CN114881157A - Method, device and equipment for detecting working state of converter valve and storage medium - Google Patents

Method, device and equipment for detecting working state of converter valve and storage medium Download PDF

Info

Publication number
CN114881157A
CN114881157A CN202210534846.8A CN202210534846A CN114881157A CN 114881157 A CN114881157 A CN 114881157A CN 202210534846 A CN202210534846 A CN 202210534846A CN 114881157 A CN114881157 A CN 114881157A
Authority
CN
China
Prior art keywords
data
reconstruction
detected
integrated
converter valve
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202210534846.8A
Other languages
Chinese (zh)
Inventor
秦金锋
严伟
王清君
洪乐洲
胡忠山
黄家豪
王蒙
赵晓杰
蔡斌
叶志良
孔玮琦
王越章
许浩强
钟鑫林
张先亮
张镇
袁海
赵明
李金安
张朝斌
李凯协
张博
王国权
周逸帆
石延辉
杨洋
陈政轩
陈朋辉
黄润烽
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Guangzhou Bureau of Extra High Voltage Power Transmission Co
Original Assignee
Guangzhou Bureau of Extra High Voltage Power Transmission Co
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Guangzhou Bureau of Extra High Voltage Power Transmission Co filed Critical Guangzhou Bureau of Extra High Voltage Power Transmission Co
Priority to CN202210534846.8A priority Critical patent/CN114881157A/en
Publication of CN114881157A publication Critical patent/CN114881157A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
    • G06F18/2135—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on approximation criteria, e.g. principal component analysis
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
    • G06Q50/06—Electricity, gas or water supply
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
    • Y02E60/00—Enabling technologies; Technologies with a potential or indirect contribution to GHG emissions mitigation
    • Y02E60/60—Arrangements for transfer of electric power between AC networks or generators via a high voltage DC link [HVCD]

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • General Physics & Mathematics (AREA)
  • Business, Economics & Management (AREA)
  • Evolutionary Computation (AREA)
  • Artificial Intelligence (AREA)
  • General Engineering & Computer Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Economics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • General Health & Medical Sciences (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Strategic Management (AREA)
  • Human Resources & Organizations (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Evolutionary Biology (AREA)
  • General Business, Economics & Management (AREA)
  • Marketing (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Molecular Biology (AREA)
  • Computational Linguistics (AREA)
  • Tourism & Hospitality (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computing Systems (AREA)
  • Primary Health Care (AREA)
  • Water Supply & Treatment (AREA)
  • Public Health (AREA)
  • Probability & Statistics with Applications (AREA)
  • Development Economics (AREA)
  • Game Theory and Decision Science (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Testing And Monitoring For Control Systems (AREA)

Abstract

The application provides a method, a device, equipment and a storage medium for detecting the working state of a converter valve, which can automatically detect whether the converter valve is in an abnormal working state or not, and comprises the following steps: obtaining reconstruction errors corresponding to each type of data used for detecting the working state of the converter valve and corresponding reconstruction data, and determining the probability that the type of data is abnormal data based on whether the reconstruction errors corresponding to the type of data are in a loss interval corresponding to the type of data; obtaining reconstruction errors corresponding to the integrated reconstruction data based on the integrated reconstruction data corresponding to the various reconstruction data and the corresponding reconstruction data, and determining the probability that the integrated reconstruction data is abnormal data based on whether the reconstruction errors corresponding to the integrated reconstruction data are in a loss interval corresponding to the integrated reconstruction data; and determining whether the converter valve is in an abnormal working state or not based on the result of weighted summation of the probabilities.

Description

Method, device and equipment for detecting working state of converter valve and storage medium
Technical Field
The application relates to the technical field of state monitoring of power grid power transmission and transformation equipment, in particular to a method and a device for detecting the working state of a converter valve, computer equipment, a storage medium and a computer program product.
Background
The converter valve is used as one of power grid power transmission and transformation equipment, detects whether the converter valve is in an abnormal working state or not in real time, and has important significance for normal operation of electric power. In a traditional detection mode of the working state of the converter valve, an inspector can monitor the surface temperature of the converter valve in real time through an online thermal imager temperature measurement system so as to determine whether the converter valve is in an abnormal working state. The detection mode needs manual participation and is low in efficiency.
Disclosure of Invention
In view of the above, it is necessary to provide a method, an apparatus, a computer device, a storage medium and a computer program product for detecting the operating state of a converter valve.
A method for detecting an operating condition of a converter valve, the method comprising:
taking the multiple types of data of the converter valve in a normal state as multiple types of normal data, respectively training data reconstruction modules corresponding to the multiple types of normal data by using the multiple types of normal data, and recording losses generated in the training process to obtain loss intervals corresponding to the multiple types of normal data;
regarding to a converter valve to be detected with a working state to be detected, taking multiple types of data of the converter valve to be detected as multiple types of data to be detected, and respectively inputting each type of data to be detected of the multiple types of data to be detected into a corresponding data reconstruction module so that the data reconstruction module reconstructs the input data to be detected to obtain corresponding reconstructed data;
aiming at each type of data to be detected of the multiple types of data to be detected, obtaining a reconstruction error corresponding to the type of data to be detected based on the type of data to be detected and corresponding reconstruction data, and determining the probability that the type of data to be detected is abnormal data based on whether the reconstruction error corresponding to the type of data to be detected is in a loss interval corresponding to the type of data to be detected;
integrating the reconstruction data obtained by the various data reconstruction modules to obtain integrated reconstruction data, and inputting the integrated reconstruction data into an integration module so that the integration module reconstructs the integrated reconstruction data to obtain corresponding reconstruction data;
obtaining a reconstruction error corresponding to the integrated reconstruction data based on the integrated reconstruction data and the corresponding reconstruction data, obtaining a loss interval corresponding to the integrated reconstruction data based on a loss generated in a training process of an integrated module corresponding to the integrated reconstruction data, and determining the probability that the integrated reconstruction data is abnormal data based on whether the reconstruction error corresponding to the integrated reconstruction data is in the loss interval corresponding to the integrated reconstruction data;
and determining whether the converter valve to be detected is in an abnormal working state or not based on the result of weighted summation of all the probabilities.
A device for detecting the operating condition of a converter valve, said device comprising:
the training unit is used for taking the various types of data of the converter valve in a normal state as various types of normal data, respectively training the data reconstruction modules corresponding to the various types of data by using the various types of normal data, and recording the loss generated in the training process to obtain the loss intervals corresponding to the various types of data;
the data classification reconstruction unit is used for regarding a converter valve to be detected in a working state, taking multiple types of data of the converter valve to be detected as multiple types of data to be detected, and respectively inputting each type of data to be detected of the multiple types of data to be detected into a corresponding data reconstruction module so that the data reconstruction module reconstructs the input data to be detected to obtain corresponding reconstructed data;
the data classification judging unit is used for obtaining a reconstruction error corresponding to the class of data to be detected based on the class of data to be detected and corresponding reconstruction data aiming at each class of data to be detected of the multiple classes of data to be detected, and determining the probability that the class of data to be detected is abnormal based on whether the reconstruction error corresponding to the class of data to be detected is in a loss interval corresponding to the class;
the data integration reconstruction unit is used for integrating reconstruction data obtained by various data reconstruction modules to obtain integrated reconstruction data and inputting the integrated reconstruction data into the integration module so that the integration module reconstructs the integrated reconstruction data to obtain corresponding reconstruction data;
an integrated data discrimination unit, configured to obtain a reconstruction error corresponding to the integrated reconstruction data based on the integrated reconstruction data and corresponding reconstruction data, obtain a loss interval corresponding to the integrated reconstruction data based on a loss generated in a training process of an integrated module corresponding to the integrated reconstruction data, and determine a probability that the integrated reconstruction data is abnormal data based on whether the reconstruction error corresponding to the integrated reconstruction data is in the loss interval corresponding to the integrated reconstruction data;
and the working state detection unit is used for determining whether the converter valve to be detected is in an abnormal working state or not based on the result of weighted summation of the probabilities.
A computer device comprising a memory storing a computer program and a processor performing the above method.
A computer-readable storage medium, on which a computer program is stored, which computer program is executed by a processor for performing the above-mentioned method.
A computer program product having a computer program stored thereon, the computer program being for execution by a processor of the above method.
In the method, the device, the computer equipment, the storage medium and the computer program product for detecting the working state of the converter valve, the various types of data of the converter valve in the normal state are used as various types of normal data, various types of normal data are respectively utilized to train data reconstruction modules corresponding to various types, and loss generated in the training process is recorded to obtain loss intervals corresponding to various types; regarding to a converter valve to be detected with a working state to be detected, taking multiple types of data of the converter valve to be detected as multiple types of data to be detected, and respectively inputting each type of data to be detected of the multiple types of data to be detected into a corresponding data reconstruction module so that the data reconstruction module reconstructs the input data to be detected to obtain corresponding reconstructed data; aiming at each type of data to be detected of the multiple types of data to be detected, obtaining a reconstruction error corresponding to the type of data to be detected based on the type of data to be detected and corresponding reconstruction data, and determining the probability that the type of data to be detected is abnormal data based on whether the reconstruction error corresponding to the type of data to be detected is in a loss interval corresponding to the type of data to be detected; integrating the reconstruction data obtained by the various data reconstruction modules to obtain integrated reconstruction data, and inputting the integrated reconstruction data into an integration module so that the integration module reconstructs the integrated reconstruction data to obtain corresponding reconstruction data; obtaining a reconstruction error corresponding to the integrated reconstruction data based on the integrated reconstruction data and the corresponding reconstruction data, obtaining a loss interval corresponding to the integrated reconstruction data based on a loss generated in a training process of an integrated module corresponding to the integrated reconstruction data, and determining the probability that the integrated reconstruction data is abnormal data based on whether the reconstruction error corresponding to the integrated reconstruction data is in the loss interval corresponding to the integrated reconstruction data; and determining whether the converter valve to be detected is in an abnormal working state or not based on the result of weighted summation of the probabilities. According to the scheme, various data are reconstructed respectively, the probability that the various data are abnormal data is determined based on whether reconstruction errors are in corresponding loss intervals or not, then the various reconstruction data are integrated, the probability that the integrated reconstruction data are abnormal data is determined based on whether reconstruction errors corresponding to the integrated reconstruction data are in the loss intervals corresponding to the integrated reconstruction data or not, whether the converter valve is in an abnormal working state or not is automatically detected based on results obtained by weighting the probabilities, the multi-class data and the integrated data corresponding to the multi-class data are integrated, and the detection accuracy is improved.
Drawings
FIG. 1 is a schematic interface diagram of monitoring based on an online thermal imager temperature measurement system in one embodiment;
FIG. 2 is a schematic flow chart illustrating a method for detecting the operating condition of a converter valve according to one embodiment;
FIG. 3 is an architectural diagram of a detection model in one embodiment;
FIG. 4 is a schematic diagram of a network architecture of a data reconstruction module in one embodiment;
FIG. 5 is a schematic diagram of a network architecture of an integration module in one embodiment;
FIG. 6 is a graphical representation of the results of risk prediction in one embodiment;
fig. 7 is a block diagram showing a configuration of a device for detecting an operating state of a converter valve according to an embodiment;
FIG. 8 is a diagram illustrating an internal structure of a computer device according to an embodiment.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
Reference in the specification to "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the specification. The appearances of the phrase in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments. It is explicitly and implicitly understood by one skilled in the art that the embodiments described herein can be combined with other embodiments.
The inspection personnel can monitor the working state of the converter valve in real time based on the online thermal imager temperature measuring system, and the online thermal imager temperature measuring system can present a detection interface shown in FIG. 1 so that the inspection personnel can check the temperature on the surface of the converter valve and determine which converter valves are in abnormal working states.
In an embodiment, as shown in fig. 2, a method for detecting an operating state of a converter valve is provided, which is described by taking the method as an example applied to a server, and includes the following steps:
and S200, taking the multiple types of data of the converter valve in the normal state as multiple types of normal data, respectively training data reconstruction modules corresponding to the multiple types of normal data by using the multiple types of normal data, and recording the loss generated in the training process to obtain loss intervals corresponding to the multiple types of normal data.
Step S201, regarding to the converter valve to be detected in a working state, regarding the converter valve to be detected, taking the multi-class data of the converter valve to be detected as the multi-class data to be detected, and inputting each class of data to be detected of the multi-class data to be detected into a corresponding data reconstruction module respectively, so that the data reconstruction module reconstructs the input data to be detected to obtain corresponding reconstructed data.
Step S202, aiming at each type of data to be detected of the multiple types of data to be detected, obtaining a reconstruction error corresponding to the type of data to be detected based on the type of data to be detected and corresponding reconstruction data, and determining the probability that the type of data to be detected is abnormal data based on whether the reconstruction error corresponding to the type of data to be detected is in a loss interval corresponding to the type of data to be detected.
Wherein, the converter valve can be the converter valve of straight-flow, and further, all kinds of data include at least: the method comprises the steps of obtaining temperature data of a converter valve to be detected, audio data of a valve hall where the converter valve to be detected is located and working condition data of the converter valve to be detected.
The main noise sources of the converter station include converter valves, cooling equipment for converter valves, converter transformers, air conditioning equipment, and filter devices, among which the noise is the highest in valve halls. It has been found that the sound level in the valve hall of the converter station is up to 105db (a), while the sound level in the main control room adjacent to the valve hall is also up to about 80db (a).
The noise of a valve hall where the converter valve is located is formed quite complicated, the noise source of the converter valve comprises noise caused by electromagnetic stress vibration of a valve reactor besides noise caused by vibration and heating of the valve body, the sound level of the valve hall in a single-pole and double-pole operation mode is more than 100dB (A) and can reach 105dB (A) at most, the noise frequency harmonic distribution of the converter valve is greatly different from the frequency spectrum distribution of common alternating-current equipment such as a transformer, a reactor and the like, the frequency of the highest sound level in the frequency spectrum distribution of the latter is 2 times of the power frequency, and the noise frequency spectrum of the valve hall has larger 1 kHz and 5kHz higher harmonic components.
The measured results also show that the noise level of the converter valves seems to be less sensitive to the change of the DC power level, which phenomenon is also confirmed in the sound level measurement with one DC voltage basic stabilization and DC current step-up commissioning procedure, when the DC voltage is stabilized at 450kV and the current is stepped up from 495A to 1218A, the change of the sound level is only about 1-2 db (a), while the commutation noise is much more sensitive to the change of the voltage, which is a function of the voltage.
Training corresponding process of a data reconstruction module by using training sample data acquired by a converter valve in a normal working state, and recording loss generated in the training process to obtain a corresponding loss interval; taking training a data reconstruction module corresponding to temperature data of the converter valve as an example: when the converter valve is in a normal working state, temperature data of the converter valve is collected and used as training sample data, loss between the temperature data and reconstruction data which are used as the training sample data is recorded in the process of training a data reconstruction module by using the training sample data, the loss is used as loss generated in the training process, and a loss interval corresponding to the temperature data is obtained.
For a data reconstruction module corresponding to the temperature data after the training is finished, a group of temperature data is newly acquired, after the data reconstruction module reconstructs the temperature data, if reconstruction loss between the reconstructed data and the newly acquired temperature data is within the loss interval and is not close to the maximum value of the loss interval, the newly acquired temperature data is considered to have no risk or have low risk, and if reconstruction loss between the reconstructed data and the newly acquired temperature data is within the loss interval or is not close to the maximum value of the loss interval, the newly acquired temperature data is considered to be not in accordance with the 'rule' of normal data acquired in the previous training process, namely the newly acquired temperature data does not appear, the newly acquired temperature data is considered to have risk, and the newly acquired temperature data is required to be checked.
Further, determining the probability that the type of data to be detected is abnormal data based on whether the reconstruction error corresponding to the type of data to be detected is in the loss interval corresponding to the type of data to be detected specifically may include: when the reconstruction error corresponding to the type of data to be detected is in the loss interval corresponding to the type and is not close to the maximum value of the loss interval, taking the probability lower than the threshold value as the probability that the type of data to be detected is abnormal data; and when the reconstruction error corresponding to the type of data to be detected is not in the loss interval corresponding to the type of data or is close to the maximum value of the loss interval, taking the probability higher than the threshold value as the probability that the type of data to be detected is abnormal data.
Step S203, integrating the reconstruction data obtained by the various data reconstruction modules to obtain integrated reconstruction data, and inputting the integrated reconstruction data into an integration module so that the integration module reconstructs the integrated reconstruction data to obtain corresponding reconstruction data.
Step S204, obtaining a reconstruction error corresponding to the integrated reconstruction data based on the integrated reconstruction data and the corresponding reconstruction data, obtaining a loss interval corresponding to the integrated reconstruction data based on a loss generated in a training process of an integrated module corresponding to the integrated reconstruction data, and determining the probability that the integrated reconstruction data is abnormal data based on whether the reconstruction error corresponding to the integrated reconstruction data is in the loss interval corresponding to the integrated reconstruction data.
Further, determining the probability that the integrated reconstruction data is abnormal data based on whether the reconstruction error corresponding to the integrated reconstruction data is in the loss interval corresponding to the integrated reconstruction data may specifically include: when the reconstruction error corresponding to the integrated reconstruction data is in a loss interval corresponding to the integrated reconstruction data and is not close to the maximum value of the loss interval, taking the probability lower than the threshold value as the probability that the integrated reconstruction data is abnormal data; and when the reconstruction error corresponding to the integrated reconstruction data is not in the loss interval or close to the maximum value of the loss interval corresponding to the integrated reconstruction data, taking the probability higher than the threshold value as the probability that the integrated reconstruction data is abnormal data.
And S205, determining whether the converter valve to be detected is in an abnormal working state or not based on the result of weighted summation of the probabilities.
According to the method for detecting the working state of the converter valve, various types of data are reconstructed respectively, the probability that the various types of data are abnormal data is determined based on whether the reconstruction error is in the corresponding loss interval, then the various types of reconstruction data are integrated, the probability that the integrated reconstruction data are abnormal data is determined based on whether the reconstruction error corresponding to the integrated reconstruction data is in the loss interval corresponding to the integrated reconstruction data, whether the converter valve is in the abnormal working state is automatically detected based on the result obtained by weighting the probabilities, the multi-type data and the integrated data corresponding to the multi-type data are integrated, and the detection accuracy is improved.
In one embodiment, various types of data of the converter valve can be acquired based on a network communication mode, the data are used for detecting the working state and transmitting the data to a control center of a system, the acquisition, transmission and recording of the working state of conventional converter valve equipment are completed, then, based on the data, PCA (principal component analysis technology) is combined to generate a countermeasure network and a self-encoder, based on a universal and extensible model architecture, the risk prediction evaluation is carried out on the working state of the conventional converter valve and the acquired data, the risk scores are predicted respectively for different types of data, finally, the data are integrated to obtain an overall risk score, a final expected result is obtained through weighting, and the analysis and evaluation of the working state of the conventional converter valve are completed.
(1) The overall framework of the detection model is shown in fig. 3, and the model may include a module one (a data reconstruction module corresponding to a first kind of data), a module two (a data reconstruction module corresponding to a second kind of data), and an integrated module, and the model may be added based on an extensible module (which is a data reconstruction module) when the data category increases. And each data reconstruction module reconstructs the input data and judges the probability that the data is abnormal data according to the reconstruction error and the intermediate variable error. The integration module integrates the data reconstructed by the data reconstruction module, inputs the data into an encoder for final reconstruction, obtains the probability that the integrated reconstruction data are abnormal data according to the reconstruction error, and finally weights the probability with each probability to obtain the final probability so as to evaluate whether the converter valve is in an abnormal working state.
(2) As shown in fig. 4, the data reconstruction module is a generation countermeasure network (VAE-GAN) based on a variational self-encoder architecture, that is, the data reconstruction module includes a variational self-encoder and a generation countermeasure network; the VAE comprises an encoder and a decoder, the encoder is used for encoding input data and mapping high-dimensional data into low-dimensional data, the decoder is used for mapping the encoded vector into a high-dimensional vector, the generated data is the same as the input data as much as possible, and the VAE increases KL divergence loss of the encoding vector and the standard which are distributed too much in the encoding process so that the encoding result is more continuous.
The data reconstruction module receives the data x after PCA preprocessing as input, and obtains a low-dimensional vector z through an encoder, namely z ═ En1(x), wherein the output of the encoder is the mean value mu and the variance sigma of Gaussian distribution 2 And z is obtained through resampling. Then z gets the reconstructed data through the decoder
Figure BDA0003647374440000081
Namely, it is
Figure BDA0003647374440000082
x and
Figure BDA0003647374440000083
the closer the better, so the reconstruction penalty is:
Figure BDA0003647374440000084
in order to make the encoding result more accurate, an encoder is additionally added to the reconstructed data
Figure BDA0003647374440000085
The re-encoding is performed with the re-encoding result being as close to z as possible, so the penalty is defined as:
Figure BDA0003647374440000086
that is, the losses used by the encoder to train the variational self-encoder include: and loss between a re-encoding result obtained by re-encoding the reconstruction data generated by the decoder of the variational self-encoder by using other encoders and the reconstruction data.
The VAE-based encoder requires that the vector P (z | x) resulting from the encoding is close to the normal positive distribution N (0, I), where KL divergence, i.e., KL (N (μ, σ) with the encoded distribution on the normal positive distribution, is used 2 ) N (0, I)) as loss, the loss calculation results are:
Figure BDA0003647374440000087
finally, the data is reconstructed
Figure BDA0003647374440000088
The original data x is input into the discriminator, and the resistance loss is:
Figure BDA0003647374440000089
wherein D (-) represents the output of the discriminator, thenThe net final loss is:
Figure BDA00036473744400000810
wherein λ 1 ,λ 2 ,λ 3 ,λ 4 The weight that each loss occupies is represented as a hyperparameter.
(3) As shown in fig. 5, the integration module may be composed of a common self-encoder, which is more compact than a variational self-encoder; the integration module receives as input the reconstruction data generated by the previous modules (module one, module two, scalable module, etc.) and combines them together, obtains the implicit spatial coding z, i.e. z ═ en (x), via the encoder, and then obtains the reconstruction data via the decoder
Figure BDA0003647374440000091
According to the final reconstruction error
Figure BDA0003647374440000092
And training the network, and when the error is larger, indicating that the input data has larger probability of being abnormal data.
(4) For training of the detection model shown in fig. 3, considering that the model is composed of a plurality of modules, each module includes a corresponding neural network, in order to reduce the amount of calculation, the collected data is firstly subjected to dimensionality reduction processing, low-dimensional data is obtained and then is used as the input of each network, and the network can be trained more quickly. The first module, the second module and the extensible module are trained in parallel, each module is trained according to a training mode of generating the confrontation network, the judgers are frozen when the generators are trained, and the generators are frozen when the judgers are trained. For the whole model, when the integrated module is trained, other modules are frozen, when other modules are trained, the integrated module is frozen, when the training times reach the specified number, the training is stopped, and at the moment, the whole model can be used for risk prediction.
(5) In the training process, a loss interval formed by corresponding training loss can be recorded, for a module after training is finished, a group of data is newly acquired, after the data is reconstructed by a data reconstruction module, if the reconstruction loss between the reconstructed data and the newly acquired data is in the loss interval and is not close to the maximum value of the loss interval, the probability that the newly acquired data is abnormal data is considered to be low, and if the reconstruction loss between the reconstructed data and the newly acquired data is in the loss interval or is not close to the maximum value of the loss interval, the newly acquired data is considered to be not in accordance with the 'rule' of normal data acquired in the previous training process, namely the newly acquired temperature data is not appeared, and the probability that the newly acquired temperature data is abnormal data is considered to be high. Referring to fig. 6, the magnitude of the value characterizes the probability that the corresponding data is abnormal data.
In one embodiment, the present application provides a system for detecting an operating condition of a converter valve, the system comprising: the detection system is connected with the directional audio sensor and the infrared temperature measurement system to obtain various types of data for monitoring the working state of the converter valve. Temperature data for monitoring the operating condition of the converter valves may be obtained based on the zone maximum temperature, the zone average temperature, the zone minimum temperature, and the point temperature.
It should be understood that, although the steps in the flowcharts related to the embodiments as described above are sequentially displayed as indicated by arrows, the steps are not necessarily performed sequentially as indicated by the arrows. The steps are not performed in the exact order shown and described, and may be performed in other orders, unless explicitly stated otherwise. Moreover, at least a part of the steps in the flowcharts related to the embodiments described above may include multiple steps or multiple stages, which are not necessarily performed at the same time, but may be performed at different times, and the execution order of the steps or stages is not necessarily sequential, but may be rotated or alternated with other steps or at least a part of the steps or stages in other steps.
In one embodiment, as shown in fig. 7, there is provided a device for detecting an operating state of a converter valve, including:
the training unit 700 is configured to use multiple types of data of the converter valve in a normal state as multiple types of normal data, train data reconstruction modules corresponding to the multiple types of data by using the multiple types of normal data, and record loss generated in a training process to obtain loss intervals corresponding to the multiple types of data;
the data classification reconstruction unit 701 is used for regarding a converter valve to be detected in a working state, regarding multiple types of data of the converter valve to be detected as multiple types of data to be detected, inputting each type of data to be detected of the multiple types of data to be detected into a corresponding data reconstruction module respectively, and enabling the data reconstruction module to reconstruct the input data to be detected to obtain corresponding reconstruction data;
a data classification judging unit 702, configured to, for each type of to-be-detected data of the multiple types of to-be-detected data, obtain a reconstruction error corresponding to the type of to-be-detected data based on the type of to-be-detected data and corresponding reconstruction data, and determine a probability that the type of to-be-detected data is abnormal data based on whether the reconstruction error corresponding to the type of to-be-detected data is in a loss interval corresponding to the type of to-be-detected data;
a data integration reconstruction unit 703, configured to integrate the reconstruction data obtained by the various data reconstruction modules to obtain integrated reconstruction data, and input the integrated reconstruction data into the integration module, so that the integration module reconstructs the integrated reconstruction data to obtain corresponding reconstruction data;
an integrated data discrimination unit 704, configured to obtain a reconstruction error corresponding to the integrated reconstruction data based on the integrated reconstruction data and corresponding reconstruction data, obtain a loss interval corresponding to the integrated reconstruction data based on a loss generated in a training process of an integrated module corresponding to the integrated reconstruction data, and determine a probability that the integrated reconstruction data is abnormal data based on whether a reconstruction error corresponding to the integrated reconstruction data is in the loss interval corresponding to the integrated reconstruction data;
and the working state detection unit 705 is configured to determine whether the converter valve to be detected is in an abnormal working state based on a result of weighted summation of the probabilities.
In an embodiment, the data classification determining unit 702 is further configured to, when a reconstruction error corresponding to the type of data to be detected is in the loss interval corresponding to the type and is not close to the maximum value of the loss interval, use a probability lower than a threshold as a probability that the type of data to be detected is abnormal data; and when the reconstruction error corresponding to the type of data to be detected is not in the loss interval corresponding to the type of data or is close to the maximum value of the loss interval, taking the probability higher than the threshold value as the probability that the type of data to be detected is abnormal data.
In an embodiment, the integrated data determining unit 704 is further configured to, when a reconstruction error corresponding to the integrated reconstruction data is in a loss interval corresponding to the integrated reconstruction data and is not close to a maximum value of the loss interval, take a probability lower than a threshold as a probability that the integrated reconstruction data is abnormal data; and when the reconstruction error corresponding to the integrated reconstruction data is not in the loss interval or close to the maximum value of the loss interval corresponding to the integrated reconstruction data, taking the probability higher than the threshold value as the probability that the integrated reconstruction data is abnormal data.
In one embodiment, the types of data include at least: the method comprises the steps of obtaining temperature data of a converter valve to be detected, audio data of a valve hall where the converter valve to be detected is located and working condition data of the converter valve to be detected.
In one embodiment, the data reconstruction module is a generation countermeasure network based on a variational self-encoder, wherein the encoder of the variational self-encoder is the generator of the generation countermeasure network.
In one embodiment, the losses used by an encoder to train the variational self-encoder include: and loss between a re-encoding result obtained by re-encoding the reconstruction data generated by the decoder of the variational self-encoder by using other encoders and the reconstruction data.
For specific limitations of the detecting device for the operating condition of the converter valve, reference may be made to the above limitations of the detecting method for the operating condition of the converter valve, and details thereof are not repeated herein. All or part of the modules in the device for detecting the working state of the converter valve can be realized by software, hardware and a combination thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.
In one embodiment, a computer device is provided, the internal structure of which may be as shown in FIG. 8. The computer device includes a processor, a memory, and a network interface connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The database of the computer device is used for storing detection data of the working state of the converter valve. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer equipment also comprises an input/output interface, wherein the input/output interface is a connecting circuit for exchanging information between the processor and external equipment, and is connected with the processor through a bus, and the input/output interface is called an I/O interface for short. The computer program is executed by a processor to implement a method of detecting an operating condition of a converter valve.
Those skilled in the art will appreciate that the architecture shown in fig. 8 is merely a block diagram of some of the structures associated with the disclosed aspects and is not intended to limit the computing devices to which the disclosed aspects apply, as particular computing devices may include more or less components than those shown, or may combine certain components, or have a different arrangement of components.
In one embodiment, a computer device is provided, comprising a memory and a processor, the memory storing a computer program, the processor implementing the steps of the above-described method embodiments when executing the computer program.
In an embodiment, a computer-readable storage medium is provided, on which a computer program is stored, which computer program, when being executed by a processor, carries out the steps of the respective method embodiment as described above.
In an embodiment, a computer program product is provided, having a computer program stored thereon, the computer program being executed by a processor for performing the steps of the various method embodiments described above.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above may be implemented by hardware instructions of a computer program, which may be stored in a non-volatile computer-readable storage medium, and when executed, may include the processes of the embodiments of the methods described above. Any reference to memory, storage, database or other medium used in the embodiments provided herein can include at least one of non-volatile and volatile memory. Non-volatile Memory may include Read-Only Memory (ROM), magnetic tape, floppy disk, flash Memory, optical storage, or the like. Volatile Memory can include Random Access Memory (RAM) or external cache Memory. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM), among others.
The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features.
The above examples only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent shall be subject to the appended claims.

Claims (10)

1. A method for detecting the working state of a converter valve is characterized by comprising the following steps:
taking the multiple types of data of the converter valve in a normal state as multiple types of normal data, respectively training data reconstruction modules corresponding to the multiple types of normal data by using the multiple types of normal data, and recording losses generated in the training process to obtain loss intervals corresponding to the multiple types of normal data;
regarding to a converter valve to be detected with a working state to be detected, taking multiple types of data of the converter valve to be detected as multiple types of data to be detected, and respectively inputting each type of data to be detected of the multiple types of data to be detected into a corresponding data reconstruction module so that the data reconstruction module reconstructs the input data to be detected to obtain corresponding reconstructed data;
aiming at each type of data to be detected of the multiple types of data to be detected, obtaining a reconstruction error corresponding to the type of data to be detected based on the type of data to be detected and corresponding reconstruction data, and determining the probability that the type of data to be detected is abnormal data based on whether the reconstruction error corresponding to the type of data to be detected is in a loss interval corresponding to the type of data to be detected;
integrating the reconstruction data obtained by the various data reconstruction modules to obtain integrated reconstruction data, and inputting the integrated reconstruction data into an integration module so that the integration module reconstructs the integrated reconstruction data to obtain corresponding reconstruction data;
obtaining a reconstruction error corresponding to the integrated reconstruction data based on the integrated reconstruction data and the corresponding reconstruction data, obtaining a loss interval corresponding to the integrated reconstruction data based on a loss generated in a training process of an integrated module corresponding to the integrated reconstruction data, and determining the probability that the integrated reconstruction data is abnormal data based on whether the reconstruction error corresponding to the integrated reconstruction data is in the loss interval corresponding to the integrated reconstruction data;
and determining whether the converter valve to be detected is in an abnormal working state or not based on the result of weighted summation of all the probabilities.
2. The method according to claim 1, wherein determining the probability that the type of data to be inspected is abnormal data based on whether the reconstruction error corresponding to the type of data to be inspected is in the loss interval corresponding to the type of data to be inspected comprises:
when the reconstruction error corresponding to the type of data to be detected is in the loss interval corresponding to the type and is not close to the maximum value of the loss interval, taking the probability lower than the threshold value as the probability that the type of data to be detected is abnormal data;
and when the reconstruction error corresponding to the type of data to be detected is not in the loss interval corresponding to the type of data or is close to the maximum value of the loss interval, taking the probability higher than the threshold value as the probability that the type of data to be detected is abnormal data.
3. The method according to claim 1, wherein the determining the probability that the integrated reconstruction data is abnormal data based on whether the reconstruction error corresponding to the integrated reconstruction data is in the loss interval corresponding to the integrated reconstruction data comprises:
when the reconstruction error corresponding to the integrated reconstruction data is in a loss interval corresponding to the integrated reconstruction data and is not close to the maximum value of the loss interval, taking the probability lower than the threshold value as the probability that the integrated reconstruction data is abnormal data;
and when the reconstruction error corresponding to the integrated reconstruction data is not in the loss interval or close to the maximum value of the loss interval corresponding to the integrated reconstruction data, taking the probability higher than the threshold value as the probability that the integrated reconstruction data is abnormal data.
4. The method according to any one of claims 1 to 3, wherein the types of data at least include: the method comprises the steps of obtaining temperature data of a converter valve to be detected, audio data of a valve hall where the converter valve to be detected is located and working condition data of the converter valve to be detected.
5. The method of claim 1, wherein the data reconstruction module is a generative countermeasure network based on a variational self-encoder, wherein the encoder of the variational self-encoder is the generator of the generative countermeasure network.
6. The method of claim 5, wherein the loss used to train an encoder of the variational self-encoder comprises: and loss between a re-encoding result obtained by re-encoding the reconstruction data generated by the decoder of the variational self-encoder by using other encoders and the reconstruction data.
7. A device for detecting the operating condition of a converter valve, said device comprising:
the training unit is used for taking the various types of data of the converter valve in a normal state as various types of normal data, respectively training the data reconstruction modules corresponding to the various types of data by using the various types of normal data, and recording the loss generated in the training process to obtain the loss intervals corresponding to the various types of data;
the data classification reconstruction unit is used for regarding a converter valve to be detected in a working state, taking multiple types of data of the converter valve to be detected as multiple types of data to be detected, and respectively inputting each type of data to be detected of the multiple types of data to be detected into a corresponding data reconstruction module so that the data reconstruction module reconstructs the input data to be detected to obtain corresponding reconstructed data;
the data classification judging unit is used for obtaining a reconstruction error corresponding to the class of data to be detected based on the class of data to be detected and corresponding reconstruction data aiming at each class of data to be detected of the multiple classes of data to be detected, and determining the probability that the class of data to be detected is abnormal based on whether the reconstruction error corresponding to the class of data to be detected is in a loss interval corresponding to the class;
the data integration reconstruction unit is used for integrating reconstruction data obtained by various data reconstruction modules to obtain integrated reconstruction data and inputting the integrated reconstruction data into the integration module so that the integration module reconstructs the integrated reconstruction data to obtain corresponding reconstruction data;
an integrated data discrimination unit, configured to obtain a reconstruction error corresponding to the integrated reconstruction data based on the integrated reconstruction data and corresponding reconstruction data, obtain a loss interval corresponding to the integrated reconstruction data based on a loss generated in a training process of an integrated module corresponding to the integrated reconstruction data, and determine a probability that the integrated reconstruction data is abnormal data based on whether the reconstruction error corresponding to the integrated reconstruction data is in the loss interval corresponding to the integrated reconstruction data;
and the working state detection unit is used for determining whether the converter valve to be detected is in an abnormal working state or not based on the result of weighted summation of the probabilities.
8. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the method of any one of claims 1 to 6 when executing the computer program.
9. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the method of any one of claims 1 to 6.
10. A computer program product comprising a computer program, characterized in that the computer program realizes the method of any of claims 1 to 6 when executed by a processor.
CN202210534846.8A 2022-05-17 2022-05-17 Method, device and equipment for detecting working state of converter valve and storage medium Pending CN114881157A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210534846.8A CN114881157A (en) 2022-05-17 2022-05-17 Method, device and equipment for detecting working state of converter valve and storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210534846.8A CN114881157A (en) 2022-05-17 2022-05-17 Method, device and equipment for detecting working state of converter valve and storage medium

Publications (1)

Publication Number Publication Date
CN114881157A true CN114881157A (en) 2022-08-09

Family

ID=82676626

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210534846.8A Pending CN114881157A (en) 2022-05-17 2022-05-17 Method, device and equipment for detecting working state of converter valve and storage medium

Country Status (1)

Country Link
CN (1) CN114881157A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115717590A (en) * 2022-11-22 2023-02-28 西安交通大学 Intelligent abnormity detection method for compressor and related device
CN116700213A (en) * 2023-06-13 2023-09-05 无锡物联网创新中心有限公司 Industrial equipment abnormality detection method and related device based on gating circulation unit

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2021189904A1 (en) * 2020-10-09 2021-09-30 平安科技（深圳）有限公司 Data anomaly detection method and apparatus, and electronic device and storage medium
CN113762486A (en) * 2021-11-11 2021-12-07 中国南方电网有限责任公司超高压输电公司广州局 Method and device for constructing fault diagnosis model of converter valve and computer equipment
CN114037478A (en) * 2021-11-18 2022-02-11 北京明略软件系统有限公司 Advertisement abnormal flow detection method and system, electronic equipment and readable storage medium
CN114186618A (en) * 2021-11-24 2022-03-15 国网福建省电力有限公司检修分公司 Method for judging working state of converter valve

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2021189904A1 (en) * 2020-10-09 2021-09-30 平安科技（深圳）有限公司 Data anomaly detection method and apparatus, and electronic device and storage medium
CN113762486A (en) * 2021-11-11 2021-12-07 中国南方电网有限责任公司超高压输电公司广州局 Method and device for constructing fault diagnosis model of converter valve and computer equipment
CN114037478A (en) * 2021-11-18 2022-02-11 北京明略软件系统有限公司 Advertisement abnormal flow detection method and system, electronic equipment and readable storage medium
CN114186618A (en) * 2021-11-24 2022-03-15 国网福建省电力有限公司检修分公司 Method for judging working state of converter valve

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115717590A (en) * 2022-11-22 2023-02-28 西安交通大学 Intelligent abnormity detection method for compressor and related device
CN115717590B (en) * 2022-11-22 2024-03-29 西安交通大学 Intelligent abnormality detection method and related device for compressor
CN116700213A (en) * 2023-06-13 2023-09-05 无锡物联网创新中心有限公司 Industrial equipment abnormality detection method and related device based on gating circulation unit
CN116700213B (en) * 2023-06-13 2024-03-29 无锡物联网创新中心有限公司 Industrial equipment abnormality detection method and related device based on gating circulation unit

Similar Documents

Publication Publication Date Title
US20220237060A1 (en) Abnormality detection system, abnormality detection method, abnormality detection program, and method for generating learned model
CN114881157A (en) Method, device and equipment for detecting working state of converter valve and storage medium
CN112766342A (en) Abnormity detection method for electrical equipment
CN111340367B (en) Oil-shortage equipment state evaluation method and system based on multi-dimensional relevance data analysis
CN112596016A (en) Transformer fault diagnosis method based on integration of multiple one-dimensional convolutional neural networks
CN115392782A (en) Method and system for monitoring and diagnosing health state of process system of nuclear power plant
CN115374822A (en) Fault diagnosis method and system based on multi-level feature fusion
CN113782113B (en) Method for identifying gas fault in transformer oil based on deep residual error network
CN114511025A (en) Fan fault diagnosis method and device based on weighted multi-sensor fusion filtering
CN113435103A (en) Power distribution room abnormity detection method, system, server, edge gateway and medium
CN113295399A (en) Converter valve element state evaluation method and device, electronic equipment and storage medium
CN113887866A (en) Method and device for generating human living environment evaluation index
CN115618732B (en) Nuclear reactor digital twin key parameter autonomous optimization data inversion method
CN117371207A (en) Extra-high voltage converter valve state evaluation method, medium and system
CN116627116A (en) Process industry fault positioning method and system and electronic equipment
CN111474476A (en) Motor fault prediction method
Jin et al. Deep Machine Learning-Based Asset Management Approach for Oil-Immersed Power Transformers Using Dissolved Gas Analysis
CN114154779A (en) Health assessment and fault diagnosis method and system based on EMD-SOM
CN114186618A (en) Method for judging working state of converter valve
RU2648413C1 (en) Method of the modes control based on the neural network diagnostics of faults and technical condition of the electric-drive gas-compressor unit
CN115545191B (en) Current noise reduction network model training method and fault current limiter current noise reduction method
CN116259161B (en) Power failure early warning system
Bhattacharyya et al. Adaptive, online models to detect and estimate gross error in SPNDs
KR102212022B1 (en) Method of automatically determining condition of hydro turbine in hydroelectric power plant and system for the same
CN116401533B (en) Fault diagnosis method and device for on-load tap-changer

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination