CN113658603A - Intelligent fault diagnosis method for belt conveyor carrier roller based on audio frequency - Google Patents

Intelligent fault diagnosis method for belt conveyor carrier roller based on audio frequency Download PDF

Info

Publication number
CN113658603A
CN113658603A CN202110782579.1A CN202110782579A CN113658603A CN 113658603 A CN113658603 A CN 113658603A CN 202110782579 A CN202110782579 A CN 202110782579A CN 113658603 A CN113658603 A CN 113658603A
Authority
CN
China
Prior art keywords
data
carrier roller
audio
fault diagnosis
roller
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
CN202110782579.1A
Other languages
Chinese (zh)
Inventor
彭晨
李志朋
杨明锦
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.)
University of Shanghai for Science and Technology
Original Assignee
University of Shanghai for Science and Technology
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 University of Shanghai for Science and Technology filed Critical University of Shanghai for Science and Technology
Priority to CN202110782579.1A priority Critical patent/CN113658603A/en
Publication of CN113658603A publication Critical patent/CN113658603A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L19/00—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis
    • G10L19/02—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis using spectral analysis, e.g. transform vocoders or subband vocoders
    • G10L19/0212—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis using spectral analysis, e.g. transform vocoders or subband vocoders using orthogonal transformation
    • G10L19/0216—Speech or audio signals analysis-synthesis techniques for redundancy reduction, e.g. in vocoders; Coding or decoding of speech or audio signals, using source filter models or psychoacoustic analysis using spectral analysis, e.g. transform vocoders or subband vocoders using orthogonal transformation using wavelet decomposition
    • 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
    • G07—CHECKING-DEVICES
    • G07C—TIME OR ATTENDANCE REGISTERS; REGISTERING OR INDICATING THE WORKING OF MACHINES; GENERATING RANDOM NUMBERS; VOTING OR LOTTERY APPARATUS; ARRANGEMENTS, SYSTEMS OR APPARATUS FOR CHECKING NOT PROVIDED FOR ELSEWHERE
    • G07C3/00—Registering or indicating the condition or the working of machines or other apparatus, other than vehicles
    • G07C3/005—Registering or indicating the condition or the working of machines or other apparatus, other than vehicles during manufacturing process
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
    • G10L25/03—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the type of extracted parameters
    • G10L25/18—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the type of extracted parameters the extracted parameters being spectral information of each sub-band
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
    • G10L25/27—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the analysis technique
    • G10L25/30—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the analysis technique using neural networks
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
    • G10L25/48—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use
    • G10L25/51—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use for comparison or discrimination
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04Q—SELECTING
    • H04Q11/00—Selecting arrangements for multiplex systems
    • H04Q11/0001—Selecting arrangements for multiplex systems using optical switching
    • H04Q11/0005—Switch and router aspects
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04Q—SELECTING
    • H04Q11/00—Selecting arrangements for multiplex systems
    • H04Q11/0001—Selecting arrangements for multiplex systems using optical switching
    • H04Q11/0062—Network aspects
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04Q—SELECTING
    • H04Q9/00—Arrangements in telecontrol or telemetry systems for selectively calling a substation from a main station, in which substation desired apparatus is selected for applying a control signal thereto or for obtaining measured values therefrom
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04Q—SELECTING
    • H04Q2209/00—Arrangements in telecontrol or telemetry systems
    • H04Q2209/80—Arrangements in the sub-station, i.e. sensing device
    • H04Q2209/84—Measuring functions

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Computational Linguistics (AREA)
  • Health & Medical Sciences (AREA)
  • Acoustics & Sound (AREA)
  • Multimedia (AREA)
  • Human Computer Interaction (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Signal Processing (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Theoretical Computer Science (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Computation (AREA)
  • Spectroscopy & Molecular Physics (AREA)
  • General Physics & Mathematics (AREA)
  • Computing Systems (AREA)
  • Software Systems (AREA)
  • Mathematical Physics (AREA)
  • Manufacturing & Machinery (AREA)
  • General Engineering & Computer Science (AREA)
  • Molecular Biology (AREA)
  • General Health & Medical Sciences (AREA)
  • Data Mining & Analysis (AREA)
  • Biophysics (AREA)
  • Biomedical Technology (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)

Abstract

The invention provides an intelligent fault diagnosis method for a belt conveyor carrier roller based on audio. The method comprises the following steps: firstly, an audio sensor is arranged beside a carrier roller to acquire an audio signal of the carrier roller, and the audio signal is sent to a server through an optical fiber; secondly, preprocessing the transmitted carrier roller audio signals on a processor; then, extracting a mean value of each frequency band as a data characteristic of the frequency band; and finally, classifying the data after the characteristic extraction by using a convolutional neural network. And finally, displaying the fault diagnosis on an upper computer interface for a user to check after the fault diagnosis is finished by the diagnosis algorithm. The invention can detect the roller fault in real time without depending on manpower, and improves the economic benefit and the intelligent level of the coal preparation plant.

Description

Intelligent fault diagnosis method for belt conveyor carrier roller based on audio frequency
Technical Field
The invention relates to the field of fault diagnosis of large mechanical equipment of an industrial automatic production line, in particular to an intelligent fault diagnosis method for a belt conveyor carrier roller based on audio frequency.
Background
The carrier roller is the most used, the most trouble, the most maintenance part among the coal preparation factory belt conveyor. The carrier roller is easy to have the faults of eccentricity, jamming, breakage and the like in work, so that the belt is deviated, and the normal use of factory equipment is seriously influenced.
The traditional fault diagnosis of the carrier roller is a manual inspection method, and workers need to be arranged to regularly inspect the carrier roller. The manual inspection method is time-consuming and labor-consuming, and fails to find faults in time. In recent years, fault diagnosis methods have been rapidly developed, and the conventional manual diagnosis has been continuously developed to intelligent diagnosis. The basic process of the data-based fault diagnosis method mainly comprises signal processing, feature extraction and fault classification. Common signal processing methods are Empirical Mode Decomposition (EMD), fourier transform, wavelet packet transform, and the like. The fault classification is to classify faults by using preprocessed data, and common classification methods include a support vector machine, a cluster, a neural network and the like.
In the case of the hoggots and the like, the carrier roller faults are diagnosed by performing time domain analysis on carrier roller audio data to extract root mean square, positive peak values, negative peak values and the like and performing FFT (fast Fourier transform) peak value detection, but the FFT is not suitable for analyzing non-stationary signals and cannot perform good time frequency analysis, and the carrier roller working site environment is complex and has a large number of non-stationary signals.
Li Wei et al diagnose idler jam faults using a wavelet packet and a Support Vector Machine (SVM). The wavelet packet conversion has good time-frequency resolution, and can effectively analyze the signals of the carrier roller; but SVMs have difficulty in processing large-scale sample data and performing multi-classification.
Therefore, the conventional roller fault diagnosis method cannot simultaneously diagnose multiple faults and has poor real-time performance.
Disclosure of Invention
The invention aims to provide an intelligent fault diagnosis method for a belt conveyor carrier roller based on audio frequency, and aims to solve the problems that various carrier roller faults cannot be detected, the real-time performance is low and the like in the prior art.
In order to achieve the above object, the idea of the present invention is:
the invention provides an intelligent fault diagnosis method for a belt conveyor carrier roller based on audio. Firstly, an audio sensor is arranged beside a carrier roller to acquire an audio signal of the carrier roller, and the audio signal is sent to a server through an optical fiber; secondly, the transmitted carrier roller audio signals are preprocessed on a processor, namely carrier roller audio data are decomposed through wavelet packets, and the carrier roller audio data are decomposed into 256 frequency bands by using 8 layers of second-order Daubechies wavelets as wavelet basis functions. Because the roller fault information is mainly shown in high-frequency data, the low-frequency data in the roller data accounts for too much. Therefore, the energy ratio of the low-frequency data is reduced, and then an average value is extracted from each frequency band to serve as the data characteristic of the frequency band; and finally, classifying the data after the characteristic extraction by using a convolutional neural network. Since the convolutional neural network requires two-dimensional data input, 256 data in one dimension are first changed into 16 × 16 data in two dimensions. The convolutional neural network used in the method is 5 layers and comprises two convolutional layers, two pooling layers and an output layer, wherein the output layer is provided with 3 nodes and represents that carrier roller audio data are divided into three types, namely normal, abnormal and fault. And after the fault diagnosis is completed, the diagnosis algorithm displays the fault diagnosis on an upper computer interface for a user to check.
In terms of hardware, an Arduino development board is used for controlling an LM386 sound sensor to collect sound information, the LM386 is a common audio sensor and can collect audio data of 50-20000 Hz; then, data is transmitted to the switch by using an Arduino Ethernet W5100, wherein the Arduino Ethernet W5100 is a network module based on ATmega328, the function of the network module is to connect an Arduino development board to the Internet, and the digital and analog port data of the Arduino are read through the network; after receiving the data of the sensors, the switch sends the data to the router through the optical fiber and then forwards the data to the server for data processing. The optical fiber is a common communication medium in network transmission, the broadcast is used as an information carrier, information to be transmitted is converted into an optical signal to be transmitted, the optical fiber selected by the method is a GJYXCH-2SC-100S single-mode single-core optical fiber, and the optical fiber has the advantages of long transmission distance, good flexibility, large communication capacity, small signal crosstalk and good confidentiality.
The method uses 10 audio sensors in total, and divides the audio sensors into 2 groups to be connected with 2 switches, 5 audio acquisition parts in each group are sequentially connected to 5 ports of the switch through RJ-45 serial ports by using network cables, and then are connected to an optical fiber transceiver through one port of the switch in a centralized manner by using the network cables; and finally, the two groups of optical fiber transceivers transmit the signals acquired by the acquisition parts of the 10 audio sensors to an upper computer through one router for processing.
According to the conception, the technical scheme of the invention is as follows:
an intelligent fault diagnosis method for a belt conveyor carrier roller based on audio comprises the following steps:
step 1: installing an LM386 and an Arduino Ethernet W1500 on the Arduino, and then installing each peripheral Arduino beside a carrier roller;
step 2: connecting Arduino to an exchanger through an optical fiber, so that a plurality of sensors are connected to the exchanger together, and then transmitting signals;
and step 3: transmitting the carrier roller audio data to a terminal server through a switch and a router, and simultaneously storing the carrier roller audio data in a database;
and 4, step 4: pretreatment of carrier roller audio data: preprocessing carrier roller audio data by utilizing a wavelet packet algorithm, and dividing the audio data into a plurality of frequency bands;
and 5: carrying out characteristic extraction on carrier roller audio data: firstly, adjusting data of the lowest frequency band, and then extracting an average value of each frequency band after the carrier roller wavelet packet conversion as a data characteristic;
step 6: and (3) judging the carrier roller state: 256 characteristic values are extracted from each group of carrier roller data, the 256 characteristic values are input into a 5-layer convolutional neural network for judgment, and the state of the carrier roller is judged through the output of the convolutional neural network;
and 7: and outputting the real-time state of the carrier roller to an upper computer interface.
The step 1 comprises the following steps:
step 1.1: firstly, writing an LM386 control program and a data acquisition program, then burning the LM386 control program and the data acquisition program into Arduino, and connecting the LM386 to a proper port;
step 1.2: connecting an Arduino board with a power supply, and installing an LM386 beside the carrier roller so as to clearly acquire carrier roller sound data in real time;
step 1.3: installing an Arduino Ethernet W5100 network expansion module on the Arduino and connecting an optical fiber so as to convert the carrier roller sound data into a network signal and transmit the network signal;
step 1.4: encapsulate Arduino board and peripheral hardware in a box, then will wholly install on the belt feeder, real-time collection reaches the sound data of bearing roller operation to go out data transmission through optic fibre.
The step 2 comprises the following steps:
step 2.1: transmitting the output of each switch to a router through an optical fiber;
step 2.2: and connecting the router to a terminal server, receiving the data of all the carrier rollers through the terminal server, and performing fault diagnosis.
The step 3 comprises the following steps:
step 3.1: connecting each switch to a terminal server through an optical fiber, so that all carrier roller audio data are transmitted to the terminal server for fault diagnosis;
step 3.2: the terminal server stores the carrier roller audio data in a database for subsequent use;
step 3.3: and diagnosing the roller state by the roller audio data every 5 minutes.
The step 4 comprises the following steps:
step 4.1: firstly, acquiring sound data of a carrier roller on a terminal server, wherein the frequency of the acquired sound data is 44100Hz, and diagnosing the data once every 5 minutes; through observation of different kinds of data, when the carrier roller has a fault, the high-frequency signal of sound is increased, and the amplitude of the sound signal is also increased, so that the high-frequency part of the signal contains a lot of fault information;
step 4.2: carrying out pretreatment on each group of carrier roller audio data by using wavelet packet transformation; adopting 8 layers of wavelet packet transformation, and obtaining 256 frequency bands for each group of carrier roller audio data; the wavelet packet transforms all wavelet basis functions into the db2 wavelet in Daubechies wavelets, i.e., using second-order vanishing moments.
The step 5 comprises the following steps:
step 5.1: after 8 layers of wavelet packet transformation, each group of audio data can obtain 256 frequency band data, and a characteristic value of the data is extracted for each frequency band;
step 5.2: the data of the lowest frequency band after wavelet packet transformation is adjusted to be the same as the data of the second lowest frequency band, the occupation ratio of the low frequency part in the adjusted wavelet packet frequency spectrum is obviously reduced, the influence of the low frequency part on signal analysis is weakened, and the high frequency part can be more accurately analyzed;
step 5.3: when the fault of the carrier roller occurs, the fluctuation of data becomes fast, the fluctuation amplitude becomes large, and therefore the average value of the data changes; the average value of the audio data when the carrier roller fails is different from that when the carrier roller is normal; and extracting the average value of each frequency band after wavelet packet transformation as the data characteristic of each frequency band.
The step 6 comprises the following steps:
step 6.1: each group of carrier roller audio data can obtain 256 frequency bands after wavelet packet conversion, and a characteristic value is extracted from each frequency band to obtain 256 data; transforming the input data into two dimensions, i.e. 16 x 16;
step 6.2: inputting 16 × 16 data into a convolutional neural network, wherein the convolutional neural network structure is improved based on Lenet-5 and is a convolutional neural network with 5 layers, the first layer is a convolutional layer with 3 × 3, the second layer is a pooling layer with 2 × 2, the third layer is a convolutional layer with 4 × 4, and the fifth layer is a fully-connected layer, and outputting a group of 0/1 data with 1 × 3;
step 6.3: and judging the state of the carrier roller at the moment according to the output of the convolutional neural network, wherein 001 represents a normal state, 010 represents an abnormal state and 100 represents a fault state.
The step 7 comprises the following steps:
step 7.1: after the convolutional neural network diagnoses the faults of the carrier rollers, transmitting the state information of each carrier roller to upper computer software;
step 7.2: the upper computer interface comprises various information and data of roller fault diagnosis, and simultaneously displays the state information of each roller and the fault information of the current day and month in real time;
step 7.3: when a carrier roller fails, the upper computer software can send alarm information to prompt a user that the carrier roller fails, and the position information and the state of the failed carrier roller are displayed on an interface.
Compared with the prior art, the invention has the following advantages:
1. compared with the traditional manual detection, the carrier roller fault diagnosis efficiency is greatly improved, and the carrier roller faults can be divided into various types by using a convolutional neural network algorithm. According to the system, through analyzing the energy spectrum of each frequency band after wavelet packet decomposition, the data of the lowest frequency band is creatively adjusted, the influence of a low-frequency part containing less fault information on the classification effect is reduced, and the accuracy of carrier roller fault classification is improved.
Drawings
FIG. 1 is a general flow diagram of the present invention.
Fig. 2 is a diagram of a sound collection device used in the present invention.
Fig. 3 is a hardware configuration diagram of the present invention.
Fig. 4 is an energy spectrum of the idler audio data after wavelet packet transformation.
Fig. 5 is a diagram of a convolutional neural network used in the present invention.
Detailed Description
The following describes in further detail specific embodiments of the present invention with reference to the accompanying drawings.
The invention provides an intelligent fault diagnosis method for a belt conveyor carrier roller based on audio, which comprises the following steps of:
step 1: installing an LM386 and an Arduino Ethernet W1500 on the Arduino, and then installing each peripheral Arduino beside a carrier roller;
step 2: connecting Arduino to an exchanger through an optical fiber, so that a plurality of sensors are connected to the exchanger together, and then transmitting signals;
and step 3: transmitting the carrier roller audio data to a terminal server through a switch and a router, and simultaneously storing the carrier roller audio data in a database;
and 4, step 4: pretreatment of carrier roller audio data: preprocessing carrier roller audio data by utilizing a wavelet packet algorithm, and dividing the audio data into a plurality of frequency bands;
and 5: carrying out characteristic extraction on carrier roller audio data: firstly, adjusting data of the lowest frequency band, and then extracting an average value of each frequency band after the carrier roller wavelet packet conversion as a data characteristic;
step 6: and (3) judging the carrier roller state: 256 characteristic values are extracted from each group of carrier roller data, the 256 characteristic values are input into a 5-layer convolutional neural network for judgment, and the state of the carrier roller is judged through the output of the convolutional neural network;
and 7: and outputting the real-time state of the carrier roller to an upper computer interface.
On the basis of the scheme, the specific steps of the step 1 are as follows:
step 1.1: firstly, writing an LM386 control program and a data acquisition program, then burning the LM386 control program and the data acquisition program into Arduino, and connecting the LM386 to a proper port;
step 1.2: connecting an Arduino board with a power supply, and installing an LM386 beside the carrier roller so as to clearly acquire carrier roller sound data in real time;
step 1.3: installing an Arduino Ethernet W5100 network expansion module on the Arduino and connecting an optical fiber so as to convert the carrier roller sound data into a network signal and transmit the network signal;
step 1.4: the Arduino board and the peripheral equipment are packaged into a box, then the whole body is installed on the belt conveyor, and the optical fiber is pulled out, so that the wireless communication is easily interfered due to more field interference of the carrier roller in work, and the stability and the safety of the communication are ensured by using the optical fiber for wired communication; as shown in figure 2, the sound sensor assembly is arranged near the carrier roller, so that sound data of carrier roller operation can be collected in real time, and the data can be transmitted out through an optical fiber.
On the basis of the scheme, the specific steps of the step 2 are as follows:
step 2.1: transmitting the output of each switch to a router through an optical fiber;
step 2.2: and the router is connected to the terminal server, so that the data of all the carrier rollers can be received by one terminal server, and fault diagnosis is carried out. The complete hardware architecture of the present invention is shown in fig. 3.
On the basis of the scheme, the specific steps of the step 3 are as follows:
step 3.1: connecting each switch to a terminal server through an optical fiber, so that all carrier roller audio data can be transmitted to the terminal server for fault diagnosis;
step 3.2: the transmitted carrier roller audio data are stored in a database on a terminal server for subsequent use, and the used database is a MySQL database;
step 3.3: and diagnosing the carrier roller state through carrier roller audio data every 5 min.
On the basis of the scheme, the specific steps of the step 4 are as follows:
step 4.1: firstly, acquiring sound data of a carrier roller on a terminal server, wherein the frequency of the acquired sound data is 44100Hz, and diagnosing the data every 5 min; through observation of different types of data, the carrier roller fault can be found, the high-frequency signal of sound can be increased, the amplitude of the sound signal can also be increased, and therefore the high-frequency part of the signal contains a lot of fault information.
Step 4.2: and preprocessing each group of carrier roller audio data by using wavelet packet transformation. The wavelet packet transform can decompose the high-frequency part which is not subdivided by the wavelet transform, and can extract the information of the high-frequency part of the signal. The wavelet packet transform can also select a frequency band that fits the signal spectrum based on the analyzed data characteristics. Two frequency bands, namely a low-frequency part and a high-frequency part, can be obtained after the wavelet packet transformation of the first layer, and when the second layer is decomposed, the data of the two frequency bands obtained by the first layer are subjected to similar decomposition and are sequentially decomposed. In the method, 8 layers of wavelet packet transformation are adopted, and 256 frequency bands can be obtained by each group of carrier roller audio data. The wavelet packet transforms all wavelet basis functions into the db2 wavelet in Daubechies wavelets, i.e., using second-order vanishing moments.
On the basis of the scheme, the specific steps of the step 5 are as follows:
step 5.1: after 8 layers of wavelet packet transformation, each group of audio data can obtain 256 frequency band data, if the frequency band data are directly used for classification, the data volume is too large, and the data characteristics are not obvious enough, so that the characteristic value of the data is extracted for each frequency band.
Step 5.2: as can be seen from the wavelet packet transform energy spectrum shown in fig. 4, the energy content of the low frequency data is too large. However, since most of the failure information of the carrier roller is included in the high-frequency component, the data of the lowest frequency band after wavelet packet conversion is adjusted to be the same as the data of the second lowest frequency band. The occupation ratio of the low-frequency part in the adjusted wavelet packet frequency spectrum is obviously reduced, the influence of the low-frequency part on signal analysis is weakened, and more accurate analysis on the high-frequency part is facilitated.
Step 5.3: when the idler failure happens, the fluctuation of data becomes fast, the fluctuation amplitude becomes large, and therefore the average value of the data can be changed. The average value can represent the centralized trend of the data and reflect the average value of the data of each frequency band after wavelet packet decomposition. The average value of the audio data when the carrier roller fails is different from that when the carrier roller is normal. Therefore, the method extracts the mean value of each frequency band after wavelet packet transformation as the data characteristic of each frequency band.
On the basis of the scheme, the specific steps of the step 6 are as follows:
step 6.1: each group of carrier roller audio data can obtain 256 frequency bands after wavelet packet conversion, and a characteristic value is extracted from each frequency band to obtain 256 data. However, the data at this time is one-dimensional, and the convolutional neural network used in the method needs two-dimensional data input, so that the input data needs to be firstly changed into two-dimensional data, namely 16 × 16;
step 6.2: inputting 16 × 16 data into a convolutional neural network, wherein the structure of the convolutional neural network used in the method is improved based on Lenet-5, as shown in FIG. 5, the method uses a convolutional neural network with 5 layers, the first layer is a convolutional layer with 3 × 3, the second layer is a pooling layer with 2 × 2, the third layer is a convolutional layer with 4 × 4, and the fifth layer is a fully-connected layer, and a set of 0/1 data with 1 × 3 can be output;
step 6.3: the state of the carrier roller at the moment can be judged according to the output of the convolutional neural network, wherein 001 represents a normal state, 010 represents an abnormal state, and 100 represents a fault state.
On the basis of the scheme, the specific steps of the step 7 are as follows:
step 7.1: after the convolutional neural network diagnoses the roller fault, the roller fault needs to be displayed on an upper computer interface for a user to check, and the state information of each roller is transmitted to upper computer software.
Step 7.2: the upper computer interface comprises various information and data of fault diagnosis of the carrier roller, wherein the information comprises position information of a belt conveyor, a carrier roller model, a sensor and the like; meanwhile, the state information of each carrier roller and the fault information of the current day and the current month can be displayed in real time.
Step 7.3: when a carrier roller fails, the upper computer software can send alarm information to prompt a user that the carrier roller fails, and the position information and the state of the failed carrier roller are displayed on an interface.
The steps 1 to 7 finish the diagnosis of the screen plate falling fault, and simultaneously determine the position of the fault point.
Finally, it should be noted that: the above embodiments are only used to illustrate the technical solution of the present invention, and not to limit the same; while the invention has been described in detail and with reference to the foregoing embodiments, it will be understood by those skilled in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some or all of the technical features may be equivalently replaced; and the modifications or the substitutions do not make the essence of the corresponding technical solutions depart from the scope of the technical solutions of the embodiments of the present invention.

Claims (8)

1. An intelligent fault diagnosis method for a belt conveyor carrier roller based on audio is characterized by comprising the following steps:
step 1: installing an LM386 and an Arduino Ethernet W1500 on the Arduino, and then installing each peripheral Arduino beside a carrier roller;
step 2: connecting Arduino to an exchanger through an optical fiber, so that a plurality of sensors are connected to the exchanger together, and then transmitting signals;
and step 3: transmitting the carrier roller audio data to a terminal server through a switch and a router, and simultaneously storing the carrier roller audio data in a database;
and 4, step 4: pretreatment of carrier roller audio data: preprocessing carrier roller audio data by utilizing a wavelet packet algorithm, and dividing the audio data into a plurality of frequency bands;
and 5: carrying out characteristic extraction on carrier roller audio data: firstly, adjusting data of the lowest frequency band, and then extracting an average value of each frequency band after the carrier roller wavelet packet conversion as a data characteristic;
step 6: and (3) judging the carrier roller state: 256 characteristic values are extracted from each group of carrier roller data, the 256 characteristic values are input into a 5-layer convolutional neural network for judgment, and the state of the carrier roller is judged through the output of the convolutional neural network;
and 7: and outputting the real-time state of the carrier roller to an upper computer interface.
2. The intelligent fault diagnosis method for the idler rollers of the audio-based belt conveyor according to claim 1, wherein the step 1 comprises the following steps:
step 1.1: firstly, writing an LM386 control program and a data acquisition program, then burning the LM386 control program and the data acquisition program into Arduino, and connecting the LM386 to a proper port;
step 1.2: connecting an Arduino board with a power supply, and installing an LM386 beside the carrier roller so as to clearly acquire carrier roller sound data in real time;
step 1.3: installing an Arduino Ethernet W5100 network expansion module on the Arduino and connecting an optical fiber so as to convert the carrier roller sound data into a network signal and transmit the network signal;
step 1.4: encapsulate Arduino board and peripheral hardware in a box, then will wholly install on the belt feeder, real-time collection reaches the sound data of bearing roller operation to go out data transmission through optic fibre.
3. The audio-based intelligent fault diagnosis method for belt conveyor idler rollers according to claim 1, characterized in that the step 2 comprises the following steps:
step 2.1: transmitting the output of each switch to a router through an optical fiber;
step 2.2: and connecting the router to a terminal server, receiving the data of all the carrier rollers through the terminal server, and performing fault diagnosis.
4. The audio-based intelligent fault diagnosis method for belt conveyor idler rollers according to claim 1, characterized in that said step 3 comprises the following steps:
step 3.1: connecting each switch to a terminal server through an optical fiber, so that all carrier roller audio data are transmitted to the terminal server for fault diagnosis;
step 3.2: the terminal server stores the carrier roller audio data in a database for subsequent use;
step 3.3: and diagnosing the roller state by the roller audio data every 5 minutes.
5. The audio-based intelligent fault diagnosis method for belt conveyor idler rollers according to claim 1, characterized in that the step 4 comprises the following steps:
step 4.1: firstly, acquiring sound data of a carrier roller on a terminal server, wherein the frequency of the acquired sound data is 44100Hz, and diagnosing the data once every 5 minutes; through observation of different kinds of data, when the carrier roller has a fault, the high-frequency signal of sound is increased, and the amplitude of the sound signal is also increased, so that the high-frequency part of the signal contains a lot of fault information;
step 4.2: carrying out pretreatment on each group of carrier roller audio data by using wavelet packet transformation; adopting 8 layers of wavelet packet transformation, and obtaining 256 frequency bands for each group of carrier roller audio data; the wavelet packet transforms all wavelet basis functions into the db2 wavelet in Daubechies wavelets, i.e., using second-order vanishing moments.
6. The audio-based intelligent fault diagnosis method for belt conveyor idler rollers according to claim 1, characterized in that said step 5 comprises the following steps:
step 5.1: after 8 layers of wavelet packet transformation, each group of audio data can obtain 256 frequency band data, and a characteristic value of the data is extracted for each frequency band;
step 5.2: the data of the lowest frequency band after wavelet packet transformation is adjusted to be the same as the data of the second lowest frequency band, the occupation ratio of the low frequency part in the adjusted wavelet packet frequency spectrum is obviously reduced, the influence of the low frequency part on signal analysis is weakened, and the high frequency part can be more accurately analyzed;
step 5.3: when the fault of the carrier roller occurs, the fluctuation of data becomes fast, the fluctuation amplitude becomes large, and therefore the average value of the data changes; the average value of the audio data when the carrier roller fails is different from that when the carrier roller is normal; and extracting the average value of each frequency band after wavelet packet transformation as the data characteristic of each frequency band.
7. The audio-based intelligent fault diagnosis method for belt conveyor idler rollers according to claim 1, characterized in that said step 6 comprises the following steps:
step 6.1: each group of carrier roller audio data can obtain 256 frequency bands after wavelet packet conversion, and a characteristic value is extracted from each frequency band to obtain 256 data; transforming the input data into two dimensions, i.e. 16 x 16;
step 6.2: inputting 16 × 16 data into a convolutional neural network, wherein the convolutional neural network structure is improved based on Lenet-5 and is a convolutional neural network with 5 layers, the first layer is a convolutional layer with 3 × 3, the second layer is a pooling layer with 2 × 2, the third layer is a convolutional layer with 4 × 4, and the fifth layer is a fully-connected layer, and outputting a group of 0/1 data with 1 × 3;
step 6.3: and judging the state of the carrier roller at the moment according to the output of the convolutional neural network, wherein 001 represents a normal state, 010 represents an abnormal state and 100 represents a fault state.
8. The audio-based intelligent fault diagnosis method for belt conveyor idler rollers according to claim 1, characterized in that said step 7 comprises the following steps:
step 7.1: after the convolutional neural network diagnoses the faults of the carrier rollers, transmitting the state information of each carrier roller to upper computer software;
step 7.2: the upper computer interface comprises various information and data of roller fault diagnosis, and simultaneously displays the state information of each roller and the fault information of the current day and month in real time;
step 7.3: when a carrier roller fails, the upper computer software can send alarm information to prompt a user that the carrier roller fails, and the position information and the state of the failed carrier roller are displayed on an interface.
CN202110782579.1A 2021-07-12 2021-07-12 Intelligent fault diagnosis method for belt conveyor carrier roller based on audio frequency Pending CN113658603A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110782579.1A CN113658603A (en) 2021-07-12 2021-07-12 Intelligent fault diagnosis method for belt conveyor carrier roller based on audio frequency

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110782579.1A CN113658603A (en) 2021-07-12 2021-07-12 Intelligent fault diagnosis method for belt conveyor carrier roller based on audio frequency

Publications (1)

Publication Number Publication Date
CN113658603A true CN113658603A (en) 2021-11-16

Family

ID=78477219

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110782579.1A Pending CN113658603A (en) 2021-07-12 2021-07-12 Intelligent fault diagnosis method for belt conveyor carrier roller based on audio frequency

Country Status (1)

Country Link
CN (1) CN113658603A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115096375A (en) * 2022-08-22 2022-09-23 启东亦大通自动化设备有限公司 Carrier roller running state monitoring method and device based on carrier roller carrying trolley detection

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115096375A (en) * 2022-08-22 2022-09-23 启东亦大通自动化设备有限公司 Carrier roller running state monitoring method and device based on carrier roller carrying trolley detection
CN115096375B (en) * 2022-08-22 2022-11-04 启东亦大通自动化设备有限公司 Carrier roller running state monitoring method and device based on carrier roller carrying trolley detection

Similar Documents

Publication Publication Date Title
CN112989712B (en) Aeroengine fault diagnosis method based on 5G edge calculation and deep learning
CN111637964B (en) Rail corrugation identification method
CN111947928B (en) Multi-source information fusion bearing fault prediction system and method
CN111623868B (en) Convolutional neural network construction method for rail corrugation identification
WO2019015311A1 (en) Vibration signal support vector machine-based gil fault online monitoring system
CN104386449B (en) On-line checking intelligent protection device is taken turns end to end for mining belt conveyer
CN109084835A (en) A kind of electric machines test and remote monitoring system based on PXIe test platform
CN110503977A (en) A kind of substation equipment audio signal sample analysis system
CN113298134B (en) System and method for remotely and non-contact health monitoring of fan blade based on BPNN
CN112986870A (en) Distributed power transformer winding state monitoring method and system based on vibration method
CN112964304A (en) Remote operation and maintenance method and system for key parts of rotary machine
CN115238785A (en) Rotary machine fault diagnosis method and system based on image fusion and integrated network
CN113551765A (en) Sound spectrum analysis and diagnosis method for equipment fault
CN113567162A (en) Fan fault intelligent diagnosis device and method based on acoustic sensor
CN111401661B (en) Predictive maintenance method and maintenance system for mechanical equipment
CN110837532A (en) Method for detecting electricity stealing behavior of charging pile based on big data platform
CN113962308A (en) Aviation equipment fault prediction method
CN113658603A (en) Intelligent fault diagnosis method for belt conveyor carrier roller based on audio frequency
CN111878322A (en) Wind power generator device
CN208076016U (en) A kind of GIL On-line Faults monitoring system based on vibration signal support vector machines
CN117782198A (en) Highway electromechanical equipment operation monitoring method and system based on cloud edge architecture
CN112801033A (en) AlexNet network-based construction disturbance and leakage identification method along long oil and gas pipeline
CN116680641A (en) User electricity consumption abnormality detection method based on machine learning algorithm
CN214407560U (en) Electromechanical device detection and maintenance system
CN115711206B (en) Wind driven generator blade icing state monitoring system based on clustering weight

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