CN-122020210-A - Intelligent measuring switch with fault self-identification and risk self-early warning functions

CN122020210ACN 122020210 ACN122020210 ACN 122020210ACN-122020210-A

Abstract

The invention provides an intelligent measuring switch with fault self-identification and risk self-warning functions, which is characterized in that a transient response waveform is accurately obtained and preprocessed by dynamically increasing the sampling rate under the triggering of a natural operation event, the characteristic of an intrinsic response is extracted by adopting multi-level wavelet packet decomposition and energy distribution analysis, a typical fingerprint and a characteristic boundary for identifying the state of equipment are generated by fingerprint clustering and dispersion evaluation, and the dynamic self-adaptive adjustment of a fault judgment threshold value is realized by combining a preset lookup table, so that an online self-calibration diagnosis model is realized.

Inventors

Tian hongfeng
YUE XING
PENG RUI
Zou Changrun
GAO YUXUAN
LIANG CHENGKE
GENG YUE
JIANG YUAN
QI FAN
Mao xinyi

Assignees

莱芜鲁能开源集团电器有限公司

Dates

Publication Date: 20260512
Application Date: 20260414

Claims (10)

1. The intelligent measuring switch with the fault self-identification and risk self-early warning functions is characterized by realizing diagnosis early warning through the following modes: S1, acquiring a high sampling rate electrical transient response waveform of an intelligent measuring switch under the triggering of a natural operation event to form an original transient signal sequence to be processed; s2, performing wavelet packet decomposition processing of a fixed order on the original transient signal sequence to generate a decomposition coefficient set; s3, extracting energy distribution entropy, main frequency offset and first half-wave rise time consistency indexes in a preset frequency band based on the decomposition coefficient set, and constructing an intrinsic response fingerprint; s4, inputting a plurality of continuous intrinsic response fingerprints into a local fingerprint clustering device to perform aggregation operation, and generating a typical fingerprint cluster center and a dispersion boundary; S5, inquiring a solidified fingerprint-threshold calibration lookup table according to the numerical value of the dispersion boundary to generate a dynamic fault determination tolerance parameter; s6, updating the register configuration of the fault identification module by utilizing the dynamic fault determination tolerance parameter to replace the original static characteristic threshold value and complete the online self calibration of the diagnosis model; s7, comparing and judging the electrical characteristic data acquired in real time based on the updated dynamic fault judgment tolerance parameters, and outputting a preliminary diagnosis result; and S8, monitoring the stability of the preliminary diagnosis result and feeding back to the local fingerprint clustering device so as to drive the accumulated updating of the intrinsic response fingerprint of the next round and the dynamic correction of the dispersion boundary.
2. The intelligent measuring switch with fault self-identification and risk self-early warning functions according to claim 1, wherein the wavelet packet decomposition comprises synchronous multi-layer decomposition of a current channel and a voltage channel, a full-band structured characteristic coefficient matrix is generated, and a standardized band energy vector is obtained through band energy normalization calculation.
3. The intelligent measuring switch with fault self-identification and risk self-warning functions according to claim 1, wherein step S3 specifically comprises: Normalizing the energy distribution information of each frequency band contained in the analysis coefficient set to generate a normalized frequency band energy probability distribution sequence; Performing shannon entropy calculation based on the normalized frequency band energy probability distribution sequence to generate an energy distribution entropy value; performing spectrum centroid offset detection by using a peak frequency position in a preset main frequency band in the decomposition coefficient set, and generating main frequency offset data; Performing zero-crossing point and peak point time difference measurement for the first half-wave period in an original transient signal sequence, generating a first half-wave rise time original measurement value, processing the first half-wave rise time original measurement value of continuous multiple events through a sliding variance algorithm, and generating a first half-wave rise time consistency index; and vector splicing and packaging the energy distribution entropy value, the main frequency offset data and the first half-wave rise time consistency index to generate an intrinsic response fingerprint vector.
4. The intelligent measuring switch with fault self-identification and risk self-early warning functions according to claim 3, wherein the energy distribution entropy value represents signal frequency spectrum complexity and disorder degree, the main frequency offset data reflects equipment inductance and capacitance parameter micro-variation, and the first half-wave rise time consistency index represents waveform repeatability and consistency.
5. The intelligent measuring switch with fault self-identification and risk self-warning functions according to claim 1, wherein step S4 specifically comprises: acquiring a plurality of continuous eigenvalue fingerprint vectors, and executing sliding time window screening processing on the eigenvalue fingerprint vectors to generate an effective fingerprint data set; Initializing iterative calculation parameters of a local fingerprint clustering device based on the effective fingerprint data set, and executing aggregation convergence operation on multidimensional feature coordinates in the effective fingerprint data set to generate typical fingerprint cluster center coordinates; calculating Euclidean distance distribution from each intrinsic response fingerprint vector in the effective fingerprint data set to the center coordinate of the typical fingerprint cluster by using the center coordinate of the typical fingerprint cluster as a reference standard, and generating an original distance deviation sequence; And carrying out statistical variance analysis and confidence interval estimation on the original distance deviation sequence, extracting an upper limit boundary value of the distance distribution according to a preset confidence probability threshold value, and generating a dispersion boundary value.
6. The intelligent measuring switch with the fault self-identification and risk self-early warning functions according to claim 5, wherein the center coordinates of the typical fingerprint clusters represent average states of electrical characteristics of the current equipment in an aging stage and at a temperature rise level, and the dispersion boundary values represent fluctuation ranges and stability of the electrical characteristics of the current equipment.
7. The intelligent measurement switch with fault self-identification and risk self-warning function according to claim 5, wherein step S4 further comprises: And carrying out structured packaging treatment on the center coordinates of the typical fingerprint clusters and the dispersion boundary values to generate an aggregation operation result packet.
8. The intelligent measuring switch with fault self-identification and risk self-warning functions according to claim 1, wherein step S5 specifically comprises: acquiring a current typical fingerprint cluster dispersion boundary value output by a local fingerprint cluster, and performing normalized quantization processing on the current typical fingerprint cluster dispersion boundary value to generate a normalized drift degree level index value; performing address mapping and data retrieval operations in a cured fingerprint threshold calibration lookup table based on the drift degree level index value, and extracting an original tolerance correction coefficient set uniquely corresponding to the drift degree level index value; Performing linear weighting operation on a basic static characteristic threshold value of the fault identification module by using the original tolerance correction coefficient set to generate an intermediate state dynamic fault determination tolerance parameter; And executing boundary clipping and validity verification on the intermediate state dynamic fault determination tolerance parameter to generate a final effective dynamic fault determination tolerance parameter.
9. The intelligent measurement switch with fault self-identification and risk self-warning function according to claim 8, wherein step S5 further comprises: and packaging the final effective dynamic fault determination tolerance parameters into a standard register configuration instruction format to generate a dynamic fault determination tolerance parameter configuration package.
10. The intelligent measuring switch with the fault self-identification and risk self-warning functions according to claim 1, wherein in the step S7, based on updated dynamic fault determination tolerance parameters, comparison and judgment are performed on electrical characteristic data collected and preprocessed in real time, a binary state flag bit is generated by comparing an overcurrent rise rate, a zero offset, a harmonic distortion rate and an adaptive threshold boundary interval, and then a primary diagnosis result is output through fusion judgment.

Description

Intelligent measuring switch with fault self-identification and risk self-early warning functions Technical Field The invention relates to the technical field of intelligent power equipment measurement and self-adaptive fault diagnosis, in particular to an intelligent measurement switch with fault self-identification and risk self-early warning functions. Background In the field of fault diagnosis of the current intelligent measuring switch, an identification scheme based on an electrical characteristic threshold is generally adopted, and abnormal state judgment is realized mainly through comparison of key electrical parameters such as overcurrent rising rate, zero point offset, harmonic distortion rate and the like with a preset static threshold. Most of the methods complete multi-working condition calibration and threshold solidification before the equipment leaves the factory, and part of schemes are optimized by assisting with environmental compensation or sliding statistics correction, but the whole method still highly depends on external parameter input and manual experience setting, lacks self-adaptive adjustment capability for long-term running state of the equipment, and is difficult to meet long-term, stable and autonomous monitoring requirements on site; In practical engineering applications, the prior art presents significant limitations: Firstly, the diagnosis model excessively depends on environment sensing data such as external temperature and humidity, load rate and the like and cloud modeling calculation, and a large amount of data needs to be continuously uploaded, so that the calculation cost is high, the response delay is high, and the light-weight and localized autonomous operation cannot be realized at the embedded terminal; secondly, the whole fault identification process adopts a static threshold value, cannot be dynamically adjusted along with the electrical intrinsic characteristic change caused by factors such as equipment aging, temperature rise drift and the like, is extremely easy to cause misjudgment and missed judgment, and is difficult to ensure the diagnosis precision and reliability; The problems commonly cause poor adaptability and insufficient stability of the traditional fault diagnosis method in long-term operation, and the engineering requirements of the intelligent measuring switch on high reliability, high real-time performance and high self-adaptability are difficult to meet. Disclosure of Invention The invention aims to solve the technical problems and provides an intelligent measuring switch with functions of fault self-identification and risk self-early warning. The technical scheme of the invention is realized by an intelligent measuring switch with fault self-identification and risk self-early warning functions, wherein the intelligent measuring switch realizes diagnosis early warning by the following modes: S1, acquiring a high sampling rate electrical transient response waveform of an intelligent measuring switch under the triggering of a natural operation event so as to form an original transient signal sequence to be processed; s2, performing wavelet packet decomposition processing of a fixed order on the original transient signal sequence to generate a decomposition coefficient set containing multiband energy distribution information; S3, extracting energy distribution entropy, main frequency offset and first half-wave rise time consistency indexes in a preset frequency band based on the decomposition coefficient set so as to construct an intrinsic response fingerprint representing the characteristics of a single event; S4, inputting a plurality of continuous intrinsic response fingerprints into a local fingerprint clustering device to perform aggregation operation so as to generate a typical fingerprint cluster center and a dispersion boundary which characterize the running state of the current equipment; S5, inquiring a solidified fingerprint-threshold calibration lookup table according to the numerical value of the dispersion boundary to generate a dynamic fault determination tolerance parameter adapting to the current electrical characteristic drift degree; S6, updating the register configuration of the fault identification module by utilizing the dynamic fault determination tolerance parameter to replace the original static characteristic threshold value and complete the online self-calibration of the diagnosis model; S7, comparing and judging the electrical characteristic data acquired in real time based on the updated dynamic fault judgment tolerance parameters so as to output a preliminary diagnosis result with environmental adaptability; And S8, monitoring the stability of the preliminary diagnosis result and feeding back to a local fingerprint clustering device so as to drive the accumulated updating of the intrinsic response fingerprint of the next round and the dynamic correction of the dispersion boundary. The invention also provides an intel