A method for automatic processing of transmission line construction data based on data processing

Abstract

The present invention belongs to the technical field of data processing, and specifically discloses a method for automatically processing transmission line construction data based on data processing. By performing correlation analysis on monitoring quantities generated by transmission line construction, the data is divided into a combination of strongly correlated monitoring quantities and a single monitoring quantity, and abnormalities are identified according to their respective characteristics. This method improves the pertinence and applicability of abnormality identification, not only enhances the timeliness of identification, but also greatly reduces the possibility of misidentification, thereby improving the reliability and robustness of abnormal identification of construction data. At the same time, after abnormal identification of the monitoring quantities generated by transmission line construction, operation fault analysis is performed by retrieving the operation indicators of the abnormal monitoring quantities, thereby classifying the abnormal monitoring quantities, and realizing abnormal tracing of the abnormal monitoring quantities. In this way, the pertinence of abnormal processing is improved by means of detailed analysis of the specific causes of the abnormalities, which helps to ensure the continuous and effective data quality.

Description

Technical Field

[0001] The invention belongs to the technical field of data processing, and specifically discloses a method for automatically processing transmission line construction data based on data processing. Background Art

[0002] With the rapid development of science and technology, power facilities have become the basic guarantee for production and life, especially basic power facilities such as transmission lines. In recent years, the number and scale of transmission line projects have been increasing, and ensuring construction quality and safety has become the key. However, the current assessment of construction quality and safety mainly relies on data collected at the construction site, which comes from various sensor records. However, due to equipment failures and external factors, data quality is often unsatisfactory. Such data quality issues greatly affect the accuracy of construction quality and safety assessments. Therefore, it is crucial to effectively identify anomalies in construction data.

[0003] The existing Chinese patent publication number CN113420063B discloses a system for automatically identifying and processing abnormal data collected for construction monitoring. The system pre-processes the time series data collected by sensors in the construction field, such as temperature, strain, displacement, acceleration, etc., to identify and delete the burrs in the time series data by using a data pre-processing system, so as to make the monitoring data more accurate, true and objective. When performing burr identification on construction data, this scheme only analyzes the time series of a single data, and ignores the correlation between the data. If a certain data has an intrinsic connection with other data, the abnormality identification should be performed based on this correlation. Ignoring these intrinsic connections and performing abnormality identification on all data separately will lead to an overly rigid identification process, lack of pertinence and applicability, thereby affecting the timeliness and accuracy of abnormality identification. In addition, the time series analysis of a single data may produce false alarms because the influence of other related data is not taken into account. For example, an abnormality in a certain data may be caused by normal fluctuations in another data, but if this relationship is not considered, it may be mistakenly identified as an abnormality, which is not conducive to the reliability and robustness of abnormality identification of construction data. Summary of the invention

[0004] To this end, one purpose of an embodiment of the present application is to provide a method for automatic processing of transmission line construction data based on data processing, by combining strongly correlated monitoring quantities and dividing single monitoring quantities for the construction data generated by the transmission line, thereby adopting different abnormality identification methods for different monitoring quantities, thereby effectively solving the problems mentioned in the background technology.

[0005] The purpose of the present invention can be achieved through the following technical solutions: a method for automatically processing transmission line construction data based on data processing, comprising the following steps: collecting various monitoring quantities generated during the transmission line construction process.

[0006] The historical time series data of various monitoring quantities are retrieved to analyze the correlation between the monitoring quantities, and based on the analysis results, the monitoring quantities generated by the transmission line construction are divided into a combination of strongly correlated monitoring quantities and a single monitoring quantity.

[0007] A monitoring period is set, and time series data of strongly correlated monitoring quantity combinations are collected within this period for anomaly identification.

[0008] The historical time series data of a single monitoring quantity is used to analyze the regularity of time distribution. When the regularity of time distribution of a single monitoring quantity exists, the characteristics of the time distribution regularity are extracted.

[0009] Collect the time series data of a single monitoring quantity within the monitoring period. When analyzing the time distribution regularity of a single monitoring quantity, identify anomalies in the time series data of the monitoring quantity within the monitoring period based on the extracted time distribution regularity characteristics.

[0010] When there is no regularity in the time distribution of a single monitoring quantity, mutation analysis is performed on the time series data of the single monitoring quantity during the monitoring period, and anomalies are identified based on the analysis results.

[0011] The identified abnormal monitoring quantity is marked as an abnormal monitoring quantity, and the time when the abnormality occurs is recorded. Then, the operating indicators of the equipment to which the abnormal monitoring quantity belongs at the time of the abnormality occurrence are retrieved to perform fault risk analysis.

[0012] The abnormal monitoring quantity is classified according to the failure risk analysis result of the equipment to which the abnormal monitoring quantity belongs, and the classification result is fed back and warned.

[0013] Combining all the above-mentioned technical solutions, the positive effects of the present invention are as follows: 1. The present invention divides the data into a combination of strongly correlated monitoring quantities and a single monitoring quantity by performing correlation analysis on the monitoring quantities generated by the construction of transmission lines, and performs anomaly identification according to their respective characteristics. This method improves the pertinence and applicability of anomaly identification, not only enhances the timeliness of identification, but also greatly reduces the possibility of misidentification, thereby improving the reliability and robustness of anomaly identification in construction data.

[0014] 2. After identifying the abnormalities of the monitoring quantities generated by the construction of the transmission line, the present invention performs operation fault analysis by retrieving the operation indicators of the abnormal monitoring quantities, thereby classifying the abnormal monitoring quantities and realizing the abnormal tracing of the abnormalities of the abnormal monitoring quantities. In this way, the specific reasons leading to the abnormalities are analyzed in detail to improve the pertinence of the abnormal processing, which helps to ensure the continuous effectiveness of data quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The present invention is further described using the accompanying drawings, but the embodiments in the accompanying drawings do not constitute any limitation to the present invention. A person skilled in the art can obtain other drawings based on the following drawings without creative work.

[0016] Figure 1 It is a diagram of the steps for implementing the method of the present invention.

[0017] Figure 2 It is a schematic diagram of marking the central vertical line and intersection points of the histogram shape outline in the present invention.

[0018] Figure 3 The present invention is a flowchart for implementing anomaly identification by collecting time series data of a combination of strongly correlated monitoring quantities.

[0019] Description of the accompanying drawings: 1 - central vertical line, 2 - dividing point, 3 - intersection point, 4 - histogram shape outline. DETAILED DESCRIPTION

[0020] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0021] See also Figure 1 As shown, the present invention proposes an automatic processing method for transmission line construction data based on data processing, comprising the following steps: collecting various monitoring quantities generated during the transmission line construction process.

[0022] In the embodiments applied to the above scheme, the monitoring quantities generated by the construction of the transmission line cover electrical parameters, environmental parameters, mechanical parameters, geological parameters, etc., among which the electrical parameters include voltage, current, frequency, etc., the environmental parameters include temperature, humidity, wind speed, etc., the mechanical parameters include conductor tension, conductor vibration frequency, etc., and the geological parameters include soil moisture, geological settlement displacement, etc.

[0023] It should be understood that the above parameters reflect the operating status and health of the transmission line in different dimensions respectively. Among them, electrical parameters directly reflect the electrical performance of the transmission line, environmental parameters directly reflect the electrical performance of the transmission line, mechanical parameters reflect the physical state of the transmission line and its supporting structure, and geological parameters reflect the basic conditions of the transmission line and its supporting structure.

[0024] The historical time series data of various monitoring quantities are retrieved to analyze the correlation between the monitoring quantities, and based on the analysis results, the monitoring quantities generated by the transmission line construction are divided into a combination of strongly correlated monitoring quantities and a single monitoring quantity.

[0025] It should be added that when retrieving the historical time series data of various monitoring quantities for correlation analysis, the breadth of the historical period must be ensured. This is because a broad historical period can cover more working conditions and operating conditions, and can provide more data samples to improve the credibility of statistical analysis. At the same time, the historical period must be limited to the current time. This is because the newer data before the current time can better reflect the current construction status of the transmission line. In addition, older data may have problems such as inconsistent formats and different sampling frequencies. Using recent data can ensure the consistency and reliability of the data. Furthermore, limiting the current time can reduce the workload of data processing and improve the efficiency of correlation analysis.

[0026] In the preferred implementation of the above scheme, the analysis of the correlation between the various monitoring quantities is implemented as follows: various types of monitoring quantities are grouped in pairs to form a number of monitoring quantity combinations.

[0027] It should be pointed out that when combining various types of monitoring quantities in pairs, the corresponding monitoring quantities under the same parameters are preferentially combined. When performing correlation analysis on the monitoring quantity combinations in the future, the corresponding monitoring quantity combinations under the same parameters are preferentially analyzed. This is because the corresponding monitoring quantity combinations under the same parameters are more likely to be correlated. The main reason is that the monitoring quantities under the same parameters usually reflect the same physical or environmental properties. For example, electrical parameters such as voltage, current, and power factor usually affect each other in the same physical process or equipment, and the correlation between them is high. For another example, mechanical parameters such as wire tension and wire vibration displacement in mechanical parameters usually show correlation on the same structure or equipment. Therefore, giving priority to analyzing the monitoring quantity combinations under the same parameters can improve the analysis efficiency. In addition, the monitoring quantity combinations under the same parameters are easier to understand and explain. For example, analyzing the relationship between current and voltage is easier to explain than analyzing the relationship between current and wind speed, because the former is a basic concept in electrical engineering.

[0028] The historical time series data of two monitoring quantities in each monitoring quantity combination are used to draw histograms respectively, and the shape contour of the histogram is extracted and overlapped with the contour of the standard normal distribution curve to calculate the overlap degree of the normal contour of the data distribution corresponding to the monitoring quantity. The specific calculation of the overlap degree of the normal contour of the data distribution can be obtained by obtaining the length of the overlap contour and dividing it by the length of the contour of the standard normal distribution curve.

[0029] It is important to know that the standard normal distribution is a special normal distribution with a mean of 0 and a standard deviation of 1.

[0030] It should be added that before drawing the histogram of the historical time series data after normalization of the two monitoring quantities, the data needs to be converted into a form with a mean of 0 and a standard deviation of 1. This processing method is called standardization, which can make the shape of the drawn histogram close to the standard normal distribution, so that it can be overlapped and compared with the standard normal distribution curve at the same scale, thereby improving the comparability of the overlap comparison.

[0031] See also Figure 2 As shown, a central vertical line is marked from the shape outline of the histogram, and the central vertical line is evenly divided into several points, and then a horizontal line perpendicular to the central vertical line is drawn at each dividing point. These horizontal lines intersect with the curves on both sides of the histogram to form two left and right intersections.

[0032] Mark the left and right intersection points corresponding to each division point, and calculate the distance between the intersection point and the division point. Then compare the left and right distances to get the left and right gaps corresponding to each division point, and calculate the data distribution symmetry corresponding to the monitoring amount. The calculation formula for data distribution symmetry is: , where Indicates The left and right gaps corresponding to the dividing points, Indicates the partition point number, , represents the number of partition points, Represents a natural constant. From this formula, we can see that the smaller the left and right gaps corresponding to each dividing point are, the greater the symmetry of the data distribution is, and the more symmetrical the data distribution is.

[0033] The normal distribution conformity of each monitoring quantity is obtained by weighted averaging the normal contour coincidence and symmetry of the data distribution corresponding to each monitoring quantity in each monitoring quantity combination, and compared with the set threshold. If the normal distribution conformity corresponding to a monitoring quantity reaches the set threshold, it is judged that the monitoring quantity conforms to the normal distribution.

[0034] It should be added that, when setting the weight values ​​corresponding to the normal contour overlap of data distribution and the symmetry of data distribution in weighted average, it should be taken into account that the contour overlap reflects the similarity between the overall shape of data distribution and the standard normal distribution, which is usually related to the overall distribution shape of data, and the symmetry reflects the symmetry between the left and right sides of data distribution, which is usually related to the local details of data. The overall correlation is greater than the local correlation. In addition, the contour overlap is more resistant to the influence of outliers and noise in the data. Even if there are a small number of outliers, the overall distribution of data can still match the standard normal distribution contour, while the symmetry may be affected by individual extreme values. In summary, it can be seen that a higher weight can be given to the contour overlap. By way of example, the weight values ​​corresponding to the normal contour overlap of data distribution and the symmetry of data distribution can be 0.6 and 0.4, which not only meets the demand for focusing on contour overlap, but also does not over-emphasize contour overlap.

[0035] It should be further added that, since the values ​​of the normal contour overlap and the data distribution symmetry are both between 0 and 1, the normal distribution conformity value obtained after weighted averaging is also between 0 and 1, and the normal distribution conformity threshold can be set to 0.7. The purpose of setting the normal distribution conformity threshold is to assist in judging whether the monitored quantity conforms to the normal distribution.

[0036] It should be explained that the above-mentioned normal distribution judgment of each monitoring quantity in the monitoring quantity combination is for the purpose of screening suitable correlation analysis methods. The commonly used correlation analysis methods include Pearson correlation coefficient and Spearman rank correlation coefficient, but these two analysis methods have different applicable scenarios. The Pearson correlation coefficient is applicable to data that conforms to or is close to normal distribution, and the Spearman rank correlation coefficient is applicable to data that does not conform to normal distribution. Therefore, it is necessary to judge the conformity of normal distribution of the monitoring quantity in order to select a more appropriate and targeted correlation analysis method to achieve the purpose of accurate and reliable correlation analysis results. In addition, the above-mentioned scheme integrates the overlap of the normal contour of the data distribution and the symmetry of the data distribution in the judgment of the conformity of the normal distribution of the monitoring quantity. The comprehensive judgment can more comprehensively reflect the characteristics of the data distribution, thereby more accurately evaluating whether the data conforms to the normal distribution.

[0037] The adaptive correlation analysis method is selected according to the judgment result of whether the two monitored quantities in each monitoring quantity combination conform to the normal distribution. In the specific selection, if the two monitored quantities in a certain monitoring quantity combination conform to the normal distribution, the Pearson correlation coefficient is used as the adaptive correlation analysis method for the monitoring quantity combination. If the two monitored quantities in a certain monitoring quantity combination do not conform to the normal distribution, the Spearman rank correlation coefficient is used as the adaptive correlation analysis method for the monitoring quantity combination. If one monitoring quantity in a certain monitoring quantity combination conforms to the normal distribution and the other does not conform to the normal distribution, the Pearson correlation coefficient and the Spearman rank correlation coefficient are used to perform correlation analysis at the same time. In this case, there are two correlation analysis results, and the analysis results can be compared. If the difference in the analysis results is small, the average of the two correlation analysis results is taken as the final correlation analysis result. If the difference in the analysis results is large, it can be selected according to actual needs. For example, the monitoring quantity that conforms to the normal distribution is more important in actual applications, and the Pearson correlation coefficient can be selected as the final result; the monitoring quantity that does not conform to the normal distribution is more important in actual applications, and the Spearman rank correlation coefficient can be selected as the final result.

[0038] The historical time series data of two monitoring quantities in each monitoring quantity combination are mapped according to time to obtain a number of paired time series data points, and then a correlation analysis is performed on the several time series data points according to the adapted correlation analysis method to obtain the correlation coefficient corresponding to each monitoring quantity combination.

[0039] It should be understood that when pairing the historical time series data of two monitoring quantities in each monitoring quantity combination through time series mapping, it is necessary to ensure that the timestamps of the paired data are consistent, that is, pairing the two monitoring quantity data at the same time point. In fact, the monitoring frequencies of the two monitoring quantities may not be consistent. In order to achieve better pairing, the historical time series data of the two monitoring quantities can be interpolated and filled first. The specific operation process is: find the common time range in the two monitoring quantity data sets, and create a timestamp list containing all time points in this time range. Use the selected interpolation method to fill the historical time series data of the two monitoring quantities so that each timestamp has a corresponding data value. This ensures that each time point has corresponding data, thereby avoiding data loss or pairing errors caused by inconsistent timestamps.

[0040] A 4x4 correlation matrix is ​​created using the correlation coefficients corresponding to each monitoring quantity combination, and a statistical test is performed based on the correlation matrix to obtain the analysis results of whether each monitoring quantity combination has correlation.

[0041] It needs to be explained that the statistical test after obtaining the correlation coefficient using the correlation analysis method is to verify whether the obtained correlation coefficient is statistically significant, so as to confirm whether the association between the monitored quantities is real and reliable.

[0042] It should be added that the statistical test mentioned above is performed by setting a significance level and checking whether the p-value is lower than this level. If the p-value is lower than the significance level, the correlation is considered significant. For example, assuming the significance level is 0.05, if the p-value is lower than 0.05, the correlation is considered significant. If the p-value is greater than or equal to 0.05, the correlation is considered not significant.

[0043] In a further preferred implementation of the above scheme, the monitoring quantities generated by the transmission line construction are divided into strongly correlated monitoring quantity combinations and single monitoring quantities, referring to the following process: when analysis finds that a certain monitoring quantity combination is correlated, the monitoring quantity combination is regarded as a strongly correlated monitoring quantity combination; when a certain monitoring quantity does not show correlation in all monitoring quantity combinations, the monitoring quantity is regarded as a single monitoring quantity.

[0044] In the above implementation example, if the monitoring quantity combination (A, B) has a significant correlation, then (A, B) is marked as a strongly correlated monitoring quantity combination; if the monitoring quantity A does not show a significant correlation in all monitoring quantity combinations, then A is marked as a single monitoring quantity.

[0045] A monitoring period is set, and time series data of strongly correlated monitoring quantity combinations are collected within this period for anomaly identification.

[0046] The above-mentioned setting of the monitoring cycle is intended to overcome the high false alarm rate problem inherent in anomaly identification based on a single data point. By stipulating a certain data collection window, it can be ensured that each anomaly detection is based on a set of data rather than an isolated value, thereby improving the accuracy and reliability of identification. In addition, setting a monitoring cycle also helps to: first, standardize data collection: ensure that the data collection process is carried out at predetermined time intervals, eliminate randomness and disorder, and facilitate system management and automated operations; second, ensure data timeliness: regularly collect the latest data to ensure that the monitoring system can capture changes in system status in a timely manner; third, optimize resource utilization: reasonably arrange the frequency of data collection, which can not only avoid the storage and processing burden caused by data redundancy, but also prevent information loss caused by insufficient data.

[0047] In the above implementation example, the setting of the monitoring cycle can be set based on the monitoring intervals of the two monitoring quantities in the monitoring quantity combination. Specifically, if the monitoring interval of the monitoring quantity is short, the monitoring cycle can be set shorter to ensure that a sufficient amount of data is obtained for analysis in a short time. On the contrary, if the monitoring interval of the monitoring quantity is long, the monitoring cycle can be extended accordingly to ensure that an appropriate amount of data is collected in each cycle.

[0048] Among the ways in which the above scheme can be implemented, see Figure 3As shown, the time series data of the strongly correlated monitoring quantity combination is collected for anomaly identification as follows: the correlation direction corresponding to the strongly correlated monitoring quantity combination is obtained based on the correlation coefficient corresponding to the strongly correlated monitoring quantity combination.

[0049] It should be noted that since the correlation coefficient obtained by correlation analysis takes values ​​between -1 and 1, when the correlation coefficient is negative, it means that the correlation direction of the combination of strongly correlated monitoring quantities is negatively correlated, and when the correlation coefficient is positive, it means that the correlation direction of the combination of strongly correlated monitoring quantities is positively correlated.

[0050] The time series data corresponding to the two monitoring quantities in the strongly correlated monitoring quantity combination are marked with several points in a two-dimensional coordinate system constructed with time as the horizontal axis and the monitoring quantity data as the vertical axis to form a data change curve corresponding to the two monitoring quantities.

[0051] A number of time points are evenly set on the horizontal axis of the data change curve, and then the tangent slope of each time point is obtained on the data change curve corresponding to the two monitoring quantities, and the tangent slopes of the same time point are formed into a slope group.

[0052] The signs of the two tangent slopes in the slope group formed at each time point are extracted and compared with the correlation direction corresponding to the strong correlation monitoring combination to identify whether they meet the correlation direction, and the time points that do not meet the correlation direction are recorded as abnormal time points, and then the occurrence proportion of abnormal time points is counted.

[0053] In the above implementation example, when the correlation direction is positively correlated, the signs of the two tangent slopes formed at a certain time point are one positive and one negative, and at this time it is identified that the time point does not conform to the correlation direction.

[0054] The occurrence ratio of abnormal time points is compared with the preset critical ratio. For example, the critical ratio is 0.4. If the occurrence ratio of abnormal time points reaches the critical ratio, it is identified that the two monitoring quantities in the strongly correlated monitoring quantity combination are abnormal.

[0055] The present invention uses the correlation direction of the monitoring quantity combination to judge the abnormality, that is, if two strongly correlated monitoring quantities keep changing in the same direction (positive correlation) or in the opposite direction (negative correlation) for most of the time, then if there is an opposite change trend at certain time points, it can be regarded as an abnormality. The advantages of this method are: first, strong pertinence: using known correlation directions to detect abnormalities, it is highly pertinent; second, high reliability: based on the existing strong correlation, the possibility of false alarms is reduced.

[0056] The historical time series data of a single monitoring quantity is used to analyze the regularity of time distribution. When the regularity of time distribution of a single monitoring quantity exists, the characteristics of the time distribution regularity are extracted.

[0057] Specifically, the time distribution regularity analysis using the historical time series data of a single monitoring quantity is implemented as follows: a number of points of the historical time series data of a single monitoring quantity are marked in a two-dimensional coordinate system constructed with time as the horizontal axis and the monitoring quantity data as the vertical axis to form a data change curve corresponding to the single monitoring quantity.

[0058] The data change curve corresponding to a single monitoring quantity is transformed into a spectrum using Fourier transform, and the peak value in the spectrum diagram is analyzed to identify the periodic component. If the periodic component can be identified, it is determined that the single monitoring quantity has a regular time distribution. Otherwise, it is determined that the single monitoring quantity does not have a regular time distribution.

[0059] It should be pointed out that in the spectrum obtained by Fourier transform, each frequency component corresponds to an amplitude value. If there is a periodic component of a certain frequency in the signal, a corresponding peak will appear in the spectrum. This is because the periodic component will produce a significant energy concentration in the spectrum, that is, the peak in the spectrum. Therefore, by analyzing the peak position in the spectrum, the periodic component in the signal can be determined. For example, if a peak at 1Hz is observed in the spectrum, it means that there is a component with a period of 1 second in the original signal.

[0060] Furthermore, the characteristics of the time distribution law are refined as follows: the periodic frequency is identified from the spectrum diagram formed by a single monitoring quantity, and the periodic length is calculated based on this, and then a single complete periodic data curve is intercepted from the data change curve corresponding to the single monitoring quantity based on the periodic length.

[0061] Inflection points are extracted from a single complete cycle data curve, thereby dividing the cycle length into horizontal segments, rising segments, and falling segments, and the duration of the corresponding segments is obtained at the same time, and then the divided horizontal segments, rising segments, falling segments and the duration of the corresponding segments are used as the characteristics of the time distribution law.

[0062] In the embodiment of the above scheme, it is assumed that the cycle length of a single monitoring quantity is 24 hours, the duration of the ascending segment is 6:00-12:00, the duration of the descending segment is 12:00-18:00, and the duration of the horizontal segment is 18:00-6:00.

[0063] It is important to know that the inflection point is usually the location of the local maximum or minimum in the data curve. Once the inflection point is determined, the cycle can be divided into different sections based on these points, where the horizontal section refers to the part where the slope of the curve is close to zero, the rising section refers to the part where the slope of the curve increases, and the falling section refers to the part where the slope of the curve decreases.

[0064] Collect the time series data of a single monitoring quantity within the monitoring period. When analyzing the time distribution regularity of a single monitoring quantity, identify anomalies in the time series data of the monitoring quantity within the monitoring period based on the extracted time distribution regularity characteristics.

[0065] Preferably, anomalies in the time series data of the monitored quantity within the monitoring period are identified based on the extracted time distribution law characteristics. See the following process: extract the monitoring time corresponding to each time series data in the time series data of a single monitored quantity within the monitoring period, and classify it according to the time distribution law characteristics to form a time series data set corresponding to the horizontal segment, rising segment, and falling segment.

[0066] The maximum and minimum data are extracted from the time series data set corresponding to the horizontal segment, and the maximum data is subtracted from the minimum data and then divided by the duration of the horizontal segment to obtain the average change amplitude of the horizontal segment.

[0067] The time series data set corresponding to the rising segment is subjected to rising rate analysis, while the time series data set corresponding to the falling segment is subjected to falling rate analysis.

[0068] In particular, when performing rise rate and fall rate analysis, the rise rate and fall rate can be obtained by dividing the difference between adjacent data points of the time series data set corresponding to the rising segment by the mean of the time interval.

[0069] The average change amplitude of the time series data set corresponding to the horizontal segment, the rising rate of the time series data set corresponding to the rising segment, and the falling rate of the time series data set corresponding to the falling segment are compared with the reference values ​​respectively. If the average change amplitude of the time series data set corresponding to all horizontal segments, the rising rate of all rising segments, and the falling rate of all falling segments are in line with the reference values, then it is identified that there is no abnormality in the single monitoring quantity, otherwise it is identified that there is an abnormality in the single monitoring quantity.

[0070] It should be pointed out that the reference values ​​include the change amplitude reference value, the rise rate reference value and the decrease rate reference value. The change amplitude reference value can be calculated from the time distribution law characteristics of a single monitored quantity to obtain the average change amplitude of the horizontal segment as the change amplitude reference value. The rise rate reference value and the decrease rate reference value are calculated from the time distribution law characteristics of a single monitored quantity to obtain the rise rate of the rising segment and the decrease rate of the descending segment as the rise rate reference value and the decrease rate reference value respectively.

[0071] The present invention analyzes the time distribution regularity of a single monitoring quantity and extracts the time distribution regularity characteristics when analyzing the existence of time distribution regularity. It can more accurately identify the behavior pattern of the monitoring quantity under normal circumstances, and thus identify anomalies in the time series data of the single monitoring quantity in the monitoring period based on the time distribution regularity characteristics. The current monitoring data is compared with the known time distribution regularity characteristics, thereby minimizing false alarms (false positives) caused by normal data fluctuations, thereby improving the effectiveness and reliability of the detection results.

[0072] When there is no regularity in the time distribution of a single monitoring quantity, mutation analysis is performed on the time series data of the single monitoring quantity during the monitoring period, and anomalies are identified based on the analysis results.

[0073] In the specific implementation of the above scheme, the mutation analysis is performed on the time series data of a single monitoring quantity in a monitoring period in the following process: the time series data of a single monitoring quantity in a monitoring period and its data at the previous moment form an adjacent data group.

[0074] The mutation degree of each adjacent data group is calculated. The specific calculation formula is to divide the absolute value of the difference between the adjacent data groups by the first data in the adjacent data group to obtain a percentage, and compare it with a reasonable mutation degree. For example, a reasonable percentage is 20%. If the mutation degree of an adjacent data group is greater than the reasonable mutation degree, the adjacent data group is recorded as an abnormal data group.

[0075] It should be supplemented that the above-mentioned reasonable mutation degree can be calculated by obtaining monitoring data collected at adjacent times when the single monitoring quantity equipment is operating normally.

[0076] The proportion of abnormal data groups is counted and compared with the preset critical proportion. If the proportion of abnormal data groups reaches the critical proportion, it is identified that a single monitoring quantity is abnormal.

[0077] It should be understood that when analyzing the absence of regularity in the time distribution of a single monitoring quantity, the abnormality identification is performed by performing mutation analysis on the values ​​of a single monitoring quantity obtained at adjacent times. This is because the change of the monitoring quantity is usually slow and continuous under normal operating conditions. In other words, in most cases, the difference in data between adjacent time points will not be large. When the monitoring quantity is abnormal, sudden and drastic changes may occur. Therefore, by analyzing the difference between the data of adjacent time points, mutations that do not conform to the normal change law can be identified, and then it can be determined whether an abnormal situation has occurred. This method is particularly suitable for data without obvious periodicity or regularity.

[0078] The present invention conducts correlation analysis on the monitoring quantities generated by the construction of transmission lines, divides the data into a combination of strongly correlated monitoring quantities and a single monitoring quantity, and identifies anomalies based on their respective characteristics. This method improves the pertinence and applicability of anomaly identification, not only enhances the timeliness of identification, but also greatly reduces the possibility of misidentification, thereby improving the reliability and robustness of construction data anomaly identification.

[0079] The identified abnormal monitoring quantity is marked as an abnormal monitoring quantity, and the time when the abnormality occurs is recorded. Then, the operating indicators of the equipment to which the abnormal monitoring quantity belongs at the time of the abnormality occurrence are retrieved to perform fault risk analysis.

[0080] In the specific implementation of the above scheme, when there is a regularity in the time distribution of the strongly correlated monitoring quantity combination and the single monitoring quantity, the operating index of the device to which the abnormal monitoring quantity obtained by the abnormal call belongs at the time of the abnormal occurrence is specifically the operating index of the abnormal monitoring quantity within the monitoring period; when there is no regularity in the time distribution of the single monitoring quantity, the operating index of the device to which the abnormal monitoring quantity obtained by the abnormal call belongs at the time of the abnormal occurrence is specifically the operating index when there is a sudden change in the monitoring quantity.

[0081] In a further example, the device to which the abnormal monitoring quantity belongs is specifically a temperature sensor when the abnormal monitoring quantity is temperature, and a tension sensor when the abnormal monitoring quantity is wire tension.

[0082] The operation indicators of the equipment to which the abnormal monitoring quantity belongs are retrieved for fault risk analysis as follows: the specification model of the equipment to which the abnormal monitoring quantity belongs is obtained, and the normal operating indicators of the equipment to which the abnormal monitoring quantity belongs are extracted from the user manual of the equipment.

[0083] The operating indicators of the equipment to which the abnormal monitoring quantity belongs at the time of the abnormality occurrence are compared with the normal operating indicators to identify whether there is an operating deviation. If there is an operating deviation, the analysis equipment is at risk of failure. Otherwise, the analysis equipment is not at risk of failure.

[0084] Specifically, the implementation process of operation deviation identification is to compare the operation indicators of the equipment to which the abnormal monitoring quantity belongs at the time of the abnormality occurrence with the normal operation indicators. If the operation indicators of the equipment to which the abnormal monitoring quantity belongs at the time of the abnormality occurrence are all within the normal operation indicator range, it is identified that there is no operation deviation, otherwise it is identified that there is an operation deviation.

[0085] S8. Classify the abnormal monitoring quantity according to the failure risk analysis result of the equipment to which the abnormal monitoring quantity belongs, and provide feedback and early warning on the classification result.

[0086] Specifically, the abnormal monitoring quantity is classified as follows: if the equipment to which the abnormal monitoring quantity belongs has a failure risk, the abnormal monitoring quantity is classified as a monitoring quantity abnormality caused by the equipment abnormality; otherwise, the abnormal monitoring quantity is classified as a monitoring quantity abnormality caused by external factors.

[0087] After identifying the abnormalities of the monitoring quantities generated by the construction of the transmission line, the present invention performs operation fault analysis by retrieving the operation indicators of the abnormal monitoring quantities, thereby classifying the abnormal monitoring quantities and realizing the abnormal tracing of the abnormalities of the abnormal monitoring quantities. In this way, the specific reasons leading to the abnormalities are analyzed in detail to improve the pertinence of the abnormal processing, which helps to ensure the continuous effectiveness of data quality.

[0088] The above contents are merely examples and explanations of the structure of the present invention. The technicians in this technical field may make various modifications or additions to the specific embodiments described or replace them in a similar manner. As long as they do not deviate from the structure of the invention or exceed the scope defined by the present invention, they should all fall within the protection scope of the present invention.

Claims

1. A method for automatically processing transmission line construction data based on data processing, characterized in that: The following steps are involved: Collect various monitoring quantities generated during the construction process of transmission lines; The historical time series data of various monitoring quantities are used to analyze the correlation between the monitoring quantities, thereby dividing the monitoring quantities into strongly correlated combinations and single monitoring quantities. For the monitoring quantities of the strongly correlated combination, the time series data is collected within the set monitoring period to identify anomalies; For a single monitored quantity, use historical time series data to analyze the regularity of time distribution. If there is a regularity in time distribution, extract the characteristics of the time distribution regularity, and use this to identify anomalies in the time series data of the single monitored quantity within the monitoring period. If there is no regularity in time distribution, perform mutation analysis of the time series data within the monitoring period to identify anomalies. Mark the abnormal monitoring quantity and record the abnormal time, and then retrieve the operating indicators of the equipment to which the abnormal monitoring quantity belongs at the abnormal time to conduct fault risk analysis; Classify abnormal monitoring quantities according to the failure risk analysis results of the equipment to which the abnormal monitoring quantities belong, and provide feedback and early warning on the classification results; The analysis of the correlation between the monitored quantities is implemented as follows: Group various monitoring quantities in pairs to form several monitoring quantity combinations; The historical time series data of two monitoring quantities in each monitoring quantity combination are used to draw histograms respectively, and the shape contour of the histogram is extracted and overlapped with the contour of the standard normal distribution curve to calculate the overlap degree of the normal contour of the data distribution corresponding to the monitoring quantity; Mark the central vertical line from the shape outline of the histogram, and divide the central vertical line evenly into several points, and then draw a horizontal line perpendicular to the central vertical line at each dividing point. These horizontal lines intersect with the curves on both sides of the histogram to form two left and right intersection points; Mark the left and right intersection points corresponding to each division point, and calculate the distance between the intersection point and the division point, and then compare the left and right distances to obtain the left and right gaps corresponding to each division point, so as to calculate the data distribution symmetry corresponding to the monitoring amount; The normal distribution conformity of each monitoring quantity is calculated by weighted average calculation of the normal contour coincidence and the data distribution symmetry of each monitoring quantity in each monitoring quantity combination, and compared with the set threshold. If the normal distribution conformity of a monitoring quantity reaches the set threshold, it is judged that the monitoring quantity conforms to the normal distribution. Select the appropriate correlation analysis method based on the judgment result of whether the two monitoring quantities in each monitoring quantity combination conform to the normal distribution; The historical time series data of two monitoring quantities in each monitoring quantity combination are mapped according to time to obtain a number of paired time series data points, and then a correlation analysis is performed on the several time series data points according to the adapted correlation analysis method to obtain the correlation coefficient corresponding to each monitoring quantity combination; A 4x4 correlation matrix is ​​created using the correlation coefficients corresponding to each monitoring quantity combination, and a statistical test is performed based on the correlation matrix to obtain the analysis results of whether each monitoring quantity combination has correlation.

2. The method for automatically processing transmission line construction data based on data processing according to claim 1, characterized in that: The combination of strongly correlated monitoring quantities and single monitoring quantities refer to the following division process: When the analysis finds that a certain monitoring quantity combination exists correlation, this monitoring quantity combination is regarded as a strongly correlated monitoring quantity combination. When a certain monitoring quantity does not show correlation in all monitoring quantity combinations, this monitoring quantity is regarded as a single monitoring quantity.

3. The method for automatically processing transmission line construction data based on data processing according to claim 1, characterized in that: The time series data of the strongly correlated monitoring quantity combination is collected to identify anomalies as follows: Based on the correlation coefficient corresponding to the combination of strongly correlated monitoring quantities, the correlation direction corresponding to the combination of strongly correlated monitoring quantities is obtained; Mark several points of the time series data corresponding to the two monitoring quantities in the strongly correlated monitoring quantity combination in a two-dimensional coordinate system with time as the horizontal axis and the monitoring quantity data as the vertical axis to form a data change curve corresponding to the two monitoring quantities; A number of time points are evenly set on the horizontal axis of the data change curve, and then the tangent slope of each time point is obtained on the data change curve corresponding to the two monitoring quantities, and the tangent slopes at the same time point are formed into a slope group; The signs of the two tangent slopes in the slope group formed at each time point are extracted and compared with the correlation direction corresponding to the strong correlation monitoring combination to identify whether they meet the correlation direction, and the time points that do not meet the correlation direction are recorded as abnormal time points, and then the occurrence ratio of abnormal time points is counted; The occurrence ratio of abnormal time points is compared with the preset critical ratio. If the occurrence ratio of abnormal time points reaches the critical ratio, it is identified that there is an abnormality in the monitoring amount. Otherwise, it is identified that there is no abnormality in the monitoring amount.

4. The method for automatically processing transmission line construction data based on data processing according to claim 1, characterized in that: The time distribution regularity analysis using the historical time series data of a single monitoring quantity is implemented as follows: The historical time series data of a single monitoring quantity is marked with several points in a two-dimensional coordinate system with time as the horizontal axis and the monitoring quantity data as the vertical axis to form a data change curve corresponding to the single monitoring quantity; The data change curve corresponding to a single monitoring quantity is transformed into a spectrum using Fourier transform, and the peak value in the spectrum diagram is analyzed to identify the periodic component. If the periodic component can be identified, it is determined that the single monitoring quantity has a regular time distribution. Otherwise, it is determined that the single monitoring quantity does not have a regular time distribution.

5. The method for automatically processing transmission line construction data based on data processing as claimed in claim 4, characterized in that: The time distribution regularity characteristics are as follows: Identify the periodic frequency from the spectrum formed by a single monitoring quantity, and calculate the periodic length based on it, and then intercept a single complete periodic data curve from the data change curve corresponding to the single monitoring quantity according to the periodic length; Inflection points are extracted from a single complete cycle data curve, thereby dividing the cycle length into horizontal segments, rising segments, and falling segments, and the duration of the corresponding segments is obtained at the same time, and then the divided horizontal segments, rising segments, falling segments and the duration of the corresponding segments are used as the characteristics of the time distribution law.

6. A method for automatically processing transmission line construction data based on data processing as claimed in claim 5, characterized in that: The abnormal identification of the time series data of the monitoring quantity within the monitoring period based on the extracted time distribution regularity characteristics is shown in the following process: Extract the monitoring time corresponding to each time series data in the time series data of a single monitoring quantity within the monitoring period, and classify it according to the characteristics of the time distribution law to form a time series data set corresponding to the horizontal segment, rising segment, and falling segment; Calculate the average change amplitude of the time series data set corresponding to the horizontal segment; Perform an ascending rate analysis on the time series data set corresponding to the ascending segment, and perform a descending rate analysis on the time series data set corresponding to the descending segment; The average change amplitude of the time series data set corresponding to the horizontal segment, the rising rate of the time series data set corresponding to the rising segment, and the falling rate of the time series data set corresponding to the falling segment are compared with the reference values ​​respectively. If the average change amplitude of the time series data set corresponding to all horizontal segments, the rising rate of all rising segments, and the falling rate of all falling segments are in line with the reference values, then it is identified that there is no abnormality in the single monitoring quantity, otherwise it is identified that there is an abnormality in the single monitoring quantity.

7. The method for automatically processing transmission line construction data based on data processing according to claim 1, characterized in that: The process of performing mutation analysis on the time series data of a single monitoring quantity in a monitoring period is as follows: The time series data of a single monitoring quantity within the monitoring period and its data at the previous moment constitute an adjacent data group; The mutation degree of each adjacent data group is calculated and compared with the reasonable mutation degree. If the mutation degree of an adjacent data group is greater than the reasonable mutation degree, the adjacent data group is recorded as an abnormal data group; The proportion of abnormal data groups is counted and compared with the preset critical proportion. If the proportion of abnormal data groups reaches the critical proportion, it is identified that a single monitoring quantity is abnormal.

8. The method for automatically processing transmission line construction data based on data processing according to claim 1, characterized in that: The operation indicators of the equipment to which the abnormal monitoring quantity belongs at the time of the abnormality occurrence are retrieved to perform fault risk analysis as follows; Obtain the specification model of the equipment to which the abnormal monitoring quantity belongs, thereby obtaining the normal operating index of the equipment to which the abnormal monitoring quantity belongs; The operating indicators of the equipment to which the abnormal monitoring quantity belongs at the time of the abnormality occurrence are compared with the normal operating indicators to identify whether there is an operating deviation. If there is an operating deviation, the analysis equipment is at risk of failure. Otherwise, the analysis equipment is not at risk of failure.

9. The method for automatically processing transmission line construction data based on data processing according to claim 8, characterized in that: The process of classifying the abnormal monitoring quantity is as follows: If the equipment to which the abnormal monitoring quantity belongs has a risk of failure, the abnormal monitoring quantity is classified as a monitoring quantity abnormality caused by the equipment abnormality; otherwise, the abnormal monitoring quantity is classified as a monitoring quantity abnormality caused by external factors.

Images