CN112033656A - Mechanical system fault detection method based on broadband spectrum processing - Google Patents
Mechanical system fault detection method based on broadband spectrum processing Download PDFInfo
- Publication number
- CN112033656A CN112033656A CN202010878023.8A CN202010878023A CN112033656A CN 112033656 A CN112033656 A CN 112033656A CN 202010878023 A CN202010878023 A CN 202010878023A CN 112033656 A CN112033656 A CN 112033656A
- Authority
- CN
- China
- Prior art keywords
- voiceprint
- signal
- fault
- normal state
- library
- 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
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
- G01M13/00—Testing of machine parts
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/02—Preprocessing
- G06F2218/04—Denoising
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/12—Classification; Matching
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
- Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)
Abstract
The invention discloses a mechanical system fault detection method based on broadband spectrum processing. The method comprises the following steps: step S1: acoustic sensor collects mechanical system running state voiceprint signalss(n) (ii) a Step S2: for collected signalss(n) Performing noise suppression processingss(n) (ii) a Step S3: the normal state signals are collected in a delayed manner to form a normal state voiceprint library through step S2N_S(n) (ii) a Step S4: based onN_S(n) To pairss(n) Performing voiceprint matching, and updating the normal state voiceprint library if the voiceprint matching is performedN_S(n) If not, updating the failure state voiceprint library and the failure label edition; step S5: and if the fault is found, carrying out early warning prompt, returning to the step S1, and if the fault is not found, directly returning to the step S1. The invention does not depend on prior knowledge of fault signals, has high information utilization rate of the signals, improves the signal-to-noise ratio of the signals and has more accurate detection result.
Description
Technical Field
The invention belongs to the field of signal processing, and particularly relates to a mechanical system fault detection method based on broadband spectrum processing.
Background
The early diagnosis of the fault of the mechanical system mainly depends on the experience of related personnel, and whether the mechanical equipment has the fault is judged through the characteristics of vibration, sound and the like generated when the mechanical equipment operates. Until the 60's of the 20 th century, mechanical system fault diagnosis was not really developed gradually as a door system discipline. The technology adopts a signal analysis and diagnosis method to monitor and evaluate the operation state of a mechanical system and establish a mechanical equipment maintenance system matched with the operation state, thereby effectively reducing the occurrence of accidents, ensuring the normal operation and safe production of equipment and fundamentally solving the problems of insufficient maintenance and excessive maintenance in the regular maintenance of the equipment.
With the development of signal processing, artificial intelligence, pattern recognition and other technologies, various new fusion methods are also continuously introduced into fault detection. Common mechanical system fault detection methods include empirical mode decomposition, independent component analysis, wavelet analysis, and the like. The method is sensitive to the influence of interference noise, directly influences the accuracy of a detection result, and has low information utilization rate of signals. Due to the complex working environment of the actual mechanical system, the interference of background noise and the interaction of signals of various parts of the machine, it is difficult to detect and pick up a really useful signal in the signal acquisition process, which is also a reason that the early failure of the mechanical equipment is not easy to be found.
Disclosure of Invention
In view of the above-mentioned shortcomings of the prior art, an object of the present invention is to provide a method for detecting a fault of a mechanical system based on broadband spectrum processing, so as to solve the problems of low information utilization rate and poor interference resistance of signals in the prior art.
In order to achieve the above object, the present invention provides a method for detecting a fault of a mechanical system based on broadband spectrum processing, the method comprising the following steps:
step S1: the acoustic sensor collects a mechanical system running state voiceprint signal s (n).
Step S2: the acquired signal s (n) is subjected to noise suppression processing ss (n).
Step S3: the normal state signal is collected in a delayed manner, and then processed by step S2 to form a normal state voiceprint library N _ S (N).
Step S4: and (c) performing voiceprint matching on ss (N) based on N _ S (N), updating a normal state voiceprint library N _ S (N) if matching is performed, and updating a fault state voiceprint and fault label editing if not matching is performed.
Step S5: and if the fault is found, carrying out early warning prompt, returning to the step S1, and if the fault is not found, directly returning to the step S1.
The invention has the beneficial effects that: on one hand, the invention is a mechanical system fault detection method based on broadband spectrum processing, which improves the information utilization rate of signals, has strong anti-interference capability and more accurate detection result. On the other hand, the method does not need to know the prior knowledge of the fault signal, and judges whether the mechanical equipment has the fault or not by utilizing the difference between the normal operation voiceprint signal of the mechanical equipment and the current operation state voiceprint signal so as to monitor the operation state of the mechanical equipment; finally, the invention stores and records the fault voiceprint signal data through the data communication module, and accumulates the original data for the subsequent optimization of the detection method.
Drawings
Fig. 1 is a schematic diagram of a mechanical system fault detection method based on broadband spectral processing.
Fig. 2 is a schematic diagram of voiceprint signal matching comparison in the operation state of the mechanical equipment.
Fig. 3a is a waveform diagram of a noise reduction signal.
Fig. 3b is a noise reduction signal spectrum diagram.
Fig. 4 is a comparison simulation diagram of the voiceprint signal of the mechanical equipment operation state.
Detailed Description
The technical solution provided by the present invention will be further explained with reference to the accompanying drawings.
The invention provides a mechanical system fault detection method based on broadband spectrum processing. According to the method, acoustic sensors are used for picking up voiceprint signals of the mechanical system in the running state, detection, extraction and identification of mechanical system fault voiceprint information are carried out through pattern recognition technologies such as background noise suppression and voiceprint matching comparison, and therefore monitoring and fault detection of the mechanical system in the running state are achieved. FIG. 1 is a schematic diagram of the detection method.
Method for restraining background interference noise
Assuming that the collected signal is s (t), the noise is n (t), and the mechanical equipment operation state voiceprint signal is x (t), then:
s(t)＝x(t)+n(t) (1)
its autocorrelation function is:
the autocorrelation is actually a measure of the similarity of different parts of the signal, but the result is of course only a periodic part, and white noise is suppressed. So that the principle is used for noise suppression
Because the interference noise under the actual condition has very complex composition, not only contains stationary noise part, but also contains non-stationary noise part, the self-correlation method has not ideal suppression effect on the non-stationary noise, so that the suppression on the non-stationary noise is carried out by using Wavelet Transform (WT), and the wavelet transform has good time-frequency localization characteristic and better analysis capability on the non-stationary signal.
Setting function
Fourier transform of
The conditions are satisfied:
basic function
After the telescopic transformation and the translation transformation, the following functions are obtained:
balance
For analyzing wavelets or continuous wavelets, wherein a is a scale factor, b is a translation factor, a, b belongs to R, and a is not equal to 0. The continuous wavelet transform is:
inverse transform of continuous wavelet:
wherein
When a is 2-j,b＝2-jk, where j is an integer, the continuous wavelet becomes a binary discrete wavelet. The relationship is as follows:
the corresponding binary wavelet transform is defined as:
its corresponding inverse transform:
matching comparison of voiceprint signals of mechanical equipment running state
The optimal threshold value is searched by comparing Euclidean distances among normal state signals, then the threshold value is used as a measuring standard to identify the existence of fault signals, if the detection signals are within the threshold value range, the normal signals are obtained, and if the detection signals are not within the threshold value range, the fault signals are obtained.
The euclidean distance expression is as follows:
where n is the dimension of the space, dst (X, Y) represents the Euclidean distance between two n-dimensional vectors X and Y, XiAnd yiRepresenting the ith values of sequences X and Y, respectively.
If the normal state voiceprint library is N _ S (N), and the sum of Euclidean distances between each sample and other samples is d, the expression of d is as follows:
wherein N is the number of samples in the background voiceprint library.
Defining the distance sequence as D ═ D1,d2,…,dnThen, a maximum value dmax is obtained based on D, and is used as a detection threshold, and the judgment form is as follows:
a flow chart of the threshold detection method is shown in fig. 2.
Third, performance analysis
In the simulation, a sine wave signal with the frequency of 100Hz and 360Hz is used for simulating a voiceprint signal of a machine in a normal running state, a sine wave with the frequency of 1000Hz and white noise are added as interference noise, the signal to noise ratio is-5 db, the sampling rate is 5120Hz, and the number of sampling points is 2048, so that the performance of the background noise suppression algorithm is verified.
From fig. 3a and fig. 3b, it can be seen that the interference noise is well suppressed by combining the autocorrelation noise suppression with the wavelet denoising, the time domain characteristics and the frequency domain characteristics of the useful signal are clearer, and from the simulation result, the characteristic frequency of the denoised signal is 100Hz and 360Hz, which respectively correspond to the original signal 100Hz and 360Hz, and the signal-to-noise ratio of the denoised signal is 15db, which proves that the noise suppression method has a better effect on the suppression of the interference noise.
For detecting the voiceprint signal of the mechanical equipment operation state, the Euclidean distance is used for measuring the similarity between samples. And calculating the sum of Euclidean distances between each sample and other samples in the sample library, and selecting an optimal detection threshold value through multiple comparisons so as to detect the fault signal. In the simulation, sine wave signals with the frequencies of 100Hz and 360Hz are used for simulating a normal-state voiceprint signal, 30 samples are simulated, a sine wave signal with the frequency of 430Hz is used for simulating a fault voiceprint signal, and the fault voiceprint signal is mixed in the normal-state voiceprint signal to serve as a detection signal. Thereby verifying the performance of the threshold detection method.
It can be seen from fig. 4 that the euclidean distance sum between the normal signal 1 and the normal signal 2 as the detection signal and the normal state voiceprint library can be regarded as not exceeding the detection threshold within the allowable error range, while the euclidean distance sum between the background voiceprint signal containing the fault signal and the normal state voiceprint library as the detection signal and the euclidean distance sum far exceeding the allowable error range significantly exceeds the detection threshold, i.e. is regarded as the fault signal, thus proving the effectiveness of the method.
The above description of the embodiments is only intended to facilitate the understanding of the method of the invention and its core idea. It should be noted that, for those skilled in the art, it is possible to make various improvements and modifications to the present invention without departing from the principle of the present invention, and those improvements and modifications also fall within the scope of the claims of the present invention.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (5)
1. A mechanical system fault detection method based on broadband spectrum processing is characterized by comprising the following steps:
step S1: acoustic sensor collects mechanical system running state voiceprint signalss(n)；
Step S2: for collecteds(n) Is subjected to noise suppression processing to obtainss(n)；
Step S3: the normal state signal is collected in a delayed way, and a normal state voiceprint library is formed through the step S2N_S(n)；
Step S4: based onN_S(n) To pairss(n) Performing voiceprint matching, and updating the normal state voiceprint library if the voiceprint matching is performedN_S(n) If not, updating the failure state voiceprint library and the failure label edition;
step S5: and if the fault is found, carrying out early warning prompt, returning to the step S1, and if the fault is not found, directly returning to the step S1.
2. The method of claim 1, wherein the noise suppression is performed by first denoising the wideband spectral processing based mechanical system fault by using an autocorrelation method, and then denoising the wideband spectral processing based mechanical system fault by combining with a wavelet denoising method.
3. The method of claim 1, wherein the normal state voiceprint library is a library of normal state voiceprintsN_S(n) If it is variable, the step S4 is executed, and if the detected signal is determined to be a normal voiceprint signal, the normal voiceprint library is updatedN_S(n)。
4. The method of claim 1, wherein the voiceprint signal recognition and detection of the operating status of the machine is based on a normal state voiceprint libraryN_S(n) Set upDetecting threshold value, using Euclidean distance as measurement standard, if the signal to be detected is in the detection threshold value, then considering it as normal state voiceprint signal, then updating normal state voiceprint libraryN_S(n) Continuing to monitor; and if the detection threshold value is exceeded, the voice print signal is regarded as a fault voice print signal.
5. The method of claim 1, wherein in step S4, if the detected signal is determined to be a faulty voiceprint signal, the faulty voiceprint signal is compared with a faulty voiceprint library, if the detected signal is not recorded, and if the detected signal is not recorded, the faulty voiceprint library is updated to perform an early warning.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010878023.8A CN112033656A (en) | 2020-08-27 | 2020-08-27 | Mechanical system fault detection method based on broadband spectrum processing |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010878023.8A CN112033656A (en) | 2020-08-27 | 2020-08-27 | Mechanical system fault detection method based on broadband spectrum processing |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112033656A true CN112033656A (en) | 2020-12-04 |
Family
ID=73585850
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010878023.8A Pending CN112033656A (en) | 2020-08-27 | 2020-08-27 | Mechanical system fault detection method based on broadband spectrum processing |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112033656A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113436645A (en) * | 2021-06-09 | 2021-09-24 | 杭州电子科技大学 | Electromechanical system fault on-line monitoring acoustic processing method under complex noise environment |
| CN113674447A (en) * | 2021-07-09 | 2021-11-19 | 深圳市慧友安电子技术有限公司 | System and method for detecting voiceprint fault of industrial equipment |
Citations (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102095560A (en) * | 2010-11-09 | 2011-06-15 | 中国人民解放军重庆通信学院 | Mechanical fault judgment system and method based on noise test |
| CN102840979A (en) * | 2012-09-04 | 2012-12-26 | 国电联合动力技术有限公司 | Method and device for detecting transmission chain faults of wind turbine generator set |
| JP2013013075A (en) * | 2011-06-27 | 2013-01-17 | General Electric Co <Ge> | Electrical substation fault monitoring and diagnostics |
| CN103940363A (en) * | 2014-04-30 | 2014-07-23 | 中国科学院半导体研究所 | High-precision optical fiber strain low-frequency sensing demodulation method based on wavelet cross-correlation technology |
| CN104198183A (en) * | 2014-09-17 | 2014-12-10 | 重庆大学 | Wind turbine drive chain vibration noise suppression and incipient fault feature extraction method |
| CN107631877A (en) * | 2017-08-11 | 2018-01-26 | 南京航空航天大学 | A kind of rolling bearing fault collaborative diagnosis method for casing vibration signal |
| CN107748314A (en) * | 2017-10-18 | 2018-03-02 | 国网重庆市电力公司北碚供电分公司 | Transformer Faults Analysis system based on sound wave shock detection |
| CN108847246A (en) * | 2018-06-15 | 2018-11-20 | 上海与德科技有限公司 | A kind of animation method, device, terminal and readable medium |
| CN109557412A (en) * | 2018-11-27 | 2019-04-02 | 华翔翔能电气股份有限公司 | A kind of method for recognizing sound-groove and system of transformer fault |
| US20190281398A1 (en) * | 2018-03-12 | 2019-09-12 | Panasonic Intellectual Property Management Co., Ltd. | Noise reduction device, noise reduction system, and fault detection method for noise reduction device |
| CN110319923A (en) * | 2019-07-08 | 2019-10-11 | 杭州电子科技大学 | A kind of detection device applied to acoustics |
| CN110617982A (en) * | 2019-09-19 | 2019-12-27 | 苏州时辰智能机电设备有限公司 | Rotating machinery equipment fault identification method based on voiceprint signals |
| CN110808059A (en) * | 2019-10-10 | 2020-02-18 | 天津大学 | Speech noise reduction method based on spectral subtraction and wavelet transform |
| CN111537919A (en) * | 2020-05-14 | 2020-08-14 | 莫毓昌 | Transformer fault diagnosis method based on voiceprint characteristics |
-
2020
- 2020-08-27 CN CN202010878023.8A patent/CN112033656A/en active Pending
Patent Citations (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102095560A (en) * | 2010-11-09 | 2011-06-15 | 中国人民解放军重庆通信学院 | Mechanical fault judgment system and method based on noise test |
| JP2013013075A (en) * | 2011-06-27 | 2013-01-17 | General Electric Co <Ge> | Electrical substation fault monitoring and diagnostics |
| CN102840979A (en) * | 2012-09-04 | 2012-12-26 | 国电联合动力技术有限公司 | Method and device for detecting transmission chain faults of wind turbine generator set |
| CN103940363A (en) * | 2014-04-30 | 2014-07-23 | 中国科学院半导体研究所 | High-precision optical fiber strain low-frequency sensing demodulation method based on wavelet cross-correlation technology |
| CN104198183A (en) * | 2014-09-17 | 2014-12-10 | 重庆大学 | Wind turbine drive chain vibration noise suppression and incipient fault feature extraction method |
| CN107631877A (en) * | 2017-08-11 | 2018-01-26 | 南京航空航天大学 | A kind of rolling bearing fault collaborative diagnosis method for casing vibration signal |
| CN107748314A (en) * | 2017-10-18 | 2018-03-02 | 国网重庆市电力公司北碚供电分公司 | Transformer Faults Analysis system based on sound wave shock detection |
| US20190281398A1 (en) * | 2018-03-12 | 2019-09-12 | Panasonic Intellectual Property Management Co., Ltd. | Noise reduction device, noise reduction system, and fault detection method for noise reduction device |
| CN108847246A (en) * | 2018-06-15 | 2018-11-20 | 上海与德科技有限公司 | A kind of animation method, device, terminal and readable medium |
| CN109557412A (en) * | 2018-11-27 | 2019-04-02 | 华翔翔能电气股份有限公司 | A kind of method for recognizing sound-groove and system of transformer fault |
| CN110319923A (en) * | 2019-07-08 | 2019-10-11 | 杭州电子科技大学 | A kind of detection device applied to acoustics |
| CN110617982A (en) * | 2019-09-19 | 2019-12-27 | 苏州时辰智能机电设备有限公司 | Rotating machinery equipment fault identification method based on voiceprint signals |
| CN110808059A (en) * | 2019-10-10 | 2020-02-18 | 天津大学 | Speech noise reduction method based on spectral subtraction and wavelet transform |
| CN111537919A (en) * | 2020-05-14 | 2020-08-14 | 莫毓昌 | Transformer fault diagnosis method based on voiceprint characteristics |
Non-Patent Citations (8)
| Title |
|---|
| ZHANGTING YU 等: "The research on transformer fault diagnosis method based on vibration and noise voiceprint imaging technology", 《PROCEEDINGS OF THE 2019 14TH IEEE CONFERENCE ON INDUSTRIAL ELECTRONICS AND APPLICATIONS》 * |
| 余长厅 等: "基于声纹及振动的变压器故障诊断技术研究", 《高压电器》 * |
| 刘剑: "相关函数在噪声去除方面的优劣点分析", 《科技经济市场》 * |
| 李波 等: "基于DSP的声纹识别技术的研究", 《辽宁工业大学学报（自然科学版）》 * |
| 牛礼民: "测试技术与信号处理课程案例式教学探索", 《大学教育》 * |
| 王东明 等: "基于统计特性的小波噪声抑制在低剂量CT中的应用", 《中国图象图形学报》 * |
| 王喆 等: "一种用于GIS本体机械缺陷诊断和定位的在线异声监测装置", 《电子测量技术》 * |
| 韩吉声 等: "稠油热采井下光纤压力传感信号的小波降噪方法", 《激光与光电子学进展》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113436645A (en) * | 2021-06-09 | 2021-09-24 | 杭州电子科技大学 | Electromechanical system fault on-line monitoring acoustic processing method under complex noise environment |
| CN113674447A (en) * | 2021-07-09 | 2021-11-19 | 深圳市慧友安电子技术有限公司 | System and method for detecting voiceprint fault of industrial equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112101174A (en) | LOF-Kurtogram-based mechanical fault diagnosis method | |
| CA2401685C (en) | Complex signal decomposition and modeling | |
| CN109668733B (en) | Variable-speed bearing fault diagnosis method based on decomposition of variational nonlinear mode | |
| CN109855874B (en) | Random resonance filter for enhancing detection of weak signals in vibration assisted by sound | |
| CN112033656A (en) | Mechanical system fault detection method based on broadband spectrum processing | |
| CN113345399A (en) | Method for monitoring sound of machine equipment in strong noise environment | |
| CN111585662A (en) | Classification identification and parameter estimation method and system for phase modulation signal | |
| CN113472390A (en) | Frequency hopping signal parameter estimation method based on deep learning | |
| Wang et al. | Rotating machine fault detection based on HOS and artificial neural networks | |
| CN111929489B (en) | Fault arc current detection method and system | |
| Li et al. | A Reliable Strategy for Improving Automatic First‐Arrival Picking of High‐Noise Three‐Component Microseismic Data | |
| CN117169346A (en) | High-altitude building damage identification method based on wavelet packet energy spectrum analysis | |
| CN108594156B (en) | Improved current transformer saturation characteristic identification method | |
| CN116364108A (en) | Transformer voiceprint detection method and device, electronic equipment and storage medium | |
| CN114114400A (en) | Microseism event effective signal pickup method | |
| CN115064182A (en) | Fan fault feature identification method of self-adaptive Mel filter in strong noise environment | |
| CN113889147A (en) | Mechanical system voiceprint signal matching method based on dynamic time warping algorithm | |
| CN114200232A (en) | Method and system for detecting fault traveling wave head of power transmission line | |
| CN113436645A (en) | Electromechanical system fault on-line monitoring acoustic processing method under complex noise environment | |
| CN114970647B (en) | Electromagnetic data identification method and system based on probabilistic neural network | |
| CN111273101A (en) | Fault arc detection method and device and storage medium | |
| CN112929053B (en) | Frequency hopping signal feature extraction and parameter estimation method | |
| CN115114955B (en) | Equipment fault detection method based on sound and vibration signals | |
| Yong | Bearings fault diagnosis based on HMM and fractal dimensions spectrum | |
| CN117290670B (en) | Transformer bushing insulation state estimation method based on enhancement filter algorithm |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20201204 |
|
| WD01 | Invention patent application deemed withdrawn after publication |