CN117029900B - Metering instrument detection method based on dynamic multipath synchronous detection - Google Patents
Metering instrument detection method based on dynamic multipath synchronous detection Download PDFInfo
- Publication number
- CN117029900B CN117029900B CN202311305344.9A CN202311305344A CN117029900B CN 117029900 B CN117029900 B CN 117029900B CN 202311305344 A CN202311305344 A CN 202311305344A CN 117029900 B CN117029900 B CN 117029900B
- Authority
- CN
- China
- Prior art keywords
- data
- detection
- instrument
- metering
- frequency
- 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.)
- Active
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 167
- 230000001360 synchronised effect Effects 0.000 title claims abstract description 42
- 230000000007 visual effect Effects 0.000 claims abstract description 18
- 230000002159 abnormal effect Effects 0.000 claims abstract description 16
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 15
- 238000004458 analytical method Methods 0.000 claims abstract description 14
- 238000012545 processing Methods 0.000 claims abstract description 10
- 238000011156 evaluation Methods 0.000 claims abstract description 7
- 238000007781 pre-processing Methods 0.000 claims abstract description 6
- 238000012800 visualization Methods 0.000 claims abstract description 5
- 238000000034 method Methods 0.000 claims description 42
- 230000008859 change Effects 0.000 claims description 21
- 230000011218 segmentation Effects 0.000 claims description 15
- 230000005856 abnormality Effects 0.000 claims description 11
- 230000008569 process Effects 0.000 claims description 10
- 238000012937 correction Methods 0.000 claims description 9
- 238000004891 communication Methods 0.000 claims description 7
- 230000004927 fusion Effects 0.000 claims description 6
- 238000005070 sampling Methods 0.000 claims description 6
- 238000013527 convolutional neural network Methods 0.000 claims description 5
- 238000013480 data collection Methods 0.000 claims description 4
- 230000003068 static effect Effects 0.000 claims description 4
- 230000003044 adaptive effect Effects 0.000 claims description 2
- 230000006870 function Effects 0.000 description 8
- 238000005516 engineering process Methods 0.000 description 5
- 238000005259 measurement Methods 0.000 description 5
- 238000012544 monitoring process Methods 0.000 description 5
- 238000001228 spectrum Methods 0.000 description 5
- 238000004364 calculation method Methods 0.000 description 4
- 238000011161 development Methods 0.000 description 4
- 230000018109 developmental process Effects 0.000 description 4
- 238000001914 filtration Methods 0.000 description 4
- 238000012423 maintenance Methods 0.000 description 4
- 238000010276 construction Methods 0.000 description 2
- 238000003709 image segmentation Methods 0.000 description 2
- 238000010606 normalization Methods 0.000 description 2
- 238000003908 quality control method Methods 0.000 description 2
- 230000009467 reduction Effects 0.000 description 2
- 230000008054 signal transmission Effects 0.000 description 2
- 230000003595 spectral effect Effects 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- 238000012952 Resampling Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 239000012530 fluid Substances 0.000 description 1
- 238000007499 fusion processing Methods 0.000 description 1
- 238000009776 industrial production Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 230000000737 periodic effect Effects 0.000 description 1
- 238000011112 process operation Methods 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 238000003860 storage Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01D—MEASURING NOT SPECIALLY ADAPTED FOR A SPECIFIC VARIABLE; ARRANGEMENTS FOR MEASURING TWO OR MORE VARIABLES NOT COVERED IN A SINGLE OTHER SUBCLASS; TARIFF METERING APPARATUS; MEASURING OR TESTING NOT OTHERWISE PROVIDED FOR
- G01D18/00—Testing or calibrating apparatus or arrangements provided for in groups G01D1/00 - G01D15/00
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
-
- 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
- G06N3/0464—Convolutional networks [CNN, ConvNet]
-
- 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
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Molecular Biology (AREA)
- Biomedical Technology (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Health & Medical Sciences (AREA)
- Biophysics (AREA)
- Computing Systems (AREA)
- Computational Linguistics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Arrangements For Transmission Of Measured Signals (AREA)
- Testing, Inspecting, Measuring Of Stereoscopic Televisions And Televisions (AREA)
Abstract
The invention discloses a metering device detection method based on dynamic multipath synchronous detection, which comprises the following steps: s1, constructing a visual dynamic multi-channel synchronous detection network, and collecting detection data of detection nodes in real time; s2, decomposing the detection data into signal data and image data, and respectively preprocessing; s3, synchronously correcting the main reference signal corresponding to the main reference source and the rest signal data; s4, processing and analyzing the image data by using a multi-layer recognition algorithm, and outputting instrument data; s5, fusing the signal data with the instrument data, and performing comprehensive analysis and detection; and S6, generating a performance evaluation report according to the detection result, and synchronously feeding back to the visualization platform for display. The invention utilizes the dynamic multipath synchronous detection method to collect the data of the metering instrument in the network in real time, obviously improves the detection precision and accuracy, reduces the system error and eliminates the influence of abnormal data of individual instruments on the whole detection result.
Description
Technical Field
The invention relates to the technical field of metering instrument detection, in particular to a metering instrument detection method based on dynamic multipath synchronous detection.
Background
In an industrial setting, gauges are measuring and control tools widely used in various processes and equipment. These meters are used to monitor and measure various physical quantities, such as pressure, temperature, flow, level, current, voltage, etc., to ensure process stability and quality control. The metering device plays an important role in industrial scenes, helps engineers and operators monitor process parameters, realizes automatic control, and ensures the stability, safety and high efficiency of the process. They are typically used in conjunction with data acquisition systems, control systems, and monitoring systems, to provide real-time measurement data and feedback information to support decision-making and optimizing process operation.
The development of automatic detection of metering devices stems from advances in automation and digitization techniques, as well as the need for high efficiency and quality control of industrial production processes. With the development of automation technology, the performance of sensors, actuators and control systems is continuously increasing. This provides a more accurate, reliable and efficient tool and platform for automatic metering device detection. The automatic detection of the metering device needs to collect a large amount of data in real time and transmit the data to a data processing system for analysis and processing. With the development of data acquisition and communication technologies, such as industrial internet of things (IIoT), sensor networks and wireless communication technologies, more efficient data acquisition and real-time monitoring are realized. Automated detection techniques may utilize data analysis and artificial intelligence techniques to extract valuable information, perform fault diagnosis and prediction. Techniques such as machine learning, deep learning, and pattern recognition can process large amounts of data and automatically discover potential problems and anomalies.
Although automatic detection techniques can increase the efficiency of detection, there are still problems with accuracy and precision. The accuracy of the meter may change over time, requiring periodic calibration and adjustment. In addition, the accuracy of the automatic detection system itself needs to be ensured to avoid erroneous judgment and adjustment of the metering device. For example, CN113375765a discloses a detection method, a device and a storage medium for a metering device, when controlling to introduce a fluid into the metering device at a first flow rate, determining a plurality of effective pulse waveforms collected by the metering device according to the first flow rate, and obtaining a period of each effective pulse waveform in the plurality of effective pulse waveforms; determining at least two continuous stable pulse waveforms from the plurality of effective pulse waveforms according to the period of each effective pulse waveform; and determining the detection result of the metering instrument according to the respective duty ratios of at least two stable pulse waveforms, wherein the detection result comprises passing detection or failing detection. The continuous at least two stable pulse waveforms are determined in the plurality of effective pulse waveforms which need to be collected, and the detection result of the metering instrument is determined according to the at least two stable pulse waveforms, so that the detection accuracy of the metering instrument is ensured. The detection method utilizes multi-pulse multi-waveform detection, but does not monitor and detect the metering instrument, and under the premise of not excluding the fault of the metering instrument, the multi-pulse multi-waveform detection result still has certain deception and uncertainty.
In summary, automatic detection involves a large amount of data collection and processing, so that data quality and reliability are critical to accurate analysis and judgment, and problems may come from sensor failures, signal interference, data transmission errors, and the like. Furthermore, in some cases, the system may not be able to accurately identify and locate faults. The interaction of complex failure modes and multiple failure factors may lead to erroneous or missed decisions of the system. For the problems in the related art, no effective solution has been proposed at present.
Disclosure of Invention
Aiming at the defects of the prior art, the invention provides a metering device detection method based on dynamic multipath synchronous detection, which is provided with the advantages that the data of a plurality of metering devices are collected in real time and synchronously processed and analyzed, the detection precision and accuracy are greatly improved and the system error is reduced through the comparison and calibration of multipath data, and further the problems that the precision and accuracy still exist in the traditional automatic detection technology, the precision of the metering devices possibly changes along with time, and the regular calibration and adjustment are needed are solved.
In order to achieve the advantages of improving the detection precision and accuracy and reducing the system error, the invention adopts the following specific technical scheme:
a metering device detection method based on dynamic multipath synchronous detection comprises the following steps:
s1, constructing a visual dynamic multi-channel synchronous detection network, and collecting detection data of detection nodes in real time;
s2, decomposing the detection data into signal data and image data, and respectively preprocessing;
s3, synchronously correcting the main reference signal corresponding to the main reference source and the rest signal data;
s4, processing and analyzing the image data by using a multi-layer recognition algorithm, and outputting instrument data;
s5, fusing the signal data with the instrument data, and performing comprehensive analysis and detection;
and S6, generating a performance evaluation report according to the detection result, and synchronously feeding back to the visualization platform for display.
Further, a visual dynamic multipath synchronous detection network is built, and the real-time acquisition of detection data of the detection nodes comprises the following steps:
s11, setting a topological structure of a visual dynamic multi-path synchronous detection network according to the layout of an industrial system;
s12, determining each detection node in the network according to the topological structure, selecting one of the detection nodes as a main reference source, and deploying metering devices and data acquisition equipment for the rest detection nodes;
s13, setting an initial detection period to detect the metering device in real time, and taking data periodically collected by the metering device and the data collection equipment as detection data.
Further, each detection node is configured with a static IP address, and communication is carried out between the detection nodes and a remote management center based on an RTSP protocol.
Further, the synchronization correction of the main reference signal corresponding to the main reference source and the rest signal data comprises the following steps:
s31, performing time-frequency analysis on the signal data, and extracting phase characteristics of the signal data;
s32, comparing a main reference signal corresponding to a main reference source with each signal data, and calculating a phase difference between the main reference signal and each signal data by using a time domain difference method to serve as a synchronization error;
s33, carrying out frequency tracking on each detection node by utilizing a frequency locking loop algorithm;
s34, compensating the phase and the frequency in the detection node by using the synchronous error and the frequency tracking result to realize synchronous correction between the signal data and the main reference signal;
and S35, carrying out self-adaptive adjustment on the detection period of the detection node according to the frequency tracking result.
Further, the adaptive adjustment of the detection period of the detection node according to the frequency tracking result includes the following steps:
s351, calculating a frequency change rate according to a frequency tracking result of the signal data;
s352, when the frequency change rate is larger than a preset frequency threshold, the detection period is reduced on the basis of the initial detection frequency, the sampling frequency is increased, and if the frequency change rate is smaller than or equal to the preset frequency threshold, the initial detection frequency is maintained unchanged.
Further, the image data is processed and analyzed by using a multi-layer recognition algorithm, and the output instrument data comprises the following steps:
s41, setting an image caching strategy for a remote management center to cache the image data in a queue;
s42, extracting image data at the same time as the signal data in a buffer queue;
s43, preferentially utilizing the meter identification model to identify the metering device in the image data, extracting a metering numerical value as meter data, and carrying out abnormality judgment on the metering device;
s44, calling a fault identification model to identify the image data with numerical value abnormality, so as to realize fault type detection of the metering instrument and output instrument fault data.
Further, setting an image caching strategy for the remote management center to perform queue caching on the image data comprises the following steps:
s411, configuring a buffer queue for each detection node in a remote management center;
s412, setting a preset queue length for the buffer queue, storing the image frames of the image data received in real time into the buffer queue, if the queue length is not greater than the preset queue length at this time, successfully storing the image data, and if the queue length is greater than the preset queue length at this time, rejecting the previous extracted and identified image frame and the adjacent 2N image frames.
Further, the method for preferentially identifying the metering device in the image data by using the device identification model, extracting the metering value as the device data, and judging the abnormality of the metering device comprises the following steps:
s431, dividing the image data by utilizing a dividing network model, and extracting an instrument area image;
s432, identifying an instrument area image by using a convolutional neural network model, extracting the deflection angle of a pointer of a metering instrument in the area, and calculating the value of the pointer to be used as instrument data;
s433, comparing the average value of the meter data with the average value of the first M meter data, if the difference value of the meter data and the average value is larger than a safety threshold value, judging that the metering device is abnormal, otherwise, calibrating the metering device as a normal metering device.
Further, the construction method of the segmentation network model comprises the following steps:
s4311, adopting a self-encoder as an infrastructure of a segmentation network model, wherein the self-encoder is provided with four convolution network modules, and each convolution network module comprises 3 multiplied by 3 convolution kernels;
s4312, introducing a cavity fusion layer at the output end of the self-encoder, wherein the cavity fusion layer comprises three cavity convolution kernels;
s4313, setting a loss function of the segmentation network model, wherein the loss function expression is as follows:
in the method, in the process of the invention,the value of the tag is indicated and,ythe predicted value is represented by a value of the prediction,nthe number of samples is represented and the number of samples,ithe order of the samples is represented and,qrepresenting the number of classifications.
Further, invoking a fault recognition model to recognize the image data with numerical value abnormality, realizing fault type detection of the metering instrument, and outputting instrument fault data comprises the following steps:
s441, selecting the front N image frames and the rear N image frames of the abnormal image data, and carrying out segmentation processing on the abnormal image by utilizing a segmentation network model to obtain 2N instrument sub-images;
s442, positioning and identifying meter pointers in the meter sub-images, respectively extracting deflection angles of each meter pointer, and combining the deflection angles according to a time sequence to obtain an angle set;
s443, calculating the average angle, deflection frequency and amplitude change of all deflection angles in the angle set, if the average angle exceeds an angle threshold value, determining that the pointer is in offset fault, if the deflection frequency exceeds a deflection threshold value, determining that the pointer is in vibration fault, and if the amplitude change is lower than an amplitude threshold value, determining that the pointer is in clamping fault.
Compared with the prior art, the invention provides a metering device detection method based on dynamic multipath synchronous detection, which has the following beneficial effects:
(1) The visual dynamic multi-channel synchronous detection network is constructed, the data of all metering instruments in the network are acquired in real time by utilizing a dynamic multi-channel synchronous detection method, the data are synchronously processed and analyzed, the accuracy and the accuracy of detection can be remarkably improved through comparison and calibration of multi-channel data, the system error is reduced, meanwhile, the data of a plurality of metering instruments are synchronously detected, the influence of abnormal data of individual instruments on the whole detection result can be eliminated, and the quality and the reliability of the data are improved through cross verification and correction of the multi-channel data; in addition, the invention has certain self-adaptability and compatibility, can adapt to metering instruments of different types and brands, and synchronously detects multipath data, thereby effectively interacting with various metering instruments and collecting data, and improving the application range and flexibility of the system.
(2) The method is characterized in that the pointer type metering instrument is detected, the high-precision segmentation algorithm and the pointer identification convolutional neural network model are combined to carry out image segmentation and identification, the pointer change is utilized to carry out high-precision degree and instrument fault identification, the high-precision degree calculation of the instrument reading can be realized, a more accurate measurement result is provided, the change information of the pointer is utilized, the fault mode and the preset rule are combined to carry out instrument fault identification, for example, when the pointer deviation exceeds a threshold value or the pointer vibration is abnormal, the instrument fault can be judged, and the operation and maintenance personnel can be helped to find the instrument fault in time and take corresponding maintenance measures.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings that are needed in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flowchart of a metering device detection method based on dynamic multi-path synchronous detection according to an embodiment of the present invention.
Detailed Description
According to the embodiment of the invention, a metering device detection method based on dynamic multipath synchronous detection is provided.
The invention will be further described with reference to the accompanying drawings and detailed description, as shown in fig. 1, a method for detecting a metering device based on dynamic multi-path synchronous detection according to an embodiment of the invention, the method comprises the following steps:
s1, constructing a visual dynamic multipath synchronous detection network, and collecting detection data of detection nodes in real time.
The visual dynamic multipath synchronous detection network is a method for displaying and managing synchronous detection processes of a plurality of metering devices through a graphical interface, and comprises a user interface which is realized by using a graphical interface development tool or Web technology. The user interface should have intuitive layout and easy-to-operate functions to demonstrate and control the synchronous detection process of multiple meters. The user side or the remote management center establishes communication connection with each metering instrument to ensure that the signal data can be received. The collected signal data are displayed on a user interface in a form of a chart, a curve or other visual forms, and the data of a plurality of metering instruments can be displayed on a time axis, so that the user can observe and compare conveniently.
The method for real-time acquisition of detection data of the detection nodes comprises the following steps of:
s11, setting the topological structure of the visual dynamic multi-path synchronous detection network according to the layout of the industrial system.
S12, determining each detection node in the network according to the topological structure, selecting one of the detection nodes as a main reference source, and deploying metering devices and data acquisition equipment for the rest detection nodes.
And according to the topological structure and the requirements of the detection nodes, deploying proper metering devices and data acquisition equipment for each node. The metering instrument selects proper model and specification, and can meet the requirements of accuracy and measurement range. The data acquisition equipment can be selected according to the communication protocol and interface requirements, so that stable acquisition and transmission of signal data can be ensured, the data acquisition equipment comprises a high-definition camera, an environment monitoring sensor, an electric signal device and the like, and if the metering instrument does not have a signal transmission function, the electric signal device is used for transmitting the signal data.
S13, setting an initial detection period to detect the metering device in real time, and taking data periodically collected by the metering device and the data collection equipment as detection data.
In addition, each detection node is respectively provided with a static IP address, and communication is carried out between the detection nodes and a remote management center based on an RTSP protocol.
And respectively determining the detection node and the remote management center as an RTSP server and an RTSP client. And configuring an RTSP server on each detection node, and setting the static IP address, the port number and the streaming media resource information of the node. Such information includes the encoding format, resolution, frame rate, etc. of the audio and video data to be transmitted. An RTSP client is configured on a remote management center, and IP addresses and port numbers of the respective detection nodes, and parameters of audio and video streams to be received are set. The RTSP client initiates a connection request to a specific RTSP server, and communicates using a specified URL and protocol. The connection request contains description information of the streaming media resource to be acquired.
Once the connection is established successfully, the RTSP server starts transmitting signal data and video streaming media data to the RTSP client. The data may be transmitted via the RTP (Real-TiMe Transport Protocol) protocol, while the RTCP (Real-TiMe Control Protocol) protocol may also be used for flow control and synchronization.
S2, decomposing the detection data into signal data and image data, and preprocessing the signal data and the image data respectively.
Wherein, the signal data preprocessing includes: noise reduction: noise interference in the signal is reduced using a filter or the like to improve signal quality. Sampling and resampling: the signal is suitably sampled and possibly resampled as needed to match a particular sampling rate. And (3) filtering: a digital filter is applied to remove unwanted frequency components or emphasize frequency ranges of interest. And (3) correction: the signal is corrected based on sensor characteristics or calibration information of the meter to ensure accuracy and reliability.
The image data preprocessing includes: denoising an image: noise in the image is removed using a denoising algorithm (e.g., mean filtering, median filtering, etc.). Image enhancement: and the techniques of histogram equalization, contrast enhancement and the like are applied to enhance the visual effect and detail of the image. Image normalization: and performing scale normalization, color space conversion and other operations on the image to enable the image to have uniform characteristic representation.
S3, performing synchronous correction on the main reference signal corresponding to the main reference source and the rest signal data, wherein the method comprises the following steps of:
s31, performing time-frequency analysis on the signal data, and extracting phase characteristics of the signal data.
The time-frequency analysis method used in the invention is Short-time Fourier transform (STFT), which decomposes the signal into spectral components over different time periods. Amplitude spectrum and phase spectrum information is extracted from each spectrum. The amplitude spectrum represents the energy or amplitude of the different frequency components, while the phase spectrum represents the phase angle of the different frequency components.
S32, comparing the main reference signal corresponding to the main reference source with each signal data, and calculating the phase difference between the main reference signal and each signal data by using a time domain difference method to serve as a synchronization error.
In dynamic multipath synchronization detection, a time domain difference method is used to calculate the phase difference between a main reference signal and each signal data as a measurement index of synchronization error.
In the above description, the synchronization error calculation includes the steps of: a primary reference signal and each signal data is acquired. The main reference signal and the signal data are preprocessed, including filtering, noise reduction and the like, so as to ensure the data quality. The primary reference signal and the signal data are time-domain aligned so that they correspond in time. For each time point, the synchronization error is measured by comparing the phase difference of the main reference signal and the signal data, and the phase difference calculation formula is as follows: phase difference = phase (main reference signal) -phase (signal data).
S33, frequency tracking is carried out on each detection node by utilizing a frequency locking loop algorithm.
The frequency locked loop algorithm (Frequency Lock Loop, FLL) is a commonly used algorithm for tracking frequency variations of a signal in real time and performing frequency compensation. In dynamic multipath synchronization detection, each detection node is frequency tracked by using a frequency locked loop algorithm to compensate for frequency errors.
And S34, compensating the phase and the frequency in the detection node by using the synchronous error and the frequency tracking result, and realizing synchronous correction between the signal data and the main reference signal.
In the above step, the phase compensation is performed by using the synchronization error, and the phase of the signal data of each detection node is adjusted so as to be synchronized with the main reference signal. The specific phase compensation method is adjusted according to the magnitude and direction of the synchronization error, for example, by adding a fixed phase offset or dynamically adjusting according to the proportion of the synchronization error.
And carrying out frequency compensation on the signal data of each detection node by utilizing the frequency tracking result, and eliminating frequency errors. And adjusting the sampling frequency or delay of the signal data according to the deviation magnitude and direction of the frequency tracking result to perform frequency compensation.
And comparing the signal data subjected to phase compensation and frequency compensation with a main reference signal, and verifying the effect of synchronous correction. The corrected synchronization error may be calculated, evaluated, and adjusted.
S35, carrying out self-adaptive adjustment on the detection period of the detection node according to the frequency tracking result, wherein the self-adaptive adjustment comprises the following steps:
s351, calculating the frequency change rate according to the frequency tracking result of the signal data.
For the frequency value at each time point, the amount of change in frequency can be obtained by calculating the frequency difference at the adjacent two time points. Assume that the frequency values at time point i and time point i+1 are f, respectively i And f i+1 The frequency change rate Δf can be calculated as Δf=f i+1 -f i 。
S352, when the frequency change rate is larger than a preset frequency threshold, the detection period is reduced on the basis of the initial detection frequency, the sampling frequency is increased, and if the frequency change rate is smaller than or equal to the preset frequency threshold, the initial detection frequency is maintained unchanged.
S4, processing and analyzing the image data by using a multi-layer recognition algorithm, and outputting instrument data, wherein the method comprises the following steps of:
s41, setting an image caching strategy for a remote management center to cache the image data in a queue, wherein the method comprises the following steps of:
s411, a buffer queue is configured for each detection node in a remote management center.
S412, setting a preset queue length for the buffer queue, storing the image frames of the image data received in real time into the buffer queue, if the queue length is not greater than the preset queue length at this time, successfully storing the image data, and if the queue length is greater than the preset queue length at this time, rejecting the previous extracted and identified image frame and the adjacent 2N image frames.
S42, extracting the image data at the same time as the signal data in the buffer queue.
S43, preferentially utilizing the meter identification model to identify the metering device in the image data, extracting a metering numerical value as meter data, and carrying out abnormality judgment on the metering device, wherein the method comprises the following steps of:
s431, dividing the image data by using the dividing network model, and extracting an instrument area image.
The construction method of the segmentation network model comprises the following steps:
s4311, adopting a self-encoder as an infrastructure of a split network model, wherein the self-encoder is provided with four convolution network modules, and each convolution network module comprises 3 multiplied by 3 convolution kernels.
S4312, introducing a cavity fusion layer at the output end of the self-encoder, wherein the cavity fusion layer comprises three cavity convolution kernels.
S4313, setting a loss function of the segmentation network model, wherein the loss function expression is as follows:
in the method, in the process of the invention,the value of the tag is indicated and,ythe predicted value is represented by a value of the prediction,nthe number of samples is represented and the number of samples,ithe order of the samples is represented and,qrepresenting the number of classifications.
S432, identifying an instrument area image by using a convolutional neural network model, extracting the deflection angle of a pointer of a metering instrument in the area, and calculating the pointer value as instrument data.
S433, comparing the average value of the meter data with the average value of the first M meter data, if the difference value of the meter data and the average value is larger than a safety threshold value, judging that the metering device is abnormal, otherwise, calibrating the metering device as a normal metering device.
S44, calling a fault identification model to identify the image data with numerical value abnormality, realizing fault type detection of the metering instrument, and outputting instrument fault data, wherein the method comprises the following steps:
s441, selecting the front N image frames and the rear N image frames of the abnormal image data, and performing segmentation processing on the abnormal image by using a segmentation network model to obtain 2N instrument sub-images.
S442, positioning and identifying meter pointers in the meter sub-images, respectively extracting deflection angles of each meter pointer, and combining the deflection angles according to time sequence to obtain an angle set.
S443, calculating the average angle, deflection frequency and amplitude change of all deflection angles in the angle set, if the average angle exceeds an angle threshold value, determining that the pointer is in offset fault, if the deflection frequency exceeds a deflection threshold value, determining that the pointer is in vibration fault, and if the amplitude change is lower than an amplitude threshold value, determining that the pointer is in clamping fault.
S5, fusing the signal data with the instrument data, and performing comprehensive analysis and detection.
Before the data fusion process, the signal data and meter data are aligned to ensure that they have the same time stamp or time period for subsequent analysis and comparison. Useful features are extracted from the signal data and the meter data, and key information in the data can be captured based on statistical features, spectral features, time-frequency features and the like. And finally, fusing the signal data and the instrument data, and integrating the information of the signal data and the instrument data into a comprehensive data set by adopting methods of weighted average, splicing, characteristic splicing and the like.
And analyzing and detecting the fused data by using a statistical analysis method to find information such as abnormality, trend, relevance and the like in the data, thereby realizing the functions of fault detection, performance evaluation, state monitoring and the like of the instrument.
And S6, generating a performance evaluation report according to the detection result, and synchronously feeding back to the visualization platform for display.
And evaluating the performance of the instrument according to the comprehensive analysis and detection results, wherein the performance comprises fault detection results, instrument index deviation, trend analysis, performance score and the like. Based on the detection result, a performance evaluation report is generated. The report includes information such as status summaries of the meters, anomaly detection results, fault type diagnostics, performance scores, anomaly data examples, and the like. The report should have a clear structure and easy to understand content so that the user can quickly obtain the key information.
And synchronously feeding back the generated performance evaluation report to a visual platform so that a user can intuitively view and analyze the detection result through a visual interface. The visual platform provides functions of real-time monitoring of instrument states, historical data inquiry, abnormal alarm and the like so as to support real-time tracking and analysis of the instrument performances by users. And displaying key indexes, trend graphs, abnormal data and the like of the detection result on a visual platform. By adopting the visualization modes such as charts, curves, thermodynamic diagrams and the like, a user can intuitively know the performance condition and the change trend of the instrument.
In summary, by means of the technical scheme, the visual dynamic multi-channel synchronous detection network is constructed, the data of all metering devices in the network are collected in real time by utilizing the dynamic multi-channel synchronous detection method and are synchronously processed and analyzed, the accuracy and the accuracy of detection can be remarkably improved, the system error is reduced, the data of a plurality of metering devices are synchronously detected, the influence of abnormal data of individual metering devices on the whole detection result can be eliminated, and the quality and the reliability of the data are improved through the cross verification and correction of the multi-channel data; in addition, the invention has certain self-adaptability and compatibility, can adapt to metering instruments of different types and brands, and synchronously detects multipath data, thereby effectively interacting with various metering instruments and collecting data, and improving the application range and flexibility of the system. The method is characterized in that the pointer type metering instrument is detected, the high-precision segmentation algorithm and the pointer identification convolutional neural network model are combined to carry out image segmentation and identification, the pointer change is utilized to carry out high-precision degree and instrument fault identification, the high-precision degree calculation of the instrument reading can be realized, a more accurate measurement result is provided, the change information of the pointer is utilized, the fault mode and the preset rule are combined to carry out instrument fault identification, for example, when the pointer deviation exceeds a threshold value or the pointer vibration is abnormal, the instrument fault can be judged, and the operation and maintenance personnel can be helped to find the instrument fault in time and take corresponding maintenance measures.
The foregoing description of the preferred embodiments of the invention is not intended to be limiting, but rather is intended to cover all modifications, equivalents, alternatives, and improvements that fall within the spirit and scope of the invention.
Claims (6)
1. The metering device detection method based on dynamic multipath synchronous detection is characterized by comprising the following steps of:
s1, constructing a visual dynamic multi-channel synchronous detection network, and collecting detection data of detection nodes in real time;
s2, decomposing the detection data into signal data and image data, and respectively preprocessing;
s3, synchronously correcting the main reference signal corresponding to the main reference source and the rest signal data;
s4, processing and analyzing the image data by using a multi-layer recognition algorithm, and outputting instrument data;
s5, fusing the signal data with the instrument data, and performing comprehensive analysis and detection;
s6, generating a performance evaluation report according to the detection result, and synchronously feeding back to the visualization platform for display;
the method for constructing the visual dynamic multipath synchronous detection network comprises the following steps of:
s11, setting a topological structure of a visual dynamic multi-path synchronous detection network according to the layout of an industrial system;
s12, determining each detection node in a network according to the topological structure, selecting one of the detection nodes as a main reference source, and deploying metering instruments and data acquisition equipment for the rest of the detection nodes;
s13, setting an initial detection period to detect the metering instrument in real time, and taking data periodically collected by the metering instrument and the data collection equipment as detection data;
the step of synchronously correcting the main reference signal corresponding to the main reference source and the rest signal data comprises the following steps:
s31, performing time-frequency analysis on the signal data, and extracting phase characteristics of the signal data;
s32, comparing the main reference signal corresponding to the main reference source with each signal data, and calculating a phase difference between the main reference signal and each signal data by using a time domain difference method to serve as a synchronization error;
s33, carrying out frequency tracking on each detection node by utilizing a frequency locking ring algorithm;
s34, compensating the phase and the frequency in the detection node by using the synchronous error and the frequency tracking result to realize synchronous correction between the signal data and the main reference signal;
s35, carrying out self-adaptive adjustment on the detection period of the detection node according to the frequency tracking result;
the image data is processed and analyzed by using a multi-layer recognition algorithm, and the output instrument data comprises the following steps:
s41, setting an image caching strategy for a remote management center to cache the image data in a queue;
s42, extracting the image data at the same time as the signal data in a buffer queue;
s43, preferentially utilizing an instrument identification model to identify the metering instrument in the image data, extracting a metering numerical value as instrument data, and judging the abnormality of the metering instrument;
s44, calling a fault identification model to identify the image data with numerical value abnormality, so as to realize fault type detection of the metering instrument and output instrument fault data;
the method for identifying the metering device in the image data by using the meter identification model preferentially, extracting a metering numerical value as meter data, and judging the abnormality of the metering device comprises the following steps:
s431, dividing the image data by using a division network model, and extracting an instrument area image;
s432, identifying the instrument area image by using a convolutional neural network model, extracting the deflection angle of a pointer of the metering instrument in the area, and calculating the pointer value as instrument data;
s433, comparing the instrument data with the average value of the first M instrument data, if the difference value of the instrument data and the first M instrument data is larger than a safety threshold value, judging that the metering instrument is abnormal, otherwise, calibrating the metering instrument as a normal metering instrument.
2. The metering device detection method based on dynamic multi-path synchronous detection according to claim 1, wherein the detection nodes are respectively configured with static IP addresses, and communication is performed between the detection nodes and a remote management center based on RTSP protocol.
3. The method for detecting a metering device based on dynamic multi-path synchronous detection according to claim 1, wherein the adaptive adjustment of the detection period of the detection node according to the frequency tracking result comprises the following steps:
s351, calculating a frequency change rate according to a frequency tracking result of the signal data;
and S352, when the frequency change rate is larger than a preset frequency threshold, reducing the detection period on the basis of the initial detection frequency, increasing the sampling frequency, and if the frequency change rate is smaller than or equal to the preset frequency threshold, maintaining the initial detection frequency unchanged.
4. The metering device detection method based on dynamic multi-path synchronous detection according to claim 1, wherein the step of setting an image buffer policy for a remote management center to perform queue buffer on the image data comprises the following steps:
s411, configuring a buffer queue for each detection node in the remote management center;
s412, setting a preset queue length for the buffer queue, storing the image frames of the image data received in real time into the buffer queue, if the queue length is not greater than the preset queue length at this time, successfully storing the image data, and if the queue length is greater than the preset queue length at this time, rejecting the previous extracted and identified image frame and 2N adjacent image frames.
5. The method for detecting the metering device based on the dynamic multi-path synchronous detection according to claim 1, wherein the method for constructing the split network model comprises the following steps:
s4311, adopting a self-encoder as an infrastructure of the segmentation network model, wherein the self-encoder is provided with four convolution network modules, and each convolution network module comprises 3 multiplied by 3 convolution kernels;
s4312, introducing a cavity fusion layer at the output end of the self-encoder, wherein the cavity fusion layer comprises three cavity convolution kernels;
s4313, setting a loss function of the segmentation network model, wherein the loss function expression is as follows:
；
in the method, in the process of the invention,representing a tag value;
yrepresenting the predicted value;
nrepresenting the number of samples;
irepresenting the sample order;
qrepresenting the number of classifications.
6. The method for detecting a metering device based on dynamic multi-path synchronous detection according to claim 5, wherein the calling the fault recognition model recognizes the image data with numerical value abnormality, realizes the fault type detection of the metering device, and outputs the fault data of the metering device, comprising the following steps:
s441, selecting the front N image frames and the rear N image frames of the abnormal image data, and carrying out segmentation processing on the abnormal image by utilizing the segmentation network model to obtain 2N instrument sub-images;
s442, positioning and identifying instrument pointers in the instrument sub-images, respectively extracting deflection angles of each instrument pointer, and combining the deflection angles according to a time sequence to obtain an angle set;
s443, calculating the average angle, deflection frequency and amplitude variation of all deflection angles in the angle set, judging that the pointer is in offset fault if the average angle exceeds an angle threshold value, judging that the pointer is in vibration fault if the deflection frequency exceeds a deflection threshold value, and judging that the pointer is in clamping stagnation fault if the amplitude variation is lower than an amplitude threshold value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311305344.9A CN117029900B (en) | 2023-10-10 | 2023-10-10 | Metering instrument detection method based on dynamic multipath synchronous detection |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311305344.9A CN117029900B (en) | 2023-10-10 | 2023-10-10 | Metering instrument detection method based on dynamic multipath synchronous detection |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117029900A CN117029900A (en) | 2023-11-10 |
| CN117029900B true CN117029900B (en) | 2024-01-02 |
Family
ID=88624936
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311305344.9A Active CN117029900B (en) | 2023-10-10 | 2023-10-10 | Metering instrument detection method based on dynamic multipath synchronous detection |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117029900B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117914003B (en) * | 2024-03-19 | 2024-05-24 | 沈阳智帮电气设备有限公司 | Intelligent monitoring auxiliary method and system for box-type transformer based on cloud edge cooperation |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101255947A (en) * | 2008-04-07 | 2008-09-03 | 成都康斯博科技有限责任公司 | Metered system for automatically capturing fluid flux |
| CN101403632A (en) * | 2008-11-13 | 2009-04-08 | 中国石化江汉油田分公司江汉采油厂 | Dynamic multipath synchronization detection apparatus and method for metering device |
| CN104501859A (en) * | 2014-11-11 | 2015-04-08 | 中国市政工程华北设计研究总院有限公司 | Mobile integrated gas cooker comprehensive performance experiment tester and test method |
| CN105721546A (en) * | 2016-01-20 | 2016-06-29 | 冶金自动化研究设计院 | Multi-data integration service platform based on industrial Internet of Things (IoT) communication |
| CN208270032U (en) * | 2018-06-22 | 2018-12-21 | 天信仪表集团有限公司 | A kind of volume correction instrument epidemic disaster performance test apparatus |
| CN112414519A (en) * | 2020-10-23 | 2021-02-26 | 哈尔滨鹏信科技有限责任公司 | Anti-cheating monitoring system of trade handover flow metering instrument and application thereof |
-
2023
- 2023-10-10 CN CN202311305344.9A patent/CN117029900B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101255947A (en) * | 2008-04-07 | 2008-09-03 | 成都康斯博科技有限责任公司 | Metered system for automatically capturing fluid flux |
| CN101403632A (en) * | 2008-11-13 | 2009-04-08 | 中国石化江汉油田分公司江汉采油厂 | Dynamic multipath synchronization detection apparatus and method for metering device |
| CN104501859A (en) * | 2014-11-11 | 2015-04-08 | 中国市政工程华北设计研究总院有限公司 | Mobile integrated gas cooker comprehensive performance experiment tester and test method |
| CN105721546A (en) * | 2016-01-20 | 2016-06-29 | 冶金自动化研究设计院 | Multi-data integration service platform based on industrial Internet of Things (IoT) communication |
| CN208270032U (en) * | 2018-06-22 | 2018-12-21 | 天信仪表集团有限公司 | A kind of volume correction instrument epidemic disaster performance test apparatus |
| CN112414519A (en) * | 2020-10-23 | 2021-02-26 | 哈尔滨鹏信科技有限责任公司 | Anti-cheating monitoring system of trade handover flow metering instrument and application thereof |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117029900A (en) | 2023-11-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108981781B (en) | Predictive analysis system and method for analyzing and detecting machine sensor faults | |
| CN117029900B (en) | Metering instrument detection method based on dynamic multipath synchronous detection | |
| CN101534305A (en) | Method and system for detecting network flow exception | |
| KR20150083832A (en) | A device and method for dynamically measuring an environmental quality factor | |
| CN104020190A (en) | Method for operating a measurement station and measurement station | |
| CN106247173A (en) | The method and device of pipeline leakage testing | |
| CN110095179A (en) | A kind of detection of motor oscillating and diagnostic method | |
| KR101953558B1 (en) | Apparatus and Method for Fault Management of Smart Devices | |
| US20230203753A1 (en) | Method and apparatus for anomaly detection on a doctor blade of a papermaking machine, and computing device | |
| WO2023045338A1 (en) | Method for real-time data dynamic online rapid processing at edge end | |
| CN111665066A (en) | Equipment fault self-adaptive upper and lower early warning boundary generation method based on convolutional neural network | |
| CN111444233B (en) | Method for discovering environmental monitoring abnormal data based on duplicator neural network model | |
| CN114255784A (en) | Substation equipment fault diagnosis method based on voiceprint recognition and related device | |
| CN103454485B (en) | Process control loop current verification | |
| CN115393817A (en) | Multi-sensor information fusion agricultural machinery working environment identification method and device | |
| CN113219296B (en) | Transmission line traveling wave fault data segment self-adaptive intercepting method | |
| WO2023185827A1 (en) | Mechanical fault detection method and apparatus, and server-side device and storage medium | |
| CN117406027A (en) | Distribution network fault distance measurement method and system | |
| CN115588439B (en) | Fault detection method and device of voiceprint acquisition device based on deep learning | |
| CN1186648C (en) | Oil deposit test ground data acquisition and processing system and data acquisition and processing method and equipment thereof | |
| CN116466166A (en) | Fault detection method, system and device for large-screen equipment | |
| CN111327630B (en) | Attack detection and correction method based on holohedral symmetry polycythemic theory | |
| CN115905793A (en) | Track reproduction method based on Internet of things and curve fitting | |
| CN114993671A (en) | Vibration fault diagnosis method and system based on Q factor wavelet transform | |
| CN114742103A (en) | City monitoring data processing method and device based on Internet of things and storage medium |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |