CN-121980488-A - Flexible cable fault hidden danger real-time identification and early warning method for distribution network uninterrupted operation

CN121980488ACN 121980488 ACN121980488 ACN 121980488ACN-121980488-A

Abstract

The invention discloses a flexible cable fault hidden danger real-time identification and early warning method for distribution network uninterrupted operation, which relates to the technical field of power cable fault detection and early warning, and comprises the steps of collecting cable partial discharge, temperature, dielectric loss, humidity and current data through a distributed sensor, calibrating, correcting sensitivity of the sensor to achieve data time-space alignment, combining parameter deviation degree dynamic weighting fusion characteristics to highlight abnormal parameter contribution, coupling time-space correlation analysis characteristic correlation degree to avoid single time-space point misjudgment, combining correlation degree change rate to judge hidden danger membership degree, combining cable load dynamic adjustment judgment threshold value, and finally combining short-term prediction deviation to output yellow, orange and red early warning. The method overcomes the defects of inaccurate data, fixed weight, stiff threshold and the like in the prior art, can accurately identify hidden danger in real time and perform grading early warning, provides clear basis for operation and maintenance, and ensures safe and stable operation of the flexible cable in a distribution network power failure operation scene.

Inventors

DANG ZHENGQIANG
JIN CHAO
WANG HAIYAN
GUO LINJUN
REN YICHANG
ZHANG YALI
ZHANG KUNKUN

Assignees

国网河南省电力公司嵩县供电公司

Dates

Publication Date: 20260505
Application Date: 20251209

Claims (10)

1. A flexible cable fault hidden danger real-time identification and early warning method for distribution network uninterrupted operation is characterized by comprising the following steps: S1, multi-source data acquisition and preliminary calibration are carried out, and a calibrated original data matrix is obtained; S2, carrying out space-time alignment on the multisource data corrected based on the sensitivity of the sensor to obtain aligned data; S3, combining dynamic weighting feature fusion of parameter deviation degree, and calculating fusion feature value; S4, analyzing the characteristic association degree of the coupling space-time association, and outputting the characteristic association degree; s5, judging the membership degree of the hidden danger of the fault by integrating the change rate of the relevancy, and determining the membership degree of the hidden danger; and S6, combining the hierarchical early warning of the short-term prediction deviation, generating an early warning index and pushing early warning information.
2. The method for identifying and early warning hidden trouble of flexible cable in uninterrupted power supply operation according to claim 1, wherein step S1 comprises steps of multi-source data acquisition and preliminary calibration, obtaining a calibrated original data matrix, specifically using a distributed sensor array, and acquiring partial discharge pulse amplitude values of flexible cable at fixed acquisition intervals Skin temperature Dielectric loss value of insulating layer Relative humidity of environment Real-time load current The acquisition time length is Preliminary calibration is carried out on the sensor zero drift error, and a calibrated original data matrix is defined The method comprises the following steps: ; In the formula, Is the first The calibrated original data matrix at each acquisition time comprises 5 parameters of partial discharge, temperature, dielectric loss, humidity and current, Is the first The time of the data acquisition is the same, The data is collected into a set of points, Is the partial discharge pulse amplitude of the flexible cable, Is a zero-drift calibration coefficient, and the zero-drift calibration coefficient is the zero-drift calibration coefficient, For the zero drift reference value of the partial discharge sensor, Is the first The skin temperature of the flexible cable at each acquisition time, For the zero drift calibration coefficient of the temperature sensor, Is a zero drift reference value of the temperature sensor, Is the first Dielectric loss values of the flexible cable insulating layer at the moment of acquisition, To be the zero drift calibration coefficient of the dielectric loss sensor, To be a zero drift reference value of the dielectric loss sensor, Is the first The relative humidity of the environment in which the flexible cable is located at the moment of acquisition, Is a zero drift calibration coefficient of the humidity sensor, Is a zero drift reference value of the humidity sensor, Is the first And collecting real-time load current of the flexible cable at the moment.
3. The method for identifying and early warning hidden trouble of flexible cable in uninterrupted power supply operation according to claim 1, wherein step S2 is based on sensor sensitivity corrected multi-source data space-time alignment to obtain aligned data, specifically considering sensor installation position difference, acquisition clock deviation and sensitivity difference, and comparing Performing space-time alignment, and defining and obtaining aligned data The formula is as follows: ; In the formula, Is the first The data vectors after time-space alignment at the acquisition time comprise the 5 types of parameters, 、 Are respectively the first 、 The sensitivity coefficient of the individual sensors is such that, Is the first Acquisition Point pair The spatiotemporal influence weights of the individual acquisition points, 、 Is the first The spatial coordinates of the individual sensors are used, 、 Is the first The spatial coordinates of the individual sensors are used, Is the first The time of the collection is the same as the time of the collection, For the data space-time equivalent propagation velocity, Is the average value of dielectric loss values of the insulating layer within 1 hour in the step S1, Is a tiny disturbance term.
4. The flexible cable fault hidden danger real-time identification and early warning method for distribution network uninterrupted operation according to claim 1 is characterized in that step S3 combines dynamic weighting feature fusion of parameter deviation degree to calculate fusion feature value, specifically based on the following Extracting effective characteristics of 5 types of parameters, introducing the deviation degree of each parameter and a normal threshold value as a weight adjustment factor, and calculating a fusion characteristic value The formula is as follows: ; In the formula, Is the first The fusion characteristic values of the acquisition moments are in a value range of , For the parameter class index, Corresponding to 5 parameters of partial discharge pulse amplitude, skin temperature, dielectric loss value of insulating layer, environment relative humidity and real-time load current respectively, Is the first Time of acquisition The dynamic weight of the class parameter is determined, Is the first Basic weight of class parameter is initially taken as 、、、、 , Is the first The range of the deviation degree weight coefficient of the class parameter is , Is the first Time of acquisition The degree of deviation of the class parameter, Is the first Time of acquisition The aligned data values of the class parameters, Is the first The normal running minimum of the class parameter, Is the first The normal running maximum value of the class parameter, Is the first The non-linear characteristic adjustment index of the class parameter is in the range of 。
5. The flexible cable fault hidden danger real-time identification and early warning method for distribution network uninterrupted operation according to claim 1, wherein the basic weight in the step S3 is as follows Correcting according to hidden trouble identification accuracy every 12 hours , Is the first The accuracy of the individual identification of the class parameters, After correction, the correction needs to be satisfied 。
6. The flexible cable fault hidden danger real-time identification and early warning method for distribution network uninterrupted operation according to claim 1, wherein the step S4 is characterized in that the feature association degree analysis of the coupling space-time association outputs the feature association degree, in particular combining The time series correlation of (2) and the sensor space distribution correlation, and calculating the characteristic correlation degree The coupling relation between the current feature and the historical and adjacent spatial features is quantized, and the formula is as follows: ; In the formula, Is the first The coupling characteristic association degree at each acquisition time is in the range of , Near 1 indicates a high degree of association, near-1 indicates a low degree of association, For the time backtracking window length, The value range is that the space neighborhood window length is , For the time backtracking step size, For the spatial neighborhood step size, For space-time neighborhood weight, satisfy , Is the first Each acquisition time and corresponding The average value of the feature values is fused in the space-time neighborhood, The value range is that the space-time attenuation coefficient is 。
7. The flexible cable fault hidden danger real-time identification and early warning method for distribution network uninterrupted operation according to claim 1 is characterized in that step S5 is integrated with fault hidden danger membership judgment of association degree change rate, and hidden danger membership is determined, in particular based on And (3) with Introducing a rate of change of the degree of association Quantifying hidden danger development trend, and defining fault hidden danger membership degree The method comprises the following steps: ; wherein the rate of change of the degree of association The method comprises the following steps: ; In the formula, Is the first The membership degree of fault hidden danger at each acquisition moment, The value range is that the steep coefficient of the membership curve is , The value range is as follows for the fusion of the reference threshold value of the characteristic value When (when) When the characteristic value deviates to the hidden danger direction, The value range is that the association degree change rate weight is , The value range is that the membership degree offset coefficient is , Is the first The rate of change of the degree of association at each acquisition instant, Is the first The degree of characteristic association at each acquisition time, A threshold value is judged for the membership degree of the fault hidden danger, When (1) In the time-course of which the first and second contact surfaces, When (1) In the time-course of which the first and second contact surfaces, 。
8. The method for identifying and pre-warning hidden trouble of flexible cable in uninterrupted operation of distribution network in real time according to claim 1, wherein the correlation change rate in step S5 is characterized in that If the threshold value is exceeded Triggering the secondary verification of the membership degree, wherein the verification formula is that 。
9. The flexible cable fault hidden danger real-time identification and early warning method for distribution network uninterrupted operation according to claim 1 is characterized in that step S6 combines the hierarchical early warning of short-term prediction deviation to generate early warning index and push early warning information, in particular based on the following Rate of change of Future 10-minute membership degree predicted by introducing ARIMA model ( ) Deviation from the actual value Calculating early warning index And the grading early warning is as follows: ; wherein the deviation is predicted The formula is as follows: ; wherein the ARIMA model formula is as follows: ; In the formula, Is the first Early warning indexes at each acquisition time are in the range of , Is the first Hidden danger membership change rate at each acquisition time, Is the first The membership degree of fault hidden danger at each acquisition moment, The value range is that the membership degree change rate weight coefficient is , For the deviation of the predicted membership value from the actual membership value at the present time for 10 minutes in the future, Prediction based on ARIMA model The membership degree of fault hidden danger at each acquisition moment, In order to predict the step size, To the end of Membership prediction bias maximum at each acquisition time, For predicting the deviation weight coefficient, the value range is , As the autoregressive coefficients of the ARIMA model, For the moving average coefficients of the ARIMA model, Error terms for the ARIMA model; when the early warning is classified as When the yellow warning is sent out, when An orange early warning is sent out when And sending out red early warning, wherein the early warning information comprises hidden danger positions, parameter values and predicted values, and is pushed by a distribution network dispatching cloud platform.
10. The method for identifying and pre-warning hidden trouble of flexible cable in uninterrupted operation of distribution network in real time according to claim 1, wherein the coefficient of ARIMA model in step S6 is 、 Retraining by historical membership data every 6 hours, wherein the training sample size is not less than Data points.

Description

Flexible cable fault hidden danger real-time identification and early warning method for distribution network uninterrupted operation Technical Field The invention relates to the technical field of power cable fault detection and early warning, in particular to a flexible cable fault hidden danger real-time identification and early warning method for distribution network uninterrupted operation. Background The distribution network is used as a terminal core link of a power system connection user, the safe and stable operation of the distribution network directly determines the electricity reliability of social production and living, and the uninterrupted operation technology thoroughly avoids the economic loss and inconvenience caused by power failure to industrial production and resident living, so that the distribution network becomes a main stream mode of operation, maintenance and overhaul of the current distribution network. The flexible cable is widely applied in the distribution network uninterrupted operation scene by virtue of the characteristics of strong flexibility, excellent bending performance, convenient installation and laying and the like, is mainly used for temporary power supply, equipment bypass connection and other key links, but is in an electrified operation state for a long time, is influenced by multiple factors such as real-time load fluctuation, outdoor environment humidity change, natural aging of an insulating layer, mechanical abrasion and the like, is easy to gradually generate fault hidden dangers such as abnormal amplitude of partial discharge pulse, abnormal rise of skin temperature, increase of dielectric loss value of the insulating layer and the like, and gradually develops serious faults such as cable insulation breakdown, short circuit and the like if the hidden dangers are not recognized and early-warned in time, so that uninterrupted operation is forced to be interrupted, local outage accidents of the distribution network are more likely to be caused, and the safe operation and the operation safety of personnel of a power grid are threatened. The existing identification and early warning technology for fault hidden danger of the flexible cable for distribution network uninterrupted operation still has obvious limitations and disadvantages that firstly, the traditional manual inspection method relies on experience judgment of operation and maintenance personnel, inspection efficiency is low, coverage range is limited, personnel are close to the electrified flexible cable under uninterrupted operation scenes and have electric shock risks, 24-hour real-time monitoring cannot be realized, hidden danger signals with short time and burst are difficult to capture, and part of technologies only adopt single parameter monitoring such as monitoring cable skin temperature or partial discharge, synergistic influence of environment humidity and real-time load on cable states is not comprehensively considered, and hidden danger omission is easily caused by 'one-sided monitoring'. On the one hand, the conventional multi-source data fusion type identification technology adopts fixed weight to fuse parameters such as partial discharge, temperature and the like, the weight cannot be dynamically adjusted according to the deviation degree of the parameters and a normal threshold value, so that the characteristic contribution of abnormal parameters is covered by the normal parameters to influence hidden danger identification accuracy, on the other hand, an effective processing mechanism is not designed for the problems of multi-sensor installation position difference and acquisition clock deviation, the phenomenon of data time-space asynchronism is prominent, for example, the data of different sensors at the same moment actually have second-level time difference, or the data of different sensors does not consider spatial correlation, and the subsequent analysis result is directly distorted. The hidden danger judging and early warning link adaptability is poor, the hidden danger judging is mostly adopted in the prior art, real-time load dynamic adjustment is not combined with a cable, the cable insulation tolerance capacity is reduced during high-load operation, the judging threshold value of a low-load scene is still used at the moment, the problem of 'hidden danger caused by excessively high threshold value missed judgment' or 'hidden danger caused by excessively low threshold value false alarm fault' is easily caused, meanwhile, the early warning is only based on the current monitoring state, the short-term predicting capability of hidden danger development trend is lacking, the hidden danger deterioration speed cannot be predicted in advance by operation and maintenance personnel, the hidden danger deterioration speed can only be passively treated when the hidden danger is serious, early discovery and early intervention are difficult to realize, and the safety guarantee advantage of u