Multi-parameter electric measuring instrument self-diagnosis system based on data analysis

Abstract

The present invention discloses a multi-parameter electric measuring instrument self-diagnosis system based on data analysis, which specifically relates to the field of self-diagnosis technology, including a parameter acquisition module, a rule initial diagnosis module, an algorithm re-diagnosis module, and a comprehensive judgment module. The parameter acquisition module acquires electric parameter data of the multi-parameter electric measuring instrument and transmits it to the rule initial diagnosis module. The rule initial diagnosis module presets diagnostic rules to diagnose abnormal faults. The algorithm re-diagnosis module trains a neural network model according to the electric parameter data of the abnormal faults for diagnosis and identification. The comprehensive judgment module integrates the diagnosis results based on the diagnostic rules and the neural network algorithm and generates a fault report. The decision implementation module outputs a self-diagnosis decision according to the fault report, prompting staff to manage and maintain the abnormal faults. The self-diagnosis accuracy of the multi-parameter electric measuring instrument is effectively improved while the flexibility of the self-diagnosis system is enhanced, and the problem of manually adjusting the rules and extracting features when the environmental conditions change is avoided.

Description

Technical Field

[0001] The present invention relates to the technical field of Internet of Things, and more specifically, to a multi-parameter electric measuring instrument self-diagnosis system based on data analysis. Background Art

[0002] A multi-parameter electrical measuring instrument is an instrument used to measure various electrical parameters in electrical equipment or circuits. The electrical parameter measuring instrument can simultaneously measure multiple parameters such as voltage, current, power, etc., and provide high-precision measurement and multi-functional analysis support, such as data recording, graphic display, remote monitoring, etc. In the existing self-diagnosis of multi-parameter electrical measuring instruments, specific frequency interference and nonlinear distortion faults are usually distinguished and identified using spectrum analysis methods, interference frequencies are judged by preset rules, and nonlinear distortion is diagnosed using harmonic analysis. However, the diagnostic method based on statistical analysis and preset rules is difficult to adapt to complex abnormal fault modes and behaviors, and the diagnostic accuracy of abnormal types is insufficient. In addition, when environmental conditions and working conditions change, the rules must be manually set and adjusted, and the flexibility of self-diagnosis analysis is limited. On the other hand, frequency interference and nonlinear distortion are often intertwined and related. Frequency interference generates nonlinear distortion through the nonlinear elements of the multi-parameter electrical measuring instrument. At the same time, the harmonic components generated by nonlinear distortion will also introduce new frequency interference. For such complex abnormal fault modes, the existing self-diagnosis system is difficult to ensure that the diagnostic flexibility is enhanced while improving the diagnostic accuracy, resulting in an imbalance between accuracy and flexibility in self-diagnosis analysis.

[0003] In order to solve the above defects, a technical solution is now proposed. Summary of the invention

[0004] The purpose of the present invention is to provide a multi-parameter electric measuring instrument self-diagnosis system based on data analysis to solve the problems in the above-mentioned background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions: a multi-parameter electric measuring instrument self-diagnosis system based on data analysis, comprising a parameter acquisition module, a rule initial diagnosis module, an algorithm re-diagnosis module, and a comprehensive judgment module;

[0006] The parameter acquisition module is used to collect the electrical parameter data of the multi-parameter electrical measuring instrument, and transmit the electrical parameter data to the rule initial diagnosis module after pre-processing;

[0007] The rule-based initial diagnosis module is used to diagnose abnormal faults based on preset diagnostic rules, detect abnormal frequencies through spectrum analysis, and detect nonlinear distortion through harmonic analysis;

[0008] The algorithm re-diagnosis module is used to train the neural network model based on the electrical parameter data of abnormal faults, and use the neural network algorithm to diagnose and identify frequency interference and nonlinear distortion faults;

[0009] The comprehensive judgment module is used to judge abnormal faults and generate fault reports by combining the diagnosis results based on the diagnosis rules and neural network algorithms;

[0010] The decision implementation module is used to output self-diagnosis decisions based on fault reports, prompting staff to manage and maintain abnormal faults.

[0011] Preferably, the electrical parameter data of the multi-parameter electrical measuring instrument include voltage, current, frequency information, harmonic components, time domain information, nonlinear information and physical information. The frequency information includes the fundamental frequency, frequency fluctuation and interference frequency of the multi-parameter electrical measuring instrument. The harmonic components include harmonic amplitude, total harmonic distortion and harmonic content. The time domain information includes signal peak value, signal effective value and waveform factor. The nonlinear information includes phase characteristics, Lyapunov exponent and approximate entropy. The physical information includes temperature and humidity of the multi-parameter electrical measuring instrument.

[0012] Preferably, the pretreatment method is:

[0013] Use filters to remove noise and smooth the data, and normalize the data by combining wavelet transform;

[0014] The method of normalizing data by combining wavelet transform is:

[0015] The wavelet basis function is used to decompose the data into detail coefficients and approximation coefficients at different scales. The detail coefficients are high-frequency components, and the approximation coefficients are low-frequency components. The approximation coefficients are decomposed at multiple levels to obtain detail coefficients at different scales and the final approximation coefficients.

[0016] Normalize the detail coefficients of each layer and the final approximation coefficients;

[0017] The normalized detail coefficients of each layer and the final approximation coefficients are inversely transformed to reconstruct the signal data back to the time domain.

[0018] Preferably, the logic of detecting abnormal frequencies through spectrum analysis is:

[0019] Use window function to reduce spectrum leakage, apply Fourier transform to the preprocessed signal to obtain frequency domain signal;

[0020] Calculate the amplitude of different frequency components. The calculation expression is: Where |X(f)| is the amplitude of the frequency component, Re(X(f)) is the real part of the frequency domain signal obtained by Fourier transform, and Im(X(f)) is the imaginary part of the frequency domain signal obtained by Fourier transform;

[0021] By calculating the amplitude of different frequency components, the frequency components are detected, including the fundamental frequency components, harmonic components and interference components;

[0022] Set the detection threshold, expressed as F q =N q *F e , where F q is the detection threshold, q represents the fault type, N q is the preset magnification for different fault types, F e is the fundamental frequency of the multi-parameter electrical measuring instrument, F qos is the preset threshold corresponding to different fault types. When the calculated F q Greater than or equal to F qos When , the multi-parameter electrical measuring instrument is marked as a critical state of type q fault;

[0023] When the calculated F q Less than F qos When the multi-parameter electrical measuring instrument is marked as a Class Q fault-safe state.

[0024] Preferably, the analysis method for detecting nonlinear distortion by harmonic analysis is:

[0025] After using spectrum analysis to detect abnormal frequencies, the harmonic components are extracted and the total harmonic distortion of the signal is calculated. The calculation expression is: Where X(f) is the amplitude of the frequency domain signal at frequency f, f0 is the fundamental frequency, n is the harmonic order, n≥2 and n is a positive integer, and N is the maximum harmonic order;

[0026] The total harmonic distortion thresholds of different harmonic orders are preset. When the calculated total harmonic distortion of the nth harmonic exceeds the threshold, the multi-parameter electrical measuring instrument is marked as the critical state of the nth harmonic fault.

[0027] When the calculated total harmonic distortion of the nth harmonic does not exceed the threshold value, the multi-parameter electrical measuring instrument is marked as the nth harmonic fault safe state.

[0028] Preferably, the method for diagnosis and identification using a neural network algorithm is:

[0029] The time domain signal in the electrical parameter data is converted into a time-frequency diagram through short-time Fourier transform or wavelet transform, a spectrum diagram is generated through fast Fourier transform, and the frequency characteristics and harmonic components of the signal are extracted;

[0030] Enhance the extracted data by adding random noise, shifting and scaling in time, scaling the amplitude, and processing the spectrum;

[0031] Construct a convolutional neural network model and input the spectrum graph and time-frequency graph into the convolutional neural network model. Use the ReLU activation function to increase the nonlinear expression ability of the model. Use the Softmax layer for multi-classification to output the fault mode and the corresponding probability of the fault mode.

[0032] Use the cross entropy loss function to measure the difference between the predicted value and the actual value, use the optimization algorithm to update the model parameters, use the hyperparameter tuning method to optimize the hyperparameters of the convolutional neural network model, and use cross-validation to evaluate the generalization ability of the model;

[0033] The model performance is tested by using data that is not input into the convolutional neural network model, and the classification effect of the model on different faults is analyzed by plotting the confusion matrix until the model reaches the convergence limit;

[0034] Use the converged convolutional neural network model to monitor and diagnose real-time data and generate diagnostic results.

[0035] Preferably, the logic for generating a fault report based on the diagnosis rule and the diagnosis result of the neural network algorithm is:

[0036] The first diagnostic threshold D1 and the second diagnostic threshold D2 are preset, and the first diagnostic threshold D1 is less than the second diagnostic threshold D2. If the rule initial diagnosis module marks the multi-parameter electrical measuring instrument as a critical fault state, when the calculated frequency interference fault probability P f The fault probability S is greater than or equal to the second diagnostic threshold D2 and has nonlinear distortion t When it is less than or equal to the first diagnostic threshold D1, the self-diagnosis system outputs that the multi-parameter electrical measuring instrument has a fault caused by frequency interference;

[0037] If the rule initial diagnosis module marks the multi-parameter electrical measuring instrument as a critical fault state, when the calculated nonlinear distortion fault probability S t The fault probability P is greater than or equal to the second diagnostic threshold D2 and the frequency interference f When it is less than or equal to the first diagnostic threshold D1, the self-diagnosis system outputs that the multi-parameter electrical measuring instrument has a fault caused by nonlinear distortion;

[0038] If the rule initial diagnosis module marks the multi-parameter electrical measuring instrument as a critical fault state, when the calculated first diagnosis threshold D1 is less than the fault probability S of the nonlinear distortion t is less than the second diagnostic threshold D2, and the first diagnostic threshold D1 is less than the fault probability P of the frequency interference f When it is less than the second diagnosis threshold D2, the self-diagnosis system outputs that the multi-parameter electrical measuring instrument has a fault caused by the association between frequency interference and nonlinear distortion.

[0039] Preferably, the logic of self-diagnosis decision output according to the fault report is:

[0040] When the self-diagnosis system outputs a fault caused by frequency interference in the multi-parameter electrical measuring instrument, the fault information is recorded to remind the staff of the frequency interference, inform them to shield the interference source and adjust the wiring;

[0041] If the self-diagnosis system outputs a fault caused by nonlinear distortion in the multi-parameter electrical measuring instrument, the fault information will be recorded to remind the staff of the nonlinear distortion, inform them to check the nonlinear components and reduce the load;

[0042] If the self-diagnosis system outputs a fault caused by frequency interference and nonlinear distortion in the multi-parameter electrical measuring instrument, the fault information will be recorded, the spectrum analysis and harmonic analysis results will be reported, and the staff will be prompted to conduct inspections.

[0043] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0044] The present invention establishes a preliminary diagnosis model through preset rules, uses spectrum analysis and harmonic analysis to screen and verify the electrical parameter signals of the multi-parameter electrical measuring instrument, marks and identifies the analysis data with fault risks, converts the signal data into image form through Fourier transform and wavelet transform, and inputs it into the convolutional neural network model, identifies and distinguishes the abnormal fault types according to the analysis results of the convolutional neural network model, and optimizes the self-diagnosis of the multi-parameter electrical measuring instrument by combining the analysis results based on the rule model and the neural network model. While ensuring the accuracy of self-diagnosis, the flexibility of the self-diagnosis system for data analysis under different environmental conditions is improved, the coverage of the self-diagnosis system for abnormal faults is further improved, the self-diagnosis workload of the staff is effectively reduced, the problem of manually adjusting the rule model and manually inputting feature information when the environmental conditions change is avoided, and the diagnosis accuracy of frequency interference and nonlinear distortion entanglement faults is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0046] Figure 1 It is a system module diagram of the present invention. DETAILED DESCRIPTION

[0047] 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.

[0048] Example 1: Please refer to Figure 1 As shown, the present invention is a multi-parameter electric measuring instrument self-diagnosis system based on data analysis, including a parameter acquisition module, a rule initial diagnosis module, an algorithm re-diagnosis module, and a comprehensive judgment module;

[0049] The parameter acquisition module is used to collect the electrical parameter data of the multi-parameter electrical measuring instrument, and transmit the electrical parameter data to the rule initial diagnosis module after pre-processing;

[0050] The rule-based initial diagnosis module is used to diagnose abnormal faults based on preset diagnostic rules, detect abnormal frequencies through spectrum analysis, and detect nonlinear distortion through harmonic analysis;

[0051] The algorithm re-diagnosis module is used to train the neural network model based on the electrical parameter data of abnormal faults, and use the neural network algorithm to diagnose and identify frequency interference and nonlinear distortion faults;

[0052] The comprehensive judgment module is used to judge abnormal faults and generate fault reports by combining the diagnosis results based on the diagnosis rules and neural network algorithms;

[0053] The decision implementation module is used to output self-diagnosis decisions based on fault reports, prompting staff to manage and maintain abnormal faults.

[0054] The self-diagnosis of multi-parameter electrical measuring instruments is that the instrument can automatically perform self-checking and testing during use to ensure its normal function and measurement accuracy. The self-diagnosis function identifies potential faults or performance problems of the instrument through a series of internal tests, inspections and calibration procedures, and takes corresponding measures to maintain the reliability and accuracy of the instrument;

[0055] The multi-parameter electrical measuring instrument self-diagnosis system ensures the stability and reliability of the instrument in long-term use and reduces the occurrence of equipment failures by real-time monitoring and detection of potential problems. It performs regular self-inspection and calibration to ensure that the instrument's measurement results are accurate and improve the credibility of testing and analysis. Self-diagnosis detects problems at an early stage and can avoid major equipment failures and downtime, thereby reducing repair and maintenance costs. It detects and handles problems in a timely manner to avoid continued damage to the equipment and extend the service life of the instrument. Through automatic diagnosis and alarm, it reduces the burden of manual inspection on staff and improves the ease of use and operational safety of the equipment.

[0056] Self-diagnosis systems based solely on statistical analysis and preset rules are difficult to solve the diagnosis problem of complex abnormal patterns and behaviors in multi-parameter electrical measuring instruments. Taking the composite fault of specific frequency interference and nonlinear distortion as an example, the fault manifestation is nonlinear. The nonlinear distortion or nonlinear noise of multi-parameter electrical measuring instruments is not easy to be detected by self-diagnosis systems based on linear models or simple rules. Complex fault patterns involve changes in time series data, such as periodic equipment failures or gradually developing aging phenomena of multi-parameter electrical measuring instruments. There are dynamic changes in time series, and it is challenging to capture and judge complex time series features. The data generated by multi-parameter electrical measuring instruments usually have high-dimensional features. When monitoring multiple electrical parameters at the same time, anomalies will occur simultaneously in different dimensions. The linear model used by the self-diagnosis system based on statistical analysis and preset rules has poor processing effect on complex nonlinear and high-dimensional data. Features that rely on manual selection and extraction are not enough to capture complex abnormal patterns. For complex abnormal patterns, manually designed features are difficult to fully cover various manifestations of abnormalities. When processing time series data, it may not be possible to fully capture the dynamic changes and long-term dependencies of data. The preset rules are dependent on the environment and working conditions. When the environmental conditions or working conditions change, the preset rules are difficult to adjust in time.

[0057] The parameter acquisition module is used to collect the electrical parameter data of the multi-parameter electrical measuring instrument, and transmit the electrical parameter data to the rule initial diagnosis module after pre-processing;

[0058] The electrical parameter data of the multi-parameter electrical measuring instrument include voltage, current, frequency information, harmonic components, time domain information, nonlinear information and physical information. The frequency information includes the fundamental frequency, frequency fluctuation and interference frequency of the multi-parameter electrical measuring instrument. The harmonic components include harmonic amplitude, total harmonic distortion and harmonic content. The time domain information includes signal peak value, signal effective value and waveform factor. The nonlinear information includes phase characteristics, Lyapunov exponent and approximate entropy. The physical information includes temperature and humidity of the multi-parameter electrical measuring instrument.

[0059] Preprocess the electrical parameter data of multi-parameter electrical measuring instruments. Data preprocessing can optimize the diagnostic effect of complex abnormal fault modes and ensure that the neural network model can accurately and efficiently process and analyze electrical parameter data. The preprocessing process includes denoising, signal smoothing and normalization. The preprocessing method is:

[0060] Use filters to remove noise and smooth the data, and normalize the data by combining wavelet transform;

[0061] The method of normalizing data by combining wavelet transform is:

[0062] The wavelet basis function is used to decompose the data into detail coefficients and approximation coefficients at different scales. The detail coefficients are high-frequency components, and the approximation coefficients are low-frequency components. The approximation coefficients are decomposed at multiple levels to obtain detail coefficients at different scales and the final approximation coefficients.

[0063] Normalize the detail coefficients of each layer and the final approximation coefficients;

[0064] The normalized detail coefficients of each layer and the final approximation coefficients are inversely transformed to reconstruct the signal data back to the time domain.

[0065] It should be noted that a low-pass filter is used to remove high-frequency noise, a band-pass filter is used to remove noise in a specific frequency band, the signal components in a specific frequency band are retained, and a notch filter is used to remove noise at a specific frequency. Common signal smoothing methods include moving average, exponential smoothing, etc.

[0066] Methods for normalizing detail coefficients and approximation coefficients include minimum-maximum normalization, standardization, and the like;

[0067] Normalizing data by combining wavelet transform can retain the time and frequency characteristics of the signal at the same time, and is suitable for processing non-stationary signals. Because wavelet transform has the characteristics of multi-resolution analysis, it can effectively analyze the characteristics of the signal at different time scales. Wavelet transform decomposes the signal into wavelet coefficients at different scales, and analyzes the changes of the signal at different frequency components through wavelet coefficients. Wavelet normalization uses wavelet coefficients to standardize the signal, so that the characteristics of the signal at different scales can be retained, while improving the robustness and consistency of the signal.

[0068] Wavelet normalization retains the characteristics of signals at different scales and is suitable for processing signals with different frequency components. Through normalization processing, the amplitude difference of the signal can be reduced and the stability and accuracy of subsequent analysis can be improved. For the analysis of non-stationary signals, wavelet normalization can process complex time-varying signals. In the electrical parameter data of multi-parameter electrical measuring instruments, the signal may contain various frequency components and noise. Through wavelet normalization, the characteristics and fault modes of the signal can be better analyzed.

[0069] The rule-based initial diagnosis module is used to diagnose abnormal faults based on preset diagnostic rules, detect abnormal frequencies through spectrum analysis, and detect nonlinear distortion through harmonic analysis;

[0070] Detect abnormal frequencies through spectrum analysis, use window functions to reduce spectrum leakage, and apply Fourier transform to the preprocessed signal to obtain frequency domain signals;

[0071] Calculate the amplitude of different frequency components. The calculation expression is: Where |X(f)| is the amplitude of the frequency component, Re(X(f)) is the real part of the frequency domain signal obtained by Fourier transform, and Im(X(f)) is the imaginary part of the frequency domain signal obtained by Fourier transform;

[0072] By calculating the amplitude of different frequency components, the frequency components are detected, including the fundamental frequency components, harmonic components and interference components;

[0073] Set the detection threshold, expressed as F q =N q *F e , where F q is the detection threshold, q represents the fault type, N q is the preset magnification for different fault types, F e is the fundamental frequency of the multi-parameter electrical measuring instrument, F qos is the preset threshold corresponding to different fault types. When the calculated F q Greater than or equal to F qos When , the multi-parameter electrical measuring instrument is marked as a critical state of type q fault;

[0074] When the calculated F q Less than F qos When the multi-parameter electrical measuring instrument is marked as a Class Q fault-safe state.

[0075] It should be noted that commonly used window functions include Hanning window, Hamming window, etc.

[0076] After using spectrum analysis to detect abnormal frequencies, the harmonic components are extracted and nonlinear distortion is detected by harmonic analysis. The analysis method is as follows:

[0077] Calculate the total harmonic distortion of the signal. The calculation expression is: Where X(f) is the amplitude of the frequency domain signal at frequency f, f0 is the fundamental frequency, n is the harmonic order, n≥2 and n is a positive integer, and N is the maximum harmonic order;

[0078] The total harmonic distortion thresholds of different harmonic orders are preset. When the calculated total harmonic distortion of the nth harmonic exceeds the threshold, the multi-parameter electrical measuring instrument is marked as the critical state of the nth harmonic fault.

[0079] When the calculated total harmonic distortion of the nth harmonic does not exceed the threshold value, the multi-parameter electrical measuring instrument is marked as the nth harmonic fault safe state.

[0080] Nonlinear distortion usually appears as harmonics of the fundamental frequency. Harmonics are frequency components that are integer multiples of the fundamental frequency. Harmonic analysis is used to identify the harmonic components in the signal, thereby determining the source of the nonlinear distortion. For example, rectifiers, frequency converters, and nonlinear loads in power systems often produce significant harmonic distortion. Harmonic analysis can quantify the degree of nonlinear distortion in the signal. By calculating the total harmonic distortion, the total strength of the harmonic components relative to the fundamental frequency components can be obtained, which helps to evaluate the level of nonlinear distortion of the system or equipment and facilitate the adoption of corresponding corrective measures. The harmonic components generated during the operation of multi-parameter electrical measuring instruments can reflect the health status of the multi-parameter electrical measuring instruments. An increase in the level of nonlinear distortion may indicate equipment failure or abnormal operation. For example, an increase in the harmonic distortion of a motor may indicate aging or mechanical damage to the motor coil. Harmonic analysis can be used to monitor the status of the equipment in real time and to promptly discover and deal with potential problems.

[0081] The algorithm re-diagnosis module is used to train the neural network model based on the electrical parameter data of abnormal faults, and use the neural network algorithm to diagnose and identify frequency interference and nonlinear distortion faults;

[0082] The method of using neural network algorithm for diagnosis and identification is:

[0083] The time domain signal in the electrical parameter data is converted into a time-frequency diagram through short-time Fourier transform or wavelet transform, a spectrum diagram is generated through fast Fourier transform, and the frequency characteristics and harmonic components of the signal are extracted;

[0084] Enhance the extracted data by adding random noise, shifting and scaling in time, scaling the amplitude, and processing the spectrum;

[0085] Construct a convolutional neural network model and input the spectrum graph and time-frequency graph into the convolutional neural network model. Use the ReLU activation function to increase the nonlinear expression ability of the model. Use the Softmax layer for multi-classification to output the fault mode and the corresponding probability of the fault mode.

[0086] Use the cross entropy loss function to measure the difference between the predicted value and the actual value, use the optimization algorithm to update the model parameters, use the hyperparameter tuning method to optimize the hyperparameters of the convolutional neural network model, and use cross-validation to evaluate the generalization ability of the model;

[0087] The model performance is tested by using data that is not input into the convolutional neural network model, and the classification effect of the model on different faults is analyzed by plotting the confusion matrix until the model reaches the convergence limit;

[0088] Use the converged convolutional neural network model to monitor and diagnose real-time data and generate diagnostic results.

[0089] It should be pointed out that random noise is added to simulate measurement errors in different environments. Random noise includes Gaussian noise, uniform noise, etc. The signal is shifted and scaled in time to simulate the time delay or advance of the signal and change the time length of the signal. The time shift includes forward shift, reverse shift, etc. The time scaling includes compression and expansion. The amplitude of the signal is scaled to simulate signals of different strengths. Scaling includes amplification and reduction. The spectrum of the signal is processed to enhance the frequency domain characteristics. The processing methods include frequency shift and frequency stretching.

[0090] Commonly used optimization algorithms include Adam algorithm, SGD algorithm, etc.

[0091] Data enhancement can generate diverse training samples, thereby increasing the diversity of the data set. In order for the model to perform well on different types of input data, increasing the diversity of data can help the model learn more features and patterns and reduce its dependence on specific data. The generalization ability of a model refers to its ability to maintain good performance on unseen data. Data enhancement can simulate various possible real-world scenarios to enable the model to better adapt to and cope with different input data, thereby improving the generalization ability of the model and avoiding overfitting. Data enhancement can simulate various interferences and uncertainties in the data by introducing noise, transformations, and other disturbances, making the model more robust in the face of noise and uncertainty, reducing its dependence on specific data patterns, and improving the stability and reliability of the model.

[0092] It should be noted that the time-frequency diagram is a graph that visually represents the frequency components of a signal at different time points. It can combine information in the time domain and frequency domain to show the frequency characteristics of a signal that changes over time. The time-frequency diagram can provide frequency information of a signal at different time scales, and is very effective for analyzing signals with different time scale characteristics. Many actual signals are non-stationary, that is, the frequency components of a signal change over time. The time-frequency diagram can display this change, which is difficult to do with traditional time domain or frequency domain analysis. In signal processing and pattern recognition, the time-frequency diagram can assist in identifying and extracting signal features, including signal periodicity, mutations, frequency modulation, etc.

[0093] The comprehensive judgment module is used to judge abnormal faults and generate fault reports by combining the diagnosis results based on the diagnosis rules and neural network algorithms;

[0094] The first diagnostic threshold D1 and the second diagnostic threshold D2 are preset, and the first diagnostic threshold D1 is less than the second diagnostic threshold D2. If the rule initial diagnosis module marks the multi-parameter electrical measuring instrument as a critical fault state, when the calculated frequency interference fault probability P f The fault probability S is greater than or equal to the second diagnostic threshold D2 and has nonlinear distortion tWhen it is less than or equal to the first diagnostic threshold D1, the self-diagnosis system outputs that the multi-parameter electrical measuring instrument has a fault caused by frequency interference;

[0095] If the rule initial diagnosis module marks the multi-parameter electrical measuring instrument as a critical fault state, when the calculated nonlinear distortion fault probability S t The fault probability P is greater than or equal to the second diagnostic threshold D2 and the frequency interference f When it is less than or equal to the first diagnostic threshold D1, the self-diagnosis system outputs that the multi-parameter electrical measuring instrument has a fault caused by nonlinear distortion;

[0096] If the rule initial diagnosis module marks the multi-parameter electrical measuring instrument as a critical fault state, when the calculated first diagnosis threshold D1 is less than the fault probability S of the nonlinear distortion t is less than the second diagnostic threshold D2, and the first diagnostic threshold D1 is less than the fault probability P of the frequency interference f When it is less than the second diagnosis threshold D2, the self-diagnosis system outputs that the multi-parameter electrical measuring instrument has a fault caused by the association between frequency interference and nonlinear distortion.

[0097] The decision implementation module is used to output self-diagnosis decisions based on fault reports, prompting staff to manage and maintain abnormal faults;

[0098] When the self-diagnosis system outputs a fault caused by frequency interference in the multi-parameter electrical measuring instrument, the fault information is recorded to remind the staff of the frequency interference, inform them to shield the interference source and adjust the wiring;

[0099] If the self-diagnosis system outputs a fault caused by nonlinear distortion in the multi-parameter electrical measuring instrument, the fault information will be recorded to remind the staff of the nonlinear distortion, inform them to check the nonlinear components and reduce the load;

[0100] If the self-diagnosis system outputs a fault caused by frequency interference and nonlinear distortion in the multi-parameter electrical measuring instrument, the fault information will be recorded, the spectrum analysis and harmonic analysis results will be reported, and the staff will be prompted to conduct inspections.

[0101] The present invention establishes a preliminary diagnosis model through preset rules, uses spectrum analysis and harmonic analysis to screen and verify the electrical parameter signals of the multi-parameter electrical measuring instrument, marks and identifies the analysis data with fault risks, converts the signal data into image form through Fourier transform and wavelet transform, and inputs it into the convolutional neural network model, identifies and distinguishes the abnormal fault types according to the analysis results of the convolutional neural network model, and optimizes the self-diagnosis of the multi-parameter electrical measuring instrument by combining the analysis results based on the rule model and the neural network model. While ensuring the accuracy of self-diagnosis, the flexibility of the self-diagnosis system for data analysis under different environmental conditions is improved, the coverage of the self-diagnosis system for abnormal faults is further improved, the self-diagnosis workload of the staff is effectively reduced, the problem of manually adjusting the rule model and manually inputting feature information when the environmental conditions change is avoided, and the diagnosis accuracy of frequency interference and nonlinear distortion entanglement faults is improved.

[0102] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0103] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments can be implemented in whole or in part in the form of computer program goods. The computer program goods include one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website site, computer, server or data center to another website site, computer, server or data center by wired or wireless (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.

[0104] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0105] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the method described above can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.

Claims

1. A multi-parameter electrical measuring instrument self-diagnosis system based on data analysis, characterized in that: It includes parameter acquisition module, rule initial diagnosis module, algorithm re-diagnosis module and comprehensive judgment module; The parameter acquisition module is used to collect the electrical parameter data of the multi-parameter electrical measuring instrument, and transmit the electrical parameter data to the rule initial diagnosis module after pre-processing; The rule-based initial diagnosis module is used to diagnose abnormal faults based on preset diagnostic rules, detect abnormal frequencies through spectrum analysis, and detect nonlinear distortion through harmonic analysis; The algorithm re-diagnosis module is used to train the neural network model based on the electrical parameter data of abnormal faults, and use the neural network algorithm to diagnose and identify frequency interference and nonlinear distortion faults; The comprehensive judgment module is used to judge abnormal faults and generate fault reports by combining the diagnosis results based on the diagnosis rules and neural network algorithms; The decision implementation module is used to output self-diagnosis decisions based on fault reports, prompting staff to manage and maintain abnormal faults.

2. The multi-parameter electrical measuring instrument self-diagnosis system based on data analysis according to claim 1 is characterized in that: The electrical parameter data of the multi-parameter electrical measuring instrument include voltage, current, frequency information, harmonic components, time domain information, nonlinear information and physical information. The frequency information includes the fundamental frequency, frequency fluctuation and interference frequency of the multi-parameter electrical measuring instrument. The harmonic components include harmonic amplitude, total harmonic distortion and harmonic content. The time domain information includes signal peak value, signal effective value and waveform factor. The nonlinear information includes phase characteristics, Lyapunov exponent and approximate entropy. The physical information includes temperature and humidity of the multi-parameter electrical measuring instrument.

3. The multi-parameter electrical measuring instrument self-diagnosis system based on data analysis according to claim 2 is characterized in that: The preprocessing method is: Use filters to remove noise and smooth the data, and normalize the data by combining wavelet transform; The method of normalizing data by combining wavelet transform is: The wavelet basis function is used to decompose the data into detail coefficients and approximation coefficients at different scales. The detail coefficients are high-frequency components, and the approximation coefficients are low-frequency components. The approximation coefficients are decomposed at multiple levels to obtain detail coefficients at different scales and the final approximation coefficients. Normalize the detail coefficients of each layer and the final approximation coefficients; The normalized detail coefficients of each layer and the final approximation coefficients are inversely transformed to reconstruct the signal data back to the time domain.

4. The multi-parameter electrical measuring instrument self-diagnosis system based on data analysis according to claim 1 is characterized in that: The logic of detecting abnormal frequencies through spectrum analysis is: Use window function to reduce spectrum leakage, apply Fourier transform to the preprocessed signal to obtain frequency domain signal; Calculate the amplitude of different frequency components. The calculation expression is: Where |X(f)| is the amplitude of the frequency component, Re(X(f)) is the real part of the frequency domain signal obtained by Fourier transform, and Im(X(f)) is the imaginary part of the frequency domain signal obtained by Fourier transform; By calculating the amplitude of different frequency components, the frequency components are detected, including the fundamental frequency components, harmonic components and interference components; Set the detection threshold, expressed as F q =N q *F e , where F q is the detection threshold, q represents the fault type, N q is the preset magnification for different fault types, F e is the fundamental frequency of the multi-parameter electrical measuring instrument, F qos is the preset threshold corresponding to different fault types. q Greater than or equal to F qos When , the multi-parameter electrical measuring instrument is marked as a critical state of type q fault; When the calculated F q Less than F qos When the multi-parameter electrical measuring instrument is marked as a Class Q fault-safe state.

5. The multi-parameter electric measuring instrument self-diagnosis system based on data analysis according to claim 4 is characterized in that: The analytical method for detecting nonlinear distortion through harmonic analysis is: After using spectrum analysis to detect abnormal frequencies, the harmonic components are extracted and the total harmonic distortion of the signal is calculated. The calculation expression is: Where X(f) is the amplitude of the frequency domain signal at frequency f, f0 is the fundamental frequency, n is the harmonic order, n≥2 and n is a positive integer, and N is the maximum harmonic order; The total harmonic distortion thresholds of different harmonic orders are preset. When the calculated total harmonic distortion of the nth harmonic exceeds the threshold, the multi-parameter electrical measuring instrument is marked as the critical state of the nth harmonic fault. When the calculated total harmonic distortion of the nth harmonic does not exceed the threshold value, the multi-parameter electrical measuring instrument is marked as the nth harmonic fault safe state.

6. The multi-parameter electric measuring instrument self-diagnosis system based on data analysis according to claim 5 is characterized in that: The method of using neural network algorithm for diagnosis and identification is: The time domain signal in the electrical parameter data is converted into a time-frequency diagram through short-time Fourier transform or wavelet transform, a spectrum diagram is generated through fast Fourier transform, and the frequency characteristics and harmonic components of the signal are extracted; Enhance the extracted data by adding random noise, shifting and scaling in time, scaling the amplitude, and processing the spectrum; Construct a convolutional neural network model and input the spectrum graph and time-frequency graph into the convolutional neural network model. Use the ReLU activation function to increase the nonlinear expression ability of the model. Use the Softmax layer for multi-classification to output the fault mode and the corresponding probability of the fault mode. Use the cross entropy loss function to measure the difference between the predicted value and the actual value, use the optimization algorithm to update the model parameters, use the hyperparameter tuning method to optimize the hyperparameters of the convolutional neural network model, and use cross-validation to evaluate the generalization ability of the model; The model performance is tested by using data that is not input into the convolutional neural network model, and the classification effect of the model on different faults is analyzed by plotting the confusion matrix until the model reaches the convergence limit; Use the converged convolutional neural network model to monitor and diagnose real-time data and generate diagnostic results.

7. The multi-parameter electric measuring instrument self-diagnosis system based on data analysis according to claim 6 is characterized in that: The logic of generating fault reports based on the diagnosis results of the diagnostic rules and neural network algorithms is as follows; The first diagnostic threshold D1 and the second diagnostic threshold D2 are preset, and the first diagnostic threshold D1 is less than the second diagnostic threshold D2. If the rule initial diagnosis module marks the multi-parameter electrical measuring instrument as a critical fault state, when the calculated frequency interference fault probability P f The fault probability S is greater than or equal to the second diagnostic threshold D2 and has nonlinear distortion t When it is less than or equal to the first diagnostic threshold D1, the self-diagnosis system outputs that the multi-parameter electrical measuring instrument has a fault caused by frequency interference; If the rule initial diagnosis module marks the multi-parameter electrical measuring instrument as a critical fault state, when the calculated nonlinear distortion fault probability S t The fault probability P is greater than or equal to the second diagnostic threshold D2 and the frequency interference f When it is less than or equal to the first diagnostic threshold D1, the self-diagnosis system outputs that the multi-parameter electrical measuring instrument has a fault caused by nonlinear distortion; If the rule initial diagnosis module marks the multi-parameter electrical measuring instrument as a critical fault state, when the calculated first diagnosis threshold D1 is less than the fault probability S of the nonlinear distortion t is less than the second diagnostic threshold D2, and the first diagnostic threshold D1 is less than the fault probability P of the frequency interference f When it is less than the second diagnosis threshold D2, the self-diagnosis system outputs that the multi-parameter electrical measuring instrument has a fault caused by the association between frequency interference and nonlinear distortion.

8. The multi-parameter electric measuring instrument self-diagnosis system based on data analysis according to claim 7 is characterized in that: The logic of self-diagnosis decision output based on fault report is: When the self-diagnosis system outputs a fault caused by frequency interference in the multi-parameter electrical measuring instrument, the fault information is recorded to remind the staff of the frequency interference, inform them to shield the interference source and adjust the wiring; If the self-diagnosis system outputs a fault caused by nonlinear distortion in the multi-parameter electrical measuring instrument, the fault information will be recorded to remind the staff of the nonlinear distortion, inform them to check the nonlinear components and reduce the load; If the self-diagnosis system outputs a fault caused by frequency interference and nonlinear distortion in the multi-parameter electrical measuring instrument, the fault information will be recorded, the spectrum analysis and harmonic analysis results will be reported, and the staff will be prompted to conduct inspections.

Images