CN115655455A - A Mechanical Fault Diagnosis Method Based on Adaptive Noise Transformation and Stochastic Resonance - Google Patents
A Mechanical Fault Diagnosis Method Based on Adaptive Noise Transformation and Stochastic Resonance Download PDFInfo
- Publication number
- CN115655455A CN115655455A CN202211258644.1A CN202211258644A CN115655455A CN 115655455 A CN115655455 A CN 115655455A CN 202211258644 A CN202211258644 A CN 202211258644A CN 115655455 A CN115655455 A CN 115655455A
- Authority
- CN
- China
- Prior art keywords
- signal
- stochastic resonance
- frequency
- noise
- fault diagnosis
- 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
Landscapes
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
Abstract
Description
技术领域technical field
本发明属于机械旋转部件智能故障诊断领域,具体提供一种基于自适应噪声变换和随机共振的机械故障诊断方法,此系统通过对输入信号中的噪声进行分布变换,得到最利于随机共振系统的噪声分布,然后经过随机共振系统进行去噪,进行故障诊断。The invention belongs to the field of intelligent fault diagnosis of mechanical rotating parts, and specifically provides a method for diagnosing mechanical faults based on adaptive noise transformation and stochastic resonance. The system transforms the noise in the input signal to obtain the most favorable noise for the stochastic resonance system. distribution, and then denoise through the stochastic resonance system for fault diagnosis.
背景技术Background technique
机械系统在工业化中具有不可或缺的作用,其中,旋转机械设备更是占到了大多数,旋转机械设备中的关键在于齿轮箱,由于工业环境恶劣以及其封闭的工作环境,使得齿轮箱保养困难,因此旋转机械设备中的齿轮箱经常发生各种各样的故障,并且每一次故障都可能造成不可或缺的金钱和生产力损失。在所谓的第四次工业革命、未来工厂、工业物联网时代,工业机械系统不断智能化、复杂化。因此,研究和开发数据驱动的方法和状态监测技术,能够实现快速、可靠和高质量的自动诊断是十分必要的。能够对齿轮箱早期故障进行准确预警,避免发生重大工业事故,工作人员可以做到及时进行维修,这对工业生产来说具有十分重大的意义。Mechanical systems play an indispensable role in industrialization, among which rotating mechanical equipment accounts for the majority. The key to rotating mechanical equipment lies in the gearbox. Due to the harsh industrial environment and its closed working environment, the maintenance of the gearbox is difficult , so gearboxes in rotating machinery often fail in a variety of ways, each of which can result in an indispensable loss of money and productivity. In the era of the so-called fourth industrial revolution, the factory of the future, and the Industrial Internet of Things, industrial machinery systems are becoming increasingly intelligent and complex. Therefore, research and development of data-driven methods and condition monitoring techniques that can achieve fast, reliable, and high-quality automatic diagnosis is highly necessary. It is of great significance for industrial production to be able to provide accurate early warning of early failures of gearboxes, avoid major industrial accidents, and allow staff to perform timely maintenance.
随机共振(Stochastic resonance,SR)作为一种能够从振动信号中提取微弱信号特征的非线性信号处理方法,以其利用噪声增强微弱信号而不是消除噪声的独特优势在机械故障诊断领域得到了广泛的研究。As a nonlinear signal processing method that can extract weak signal features from vibration signals, stochastic resonance (SR) has been widely used in the field of mechanical fault diagnosis for its unique advantage of using noise to enhance weak signals instead of eliminating noise. Research.
发明内容Contents of the invention
基于上述问题背景,本发明提供了一种基于自适应噪声变换和随机共振的机械故障诊断方法,通过对输入信号中的噪声进行分布变换,并且利用人工蜂群算法对参数进行快速寻优,得到最利于随机共振系统的噪声分布,然后经过随机共振系统进行信号去噪,最后通过包络谱分析,进行故障诊断。Based on the background of the above problems, the present invention provides a mechanical fault diagnosis method based on adaptive noise transformation and stochastic resonance. By transforming the distribution of the noise in the input signal and using the artificial bee colony algorithm to quickly optimize the parameters, the obtained It is most beneficial to the noise distribution of the stochastic resonance system, and then the signal is denoised through the stochastic resonance system, and finally the fault diagnosis is carried out through the envelope spectrum analysis.
本发明的具体技术方案是:Concrete technical scheme of the present invention is:
S1、利用加速度传感器采集机械旋转部件的振动数据,作为原始信号;S1. Use the acceleration sensor to collect the vibration data of the mechanical rotating parts as the original signal;
S2、将原始信号经过离散小波变换(DWT)分解为不同频带的信号,分解层数为待定系数,对分解后的信号进行重分布,重分布系数待定,重分布的目的是使原始信号中的有色噪声转化为对随机共振系统最有利的粉红噪声,将得到的重分布后的噪声再经过离散小波变换(DWT)重构得到新的包含粉红噪声分布的信号。S2, the original signal is decomposed into signals of different frequency bands through discrete wavelet transform (DWT), the number of decomposition layers is an undetermined coefficient, and the decomposed signal is redistributed, the redistribution coefficient is undetermined, and the purpose of redistribution is to make the original signal The colored noise is transformed into the most favorable pink noise for the stochastic resonance system, and the obtained redistributed noise is reconstructed by discrete wavelet transform (DWT) to obtain a new signal containing pink noise distribution.
S3、通过人工蜂群算法(ABC)对S2中的离散小波分解层数和重分布系数进行寻优,寻优目标函数为加权谱峭度(CSK),将寻优结果代入S2中得到重构后的信号。S3. Optimizing the discrete wavelet decomposition layers and redistribution coefficients in S2 through artificial bee colony algorithm (ABC), the optimization objective function is weighted spectral kurtosis (CSK), and substituting the optimization results into S2 to obtain reconstruction after the signal.
S4、将重构后的信号输入标准化后的双稳态随机共振系统,通过随机共振利用噪声能量增强低频信号能量,放大原信号中的故障特征信号,得到去噪后的信号。S4. Input the reconstructed signal into the standardized bistable stochastic resonance system, use the noise energy to enhance the low-frequency signal energy through stochastic resonance, amplify the fault characteristic signal in the original signal, and obtain the denoised signal.
S5、对最终得到的经过随机共振系统去噪后的信号进行希尔伯特包络谱分析,将包络谱峰值频率和计算出的理论故障特征频率进行比较,对机械旋转部件进行故障诊断。S5. Perform Hilbert envelope spectrum analysis on the finally obtained signal denoised by the stochastic resonance system, compare the peak frequency of the envelope spectrum with the calculated theoretical fault characteristic frequency, and perform fault diagnosis on the mechanical rotating parts.
在本发明的某一具体实例中,S2中原始信号经过离散小波变换进行信号噪声变换的方法为:In a certain specific example of the present invention, the original signal in S2 carries out the method for signal-to-noise transformation through discrete wavelet transform:
假设x(t)为原始输入信号,首先对原始信号进行离散小波变换,得到一系列的细节系数和近似系数,表达式如下:Assuming that x(t) is the original input signal, the discrete wavelet transform is first performed on the original signal to obtain a series of detail coefficients and approximate coefficients, the expressions are as follows:
其中,为尺度函数,为母小波函数,j为分解层数,j=1,2,..,J,J为最后一层,于是,我们得到了在不同频带的一些列小波系数:in, is a scaling function, is the mother wavelet function, j is the number of decomposition layers, j=1,2,...,J, J is the last layer, so we get a series of wavelet coefficients in different frequency bands:
Φ={d1,d2,…,dj,…,dJ,dJ+1} (3)Φ={d 1 ,d 2 ,…,d j ,…,d J ,d J+1 } (3)
其中dJ+1为最后一层的近似系数aJ,由于离散小波分解的本质在于构建一系列低通和高通滤波器对原始信号进行滤波,输出不同频带的信号。Where d J+1 is the approximate coefficient a J of the last layer, because the essence of discrete wavelet decomposition is to construct a series of low-pass and high-pass filters to filter the original signal and output signals of different frequency bands.
分解层数J由下式决定:The number of decomposition layers J is determined by the following formula:
其中fs为采样频率,f0为故障特征频率,将故障特征频率包含在最后一层细节系数里,而f0通常未知。Where f s is the sampling frequency, f 0 is the fault characteristic frequency, and the fault characteristic frequency is included in the detail coefficient of the last layer, and f 0 is usually unknown.
接下来,我们对各个不同频带信号中的噪声进行重分布,得到对随机共振系统最有利的粉红噪声,粉红噪声的特点是噪声强度随着频率的增加而变小,也就是说,粉红噪声主要集中在信号的低频部分,对各个小波系数进行重分布的公式如下:Next, we redistribute the noise in different frequency bands to obtain the pink noise that is most beneficial to the stochastic resonance system. The characteristic of pink noise is that the noise intensity decreases with the increase of frequency, that is to say, pink noise mainly Focusing on the low frequency part of the signal, the formula for redistribution of each wavelet coefficient is as follows:
其中,α为重分布系数,最后将重分布后的信号进行重构得到新的信号yn(t):Among them, α is the redistribution coefficient, and finally the redistributed signal is reconstructed to obtain a new signal y n (t):
在本发明的某一具体实例中,S3中使用人工蜂群算法(ABC)以新提出的相关谱峭度(CSK)为适应度函数对离散小波分解层数J和重分布系数α进行寻优的步骤如下:In a specific example of the present invention, artificial bee colony algorithm (ABC) is used in S3 to optimize the discrete wavelet decomposition layer number J and redistribution coefficient α with the newly proposed correlation spectrum kurtosis (CSK) as the fitness function The steps are as follows:
S21、初始化解空间维度和范围,种群个数和种群解,侦查蜂数量,加速度常数,蜜源最大不更新次数,以及最大迭代次数。S21. Initialize the dimension and range of the solution space, the number of populations and population solutions, the number of scout bees, the acceleration constant, the maximum number of non-updated nectar sources, and the maximum number of iterations.
S22、计算各蜜源的适应度函数值。雇佣蜂在目前蜜源的附近进行探索搜寻新的蜜源,跟随蜂根据贪婪策略选择最优的蜜源,并在蜜源附近进行探索搜索新的蜜源。S22. Calculate the fitness function value of each nectar source. Hire bees to explore and search for new nectar sources near the current nectar source, follow the bees to choose the best nectar source according to the greedy strategy, and explore near the nectar source to search for new nectar sources.
S23、重复步骤S22,如果某个蜜源未更新次数达到蜜源最大不更新次数,丢弃该蜜源,并且根据侦查蜂数量随机产生一个最优蜜源进行代替。S23. Step S22 is repeated. If the non-renewing times of a certain nectar source reach the maximum non-renewing times of the nectar source, the nectar source is discarded, and an optimal nectar source is randomly generated according to the number of scout bees to replace it.
S24、不断重复S22和S23,直到达到最大迭代次数,得到最优解。S24. Repeat S22 and S23 continuously until the maximum number of iterations is reached, and an optimal solution is obtained.
在本发明的某一具体实例中,S4中利用双稳态随机共振系统对信号进行去噪的原理如下:In a specific example of the present invention, the principle of using the bistable stochastic resonance system to denoise the signal in S4 is as follows:
双稳态过阻尼随机共振系统的朗之万方程如下:The Langevin equation for a bistable overdamped stochastic resonance system is as follows:
其中x(t)是粒子运动轨迹,a,b是非负系统参数,A0是微弱信号幅值,f0为周期信号频率,ξ(t)是零均值高斯白噪声,强度为D。Where x(t) is the trajectory of the particle, a and b are the non-negative system parameters, A 0 is the amplitude of the weak signal, f 0 is the frequency of the periodic signal, ξ(t) is the zero-mean Gaussian white noise, and the intensity is D.
为了克服小参数限制,令τ=at,因此,(7)式变为:To overcome the small parameter limitation, let τ=at, therefore, formula (7) becomes:
于是得到了双稳态随机共振系统的标准形式,并且对输入周期信号进行了频率和幅值变换,满足了小参数限制。Then the standard form of the bistable stochastic resonance system is obtained, and the frequency and amplitude of the input periodic signal are transformed to meet the small parameter limit.
在本发明的某一具体实例中,S5中希尔伯特包络谱的原理如下:In a specific example of the present invention, the principle of the Hilbert envelope spectrum in S5 is as follows:
将信号经过希尔伯特变换得到原始信号的复域部分,将原信号和其复域部分相结合得到信号的解析信号,求解析信号的模即得到希尔伯特包络信号,求幅值谱即得到希尔伯特包络谱,主要用于显示原信号的低频调制部分。The complex domain part of the original signal is obtained by Hilbert transforming the signal, the analytical signal of the signal is obtained by combining the original signal and its complex domain part, and the Hilbert envelope signal is obtained by calculating the modulus of the analytical signal, and the amplitude is obtained The spectrum is the Hilbert envelope spectrum, which is mainly used to display the low-frequency modulation part of the original signal.
与现有技术相比,本发明的有益效果在于:Compared with prior art, the beneficial effect of the present invention is:
本发明将工程中包含有大量噪声的振动信号先经过噪声变换,得到更有利于随机共振系统的粉红噪声,然后在经过随机共振系统通过低频噪声放大原始信号,得到更好的原始信号放大的效果。In the present invention, the vibration signal containing a large amount of noise in the project is first subjected to noise transformation to obtain pink noise that is more beneficial to the stochastic resonance system, and then the original signal is amplified through the stochastic resonance system through low-frequency noise to obtain a better original signal amplification effect .
附图说明Description of drawings
图1为本发明流程图;Fig. 1 is a flowchart of the present invention;
图2为本发明中轴承外圈故障的时域图;Fig. 2 is the time-domain diagram of bearing outer ring fault among the present invention;
图3为本发明中轴承外圈故障的频域图;Fig. 3 is the frequency domain diagram of bearing outer ring fault among the present invention;
图4为本发明中进过重分布后的重构信号时域图;Fig. 4 is the time-domain diagram of the reconstruction signal after entering the heavy distribution in the present invention;
图5为本发明中人工蜂群算法迭代曲线图;Fig. 5 is artificial bee colony algorithm iteration curve figure among the present invention;
图6为本发明中经过随机共振系统后的信号时域图;Fig. 6 is the time-domain diagram of the signal after the stochastic resonance system in the present invention;
图7为本发明中经过随机共振系统后的信号包络谱图。Fig. 7 is a spectrum diagram of the signal envelope after passing through the stochastic resonance system in the present invention.
具体实施方式Detailed ways
为使本发明的目的、技术方案和优点更加清晰,下面结合附图和实际实验对本发明的技术方案作进一步描述。应当理解,此处所描述的具体实施例仅仅用于解释本发明,并不用于限定本发明。In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be further described below in conjunction with the accompanying drawings and actual experiments. It should be understood that the specific embodiments described here are only used to explain the present invention, not to limit the present invention.
本发明一种基于自适应噪声变换和随机共振的机械故障诊断方法,包括如下步骤:A method for diagnosing mechanical faults based on adaptive noise transformation and stochastic resonance of the present invention comprises the following steps:
S1、利用加速度传感器采集机械旋转部件的振动数据,作为原始信号;S1. Use the acceleration sensor to collect the vibration data of the mechanical rotating parts as the original signal;
S2、将原始信号经过离散小波变换(DWT)分解为不同频带的信号,分解层数为待定系数,对分解后的信号进行重分布,重分布系数待定,重分布的目的是使原始信号中的有色噪声转化为对随机共振系统最有利的粉红噪声,将得到的重分布后的噪声再经过离散小波变换(DWT)重构得到新的包含粉红噪声分布的信号。S2, the original signal is decomposed into signals of different frequency bands through discrete wavelet transform (DWT), the number of decomposition layers is an undetermined coefficient, and the decomposed signal is redistributed, the redistribution coefficient is undetermined, and the purpose of redistribution is to make the original signal The colored noise is transformed into the most favorable pink noise for the stochastic resonance system, and the obtained redistributed noise is reconstructed by discrete wavelet transform (DWT) to obtain a new signal containing pink noise distribution.
S3、通过人工蜂群算法(ABC)对S2中的离散小波分解层数和重分布系数进行寻优,寻优目标函数为加权谱峭度,将寻优结果代入S2中得到重构后的信号。S3. Optimizing the number of discrete wavelet decomposition layers and redistribution coefficients in S2 through the artificial bee colony algorithm (ABC), the optimization objective function is the weighted spectral kurtosis, and substituting the optimization results into S2 to obtain the reconstructed signal .
S4、将重构后的信号输入标准化后的双稳态随机共振系统,通过随机共振利用噪声能量增强低频信号能量,放大原信号中的故障特征信号,得到去噪后的信号。S4. Input the reconstructed signal into the standardized bistable stochastic resonance system, use the noise energy to enhance the low-frequency signal energy through stochastic resonance, amplify the fault characteristic signal in the original signal, and obtain the denoised signal.
S5、对最终得到的经过随机共振系统去噪后的信号进行希尔伯特包络谱分析,将包络谱峰值频率和计算出的理论故障特征频率进行比较,对机械旋转部件进行故障诊断。S5. Perform Hilbert envelope spectrum analysis on the finally obtained signal denoised by the stochastic resonance system, compare the peak frequency of the envelope spectrum with the calculated theoretical fault characteristic frequency, and perform fault diagnosis on the mechanical rotating parts.
以下为本发明具体实施实例:The following are concrete implementation examples of the present invention:
本实例采用凯斯西储大学(CRWU)的轴承数据,选用轴承外圈驱动端的故障数据,故障大小为7mm,采样频率fs为12000,电机转速为1772rmp,截取数据长度N为2048,可以计算得到本实例的理论故障频率f0为159.9,轴承内圈理论故障特征频率计算公式如下:This example uses the bearing data of Case Western Reserve University (CRWU), and selects the fault data of the drive end of the outer ring of the bearing. The fault size is 7mm, the sampling frequency f s is 12000, the motor speed is 1772rmp, and the intercepted data length N is 2048, which can be calculated The theoretical fault frequency f0 of this example is obtained as 159.9, and the calculation formula of the theoretical fault characteristic frequency of the inner ring of the bearing is as follows:
其中,fi为理论内圈故障特征频率;D为轴承中径;d为滚动体直径;Z为滚动体数目;α为接触角;fn为转频。Among them, f i is the fault characteristic frequency of the theoretical inner ring; D is the pitch diameter of the bearing; d is the diameter of the rolling element; Z is the number of rolling elements; α is the contact angle; f n is the rotation frequency.
首先,利用离散小波变换对输入信号进行重构,使得噪声分布接近粉红噪声:First, the discrete wavelet transform is used to reconstruct the input signal so that the noise distribution is close to pink noise:
x(t)为原始输入信号,首先对原始信号进行离散小波变换,得到一系列的细节系数和近似系数,表达式如下:x(t) is the original input signal. First, discrete wavelet transform is performed on the original signal to obtain a series of detail coefficients and approximate coefficients. The expressions are as follows:
其中,为尺度函数,为母小波函数,j为分解层数,j=1,2,..,J,J为最后一层,于是,我们得到了在不同频带的一些列小波系数:in, is a scaling function, is the mother wavelet function, j is the number of decomposition layers, j=1,2,...,J, J is the last layer, so we get a series of wavelet coefficients in different frequency bands:
Φ={d1,d2,…,dj,…,dj,dJ+1} (3)Φ={d 1 ,d 2 ,…,d j ,…,d j ,d J+1 } (3)
其中dJ+1为最后一层的近似系数aJ,由于离散小波分解的本质在于构建一系列低通和高通滤波器对原始信号进行滤波,输出不同频带的信号。Where d J+1 is the approximate coefficient a J of the last layer, because the essence of discrete wavelet decomposition is to construct a series of low-pass and high-pass filters to filter the original signal and output signals of different frequency bands.
分解层数J由下式决定:The number of decomposition layers J is determined by the following formula:
其中fs为采样频率,f0为故障特征频率,将故障特征频率包含在最后一层细节系数里,而f0通常未知。Where f s is the sampling frequency, f 0 is the fault characteristic frequency, and the fault characteristic frequency is included in the detail coefficient of the last layer, and f 0 is usually unknown.
接下来,我们对各个不同频带信号中的噪声进行重分布,得到对随机共振系统最有利的粉红噪声,粉红噪声的特点是噪声强度随着频率的增加而变小,也就是说,粉红噪声主要集中在信号的低频部分,对各个小波系数进行重分布的公式如下:Next, we redistribute the noise in different frequency bands to obtain the pink noise that is most beneficial to the stochastic resonance system. The characteristic of pink noise is that the noise intensity decreases with the increase of frequency, that is to say, pink noise mainly Focusing on the low frequency part of the signal, the formula for redistribution of each wavelet coefficient is as follows:
其中,α为重分布系数,最后将重分布后的信号进行重构得到新的信号yn(t):Among them, α is the redistribution coefficient, and finally the redistributed signal is reconstructed to obtain a new signal yn(t):
利用人工蜂群算法(ABC)以相关谱峭度为适应度函数对离散小波分解层数J和重分布系数α进行寻优的步骤如下:Using the artificial bee colony algorithm (ABC) with the correlation spectrum kurtosis as the fitness function to optimize the discrete wavelet decomposition layer number J and the redistribution coefficient α are as follows:
S21、初始化解空间维度为2,分解层数上下界为[1 8],重分布系数α范围为[020],种群个数为100,侦查蜂数量为100,加速度常数a=1,蜜源最大不更新次数为120,以及最大迭代次数为200。S21. The initial solution space dimension is 2, the upper and lower bounds of decomposition layers are [1 8], the range of redistribution coefficient α is [020], the population number is 100, the number of scout bees is 100, the acceleration constant a=1, and the honey source is the largest The number of no updates is 120, and the maximum number of iterations is 200.
S22、计算各蜜源的适应度函数值。雇佣蜂在目前蜜源的附近进行探索搜寻新的蜜源,跟随蜂根据贪婪策略选择最优的蜜源,并在蜜源附近进行探索搜索新的蜜源。S22. Calculate the fitness function value of each nectar source. Hire bees to explore and search for new nectar sources near the current nectar source, follow the bees to choose the best nectar source according to the greedy strategy, and explore near the nectar source to search for new nectar sources.
S23、重复步骤S22,如果某个蜜源未更新次数达到蜜源最大不更新次数,丢弃该蜜源,并且根据侦查蜂数量随机产生一个最优蜜源进行代替。S23. Step S22 is repeated. If the non-renewing times of a certain nectar source reach the maximum non-renewing times of the nectar source, the nectar source is discarded, and an optimal nectar source is randomly generated according to the number of scout bees to replace it.
S24、不断重复S22和S23,直到达到最大迭代次数,得到最优解。S24. Repeat S22 and S23 continuously until the maximum number of iterations is reached, and an optimal solution is obtained.
适应度函数加权谱峭度(CSK)的计算公式如下:The calculation formula of fitness function weighted spectral kurtosis (CSK) is as follows:
其中,x(t)为输入信号,y(t)为输出信号,yES(f)为输出信号包络谱,N为数据长度。Among them, x(t) is the input signal, y(t) is the output signal, y ES (f) is the envelope spectrum of the output signal, and N is the data length.
将得到的最优解带入离散小波变换,得到重构后的信号x_new(t)。Bring the obtained optimal solution into the discrete wavelet transform to obtain the reconstructed signal x_new(t).
将重构后的信号进行尺度变换后输入标准化的双稳态随机共振系统进行去噪,在本实例中,信号频率缩放系数为R=500,幅值缩放系数为双稳态随机共振系统表达式如下:Scale-transform the reconstructed signal and input it into a standardized bistable stochastic resonance system for denoising. In this example, the signal frequency scaling factor is R=500, and the amplitude scaling factor is The expression of the bistable stochastic resonance system is as follows:
双稳态过阻尼随机共振系统的朗之万方程如下:The Langevin equation for a bistable overdamped stochastic resonance system is as follows:
其中x(t)是粒子运动轨迹,a,b是非负系统参数,A0是微弱信号幅值,f0为周期信号频率,ξ(t)是零均值高斯白噪声,强度为D。Where x(t) is the trajectory of the particle, a and b are the non-negative system parameters, A 0 is the amplitude of the weak signal, f 0 is the frequency of the periodic signal, ξ(t) is the zero-mean Gaussian white noise, and the intensity is D.
为了克服小参数限制,令τ=at,因此,(7)式变为:To overcome the small parameter limitation, let τ=at, therefore, formula (7) becomes:
采用4阶龙格库塔算法求解双稳态过阻尼非线性系统的朗之万方程,求解过程如下:The 4th-order Runge-Kutta algorithm is used to solve the Langevin equation of the bistable overdamped nonlinear system. The solution process is as follows:
对于数据长度n=1:NFor data length n=1:N
k1=f(yn,tn)k 1 =f(y n ,t n )
k4=f(yn+h k1,tn+h)k 4 =f(y n +hk 1 ,t n +h)
其中yn为输出数据,h为时间步长,本例中设为1/fs。Where y n is the output data, h is the time step size, which is set to 1/f s in this example.
首先对信号进行时频域分析,采集到的原始信号的时域图和频域图如附图2和图3所示,由时域图可以观察到明显的信号脉冲部分,但是脉冲的周围也存在很多噪声成分。从频域图中可以观察到原信号又很多高频分量以及一些调制的边频带,低频分量几乎为0,因此,需要对信号进行如下分析。First, time-frequency domain analysis is performed on the signal. The time-domain diagram and frequency-domain diagram of the original signal collected are shown in Figure 2 and Figure 3. From the time-domain diagram, the obvious signal pulse part can be observed, but the surrounding of the pulse is also There are many noise components. From the frequency domain diagram, it can be observed that the original signal has many high-frequency components and some modulated sidebands, and the low-frequency components are almost 0. Therefore, the signal needs to be analyzed as follows.
按上述方法对采集的信号进行分析,下面介绍本具体实例使用此方法的具体过程。The collected signal is analyzed according to the above method, and the specific process of using this method in this specific example is introduced below.
首先使用离散小波变换对原信号进行分解,小波基函数选择db6小波,对分解信号进行重分布,使得原始输入信号中的噪声近似等于粉红噪声,即噪声能量集中在低频附近。分解层数和重分布系数由人工蜂群算法寻优得到,蜂群数设为100,最大迭代次数为200,适应度函数为输出信号谱峭度和输出输入信号的相关系数的乘积,解的范围分别为[1 8]和[0 20],得到最优解带入信号重构部分,重构信号如图4所示,人工蜂群算法迭代曲线如图5所示。将重构后的信号经过随机共振系统,将噪声能量加在信号低频分量上,对低频信号进行放大,放大后的输出信号如图6所示,输出信号包络谱如题7所示,观察包络谱峰值对应频率为158.2,实际计算得到的理论故障频率为159.9,由于理论故障频率并不是在实际工作环境下得到的,因此在轴承具体的工作环境下轴承会产生摩擦和滑动的现象,导致故障频率发生偏差,误差大小不影响实际故障诊断性能。可以看出本方法能够有效的加强并提取出信号的冲击成分,准确提取出故障特征频率,证明了此方法的有效性。First, the discrete wavelet transform is used to decompose the original signal. The wavelet basis function selects db6 wavelet, and the decomposed signal is redistributed, so that the noise in the original input signal is approximately equal to the pink noise, that is, the noise energy is concentrated near the low frequency. The number of decomposition layers and the redistribution coefficient are obtained by optimizing the artificial bee colony algorithm. The number of bee colonies is set to 100, and the maximum number of iterations is 200. The fitness function is the product of the kurtosis of the output signal spectrum and the correlation coefficient of the output and input signals. The solution The ranges are [1 8] and [0 20] respectively, and the optimal solution is brought into the signal reconstruction part. The reconstructed signal is shown in Figure 4, and the iteration curve of the artificial bee colony algorithm is shown in Figure 5. Pass the reconstructed signal through the stochastic resonance system, add noise energy to the low-frequency component of the signal, and amplify the low-frequency signal. The amplified output signal is shown in Figure 6, and the envelope spectrum of the output signal is shown in Question 7. Observe the package The frequency corresponding to the peak value of the network spectrum is 158.2, and the theoretical fault frequency obtained by actual calculation is 159.9. Since the theoretical fault frequency is not obtained in the actual working environment, the bearing will produce friction and sliding in the specific working environment of the bearing, resulting in The fault frequency deviates, and the size of the error does not affect the actual fault diagnosis performance. It can be seen that this method can effectively strengthen and extract the impact component of the signal, and accurately extract the fault characteristic frequency, which proves the effectiveness of this method.
以上是本发明的较佳实施例,凡依本发明技术方案所作的改变,所产生的功能作用未超出本发明技术方案的范围时,均属于本发明的保护范围。The above are the preferred embodiments of the present invention, and all changes made according to the technical solution of the present invention, when the functional effect produced does not exceed the scope of the technical solution of the present invention, all belong to the protection scope of the present invention.
Claims (5)
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211258644.1A CN115655455A (en) | 2022-10-14 | 2022-10-14 | A Mechanical Fault Diagnosis Method Based on Adaptive Noise Transformation and Stochastic Resonance |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211258644.1A CN115655455A (en) | 2022-10-14 | 2022-10-14 | A Mechanical Fault Diagnosis Method Based on Adaptive Noise Transformation and Stochastic Resonance |
Publications (1)
Publication Number | Publication Date |
---|---|
CN115655455A true CN115655455A (en) | 2023-01-31 |
Family
ID=84986931
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202211258644.1A Pending CN115655455A (en) | 2022-10-14 | 2022-10-14 | A Mechanical Fault Diagnosis Method Based on Adaptive Noise Transformation and Stochastic Resonance |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115655455A (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116222750A (en) * | 2023-03-22 | 2023-06-06 | 哈尔滨工程大学 | Stochastic resonance detector and method suitable for high-frequency narrow pulse width acoustic beacon signals |
CN116304848A (en) * | 2023-05-26 | 2023-06-23 | 广东石油化工学院 | A rolling bearing fault diagnosis system and method |
CN117905711A (en) * | 2024-03-20 | 2024-04-19 | 江苏海拓宾未来工业科技集团有限公司 | Multi-shaft centrifugal compressor and fault diagnosis method for impeller of multi-shaft centrifugal compressor |
CN119128529A (en) * | 2024-11-12 | 2024-12-13 | 西南交通大学 | A method for predicting the water content of oil in the lubrication system of mechanical equipment |
-
2022
- 2022-10-14 CN CN202211258644.1A patent/CN115655455A/en active Pending
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116222750A (en) * | 2023-03-22 | 2023-06-06 | 哈尔滨工程大学 | Stochastic resonance detector and method suitable for high-frequency narrow pulse width acoustic beacon signals |
CN116304848A (en) * | 2023-05-26 | 2023-06-23 | 广东石油化工学院 | A rolling bearing fault diagnosis system and method |
CN116304848B (en) * | 2023-05-26 | 2023-08-22 | 广东石油化工学院 | Rolling bearing fault diagnosis system and method |
CN117905711A (en) * | 2024-03-20 | 2024-04-19 | 江苏海拓宾未来工业科技集团有限公司 | Multi-shaft centrifugal compressor and fault diagnosis method for impeller of multi-shaft centrifugal compressor |
CN117905711B (en) * | 2024-03-20 | 2024-07-23 | 江苏海拓宾未来工业科技集团有限公司 | A multi-shaft centrifugal compressor and its impeller fault diagnosis method |
CN119128529A (en) * | 2024-11-12 | 2024-12-13 | 西南交通大学 | A method for predicting the water content of oil in the lubrication system of mechanical equipment |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN115655455A (en) | A Mechanical Fault Diagnosis Method Based on Adaptive Noise Transformation and Stochastic Resonance | |
Chen et al. | Fault feature extraction and diagnosis of rolling bearings based on wavelet thresholding denoising with CEEMDAN energy entropy and PSO-LSSVM | |
CN111695209B (en) | Rotary machine small sample health assessment method driven by meta-deep learning | |
CN105275833B (en) | CEEMD (Complementary Empirical Mode Decomposition)-STFT (Short-Time Fourier Transform) time-frequency information entropy and multi-SVM (Support Vector Machine) based fault diagnosis method for centrifugal pump | |
CN109033719B (en) | Wind turbine bearing fault diagnosis method | |
CN103499437B (en) | The rotating machinery fault detection method of adjustable quality factor dual-tree complex wavelet transform | |
CN111079690B (en) | Spindle and workpiece vibration prediction method based on stacked sparse auto-encoding network | |
CN109975013A (en) | Fault feature extraction method of wind turbine gearbox based on IVMD-SE | |
CN115434872B (en) | A wind turbine gearbox composite fault diagnosis method based on AVMD and improved RSSD | |
CN104677632A (en) | Rolling bearing fault diagnosis method using particle filtering and spectral kurtosis | |
Liao et al. | An automatic filtering method based on an improved genetic algorithm—with application to rolling bearing fault signal extraction | |
Han et al. | An intelligent fault diagnosis method of variable condition gearbox based on improved DBN combined with WPEE and MPE | |
CN114936575A (en) | Motor bearing fault diagnosis method based on sample generation and deep migration learning | |
Ke et al. | Fault diagnosis method of weak vibration signal based on improved VMD and MCKD | |
CN116067657B (en) | Rolling bearing fault diagnosis method and system | |
Liu et al. | An interpretable multiplication-convolution network for equipment intelligent edge diagnosis | |
CN118137485A (en) | Informer-Bi-LSTM model-based power load prediction method and system | |
Li et al. | A novel unsupervised graph wavelet autoencoder for mechanical system fault detection | |
Pan et al. | Improved complete ensemble robust local mean decomposition with adaptive noise for slewing bearings performance degradation assessment | |
Wei et al. | Few-shot bearing fault diagnosis using GAVMD–PWVD time–frequency image based on meta-transfer learning | |
Duan et al. | Bearing fault diagnosis based on state-space principal component tracking filter algorithm | |
CN118706452A (en) | A bearing fault diagnosis method and system based on sound and vibration signals | |
CN118296470A (en) | An intelligent classification method based on multi-classifier fusion | |
Chi et al. | Experimental application of stochastic resonance based on Wood–Saxon potential on fault diagnosis of bearing and planetary gearbox | |
CN116429430A (en) | A bearing fault detection method, system, device and medium based on an adaptive multi-scale enhanced dictionary learning framework |
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 | ||
CB03 | Change of inventor or designer information |
Inventor after: Chi Feng Inventor after: Wang Yong Inventor after: Huang Xiaoyan Inventor after: Gao Xiang Inventor after: Yang Qinmin Inventor after: Weng Deyu Inventor after: Chen Xu Inventor after: Cao Weiwei Inventor before: Chi Feng Inventor before: Wang Yong Inventor before: Huang Xiaoyan Inventor before: Gao Xiang Inventor before: Yang Qinmin Inventor before: Weng Deyu Inventor before: Chen Xu Inventor before: Cao Weiwei |
|
CB03 | Change of inventor or designer information |