; The mathematical model establishment module is used to establish a mathematical model of the measured grounding resistance and each feature data in the feature data set; The adjustment strategy distribution module is used to determine the type of the feature influencing factor in the target data set and execute different adjustment strategies based on different types.

7. A data intelligent analysis system for power cable earth resistance as claimed in claim 6 wherein: The mathematical model establishment module includes an associated data pair generation unit and a resistance regression model establishment unit; The associated data pair generation unit is used to extract the maximum value of the grounding resistance recorded in each set corresponding to a time period, and construct an associated data pair of the maximum value of the grounding resistance and the feature data set in the corresponding set to generate an associated data pair corresponding to each set; The resistance regression model establishment unit is used to substitute all associated data pairs corresponding to each set to calculate the corresponding regression coefficient and error term to construct a complete resistance regression model.

8. A data intelligent analysis system for power cable earth resistance as claimed in claim 7 wherein: The adjustment strategy distribution module includes a response event determination unit and a feature data distinguishing unit; The response event determination unit is used to obtain a response event applied after the model is established, which refers to an event that triggers a response mechanism when the difference between the predicted grounding resistance value and the actual measured grounding resistance is greater than or equal to the difference threshold after input data in a target period is obtained and substituted into the model to calculate the predicted grounding resistance value; The feature data differentiation unit is used to extract the input data recorded in the response event as the target data set; to judge each feature data in the target data set, and to output feature data of different types.

9. A data intelligent analysis system for power cable earth resistance as claimed in claim 7, wherein: The adjustment strategy allocation module also includes a priority analysis unit and a sorting response unit; The priority analysis unit is used to perform priority analysis on the two types of influencing factors in each response event; The sorting response unit is used to sort the two types of influencing factors in the same response event in descending order based on the comprehensive influence quantification value to generate a priority sequence. The strategies corresponding to each of the two types of influencing factors are responded to sequentially based on the priority sequence.

Images