A method and system for monitoring and early warning of nuclear power plant equipment failure risk

Abstract

The application relates to the technical field of fault prediction, in particular to a nuclear power plant equipment fault risk monitoring and early warning method and system, which specifically comprises the following steps: constructing a first fluctuation characteristic value by the random degree of the peak point occurrence time in the planetary gear box bearing vibration signal of the nuclear power plant equipment and the fluctuation characteristics of the signal sharpness; constructing a second fluctuation characteristic value by the distribution of the sideband characteristics in the envelope spectrum of the bearing vibration signal of each sampling period and comparing the center frequencies with those of the previous sampling period; comprehensively adjusting the filter length, combining the filtering algorithm, extracting the fault characteristic signals in the bearing vibration signals of each sampling period, and detecting the faults of the planetary gear box in each sampling period; the problems of improper filter length value during the filtering of the planetary gear box bearing vibration signal of the nuclear power plant are solved, the extraction accuracy of the fault characteristic signals is improved, and the accuracy of subsequent fault detection is improved.

Description

Technical Field

[0001] This application relates to the field of fault prediction technology, specifically to a method and system for monitoring and early warning of equipment fault risks in nuclear power plants. Background Technology

[0002] The stable and reliable operation of the circulating water pumps (CRF) units in nuclear power plant circulating water systems is a crucial factor in ensuring nuclear power output and improving economic efficiency. Planetary gearboxes, with their advantages of high transmission ratios and compact structure, are the core components of nuclear power circulating water pump units. Therefore, ensuring their healthy and stable operation is essential for guaranteeing the fault-free operation of CRF pump units. Nuclear equipment fault detection algorithms can accurately detect faults in planetary gearboxes and issue early warnings, providing technical support for the economical and safe operation of nuclear power plants. However, the complexity of the nuclear power operating environment leads to significant noise interference and operating condition fluctuations in monitoring data, placing higher demands on the accuracy of fault detection.

[0003] Extracting high-quality vibration signals from planetary gearbox bearings is a prerequisite for ensuring the accuracy of nuclear power equipment fault detection algorithms. The Minimum Optimal Entropy Deconvolution Algorithm (MOMEDA) is one of the commonly used fault feature extraction algorithms for planetary gearbox bearing vibration signals. It can extract weak periodic impact information features of faults. However, the deconvolution effect of MOMEDA is directly affected by the filter length. The larger the filter length, the longer the calculation time, which affects the fault detection efficiency. The smaller the filter length, the worse the deconvolution effect, resulting in the extracted fault feature signals containing more noise, which affects the fault detection accuracy. Since the vibration signals of planetary gearbox operation are easily affected by noise interference and operating condition fluctuations, if the filter length is not properly selected, it will affect the accuracy of fault feature extraction and reduce the fault detection accuracy of nuclear power plant equipment. Summary of the Invention

[0004] To address the aforementioned technical problems, the purpose of this application is to provide a method and system for monitoring and early warning of equipment failure risks in nuclear power plants. The specific technical solution adopted is as follows: In a first aspect, embodiments of this application provide a method for monitoring and early warning of equipment failure risks in nuclear power plants, the method comprising the following steps: The time sequence of vibration signals of planetary gearbox bearings in nuclear power plant equipment is collected and denoted as bearing vibration signal; Based on the randomness of the occurrence time of peak points in the bearing vibration signal of each sampling period and the fluctuation characteristics of the sharpness of the signal in the bearing vibration signal, the first fluctuation characteristic value of the bearing vibration signal of each sampling period is constructed. The center frequency of the spectrum is determined based on the kurtosis plot of the bearing vibration signal spectrum; the second fluctuation characteristic value of the bearing vibration signal spectrum in each sampling period is constructed based on the distribution of sideband features in the envelope spectrum of the bearing vibration signal in each sampling period and the difference between the center frequency of each sampling period and the previous sampling period. Based on the first and second fluctuation feature values, the filter length corresponding to each sampling period is adjusted. Combined with the filtering algorithm, the fault feature signal in the bearing vibration signal of each sampling period is extracted to perform fault detection of the planetary gearbox in each sampling period.

[0005] In one embodiment, the process of obtaining the first fluctuation feature value is as follows: Obtain the first-order difference sequence of all peak points in the bearing vibration signal corresponding to the time intervals, and calculate the coefficient of variation of the element values ​​in the first-order difference sequence; set a sliding window in the bearing vibration signal, construct a kurtosis sequence through the kurtosis of the bearing vibration signal in each sliding window, and take the mean of the absolute values ​​of the rate of change of all elements in the kurtosis sequence as the kurtosis volatility of the bearing vibration signal; construct the kurtosis compensation coefficient of the kurtosis volatility based on the numerical distribution of the rate of change of the elements in the kurtosis sequence. The fused value of the coefficient of variation, the kurtosis volatility, and the kurtosis compensation coefficient is used as the first volatility characteristic value.

[0006] In one embodiment, the kurtosis compensation coefficient is the ratio of the number of elements whose absolute rate of change is greater than a preset rate of change threshold to the total number of elements in the kurtosis sequence.

[0007] In one embodiment, the center frequency is the center of the frequency band with the largest kurtosis in the kurtosis graph.

[0008] In one embodiment, the process of obtaining the second fluctuation feature value is as follows: Based on the sideband energy analysis of the envelope spectrum, the degree of dispersion of the peak energy at the harmonic frequency of the center frequency is analyzed, and the energy concentration of the envelope spectrum is constructed. The fundamental frequency offset of each sampling period is calculated based on the difference in center frequency between adjacent sampling periods. The second fluctuation characteristic value is calculated based on the fundamental frequency offset and the energy concentration. The second fluctuation characteristic value is proportional to the fundamental frequency offset and inversely proportional to the energy concentration.

[0009] In one embodiment, the process of obtaining the energy concentration is as follows: The peak points at the first preset number of harmonic frequencies of the center frequency in the envelope spectrum are obtained. The sum of the energy of the frequency band within the half-peak width of all the peak points at the harmonic frequencies is divided by the total energy of the envelope spectrum to obtain the energy concentration of the envelope spectrum.

[0010] In one embodiment, the expression for the fundamental frequency offset is: ,in, This represents the fundamental frequency offset for the current sampling period. , These represent the center frequencies of the current sampling period and its previous sampling period, respectively.

[0011] In one embodiment, the process of obtaining the filter length is as follows: The weighted sum of the first volatility eigenvalue and the second volatility eigenvalue is denoted as the adjustment factor. The expression for the filter length is: In the formula, Adjust the filter length corresponding to the current sampling period. , These are the lower and upper limits of the filter length range, respectively; This represents the floor function.

[0012] In one embodiment, the step of extracting fault feature signals from the bearing vibration signals of each sampling period to perform fault detection of the planetary gearbox within each sampling period specifically involves: The bearing vibration signal of each sampling period is used as the input of the MOMEDA algorithm. The filter length in the MOMEDA algorithm adopts the adjusted filter length corresponding to each sampling period to obtain the fault characteristic signal of each sampling period. This signal is then used as the input of the fault detection algorithm to output the fault detection result.

[0013] Secondly, embodiments of this application also provide a nuclear power plant equipment fault risk monitoring and early warning system, including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program to implement the steps of any of the methods described above.

[0014] The embodiments of this application have at least the following beneficial effects: This application collects vibration signals from planetary gearbox bearings in nuclear power plant equipment. Based on the randomness of the peak occurrence time and the fluctuation characteristics of signal sharpness in the bearing vibration signals, a first fluctuation feature value is constructed to analyze the random fluctuation of the signal under the influence of background noise in the time domain. By analyzing the distribution of sideband features in the envelope spectrum of the bearing vibration signals in each sampling period and comparing it with the center frequency of the previous sampling period, a second fluctuation feature value is constructed to analyze the influence of background noise on fault characteristics in the frequency domain. Combining the two methods, the filter length is appropriately adjusted, and combined with the filtering algorithm, fault feature signals in the bearing vibration signals of each sampling period are extracted for fault detection of the planetary gearbox in each sampling period. This allows for adaptive setting of a suitable filter length based on the current vibration signal fluctuation, providing high flexibility. It solves the problem that the bearing vibration signals of nuclear power plant planetary gearbox equipment are easily affected by operating noise and easily drown out weak fault feature signals, improving the extraction accuracy of fault feature signals and contributing to the accuracy of subsequent fault detection. Attached Figure Description

[0015] To more clearly illustrate the technical solutions and advantages in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 A flowchart illustrating the steps of a method for monitoring and early warning of equipment failure risks in a nuclear power plant, as provided in one embodiment of this application; Figure 2 This is a schematic diagram illustrating the process of obtaining the first wave characteristic value. Detailed Implementation

[0017] To further illustrate the technical means and effects adopted by this application to achieve the intended purpose of the invention, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a nuclear power plant equipment fault risk monitoring and early warning method and system proposed in this application. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0018] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains.

[0019] The following description, in conjunction with the accompanying drawings, details the specific scheme of the nuclear power plant equipment fault risk monitoring and early warning method and system provided in this application.

[0020] Please see Figure 1 The diagram illustrates a flowchart of a method for monitoring and early warning of equipment failure risks in a nuclear power plant, according to an embodiment of this application. The method includes the following steps: Step S1: Collect the time sequence of vibration signals of the planetary gearbox bearings in the nuclear power plant equipment, and record it as the bearing vibration signal.

[0021] To address the inner ring fault of the planetary gear bearing, a single-axis accelerometer was placed directly above the gear ring of the gearbox to collect the vibration signal of the planetary gearbox bearing. The single-axis accelerometer had a sensitivity of 100mV / g, a sampling frequency of 25.6kHz, and a demodulation frequency range of 5.2kHz-5.6kHz. A 9234 acquisition card was used to convert the voltage signal into a digital signal, thereby obtaining the time sequence of the vibration signal of the planetary gearbox bearing, which is denoted as the bearing vibration signal.

[0022] It should be noted that the implementer can set the location and parameters of the accelerometer according to the actual situation, and this application does not impose specific restrictions.

[0023] The sampling period is set; in this embodiment, the sampling period is set to 10 seconds. In other embodiments of this application, the implementer can set the sampling period length according to the actual situation. Taking any sampling period as the current period, the following analysis is performed on the current period.

[0024] Step S2: Based on the randomness of the occurrence time of peak points in the bearing vibration signal of each sampling period and the fluctuation characteristics of the sharpness of the signal in the bearing vibration signal, construct the first fluctuation characteristic value of the bearing vibration signal of each sampling period.

[0025] In the early stages of a planetary gearbox failure, the bearing vibration signal usually contains weak fault characteristics. If maintenance is not performed in time, serious failures may occur over time. Therefore, it is necessary to detect potential failure risks based on the weak fault characteristics contained in the bearing vibration signal in the early stages of failure. However, planetary gearbox bearings are easily affected by strong background noise during operation, such as machine operating noise and the effects of gear meshing, which can mask weak fault characteristics and prevent timely and accurate detection of equipment failure risks.

[0026] Planetary gearbox bearings are circular in shape. If there is a minor wear or breakage, a sharp impact pulse will be generated in the bearing vibration signal per revolution. This impact pulse has a stable amplitude and periodic variation. However, when affected by noise, the amplitude of the impact pulse in the bearing vibration signal becomes unstable and its occurrence becomes more random. The vibration signal generated by the fault is recorded as the fault characteristic signal. To avoid the possibility of noise with a similar intensity to the fault characteristic signal, which could lead to small fluctuations in the pulse amplitude of the bearing vibration signal and make it difficult to analyze the fluctuation characteristics, it is necessary to analyze the periodicity of the peak points in the bearing vibration signal.

[0027] Therefore, for the bearing vibration signal of the current period, the bearing vibration signal is used as input, and a peak detection algorithm is used to obtain the peak points in the bearing vibration signal. Then, the first-order difference sequence of all peak points is obtained. In this embodiment, the first-order difference sequence of corresponding times is a sequence formed by arranging the time intervals between all adjacent peak points in the bearing vibration signal in chronological order. The coefficient of variation of the element values ​​in this first-order difference sequence is calculated. The obtained coefficient of variation represents the randomness of the pulse peak point fluctuation time of the planetary gearbox bearing vibration signal. The calculation of the first-order difference sequence and the coefficient of variation are well-known techniques, and the specific process will not be described in detail.

[0028] Due to the numerous sources of background noise and the presence of random noise, the regularity of impact pulses in the bearing vibration signal is reduced, resulting in lower amplitude stability and periodicity of peak point fluctuations. This causes fluctuations in the kurtosis of the bearing vibration signal at different times. To obtain the kurtosis fluctuation characteristics of the bearing vibration signal, a sliding window is set in the bearing vibration signal, with a time length of 3 sampling moments and a sliding step size of 1 sampling moment. The kurtosis of the bearing vibration signal within each sliding window is calculated. Because of the randomness of noise, the sharpness of the signal distribution changes drastically between adjacent sliding windows, which can significantly interfere with equipment fault risk monitoring. Therefore, the kurtosis of all sliding windows is arranged in chronological order to obtain a kurtosis sequence. Then, the rate of change of all elements in the kurtosis sequence except the first element is calculated and denoted as the kurtosis rate of change. The mean of the absolute values ​​of the kurtosis rates of change of all elements is calculated and denoted as the kurtosis fluctuation rate of the bearing vibration signal, representing the drastic sharp changes in the planetary gearbox bearing vibration signal under the influence of background noise. The calculation of kurtosis and rate of change are well-known techniques, and the specific process will not be elaborated further.

[0029] Because there may be small kurtosis rates of change between adjacent sliding windows in the bearing vibration signal, the significance of sharp changes in the bearing vibration signal under the influence of background noise is reduced. Therefore, to more accurately measure the sharp changes in the bearing vibration signal under the influence of background noise, the absolute values ​​of the kurtosis rates of change of all elements in the kurtosis sequence except the first element are used as inputs to the Otsu's inter-class variance algorithm to obtain a rate of change threshold. The number of elements in the kurtosis sequence whose absolute values ​​of kurtosis rates of change are greater than the rate of change threshold is counted. These elements are those whose kurtosis fluctuates significantly under the influence of background noise. The ratio of the number of these elements to the total number of elements in the kurtosis sequence is calculated and denoted as the kurtosis compensation coefficient. This coefficient is used to compensate for the kurtosis fluctuation rate, avoiding the problem of excessively low significance of sharp changes in the bearing vibration signal under the influence of background noise. The Otsu's inter-class variance algorithm is a well-known technique, and its specific process will not be elaborated further.

[0030] It should be noted that this application only provides one threshold segmentation method for threshold segmentation of kurtosis rate of change. There are many existing threshold segmentation methods, and implementers may also use other threshold segmentation algorithms to perform threshold segmentation of kurtosis rate of change. This application does not impose any specific restrictions.

[0031] Based on the above analysis, the fused value of the coefficient of variation, the kurtosis volatility, and the kurtosis compensation coefficient is taken as the first fluctuation characteristic value of the bearing vibration signal in the current sampling period, representing the random fluctuation of the planetary gearbox bearing vibration signal under the influence of background noise within the current sampling period. The fusion can be achieved through addition, multiplication, averaging, etc.

[0032] Preferably, in this embodiment of the application, the expression for the first fluctuation characteristic value is: In the formula, This represents the first fluctuation characteristic value of the bearing vibration signal in the current sampling period; Represents the hyperbolic tangent function; The coefficient of variation represents the peak point in the bearing vibration signal during the current sampling period. This represents the kurtosis fluctuation rate of the bearing vibration signal in the current sampling period. This represents the kurtosis compensation coefficient for the kurtosis volatility. In the above expression, the natural number 1 represents the coefficient of the kurtosis volatility before compensation. This is the coefficient after compensation for kurtosis volatility.

[0033] In other embodiments of this application, the expression for the first fluctuation characteristic value may also be: .

[0034] The larger the value of 'a', the less regular the periodicity of the peak points of the bearing vibration signal within the sampling period; the larger the values ​​of 'b' and 'c', the greater the influence of background noise, the worse the pulse periodicity of the bearing vibration signal, and the greater the sharp and drastic changes in the data distribution of the bearing vibration signal between different time periods, i.e., the increased kurtosis fluctuation rate. To avoid the bearing vibration signal's sharp and drastic changes under the influence of background noise having too low significance, a kurtosis compensation coefficient is calculated based on the difference in kurtosis change rate in different time windows to compensate for the coefficient of kurtosis change rate. Therefore, by comprehensively analyzing the pulse peak periodicity characteristics and kurtosis characteristics of the planetary gearbox bearing vibration signal in different time periods, the random fluctuation of the entire planetary gearbox bearing vibration signal under the influence of background noise can be reflected.

[0035] Step S3: Determine the center frequency of the spectrum based on the kurtosis diagram of the bearing vibration signal spectrum; construct the second fluctuation characteristic value of the bearing vibration signal spectrum for each sampling period based on the distribution of sideband features in the envelope spectrum of the bearing vibration signal for each sampling period and the difference between the center frequency of each sampling period and the previous sampling period.

[0036] When the background noise has little effect, the fault characteristic signal will cause obvious peaks at the first four harmonic frequencies in the envelope spectrum of the bearing vibration signal. After being affected by the background noise, more sideband features will be generated in the envelope spectrum.

[0037] To obtain the frequency domain characteristics of the planetary gearbox bearing vibration signal, the bearing vibration signal of the current sampling period is used as the input to a Fourier transform to obtain the spectrum of the bearing vibration signal. The kurtosis plot of the spectrum is obtained using the fast spectral kurtosis method, and the center of the frequency band with the largest kurtosis in the kurtosis plot is the center frequency (fundamental frequency). Furthermore, the bearing vibration signal is used as the input to a Hilbert transform to obtain the upper envelope signal of the bearing vibration signal, and this upper envelope signal is used as the input to a Fourier transform to obtain the envelope spectrum of the bearing vibration signal. The fast spectral kurtosis method, Hilbert transform, and Fourier transform are all well-known techniques, and their specific processes will not be elaborated further.

[0038] Furthermore, the peak points at the 1st to 4th harmonic frequencies of the center frequency in the envelope spectrum are obtained, and the frequency band range of the half-width at half-maximum (WHM) of these peak points is obtained. The total energy of the frequency band range of the WHM of these peak points is calculated. Simultaneously, the total energy of the entire envelope spectrum is calculated, and the ratio of the total energy to the total energy of the entire envelope spectrum is denoted as the energy concentration. A higher energy concentration indicates less sideband energy, a more prominent peak point of the harmonic frequencies of the center frequency within the envelope spectrum, and less influence of background noise on the peak points of the harmonic frequencies. The acquisition of the WHM and the calculation of the frequency band energy are well-known techniques, and the specific process will not be elaborated further.

[0039] If a planetary gearbox bearing experiences structural deformation failure, such as severe wear or fracture, the fundamental frequency in its bearing vibration signal spectrum will deviate significantly from the fundamental frequency under normal operating conditions. Taking wear failure as an example, before a planetary gearbox bearing experiences severe wear failure, there is slight wear. Slight wear characteristics are usually difficult to detect. Therefore, to prevent severe wear failure, it is necessary to extract the characteristics of slight wear signals to achieve accurate detection of wear failure.

[0040] To obtain the fundamental frequency shift characteristics of the planetary gearbox bearing vibration signal spectrum, the bearing vibration signal spectra of the current sampling period and the previous sampling period are used as inputs. The fast spectral kurtosis method is used to obtain the kurtosis maps of the spectra of the current sampling period and the previous sampling period, respectively. The fundamental frequency shift is then calculated based on the center frequency in the kurtosis maps, expressed as: ,in, This represents the fundamental frequency offset for the current sampling period. , These represent the center frequencies of the spectrum of the current sampling period and its previous sampling period, respectively. A larger fundamental frequency offset indicates a greater degree of fundamental frequency fluctuation in the planetary gearbox bearing vibration signal spectrum, and thus contains more fault characteristics.

[0041] Based on the above analysis, the second fluctuation characteristic value of the bearing vibration signal spectrum is calculated, representing the degree of fluctuation of the energy and fundamental frequency of the planetary gearbox bearing vibration signal spectrum under the influence of background noise. Preferably, the expression for the second fluctuation characteristic value is: In the formula, B represents the second fluctuation characteristic value of the spectrum of the bearing vibration signal in the current sampling period. Let f represent the hyperbolic tangent function, f represent the fundamental frequency offset of the bearing vibration signal spectrum in the current sampling period, and k represent the energy concentration of the bearing vibration signal spectrum in the current sampling period.

[0042] In other embodiments of this application, the expression for the second fluctuation characteristic value may also be: .

[0043] The smaller the energy concentration, the more sideband features there are and the more it is affected by background noise; the larger the fundamental frequency offset, the greater the degree of fundamental frequency fluctuation over time, and the more fault features the bearing vibration signal in the current sampling period contains; therefore, if there are fault features in the bearing vibration signal in the current sampling period, and the more it is affected by background noise, the larger the second fluctuation characteristic value of the bearing vibration signal spectrum will be.

[0044] Step S4: Adjust the filter length corresponding to each sampling period based on the first fluctuation feature value and the second fluctuation feature value, and combine the filtering algorithm to extract the fault feature signal in the bearing vibration signal of each sampling period in order to perform fault detection of the planetary gearbox in each sampling period.

[0045] To extract bearing vibration signals with minimal background noise and a high proportion of subtle fault features, the MOMEDA (multipoint optimal minimum entropy deconvolution adjusted) algorithm is primarily used to extract features from the spectrum of planetary gearbox bearing vibration signals. However, the feature extraction accuracy of this method is directly affected by the filter length. Therefore, it is necessary to set an appropriate filter length based on the fluctuation characteristics of the bearing vibration signal. Generally, the greater the random fluctuation of the vibration signal, the greater the influence of background noise on the vibration signal, the more complex the signal, and the easier it is to obscure subtle fault features. Thus, the filter length required by MOMEDA should be appropriately increased within a certain range.

[0046] In this embodiment, the range of values ​​for the filter length is set to... In other embodiments of this application, the implementer may set the range of filter length values ​​according to actual conditions.

[0047] Based on the above analysis, an adjustment factor is calculated according to the first and second wave characteristic values ​​of the planetary gearbox bearing vibration signal to adjust the filter length. Preferably, in this embodiment, the expressions for the adjustment factor and the adjusted filter length are: In the formula, Adjust the filter length corresponding to the current sampling period. This represents the adjustment factor for the current sampling period. This represents the preset weighting coefficients, where A and B are the first and second fluctuation feature values ​​for the current sampling period, respectively. , These are the lower and upper limits of the filter length range, respectively; This represents the floor function. In this embodiment, it will be... The value is set to 0.5. In other embodiments of this application, the implementer may set the value according to the actual situation. The value of .

[0048] The bearing vibration signal of the current sampling period is used as the input to MOMEDA for fault feature signal extraction. The filter length is adjusted to L based on the current sampling period. The filter length is typically 2-5 times the deconvolution period, and the deconvolution period can be a non-integer. In this embodiment, the deconvolution period is set to... Implementers can set the deconvolution period according to the actual situation, and this application does not impose any special restrictions.

[0049] The extracted fault feature signals are used as input to the fault detection algorithm, which outputs the fault detection result of the planetary gearbox for the current sampling period. The fault detection result includes whether the planetary gearbox is operating normally or abnormally. If an abnormal operation of the planetary gearbox is detected, an alarm device is triggered and an alarm is issued to promptly remind personnel to inspect and maintain it. If the planetary gearbox is detected to be operating normally, no alarm is triggered. There are many existing fault detection algorithms, such as... Anomaly detection algorithms, LOF anomaly detection algorithms, etc., are selected in this embodiment. The anomaly detection algorithm enables fault detection in planetary gearboxes. Implementers may also use other fault detection algorithms for fault detection in planetary gearboxes, and this application does not impose any special restrictions.

[0050] A schematic diagram illustrating the process of obtaining the first wave characteristic value is shown below. Figure 2 As shown.

[0051] Based on the same inventive concept as the above method, this application embodiment also provides a nuclear power plant equipment failure risk monitoring and early warning system, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of any one of the above-described nuclear power plant equipment failure risk monitoring and early warning methods.

[0052] In summary, this application provides a method for monitoring and early warning of fault risks in nuclear power plant equipment. It collects vibration signals from planetary gearbox bearings in nuclear power plant equipment. Based on the randomness of the peak occurrence time and the fluctuation characteristics of signal sharpness in the bearing vibration signals, a first fluctuation feature value is constructed to analyze the random fluctuation of the signal under the influence of background noise in the time domain. A second fluctuation feature value is constructed by analyzing the distribution of sideband features in the envelope spectrum of the bearing vibration signals in each sampling period and comparing it with the center frequency of the previous sampling period. This analyzes the influence of background noise on fault characteristics in the frequency domain. By combining these two methods, the filter length is appropriately adjusted, and a filtering algorithm is used to extract fault feature signals from the bearing vibration signals in each sampling period, enabling fault detection of the planetary gearbox within each sampling period. This method adaptively sets an appropriate filter length based on the current vibration signal fluctuation, providing high flexibility. It solves the problem that the vibration signals of bearings in nuclear power plant planetary gearbox equipment are easily affected by operating noise and can easily drown out weak fault feature signals, improving the extraction accuracy of fault feature signals and contributing to the accuracy of subsequent fault detection.

[0053] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this application. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0054] The various embodiments in this application are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0055] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the principles of this application should be included within the protection scope of this application.

Claims

1. A method for monitoring and early warning of equipment failure risks in nuclear power plants, characterized in that, The method includes the following steps: The time sequence of vibration signals of planetary gearbox bearings in nuclear power plant equipment is collected and denoted as bearing vibration signal; Based on the randomness of the occurrence time of peak points in the bearing vibration signal of each sampling period and the fluctuation characteristics of the sharpness of the signal in the bearing vibration signal, the first fluctuation characteristic value of the bearing vibration signal of each sampling period is constructed. The center frequency of the spectrum is determined based on the kurtosis plot of the bearing vibration signal spectrum; the second fluctuation characteristic value of the bearing vibration signal spectrum in each sampling period is constructed based on the distribution of sideband features in the envelope spectrum of the bearing vibration signal in each sampling period and the difference between the center frequency of each sampling period and the previous sampling period. Based on the first and second fluctuation feature values, the filter length corresponding to each sampling period is adjusted. Combined with the filtering algorithm, the fault feature signal in the bearing vibration signal of each sampling period is extracted to perform fault detection of the planetary gearbox in each sampling period.

2. The method for monitoring and early warning of equipment failure risks in nuclear power plants as described in claim 1, characterized in that, The process of obtaining the first fluctuation feature value is as follows: Obtain the first-order difference sequence of all peak points in the bearing vibration signal corresponding to the time intervals, and calculate the coefficient of variation of the element values ​​in the first-order difference sequence; set a sliding window in the bearing vibration signal, construct a kurtosis sequence through the kurtosis of the bearing vibration signal in each sliding window, and take the mean of the absolute values ​​of the rate of change of the remaining elements in the kurtosis sequence except for the first element as the kurtosis volatility of the bearing vibration signal; construct the kurtosis compensation coefficient of the kurtosis volatility based on the numerical distribution of the rate of change of the elements in the kurtosis sequence. The fused value of the coefficient of variation, the kurtosis volatility, and the kurtosis compensation coefficient is used as the first volatility characteristic value.

3. The method for monitoring and early warning of equipment failure risks in nuclear power plants as described in claim 2, characterized in that, The kurtosis compensation coefficient is the ratio of the number of elements whose absolute rate of change is greater than a preset rate of change threshold to the total number of elements in the kurtosis sequence.

4. The method for monitoring and early warning of equipment failure risks in nuclear power plants as described in claim 1, characterized in that, The center frequency is the center of the frequency band with the largest kurtosis in the kurtosis diagram.

5. The method for monitoring and early warning of equipment failure risks in nuclear power plants as described in claim 1, characterized in that, The process of obtaining the second fluctuation characteristic value is as follows: Based on the sideband energy analysis of the envelope spectrum, the degree of dispersion of the peak energy at the harmonic frequency of the center frequency is analyzed, and the energy concentration of the envelope spectrum is constructed. The fundamental frequency offset of each sampling period is calculated based on the difference in center frequency between adjacent sampling periods. The second fluctuation characteristic value is calculated based on the fundamental frequency offset and the energy concentration. The second fluctuation characteristic value is proportional to the fundamental frequency offset and inversely proportional to the energy concentration.

6. The method for monitoring and early warning of equipment failure risks in nuclear power plants as described in claim 5, characterized in that, The process of obtaining the energy concentration is as follows: The peak points at the first preset number of harmonic frequencies of the center frequency in the envelope spectrum are obtained. The sum of the energy of the frequency band within the half-peak width of all the peak points at the harmonic frequencies is divided by the total energy of the envelope spectrum to obtain the energy concentration of the envelope spectrum.

7. The method for monitoring and early warning of equipment failure risks in nuclear power plants as described in claim 5, characterized in that, The expression for the fundamental frequency offset is: ,in, This represents the fundamental frequency offset for the current sampling period. , These represent the center frequencies of the current sampling period and its previous sampling period, respectively.

8. The method for monitoring and early warning of equipment failure risks in nuclear power plants as described in claim 1, characterized in that, The process of obtaining the filter length is as follows: The weighted sum of the first volatility eigenvalue and the second volatility eigenvalue is denoted as the adjustment factor. The expression for the filter length is: In the formula, Adjust the filter length corresponding to the current sampling period. , These are the lower and upper limits of the filter length range, respectively; This represents the floor function.

9. The method for monitoring and early warning of equipment failure risks in nuclear power plants as described in claim 1, characterized in that, The step of extracting fault feature signals from the bearing vibration signals of each sampling period to perform fault detection of the planetary gearbox within each sampling period is as follows: The bearing vibration signal of each sampling period is used as the input of the MOMEDA algorithm. The filter length in the MOMEDA algorithm adopts the adjusted filter length corresponding to each sampling period to obtain the fault characteristic signal of each sampling period. This signal is then used as the input of the fault detection algorithm to output the fault detection result.

10. A nuclear power plant equipment failure risk monitoring and early warning system, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method as described in any one of claims 1-9.

Images