CN-120779178-B - Multi-mode sensing on-line monitoring method for safety partial discharge of power equipment

CN120779178BCN 120779178 BCN120779178 BCN 120779178BCN-120779178-B

Abstract

The invention relates to the technical field of safety monitoring of power equipment, and discloses a multi-mode sensing online monitoring method for safety partial discharge of the power equipment, which comprises the steps of acquiring data through ultrasonic, ultrahigh frequency and electromagnetic interference sensors, acquiring each data predicted value at a predicted moment by using a prediction model, and correcting the ultrasonic data predicted value by combining correlation analysis; and finally, obtaining a partial discharge characteristic predicted value according to the predicted correction value, the historical predicted value and the similarity, and realizing discharge state monitoring through threshold judgment. According to the method, the multi-mode data are fused, and the accuracy and the anti-interference capability of partial discharge monitoring are improved by combining correlation degree and historical data similarity analysis, so that the safety of the power equipment can be effectively ensured.

Inventors

SUN CE
YANG LANZHOU
LI WEIYUAN

Assignees

安徽中成检测有限公司

Dates

Publication Date: 20260508
Application Date: 20250730

Claims (9)

1. The method for monitoring the safety partial discharge multi-mode sensing on line of the power equipment is characterized by comprising the following steps of: collecting ultrasonic data, ultrahigh frequency data and electromagnetic interference data of external environments of partial discharge of power equipment through different types of sensors; The method comprises the steps of recording the next time of a current time as a prediction time, obtaining predicted values of ultrasonic data, ultrahigh frequency data and electromagnetic interference data at the prediction time through a prediction model for a preset number of times before the current time, obtaining different partial discharge characteristic predicted values according to the association degree of the ultrasonic data and the ultrahigh frequency data, the association degree of the ultrasonic data and the electromagnetic interference data and the predicted values of the ultrahigh frequency data and the ultrasonic data at the prediction time, and correcting the predicted values of the ultrasonic data at the prediction time by combining the different partial discharge characteristic predicted values according to the association degree of the ultrasonic data and the ultrahigh frequency data and the respective duty ratio of the correlation degree of the ultrasonic data and the electromagnetic interference data to obtain a prediction correction value; For the clustering cluster where the current moment is located, the three-dimensional characteristic points form different sequences according to the ultrasonic signal, the ultrahigh frequency signal and the electromagnetic interference signal, the sequences with different lengths are analyzed to obtain the similarity among the three-dimensional characteristic points with different sequence lengths, and then the historical predicted value of the partial discharge characteristic is obtained by combining the actual discharge characteristic value formation weight; Obtaining a partial discharge characteristic predicted value according to the similarity between the predicted correction value and the historical predicted value and the three-dimensional characteristic point; the method for obtaining the partial discharge characteristic predicted value according to the similarity between the predicted correction value and the historical predicted value and the three-dimensional characteristic point comprises the following steps: Obtaining all similar feature points under the current sequence length, and obtaining a prediction correction value and a historical prediction value of the similar feature points; Respectively differencing the predicted correction value and the history predicted value of each similar characteristic point with the actual discharge characteristic value at the next moment to form a correction difference sequence and a history difference sequence; For each difference sequence, taking the product of the variance of the difference sequence and the average value of all sequence values in the difference sequence as a weight factor; Comparing the weight factor of the corrected difference sequence with the sum of the weight factor of the corrected difference sequence and the weight factor of the historical difference sequence, and taking the obtained ratio as a weight coefficient of the predicted correction value; comparing the weight factor of the history difference sequence with the sum of the weight factors of the history difference sequence and the sum of the weight factors, obtaining a weight coefficient of the historical predicted value; and carrying out weighted addition on the prediction correction value and the historical prediction value through respective weight coefficients to obtain the partial discharge characteristic prediction value.
2. The method for monitoring the safety partial discharge multi-mode sensing on line of the power equipment according to claim 1, wherein the method for collecting the ultrasonic data, the ultrahigh frequency data and the electromagnetic interference data of the external environment of the partial discharge of the power equipment by the sensors of different types is as follows: The method comprises the steps of arranging an ultrasonic sensor and an ultrahigh frequency sensor on the surface of power equipment, arranging an electromagnetic interference sensor in the surrounding environment of the equipment, collecting data once every preset time of each sensor, collecting a plurality of data, recording the collected data at all moments as a sequence, and respectively obtaining an ultrasonic sequence, an ultrahigh frequency sequence and an electromagnetic interference sequence.
3. The method for online monitoring of safe partial discharge multi-mode sensing of power equipment according to claim 2, wherein the method for obtaining different prediction values of partial discharge characteristics according to the association degree of ultrasonic data and ultrahigh frequency data, the association degree of ultrasonic data and electromagnetic interference data, and the prediction values of the ultrahigh frequency data and the ultrasonic data at the prediction time is as follows: Calculating the association degree of the sequence values of the ultrasonic sequence and the electromagnetic interference sequence at the same time, and forming a sequence as a sound interference association sequence; the average value of all sequence values in the audio correlation sequence is marked as a first correlation degree, and the average value of all sequence values in the sound interference correlation sequence is marked as a second correlation degree; And obtaining a second partial discharge characteristic predicted value and a third partial discharge characteristic predicted value according to the first association degree, the second association degree and the predicted values of the ultrahigh frequency sequence and the electromagnetic interference sequence at the predicted time.
4. The method for online monitoring of safe partial discharge multi-mode sensing of power equipment according to claim 3, wherein the method for obtaining the second partial discharge characteristic predicted value and the third partial discharge characteristic predicted value according to the first association degree, the second association degree and the predicted values of the ultrahigh frequency sequence and the electromagnetic interference sequence at the predicted time is as follows: and multiplying the predicted value of the ultrahigh frequency sequence at the predicted time by the first association degree to obtain a second partial discharge characteristic predicted value, and multiplying the predicted value of the electromagnetic interference sequence at the predicted time by the second association degree to obtain a third partial discharge characteristic predicted value.
5. The method for online monitoring of safe partial discharge multi-mode sensing of power equipment according to claim 2, wherein the method for correcting the predicted value of the ultrasonic data at the predicted time by combining different partial discharge characteristic predicted values with the respective duty ratios of the correlation degree of the ultrasonic data and the ultrahigh frequency data and the correlation degree of the ultrasonic data and the electromagnetic interference data is as follows: calculating standard deviations of all sequence values in the audio correlation sequence, and taking the reciprocal of the standard deviation as the correlation degree of the ultrasonic sequence and the ultrahigh frequency sequence; calculating standard deviations of all sequence values in the sound interference correlation sequence, and taking the reciprocal of the standard deviation as the correlation degree of the ultrasonic wave sequence and the electromagnetic interference sequence; marking the predicted value of the ultrasonic data at the predicted moment obtained by the prediction model as a first partial discharge characteristic predicted value; Presetting a weighted value for the first partial discharge characteristic predicted value, and acquiring weighted values of the second partial discharge characteristic predicted value and the third partial discharge characteristic predicted value according to the ratio of the correlation; and weighting and adding the first partial discharge characteristic predicted value, the second partial discharge characteristic predicted value and the third partial discharge characteristic predicted value to obtain a predicted correction value.
6. The method for online monitoring of safe partial discharge multi-mode sensing of power equipment according to claim 2, wherein for the cluster where the current moment is located, the three-dimensional feature points are formed into different sequences according to ultrasonic signals, ultrahigh frequency signals and electromagnetic interference signals, the sequences with different lengths are analyzed to obtain the similarity between the three-dimensional feature points with different sequence lengths, and then the actual discharge feature value is combined to form a weight to obtain the historical predicted value of the partial discharge feature, the method comprises the following steps: marking three-dimensional feature points corresponding to the current moment as target feature points, acquiring a cluster in which the target feature points are located, and marking all feature points in the cluster as class feature points; the method comprises the steps that the class feature points select ultrasonic waves, ultrahigh frequency and electromagnetic interference signals at preset moments to obtain ultrasonic sequences, ultrahigh frequency sequences and electromagnetic interference sequences corresponding to the class feature points; Obtaining the similarity between three-dimensional feature points according to the similarity of the corresponding sequences of the class feature points and the target feature points, and determining the similar feature points; And obtaining historical predicted values of the partial discharge characteristics according to the discharge characteristic values of the corresponding similar characteristic points at the next moment under different sequence lengths.
7. The method for online monitoring of safe partial discharge multi-mode sensing of power equipment according to claim 6, wherein the method for obtaining the similarity between three-dimensional feature points according to the similarity of the corresponding sequences of the class feature points and the target feature points and determining the similarity feature points is as follows: for each class feature point, calculating the similarity between the corresponding ultrasonic sequence, the ultrahigh frequency sequence and the electromagnetic interference sequence and the ultrasonic sequence, the ultrahigh frequency sequence and the electromagnetic interference sequence at the current moment respectively; The similarity of the class feature points and the three sequences of the target feature points is averaged, and then linear normalization is carried out to obtain the similarity of the class feature points and the target feature points; and obtaining the similarity of all the class feature points and the target feature points, classifying all the similarities of all the class feature points and the target feature points through a threshold segmentation algorithm, obtaining the similarity mean value of each class after classification, and marking the class feature points in the class with large similarity mean value as the similar feature points.
8. The method for online monitoring of safe partial discharge multi-mode sensing of power equipment according to claim 6, wherein the method for obtaining the historical predicted value of the partial discharge characteristic according to the discharge characteristic value of the corresponding similar characteristic point at the next moment under different sequence lengths is as follows: For each sequence length, taking the average value of the discharge characteristic values of all similar characteristic points adjacent to the next moment as a prediction reference value of the prediction moment, and calculating the variance of the discharge characteristic values of all similar characteristic points at the next moment under the sequence length; and obtaining the historical predicted value by weighting and summing the predicted reference values under all the sequence lengths through the period weight.
9. The method for determining the preset time in the data acquired by each sensor every time the preset time passes is characterized in that the optimal value of the data acquisition time interval is acquired through statistical analysis according to the operation period of the power equipment and the fluctuation frequency of the historical partial discharge data, and the standard acquisition period corresponding to the current equipment type and the optimal value are averaged to be used as the final preset time.

Description

Multi-mode sensing on-line monitoring method for safety partial discharge of power equipment Technical Field The invention relates to the technical field of safety monitoring of power equipment, in particular to a multi-mode sensing on-line monitoring method for safety partial discharge of the power equipment. Background During operation of an electrical power system, a partial discharge phenomenon of electrical power equipment is one of important causes of equipment failure. Early monitoring and accurate early warning of partial discharge are of great significance in guaranteeing safe and stable operation of a power system. As power equipment evolves toward high voltage and large capacity, its operating environment becomes increasingly complex, and conventional methods of partial discharge monitoring face a number of challenges. In the prior art, a single-mode partial discharge monitoring method is common, for example, data acquisition and analysis are performed only by means of an ultrasonic sensor or an ultrahigh frequency sensor. However, the method can only acquire the discharge characteristic information with a single dimension, so that the real state of partial discharge is difficult to comprehensively and accurately reflect, the anti-interference capability is weak in a complex electromagnetic environment, and the deviation and even the misjudgment of a monitoring result are easy to occur. In addition, when the traditional method processes multi-source data, deep analysis of relevance among different types of data is often lacking, and complementary information among ultrasonic data, ultrahigh frequency data and electromagnetic interference data cannot be fully utilized, so that accuracy and reliability of partial discharge characteristic prediction are affected. In the existing partial discharge monitoring method, a collection mode with a fixed time interval is generally adopted on a data collection strategy, and the mode cannot be adaptively adjusted according to the actual operation period of power equipment and the fluctuation frequency of historical partial discharge data, so that the data collection efficiency is low or key information is omitted. In the aspect of prediction model construction, the prior art is used for predicting based on a single data sequence, lacks comprehensive consideration of the relevance of multi-mode data, does not effectively correct a prediction result, and is difficult to adapt to complex and changeable working conditions in the operation process of power equipment. For the utilization of historical data, the traditional method often fails to fully mine the similarity characteristics among the data, and the historical data with similar discharge characteristics cannot be effectively classified by means of cluster analysis and the like, so that the accuracy of the current discharge state prediction is difficult to improve by means of the rules of the historical data. In the face of external environment electromagnetic interference, the prior art also lacks an effective interference identification and suppression mechanism, so that the stability and reliability of a monitoring system are greatly affected. In summary, the existing power equipment partial discharge monitoring method has obvious defects in the aspects of multi-mode data fusion, anti-interference capability, prediction precision, data acquisition efficiency and the like, and a more advanced and reliable monitoring method is urgently needed to meet actual engineering requirements. Disclosure of Invention The invention aims to provide a safe partial discharge multi-mode sensing on-line monitoring method for power equipment, so as to solve the problems in the background technology. In order to achieve the purpose, the invention provides the technical scheme that the method for monitoring the safety partial discharge multi-mode sensing on line of the power equipment comprises the following steps: collecting ultrasonic data, ultrahigh frequency data and electromagnetic interference data of external environments of partial discharge of power equipment through different types of sensors; The method comprises the steps of recording the next time of a current time as a prediction time, obtaining predicted values of ultrasonic data, ultrahigh frequency data and electromagnetic interference data at the prediction time through a prediction model for a preset number of times before the current time, obtaining different partial discharge characteristic predicted values according to the association degree of the ultrasonic data and the ultrahigh frequency data, the association degree of the ultrasonic data and the electromagnetic interference data and the predicted values of the ultrahigh frequency data and the ultrasonic data at the prediction time, and correcting the predicted values of the ultrasonic data at the prediction time by combining the different partial discharge characteristic predicted values according t