CN-121980329-A - Automatic judging method, equipment and medium for full-performance detection result of metering equipment

CN121980329ACN 121980329 ACN121980329 ACN 121980329ACN-121980329-A

Abstract

The invention relates to the technical field of performance detection, and discloses a method, equipment and medium for automatically judging the full-performance detection result of metering equipment, wherein the method comprises the steps of collecting multidimensional performance parameters of the metering equipment under different test conditions, preprocessing the multidimensional performance parameters and extracting preliminary characteristics, and then carrying out intelligent comprehensive processing by using a nonlinear Gao Weijiang-dimensional algorithm to obtain comprehensive performance parameter data; the comprehensive performance parameter data is intelligently detected through a detection algorithm based on machine learning to obtain a detection result, and a judgment result is automatically generated based on the detection result and the self-adaptive threshold adjustment.

Inventors

WANG JUNRONG
XU HONGWEI
REN QIJUN
LI XIANG
CHEN QINGHUI
Tang Hezhong
LI PENGCHENG
TANG XIANMIN

Assignees

贵州电网有限责任公司

Dates

Publication Date: 20260505
Application Date: 20251218

Claims (10)

1. An automatic judging method for the full-performance detection result of metering equipment is characterized by comprising the following steps, Determining test conditions, and combining according to the test conditions to form different test scenes; under different test scenes, collecting multidimensional performance parameters of the metering equipment through a sensor, and preprocessing the multidimensional performance parameters; Extracting preliminary features of the preprocessed data to obtain a preliminary feature data set; Performing intelligent comprehensive processing on the preliminary characteristic dataset by using a nonlinear Gao Weijiang-dimensional algorithm to obtain comprehensive performance parameter data; Intelligent detection is carried out on the comprehensive performance parameter data through a detection algorithm based on machine learning, so that a detection result is obtained; based on the detection result, a judging result is automatically generated by combining an adaptive threshold adjustment mechanism, and the adaptive threshold is dynamically adjusted according to the real-time measurement error, the historical deviation, the running time variation and the environmental temperature of the equipment.
2. The method for automatically judging the full-performance detection result of the metering equipment according to claim 1, wherein the test conditions comprise external condition environmental factors and internal factor equipment states; the environmental factors include temperature, humidity, electromagnetic interference; The equipment state comprises a load level and equipment aging degree; based on the test conditions, different test scenes are formed by combination according to specific application scenes, and under the different test scenes, the sensor is used for collecting multidimensional performance parameters of the metering equipment; the multi-dimensional performance parameters include measurement error, repeatability, linearity, drift, temperature effects, response time.
3. The method for automatically judging the full-performance detection result of the metering equipment according to claim 2, wherein the preprocessing of the multi-dimensional performance parameter comprises the preprocessing of cleaning, outlier processing, standardization processing and dimension processing of the multi-dimensional performance parameter to obtain the preprocessed multi-dimensional performance parameter; And carrying out preliminary feature extraction on the preprocessed multidimensional performance parameters by utilizing a feature engineering technology to obtain a preliminary feature data set.
4. The method for automatically judging the full-performance detection result of the metering equipment according to claim 3, wherein the intelligent comprehensive processing of the preliminary feature dataset by using the nonlinear Gao Weijiang-dimensional algorithm comprises the steps of Assume that the preliminary feature dataset contains Each sample has Individual features, the preliminary feature dataset is expressed as ; Performing high-dimensional tensor mapping on the preliminary feature data to construct feature data, wherein the specific expression is as follows: Wherein, the Is a high-dimensional characteristic tensor constructed before dimension reduction, Is the first A high-dimensional feature vector of the individual samples, Is that Is to be used in the present invention, Is the operation of the tensor product, Is a non-linear transformation function that is a function of the non-linear transformation, Is an exponential decay term that is used to determine, Is the Hadamard product.
5. The method for automatically determining full-performance detection results of a metering device according to claim 4, wherein said intelligent integrated processing comprises the steps of Performing tensor decomposition, reducing computation complexity, extracting main characteristic information, performing tensor decomposition on tensor by adopting a singular value adjustment method, Wherein, the Is the characteristic data after the decomposition, Is the order of the tensor decomposition, Is the first The value of the characteristic is a value of, Is the first The left singular vectors of the number of the vectors, Is the first The right singular vectors of the two vectors are used, Is a low rank approximate projection matrix for representing the th Influence of individual feature directions; Is that Is a euclidean norm of (c), Is a regularized term that is used to determine the degree of regularization, Is an identity matrix of the unit cell, Is an adjustment parameter.
6. The method for automatically determining full-performance detection results of a metering device according to claim 5, wherein said intelligent integrated processing further comprises, based on the decomposed feature data Performing nonlinear multi-stage projective transformation, and further projecting the data into a low-dimensional subspace; adopting a multi-stage iteration mode to enable the data to gradually approach a final dimension reduction target and initially project a matrix The definition is as follows: Wherein, the Projection matrix for random initialization, at each iteration step In the method, the projection mode of the data adopts nonlinear kernel function transformation, so that the dimension reduction process is adapted to the nonlinear characteristic of the data: Wherein, the Is the characteristic data after nonlinear transformation; for the sigmoid activation function, For dynamically adjusting parameters, the method is used for controlling the influence of nonlinear components in each step of iterative process, and is determined according to an expert experience method; Is the first And (3) characteristic data after nonlinear transformation in the iteration.
7. The method for automatically determining full-performance detection results of a metering device according to claim 6, wherein the step of automatically generating the determination results in combination with the adaptive threshold adjustment mechanism comprises introducing an adaptive threshold adjustment calculation formula, determining the detection results, Wherein, the Is the current time Is the first of (2) The adaptive threshold of the class-metering device, Is the first The initial threshold of the class metering device in a standard operating environment, Is the first Sensitivity coefficient of the performance of the class meter to historical bias changes, Is the current time Is the first of (2) The class meter device measures the amount of error change, Is the current time Is the first of (2) The measurement error of the class gauge device is, Is the first The time-sensitivity coefficient of the class-metering device, Is the first The amount of change in the run time of the class metering device, Is the first The working time referenced when the class metering device is calibrated, Is the first Influence of the performance of the metering-like device the sensitivity coefficient, Is the current time Is the first of (2) The ambient temperature of the class meter device, Is the first The reference ambient temperature at the time of calibration of the class meter, Is a temperature sensitive index.
8. The method for automatically determining full-performance detection results of metering equipment according to claim 7, wherein the step of automatically generating the determination results by combining the adaptive threshold adjustment mechanism further comprises the step of combining the detection results of each type of equipment with the detection results of the current type of equipment Comparing, when the threshold value is exceeded, confirming that the equipment is unqualified, otherwise, the equipment is qualified; And if the equipment is judged to be unqualified, generating an unqualified report, marking a specific abnormal type and providing a fault checking suggestion.
9. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor, when executing the computer program, implements the steps of a method for automatically determining the full performance test results of a metering device according to any one of claims 1 to 8.
10. A computer-readable storage medium having stored thereon a computer program, characterized in that the computer program, when executed by a processor, implements the steps of a method for automatically determining the full-performance test results of a metering device according to any one of claims 1 to 8.

Description

Automatic judging method, equipment and medium for full-performance detection result of metering equipment Technical Field The invention relates to the technical field of performance detection, in particular to an automatic judging method, equipment and medium for full-performance detection results of metering equipment. Background In the modern industry, accurate metering equipment detection and performance assessment is critical to ensure stability of equipment operation and production quality. With the development of industrial technology, the types and application scenes of metering equipment become more and more complex, and the performance detection requirements of the metering equipment are correspondingly improved. However, the conventional measuring equipment performance detection methods are mostly dependent on manual experience and simple linear analysis means, and have larger limitations when processing complex and multidimensional data, and cannot effectively capture the nonlinear characteristics of the equipment under different working environments, so that the detection result is inaccurate or misjudged. In practice, the performance of a metering device is often affected by a number of factors, such as the age of the device, ambient temperature, humidity changes, load on use, etc. The conventional method often fails to consider the interaction of these factors, so that the detection result cannot sufficiently reflect the actual working state of the device. The existing performance detection method is mostly limited to the qualification judgment of equipment by a fixed threshold comparison method, and the method shows poor robustness when facing complex environments and variable conditions. Along with the extension of the service time and the change of the working condition of the equipment, the limitations of the traditional detection means are particularly prominent, and the potential performance degradation or abnormality of the equipment cannot be found in time. Disclosure of Invention The present invention has been made in view of the above-described problems occurring in the prior art. Therefore, the invention provides the automatic judging method for the full-performance detection result of the metering equipment, which can solve the technical problems of incomplete consideration of the performance parameters of the equipment, inaccurate treatment and inaccurate detection judgment. The technical scheme includes that testing conditions are determined, different testing scenes are formed by combining the testing conditions, under the different testing scenes, multidimensional performance parameters of the metering equipment are collected through a sensor, preprocessing is conducted on the multidimensional performance parameters, preliminary feature extraction is conducted on preprocessed data to obtain a preliminary feature dataset, a nonlinear Gao Weijiang-dimensional algorithm is utilized to conduct intelligent comprehensive processing on the preliminary feature dataset to obtain comprehensive performance parameter data, intelligent detection is conducted on the comprehensive performance parameter data through a machine learning-based detection algorithm to obtain a detection result, the detection result is automatically generated based on the detection result, and an adaptive threshold value adjusting mechanism is used for dynamically adjusting the adaptive threshold value according to real-time measurement errors, historical deviation, running time variation and environmental temperature of the equipment. As a preferable scheme of the automatic judging method of the full-performance detection result of the metering equipment, the testing condition comprises an external condition environment factor and an internal factor equipment state; the environmental factors include temperature, humidity, electromagnetic interference; The equipment state comprises a load level and equipment aging degree; based on the test conditions, different test scenes are formed by combination according to specific application scenes, and under the different test scenes, the sensor is used for collecting multidimensional performance parameters of the metering equipment; the multi-dimensional performance parameters include measurement error, repeatability, linearity, drift, temperature effects, response time. The method for automatically judging the full-performance detection result of the metering equipment comprises the following steps of preprocessing the multi-dimensional performance parameters, such as cleaning, outlier processing, standardization processing and dimension processing, to obtain preprocessed multi-dimensional performance parameters; And carrying out preliminary feature extraction on the preprocessed multidimensional performance parameters by utilizing a feature engineering technology to obtain a preliminary feature data set. As a preferable scheme of the automatic judging method of the full-