Multi-Sensor Information Fusion Fault Diagnosis Method for the Gearbox of the Circulating Water Pump in a Nuclear Power Plant

Abstract

The present invention belongs to the technical field of mechanical engineering, and specifically relates to a fault diagnosis method for the gearbox of a circulating water pump in a nuclear power plant based on multi-sensor data fusion. The method includes the following steps: respectively collecting vibration signals x(t), y(t), and z(t) at the vertical direction of the bearing seat at the input end of the gearbox of the circulating water pump in the nuclear power plant, the vertical direction of the gearbox housing, and the vertical direction of the bearing seat at the output end; respectively performing normalization processing on the vibration signals x(t), y(t), and z(t); performing correlation function fusion on the normalized vibration signals; performing Fourier transform on the fused signal to obtain the spectrum X(f) of the fused signal, and constructing a gearbox meshing state index ε according to the spectrum information; when the gearbox of the circulating water pump in the nuclear power plant is running, collecting vibration signals at different positions in real time and performing multi-sensor information fusion, calculating the meshing state index ε of the fused signal, and determining the health state of the circulating water pump gearbox according to the magnitude of the ε value. The advantages are: the vibration signals of three key parts of the circulating pump gearbox are fused by using the correlation function method.

Description

Technical Field

[0001] The present invention belongs to the technical field of mechanical engineering, and particularly relates to a fault diagnosis method for the gearbox of a circulating water pump in a nuclear power plant based on multi-sensor data fusion. Background Technique

[0002] The circulating water pump is an important device for heat extraction in a nuclear power plant, used to cool the exhausted steam discharged by the steam turbine. If this device fails during the operation of the unit, it will cause the unit to reduce its load by 50%, and in severe cases, it will lead to the shutdown of the unit and the reactor; at the same time, the structure of the circulating water pump is complex, and the overall disassembly and repair take a long time. Excessive maintenance of the circulating water pump will result in an increase in maintenance costs and human factor failures related to maintenance activities. As an important component of the circulating water pump, the gearbox of the circulating water pump has the characteristics of large power transmission, high technical requirements, and high machining accuracy requirements. The complexity and particularity of the working conditions pose a major challenge to the health status of the gearbox of the circulating water pump.

[0003] The gearbox of the circulating water pump has many parts and a complex structure. Moreover, due to the transmissibility of vibration, various vibration components of the gearbox are coupled with each other. The vibration information collected by a single sensor alone is not sufficient for a comprehensive and accurate fault diagnosis of the gearbox. In order to improve the accuracy and reliability of the fault diagnosis of the gearbox of the circulating water pump, vibration acceleration sensors are installed at different positions of the key components of the gearbox during monitoring, the information of multiple sensors is fused, and the fused vibration signal is used for the fault diagnosis of the gearbox of the circulating water pump.

[0004] In summary, a multi-sensor information fusion fault diagnosis method for the gearbox of a circulating water pump in a nuclear power plant is yet to be proposed. Summary of the Invention

[0005] The purpose of the present invention is to provide a multi-sensor information fusion fault diagnosis method for the gearbox of a circulating water pump in a nuclear power plant, which can solve the problem of multi-state information fault diagnosis of the gearbox of a circulating water pump in a nuclear power plant, so as to conduct a more reliable and accurate diagnosis.

[0006] The technical solution of the present invention is as follows: A multi-sensor information fusion fault diagnosis method for the gearbox of a circulating water pump in a nuclear power plant includes the following steps:

[0007] Step (1): Collect the vibration signals x(t), y(t), and z(t) at the vertical direction of the input end bearing seat, the vertical direction of the gearbox housing, and the vertical direction of the output end bearing seat of the gearbox of the circulating water pump in a nuclear power plant respectively;

[0008] Step (2): Normalize the vibration signals x(t), y(t), and z(t) respectively;

[0009] Step (3): Perform correlation function fusion on the normalized vibration signals;

[0010] Step (4): Perform Fourier transform on the fused signal to obtain the spectrum X(f) of the fused signal, and construct the gearbox meshing state index ε based on the spectrum information.

[0011] Step (5): When the gearbox of the circulating water pump in the nuclear power plant is running, collect vibration signals at different positions in real time and perform multi-sensor information fusion, calculate the meshing state index ε of the fused signal, and determine the health state of the circulating water pump gearbox according to the magnitude of the ε value.

[0012] The said step (2) is as follows:

[0013]

[0014] where, μ x , μ y and μ z represent the means of the vibration signals x(t), y(t) and z(t) respectively, and σ x , σ y and σ z represent the standard deviations of the vibration signals x(t), y(t) and z(t) respectively.

[0015] The said step (3) is as follows:

[0016] First, calculate the cross-correlation functions between the vibration signals x(t), y(t) and z(t):

[0017]

[0018] where, N is the number of data, and m = 1, 2, …, N - 1;

[0019] Perform pairwise cross-correlation operations on the signals to obtain the cross-correlation energy between the signals as:

[0020]

[0021] From this, the total cross-correlation energy between the vibration signals x(t), y(t) and z(t) and other signals can be obtained.

[0022] E x = E xy + E xz

[0023] E y = E xy + E yz

[0024] E z = E xz + E yz

[0025] Since the weight value of the signal is proportional to the related energy, the following relationship is obtained:

[0026]

[0027] Wherein, P x 、P y and P z are the weight values of the vibration signals x(t), y(t) and z(t) respectively. Thus, the fused vibration signal can be obtained as:

[0028] X(t) = P x *x(t) + P y *y(t) + P z *z(t).

[0029] The step (4) is as follows:

[0030] Take one-eighth of the entire frequency range as the low-frequency band and calculate the vibration energy of the low-frequency band:

[0031]

[0032] Wherein, f s is the sampling frequency;

[0033] It is known that the meshing frequency of the planetary gearbox of the circulating water pump is f m . When there are faults in the sun gear, internal gear ring and planetary gear, their fault frequencies are respectively

[0034]

[0035] Wherein, Z s 、Z r and Z p are the number of teeth of the sun gear, the number of teeth of the gear ring and the number of teeth of the planetary gear respectively, and N is the number of planetary gears;

[0036] Take the maximum value f max = max{f ds , f dr , f dp} of the fault characteristic frequencies of the sun gear, internal gear ring and planetary gear, and construct the meshing state index ε of the gearbox:

[0037]

[0038] In the step (5), when ε < 30%, the gear is in good running condition.

[0039] In the step (5), when 30% ≤ ε < 50%, the gear has a weak fault.

[0040] In step (5), when ε ≥ 50%, the gear has a serious fault.

[0041] The beneficial effects of the present invention are as follows: By analyzing the vibration mechanism of the circulating pump gearbox, the present invention uses the correlation function method to fuse the vibration signals of three key parts of the circulating pump gearbox. Based on the fused signal, a meshing state index is established, and the index determination rules for different health states of the gearbox are given, realizing the monitoring and diagnosis of the operating state of the circulating water pump gearbox. This strategy makes full use of all the vibration information of the circulating pump gearbox, avoiding the problems of insufficient and incomplete information existing in the information of a single sensor, and can monitor and diagnose the operating state of the gearbox more comprehensively and accurately. Description of the Drawings

[0042] Figure 1 Flow chart of the multi-sensor information fusion fault diagnosis method for the circulating water pump gearbox of a nuclear power plant proposed by the present invention;

[0043] Figure 2 Vibration signals of the fused healthy and cracked fault planetary gearboxes drawn in the embodiments of the present invention;

[0044] Figure 3 Spectrum diagram of the vibration signals of the fused healthy and cracked fault planetary gearboxes drawn in the embodiments of the present invention. Detailed Embodiments

[0045] The present invention will be further described in detail below with reference to the drawings and specific embodiments.

[0046] The present invention proposes a multi-sensor information fusion fault diagnosis method for the circulating water pump gearbox of a nuclear power plant. To verify the effectiveness of the method proposed by the present invention, a planetary gearbox vibration test platform is used for analysis. The planetary gearbox consists of a sun gear, an internal gear ring, and three evenly distributed planetary gears, and the planet carrier is connected to the output shaft.

[0047] As Figure 1 shown, a multi-sensor information fusion fault diagnosis method for the circulating water pump gearbox of a nuclear power plant includes the following specific steps:

[0048] Step (1): Collect the vibration signals x(t), y(t), and z(t) at the vertical direction of the input end bearing seat, the vertical direction of the gearbox housing, and the vertical direction of the output end bearing seat of the circulating water pump gearbox of the nuclear power plant respectively.

[0049] Step (2): Normalize the vibration signals x(t), y(t), and z(t) respectively:

[0050]

[0051] where μx , μ y and μ z represent the mean values of the vibration signals x(t), y(t), and z(t) respectively, and σ x , σ y and σ z represent the standard deviations of the vibration signals x(t), y(t), and z(t) respectively.

[0052] Step (3): Perform cross-correlation function fusion on the normalized vibration signals. The specific steps are as follows:

[0053] First, calculate the cross-correlation functions between the vibration signals x(t), y(t), and z(t):

[0054]

[0055]

[0056] where N is the number of data, m = 1, 2,..., N - 1, and R xy , R xz and R yz represent the cross-correlation functions between the vibration signals x(t) and y(t), x(t) and z(t), and y(t) and z(t) respectively.

[0057] Perform pairwise cross-correlation operations on the signals to obtain the cross-correlation energy between the signals as:

[0058]

[0059] where E xy , E xz and E yz represent the cross-correlation energies between the vibration signals x(t) and y(t), x(t) and z(t), and y(t) and z(t) respectively. Thus, the total cross-correlation energy between the vibration signals x(t), y(t), and z(t) and other signals can be obtained.

[0060] E x = E xy + E xz

[0061] E y = E xy + E yz

[0062] E z = E xz + E yz

[0063] where E x , E y and E zrespectively represent the total correlation energy between the vibration signals x(t), y(t), and z(t) and other signals. Since the weight of the signal is proportional to the correlation energy, the following relationship can be obtained:

[0064]

[0065] Among them, P x , P y and P z are the weights of the vibration signals x(t), y(t), and z(t) respectively. Thus, the fused vibration signal can be obtained as:

[0066] X(t) = P x *x(t) + P y *y(t) + P z *z(t)

[0067] Step (4): Perform Fourier transform on the fused signal to obtain the spectrum X(f) of the fused signal. According to the spectrum information, construct the gearbox meshing state index ε. The specific steps are as follows:

[0068] Take one-eighth of the entire frequency range as the low-frequency band and calculate the vibration energy of the low-frequency band:

[0069]

[0070] Among them, f s is the sampling frequency, X(f) is the spectrum of the fused signal, and P is the vibration energy of the low-frequency band.

[0071] Given that the meshing frequency of the circulating water pump planetary gearbox is f m , when there are faults in the sun gear, internal gear ring, and planetary gears, their fault frequencies are respectively

[0072]

[0073] Among them, Z s , Z r and Z p are the number of teeth of the sun gear, the number of teeth of the gear ring, and the number of teeth of the planetary gear respectively, N is the number of planetary gears, f ds , f dr and f dp are the fault characteristic frequencies when there are faults in the sun gear, internal gear ring, and planetary gears respectively.

[0074] Take the maximum value f max = max{f ds , f dr , f dp} of the fault characteristic frequencies of the sun gear, internal gear ring, and planetary gears, and construct the gearbox meshing state index ε:

[0075]

[0076] Step (5): When the nuclear power plant circulating water pump gearbox is running, vibration signals at different positions are collected in real time and multi-sensor information fusion is performed. Calculate the meshing state index ε of the fused signal, and determine the health state of the circulating water pump gearbox according to the magnitude of the ε value:

[0077] When ε < 30%, the gear is in good running condition;

[0078] When 30% ≤ ε < 50%, the gear has a minor fault;

[0079] When ε ≥ 50%, the gear has a serious fault.

[0080] Example:

[0081] In this example, a comparative analysis is performed on a healthy planetary gearbox and a planetary gearbox with a crack fault in one of the planetary gears, including the following steps:

[0082] Step (1): Collect vibration signals x1(t), y1(t), z1(t) and x2(t), y2(t), and z2(t) at the vertical direction of the input end bearing seat, the vertical direction of the gearbox housing, and the vertical direction of the output end bearing seat of the healthy and cracked planetary gearboxes respectively;

[0083] Step (2): Normalize the multi-measurement point vibration signals of the healthy and cracked planetary gearboxes respectively;

[0084] Step (3): Perform correlation function fusion on the normalized vibration signals of the healthy and cracked planetary gearboxes to obtain the fused healthy and cracked vibration signals X1(t) and X2(t);

[0085] Step (4): Perform Fourier transform on the fused healthy and cracked signals respectively to obtain the spectra X1(f) and X2(f) of the fused signals;

[0086] Step (5): Calculate the relevant characteristic frequencies of the planetary gearbox according to the test settings and the parameters of the planetary gearbox, as shown in Table 1;

[0087] Table 1 Relevant characteristic frequencies of the planetary gearbox

[0088]

[0089] Step (6): Take the maximum value f max = 90Hz of the fault characteristic frequencies of the sun gear, internal gear ring and planetary gear, and then calculate the meshing state indices ε1 and ε2 of the healthy and cracked planetary gearboxes respectively:

[0090]

[0091] It can be seen from the calculation results that according to the threshold set in the present invention, the meshing state index ε1 of the healthy planetary gearbox is less than 30%, and it can be judged that the gear is in good operating condition; while the meshing state index of the cracked planetary gearbox is 30% ≤ ε2 < 50%, and it can be judged that the gear has a weak fault, which is consistent with the actual situation, thereby the effectiveness of the diagnostic method proposed in the present invention can be judged.

Claims

1. A multi-sensor information fusion fault diagnosis method for the gearbox of a circulating water pump in a nuclear power plant, characterized in that, It includes the following steps: Step (1): Collect the vibration signals x(t), y(t), and z(t) at the vertical direction of the bearing seat at the input end of the gearbox of the nuclear power plant circulating water pump, the vertical direction of the gearbox housing, and the vertical direction of the bearing seat at the output end respectively; Step (2): Normalize the vibration signals x(t), y(t), and z(t) respectively; Step (3): Perform cross-correlation function fusion on the normalized vibration signals; Step (4): Perform Fourier transform on the fused signal to obtain the spectrum X(f) of the fused signal, and construct the gearbox meshing state index ε according to the spectrum information; The said step (4) is Take one-eighth of the entire frequency range as the low-frequency band and calculate the vibration energy in the low-frequency band: where f s is the sampling frequency; The meshing frequency of the known planetary gearbox of the circulating water pump is f m , when faults occur in the sun gear, internal gear ring and planetary gear, their fault frequencies are respectively Among them, Z s , Z r and Z p are the number of teeth of the sun gear, the number of teeth of the ring gear, and the number of teeth of the planet gear respectively, and N is the number of planet gears; Take the maximum value \(f\) of the fault characteristic frequencies of the sun gear, internal gear ring and planet gears max = max{\(f\) ds , \(f\) dr , \(f\) dp}, and construct the gearbox meshing state index \(\varepsilon\): Step (5): When the gearbox of the nuclear power plant circulating water pump is running, collect the vibration signals at different positions in real time and perform multi-sensor information fusion, calculate the meshing state index ε of the fused signal, and determine the health state of the gearbox of the circulating water pump according to the magnitude of the ε value.

2. The multi-sensor information fusion fault diagnosis method for the gearbox of the circulating water pump in a nuclear power plant according to claim 1, wherein: The said step (2) is Among them, μ x , μ y and μ z represent the means of the vibration signals x(t), y(t), and z(t) respectively, and σ x , σ y and σ z represent the standard deviations of the vibration signals x(t), y(t), and z(t) respectively.

3. The multi-sensor information fusion fault diagnosis method for the gearbox of the circulating water pump in a nuclear power plant according to claim 1, wherein: The said step (3) is First, calculate the cross-correlation functions between the vibration signals x(t), y(t), and z(t): where N is the number of data, and m = 1, 2, …, N - 1; Perform pairwise cross-correlation operations on the signals to obtain the cross-correlation energy between the signals as: Thus, the total cross-correlation energy between the vibration signals x(t), y(t), and z(t) and other signals can be obtained. E x = E xy + E xz E y = E xy + E yz E z = E xz + E yz Since the weight of the signal is proportional to the cross-correlation energy, the following relationship is obtained: Among them, P x , P y and P z are the magnitudes of the weights of the vibration signals x(t), y(t), and z(t) respectively, so that the fused vibration signal can be obtained as follows: X(t) = P x * x(t) + P y * y(t) + P z * z(t).

4. The multi-sensor information fusion fault diagnosis method for the gearbox of the circulating water pump in a nuclear power plant according to claim 1, wherein: In the said step (5), when ε < 30%, the gear running state is good.

5. The multi-sensor information fusion fault diagnosis method for the gearbox of the circulating water pump in a nuclear power plant according to claim 1, characterized in that: In the said step (5), when 30% ≤ ε < 50%, the gear has a weak fault.

6. The multi-sensor information fusion fault diagnosis method for the gearbox of the circulating water pump in a nuclear power plant according to claim 1, wherein: In the said step (5), when ε ≥ 50%, the gear has a serious fault.

Images