Abnormal state detection method and system of bucket wheel of portal bucket wheel excavator based on audio signal

Abstract

The present invention provides a method and system for detecting abnormal state of bucket wheel of portal bucket wheel machine based on audio signal. The method of the present invention collects on-site audio data in real time by installing and deploying sound sensors near the bucket wheel, and finds abnormal working conditions by filtering and frequency domain transformation of audio signals, thereby alarming the operation state of bucket wheel machine, solving the problem that the operator can timely find abnormal state of bucket wheel after the bucket wheel machine is unattended.

Description

Technical Field

[0001] The present invention relates to the technical field of thermal power generation, and in particular to a method and system for detecting an abnormal state of a bucket wheel of a portal bucket wheel machine based on an audio signal. Background Art

[0002] In accordance with the business needs of reducing costs and increasing efficiency in the thermal power generation industry, most companies are currently carrying out or have completed the unmanned transformation of bucket wheel excavators. Since the unmanned system was put into use, since there is no one operating the equipment, the abnormal working conditions of the bucket wheel part cannot be discovered by the operating personnel in the first time, which can easily lead to the overall functional failure of the bucket wheel excavator. In severe cases, it will cause the equipment to shut down for a long time, resulting in serious economic losses. Summary of the invention

[0003] According to the technical problems raised above, a method and system for detecting abnormal state of bucket wheel of portal bucket wheel excavator based on audio signal is provided. The technical means adopted by the present invention are as follows:

[0004] A method for detecting abnormal state of a bucket wheel of a portal bucket wheel excavator based on an audio signal comprises the following steps:

[0005] Install and deploy sound sensors near the bucket wheel of the gantry bucket wheel excavator to collect on-site audio data;

[0006] Preprocessing the collected audio data to convert the continuous time signal into a discrete time signal, wherein the preprocessing includes signal discretization processing, original discrete audio data framing processing, and Fourier frequency domain conversion processing, thereby obtaining the frequency domain characteristics of the audio signal;

[0007] The pre-processed frequency domain features are compared with the preset standards. If there is any abnormality, an alarm is sent to the control system.

[0008] Furthermore, the on-site audio data is connected to the main control system through the OPC ua communication method. The sound sensor collects a section of audio data in a period of 3 seconds and generates a file in PCM format.

[0009] Furthermore, before preprocessing the collected audio data, the following steps are performed:

[0010] Detect whether the bucket wheel is in the started state. If it is in the started state, the on-site audio data will be transmitted to the main control system; if it is not in the started state, the file generated by the sound sensor will not enter the main control system for analysis, and the next audio data will be recorded again.

[0011] Furthermore, during the preprocessing process, a sampling frequency of 16 kHz was selected to discretize the signal.

[0012] Further, the original discrete audio data in the preprocessing process is framed, and the specific steps are as follows:

[0013] Frame the original discrete audio data using a rectangular window:

[0014] x n [m] = w(m)x[n + m] (1 - 1)

[0015] The rectangular window therein is expressed as:

[0016]

[0017] The short - time average amplitude is expressed as:

[0018]

[0019] When the value of the short - time amplitude M n exceeds the preset value, it corresponds to the voiced segment. When the value of the short - time amplitude M n is less than the preset value, it corresponds to the unvoiced segment. When the number of voiced segments increases and the average amplitude value is too large, the pre - processed data is sent to the Fourier transform module for secondary analysis.

[0020] Further, the comparison of the frequency - domain features based on the preprocessing with the preset standard is specifically as follows: Statistically analyze the discrete frequency - domain features. When the data within 0 - 50000 HZ is recorded as valid data, sample and divide the above - mentioned frequency range, intercept 499 features with 100 Hz as the segmentation unit. The re - screening rules are as follows:

[0021] (1) Select the median M of the amplitudes of the feature points and compare from the low - frequency band to the high - frequency band;

[0022] (2) The amplitude A i (0 < i < 500); A i > M is one comparison result, and A i < M is another comparison result

[0023] (3) When one comparison result is different from the previous comparison result, record it as 1 change. After all comparisons are completed, count the number of times n that the comparison results change. When n > 30, it indicates that the sound frequency has a large disturbance, and it is judged as an abnormal sound.

[0024] Further, while alarming to the control system when an abnormality occurs, the operator reads the abnormal sound signal, and the system plays the real - time audio signal at the scene for the operator to confirm the fault.

[0025] The present invention also discloses a bucket wheel abnormal state detection system for a gantry bucket wheel machine based on audio signals, including:

[0026] Sound sensors to collect live audio data near the bucket wheel of a gantry bucket wheel machine;

[0027] The sound signal processing unit is used to convert the collected on-site audio data into discrete time signals, which specifically includes a signal discretization processing module, an original discrete audio data frame processing module, and a Fourier frequency domain conversion processing module to obtain the frequency domain characteristics of the audio signal;

[0028] The sound signal judgment unit is used to compare the frequency domain characteristics preprocessed by the sound signal processing unit with the preset standard and issue an alarm based on the abnormal judgment result.

[0029] The present invention collects on-site audio data in real time by installing and deploying sound sensors near the bucket wheel, and finds abnormal operating conditions by filtering the audio signal, transforming the frequency domain, and other processing methods, thereby alarming the operating status of the bucket wheel machine. This makes up for the problem of detecting and diagnosing abnormal working conditions of the bucket wheel after the existing bucket wheel machine is unmanned. There is no need for regular manual inspections, and the audio signal of the bucket wheel of the bucket wheel machine can be used to remotely judge whether the bucket wheel is in a normal working condition, providing reliable technical support for the unmanned bucket wheel machine. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0031] Figure 1 The present invention is a flow chart of a method for detecting abnormal state of a bucket wheel of a gantry bucket wheel excavator based on audio signals.

[0032] Figure 2 This is a frequency domain analysis diagram of the bucket wheel in normal operation in an embodiment of the present invention.

[0033] Figure 3 It is a frequency domain analysis diagram of bucket wheel abnormality (bucket wheel derailment) in an embodiment of the present invention. DETAILED DESCRIPTION

[0034] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0035] like Figure 1 As shown, the embodiment of the present invention discloses a method for detecting an abnormal state of a bucket wheel of a portal bucket wheel machine based on an audio signal, comprising the following steps:

[0036] S00. Install and deploy sound sensors near the bucket wheel of the portal bucket wheel excavator;

[0037] S01. Before use, check that the sound signal output is normal and the network communication is normal, then you can start the detection process.

[0038] S02. The PC running the algorithm obtains the start signal of the bucket wheel through the OPC ua communication method. The sound sensor collects a piece of audio data with a period of 3 seconds and generates a file in PCM format.

[0039] S03. When it is detected that the bucket wheel is not started at that time, the data of the PCM file does not enter the algorithm analysis, and the recording of the next audio data is restarted.

[0040] S04: When the bucket wheel operation signal is detected, the audio data file being recorded at that moment enters the algorithm for processing and analysis. In this embodiment, the algorithm enters the processing and analysis specifically for the frequency domain feature analysis part of steps S05 to S07.

[0041] S05. In order to facilitate the computer to analyze and process the audio, in this embodiment, it is necessary to set certain rules to sample and process the analog audio signal, which is used to convert the continuous time signal into a discrete time signal.

[0042] According to Nyquist theory, the sampling frequency is not less than twice the highest frequency of the sound signal. By analyzing the audio signals actually collected on site and comprehensively considering the actual audio restoration quality, in this embodiment, a sampling frequency of 16 kHz is selected for discretization of the signal.

[0043] The quantization accuracy of the sampling amplitude is 16-bit in this real-time example to balance the amount of calculation and the audio quality. The quantization takes up 2 bytes of computer storage space. After actual application, the analysis effect meets the use requirements. At this point, the preprocessing of the audio data is completed.

[0044] Since the amount of calculation for Fourier frequency domain conversion is very huge, in order to save the calculation cost, in this embodiment, a short-time average amplitude analysis is performed on the pre-processed audio data in advance.

[0045] Use a rectangular window to frame the raw discrete audio data:

[0046] x n [m] = w(m) x [n+m] (1-1)

[0047] The rectangular window is represented as:

[0048]

[0049] The short-time average amplitude is expressed as:

[0050]

[0051] The short-time amplitude M n If the value of exceeds the preset value, it corresponds to the voiced segment. If the value of the short-time amplitude M n is less than the preset value, it corresponds to the unvoiced segment. When the number of voiced segments increases and the average amplitude value is too large, the preprocessed data is sent to the Fourier transform module for secondary analysis.

[0052] S06 According to the DFT definition of the finite-length discrete signal x(n), n = 0, 1, 2........N-1:

[0053]

[0054] Decompose x(n) into the sum of two sequences of even and odd numbers x(n) = x1(n) + x2(n); the lengths of x1(n) and x2(n) are both N / 2, x1(n) is an odd sequence, and x2(n) is an even sequence. Then:

[0055]

[0056] Since Then

[0057]

[0058] Through the above transformation, the computational complexity of the Fourier transform is reduced, and thus the frequency-domain characteristics of the audio signal are obtained.

[0059] S07. As Figure 2 . Figure 3 shown, perform statistical analysis on the discrete frequency-domain characteristics. When it is recorded as valid data within 0 to 50000HZ, the figure shows the corresponding relationship between the sound wave amplitude (the amplitude has been normalized) corresponding to the frequency after Fourier transform and the frequency. Among them, the abscissa is the frequency, and the ordinate is the sound wave amplitude.

[0060] The abnormal judgment method is as follows:

[0061] (1) Select the median M of the characteristic point amplitudes and compare them from the low-frequency band to the high-frequency band;

[0062] (2) The amplitude A i (0 < i < 500); A i > M is one comparison result, and A i < M is another comparison result

[0063] (3) When a comparison result is different from the previous comparison result, it is recorded as one change. After all comparisons are completed, the number of times the comparison result changes is counted. When n>30, it means that the sound frequency has been greatly disturbed and it is judged as an abnormal sound.

[0064] S08. Send an alarm to the control system host computer through the OPC interface to inform the operating personnel that an abnormality in the sound detection has been found.

[0065] S09. The operator clicks the confirmation button, and the system plays the real-time audio signal on site for the operator to confirm the fault.

[0066] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for detecting abnormal state of bucket wheel of portal bucket wheel excavator based on audio signal, characterized in that: The steps include: Install and deploy sound sensors near the bucket wheel of the gantry bucket wheel excavator to collect on-site audio data; Preprocessing the collected audio data to convert the continuous time signal into a discrete time signal, wherein the preprocessing includes signal discretization processing, original discrete audio data framing processing, and Fourier frequency domain conversion processing, thereby obtaining the frequency domain characteristics of the audio signal; Compare the pre-processed frequency domain features with the preset standards, and if there is any abnormality, an alarm will be sent to the control system; The original discrete audio data frame processing in the preprocessing process specifically includes the following steps: Use a rectangular window to frame the raw discrete audio data: x n [m]=w(m)x[n+m] (1-1) The rectangular window is represented as: The short-term average amplitude is expressed as: Short-term amplitude M n If the value exceeds the preset value, it corresponds to the voiced segment, and the short-term amplitude M n If the value of is less than the preset value, it corresponds to the unvoiced segment. When the number of voiced segments increases and the average amplitude value is too large, the preprocessed data is sent to the Fourier transform module for secondary analysis. The comparison between the pre-processed frequency domain features and the preset standards is as follows: the discrete frequency domain features are statistically analyzed, and the frequency range of 0 to 50000 Hz is recorded as valid data. The above frequency range is sampled and separated, and 499 features are intercepted with 100 Hz as the segmentation unit. The screening rules are as follows: (1) Select the median M of the feature point amplitude and compare from the low frequency band to the high frequency band; (2) Amplitude A i (0 < i < 500); A i > M is a comparison result, A i < M is another comparison result (3) When a comparison result is different from the previous comparison result, it is recorded as one change. After all comparisons are completed, the number of times the comparison result changes is counted. When n>30, it means that the sound frequency has been greatly disturbed and it is judged as an abnormal sound.

2. The method for detecting abnormal state of bucket wheel of portal bucket wheel excavator based on audio signal according to claim 1, characterized in that: The on-site audio data is connected to the main control system through OPC ua communication. The sound sensor collects a segment of audio data at a cycle of 3 seconds and generates a file in PCM format.

3. The method for detecting abnormal state of bucket wheel of portal bucket wheel excavator based on audio signal according to claim 1 or 2, characterized in that: Before preprocessing the collected audio data, the following steps are performed: Detect whether the bucket wheel is in the started state. If it is in the started state, the on-site audio data will be transmitted to the main control system; if it is not in the started state, the file generated by the sound sensor will not enter the main control system for analysis, and the next audio data will be recorded again.

4. The method for detecting abnormal state of bucket wheel of portal bucket wheel excavator based on audio signal according to claim 1, characterized in that: During the preprocessing process, a sampling frequency of 16kHz was selected for signal discretization.

5. The method for detecting abnormal state of bucket wheel of portal bucket wheel excavator based on audio signal according to claim 1, characterized in that: When an abnormality occurs and an alarm is sent to the control system, the operator reads the abnormal sound signal, the system plays the real-time audio signal on site, and the operator confirms the fault.

Images