Vibration test fault diagnosis method for speed reducer of vertical ship lift

Abstract

The invention provides a vibration test fault diagnosis method for a speed reducer of a vertical ship lift, and the method comprises the steps: receiving a voltage signal transmitted by a vibration sensor through a data collection instrument, converting the voltage signal into a discrete digital value, and carrying out the signal processing through a computer. According to the fault diagnosis method, a particle swarm algorithm is used for optimizing key parameters of a swarm decomposition algorithm, a vibration signal is decomposed, kurtosis, envelope entropy and correlation coefficients of signal components are extracted to serve as feature values, spectral analysis is conducted on the component with the maximum feature value, and whether a fault occurs or not is judged. By optimizing the group decomposition algorithm and extracting the multi-dimensional features, the early fault and the complex fault of the speed reducer can be effectively identified, and the method has the advantages of high diagnosis accuracy, high adaptability, good practicability and the like, and can be widely applied to online monitoring and fault diagnosis of the ship lift speed reducer.

Description

Technical Field

[0001] The invention belongs to the technical field of waterway navigation guarantee, and particularly relates to a vibration test fault diagnosis method for a vertical ship lift reducer. Background Art

[0002] As the mainstream ship lift in the world at present, the vertical ship lift has many characteristics such as complex structure, high running precision requirements, heavy load, large lifting height, etc. These characteristics lead to the complex operation conditions and high operation intensity of the drive mechanism in the ship lift, and the maintenance tasks of personnel are also relatively heavy. Among them, each set of drive mechanism is provided with two main reducers (the reducer located outside the drive mechanism and close to the safety mechanism side is reducer A, and the reducer close to the synchronous shaft system side is reducer B). Both reducers are four-stage (five-axis) cylindrical spur gear reducers, with the same external dimensions and internal gear arrangements. The low-speed shaft of each reducer extends out of the shaft end on its inner side and is connected to the universal coupling assembly, and drives the pinion shaft through the universal coupling; the high-speed shaft of the reducer extends out on both sides of the reducer and is connected to the coupling with a brake. The intermediate shaft of reducer A extends out on both sides of the reducer. One end is connected to the bevel gear box through a safety clutch, and its power is transmitted to the safety mechanism through the intermediate shaft of the gear coupling and the bevel gear box; the other end is installed with an absolute rotary encoder of the electrical control system. The outer output shaft of the high-speed shaft of reducer B is connected to the bevel gear box through the intermediate shaft of the gear coupling, and the power is transmitted to the synchronous shaft system through the bevel gear box. As the connection device between the drive motor of each drive point and the synchronous shaft system, the reducer provides an important guarantee for the safe and stable operation of the ship chamber.

[0003] As the core component of the ship lift transmission system, the operation state of the ship lift reducer is directly related to the safety and reliability of the ship lift. Due to the long-term operation of the ship lift reducer under complex working conditions such as heavy load, low speed, and impact, faults such as gear wear, tooth breakage, and bearing damage are likely to occur. Traditional fault diagnosis methods mainly rely on spectrum analysis and time-domain analysis, and it is difficult to effectively identify early faults and complex faults. In recent years, signal decomposition algorithms have been widely used in the field of fault diagnosis, but for the special working conditions and structural characteristics of the ship lift reducer, the existing fault diagnosis methods still have certain limitations. Summary of the Invention

[0004] To solve the current technical problems, the main purpose of the present invention is to provide a vibration test fault diagnosis method for a vertical ship lift reducer, which can effectively improve the accuracy and reliability of reducer fault diagnosis and provide guarantee for the safe operation of the ship lift.

[0005] To achieve the above technical features, the object of the present invention is realized as follows: A vibration test fault diagnosis method for a vertical ship lift reducer, comprising the following steps:

[0006] S1. Collect the vibration signals of the ship lift reducer, process the vibration signals, and then obtain the corresponding vibration data;

[0007] S2. Use the particle swarm algorithm to optimize the key parameters in the group decomposition algorithm;

[0008] S3. Process the vibration data of the ship lift reducer through the group decomposition algorithm to obtain the components of the vibration data;

[0009] S4. Calculate three indexes of kurtosis, envelope entropy and correlation coefficient of the component, and combine the three indexes according to formula (1) as the eigenvalue of the signal component:

[0010]

[0011] S5. Conduct spectrum analysis on the component with the largest eigenvalue, extract its characteristic frequency to judge whether the ship lift reducer fails.

[0012] Preferably, the specific steps for collecting the vibration data of the ship lift reducer in S1 are as follows:

[0013] S11. Arrange corresponding vibration sensors at the monitoring points of each reducer, specifically set at 4 positions of the reducer, that is, arrange 4 vibration sensors in the directions perpendicular and parallel to the output shaft, capture the vibration signals of the reducer and transmit voltage signals;

[0014] S12. The data acquisition instrument receives the voltage signal, and after signal filtering and amplification conditioning, converts the voltage signal into a digital signal through the analog-to-digital converter of the data acquisition instrument;

[0015] S13. The continuous voltage signal is converted into discrete digital values through the analog-to-digital converter, and the digital values are analyzed and processed by a computer to obtain the final vibration data.

[0016] Preferably, according to the actual working conditions of the ship lift reducer, when the motor starts, stops and makes an emergency stop, the vibration of the reducer is the largest. Collect the vibration signals under these three working conditions to obtain more characteristic information.

[0017] Preferably, the vibration sensor adopts a single-axis acceleration sensor.

[0018] Preferably, the data acquisition instrument has built-in filtering and signal amplification functions.

[0019] Preferably, the group decomposition algorithm in S2 specifically includes the following steps:

[0020] S21. Determine the center frequency of the initial signal according to the power spectrum peak

[0021]

[0022] Wherein, S x (ω) represents the Welch power spectrum, q represents the q-th time the frequency serves as the center frequency during the group decomposition process, and P th is the threshold; argmax ω S x (ω) represents the frequency corresponding to the maximum value taken by S x (ω) each time;

[0023] For S22, filter the input signal x[n] to obtain the output signal y[n], and calculate the variance StD value between the output signal and the input signal:

[0024]

[0025] When StD > StD th , repeat the filtering with y[n] as the input signal until StD < StD th , and denote x i '[n] = y[n], where StD th is the variance threshold;

[0026] For S23, update the input signal;

[0027]

[0028] Wherein, R xi (τ) is the cross-correlation function (-(L - 1) ≤ τ ≤ (L - 1)), and τ is the time delay;

[0029] For S24, use the new "input" signal and repeat steps S21 - S23 until At this time, the remaining amount of the input signal is r[n];

[0030] For S25, calculate the OC component;

[0031]

[0032] Ω dominant : {ω: ω = ω dom}; (9).

[0033] Preferably, in S4, kurtosis is used to measure the impulse impact. The larger the kurtosis, the more obvious the peak of the fault frequency. The envelope entropy is used to measure the uniformity of periodic impulses. The more impulses are detected, the cleaner the envelope spectrum, and the smaller the envelope entropy. The correlation coefficient reflects the correlation degree between variables. The eigenvalue comprehensively considers the advantages of kurtosis, envelope entropy, and correlation coefficient, making the eigenvalue more sensitive to periodic impact signals. The larger the eigenvalue, the more fault information contained in the component, and the more reference value it has.

[0034] Preferably, the threshold P th and the variance threshold StD th are set in the optimized range of (0, 1);

[0035] At this time, the optimization problem is described as:

[0036]

[0037] That is, on the premise of satisfying the value range, the threshold combination that can obtain the maximum eigenvalue is the optimal parameter combination.

[0038] Preferably, the fault diagnosis of the shiplift reducer in S5 specifically includes:

[0039] The typical characteristics of gear wear are exciting the natural frequency of the gear and having sidebands at intervals of the rotational speed of the worn gear on both sides of the natural frequency of the gear. The amplitude of the gear meshing frequency may also remain unchanged. When the wear is obvious, high-amplitude sidebands appear on both sides of the gear meshing frequency, and these sidebands may be better wear indicators than the gear meshing frequency itself.

[0040] Preferably, the fault diagnosis of the shiplift reducer in S5 specifically further includes:

[0041] When a gear tooth breakage fault occurs in the reducer gear, its manifestation in the time domain is an increase in amplitude and a certain periodic impact. The amplitude of the vibration energy increases significantly, and many side frequency bands with large amplitudes and wide distributions appear near the meshing frequency; the fault location and type can be judged through the side frequency band structure. The more serious the gear tooth breakage fault and the more the number of broken teeth, the more obvious the impact phenomenon.

[0042] The present invention has the following beneficial effects:

[0043] 1. High accuracy: By optimizing the group decomposition algorithm and extracting multi-dimensional features, it can effectively identify the early faults and complex faults of the reducer, improving the diagnostic accuracy.

[0044] 2. Strong adaptability: Using the particle swarm algorithm to optimize the group decomposition parameters can meet the fault diagnosis requirements of the reducer under different working conditions.

[0045] 3. Good practicality: This method is simple to operate and easy to implement, and can be widely applied to the online monitoring and fault diagnosis of the shiplift reducer. Description of the Drawings

[0046] The present invention will be further described below with reference to the drawings and embodiments.

[0047] Figure 1 is the fault diagnosis flow chart of the present invention.

[0048] Figure 2 It is the broken tooth gear fault signal diagram in the embodiment of the present invention.

[0049] Figure 3 It is the process diagram of key parameter optimization in the embodiment of the present invention.

[0050] Figure 4 It is the schematic diagram of signal decomposition in the embodiment of the present invention.

[0051] Figure 5 It is the spectrum analysis diagram of component signals in the embodiment of the present invention.

[0052] Figure 6 It is the schematic diagram of the structure of the five-axis four-stage speed reducer of the ship lift.

[0053] Figure 7 It is the schematic diagram of the installation position of the vibration sensor in the embodiment of the present invention.

[0054] Figure 8 It is the schematic diagram of the vibration test system of the speed reducer. Specific embodiments

[0055] The following further describes the embodiments of the present invention with reference to the accompanying drawings.

[0056] Embodiment 1:

[0057] Refer to Figure 1-8 , a vibration test fault diagnosis method for the speed reducer of a vertical ship lift, comprising the following steps:

[0058] S1, collect the vibration signals of the ship lift speed reducer, and process the vibration signals to obtain corresponding vibration data;

[0059] Specifically, it includes the following steps:

[0060] S11, arrange corresponding vibration sensors at the monitoring points of each speed reducer, specifically set at 4 positions of the speed reducer, that is, arrange 4 vibration sensors in the directions perpendicular and parallel to the output shaft, capture the vibration signals of the speed reducer and transmit voltage signals;

[0061] S12, the data acquisition instrument receives the voltage signal, and after signal filtering and amplification conditioning, converts the voltage signal into a digital signal through the analog-to-digital converter of the data acquisition instrument;

[0062] S13, the continuous voltage signal is converted into discrete digital values through the analog-to-digital converter, and the digital values are analyzed and processed by a computer to obtain the final vibration data.

[0063] Furthermore,

[0064] S2, use the particle swarm algorithm to optimize the key parameters in the group decomposition algorithm;

[0065] Specifically, it includes the following steps:

[0066] S21. Determine the center frequency of the initial signal according to the peak value of the power spectrum

[0067]

[0068] In the formula, S x (ω) represents the Welch power spectrum, q represents the qth time this frequency serves as the center frequency during the group decomposition process, and p th is the threshold; argmax ω s x (ω) represents the frequency corresponding to the maximum value taken each time by S x (ω);

[0069] S22. Filter the input signal x[n] to obtain the output signal y[n], and calculate the variance StD value between the output signal and the input signal:

[0070]

[0071] When StD > StD th , repeat the filtering with y[n] as the input signal until StD < StD th , denote x i '[n] = y[n], where StD th is the variance threshold;

[0072] S23. Update the input signal;

[0073]

[0074]

[0075] In the formula, is the cross-correlation function (-(L - 1) ≤ τ ≤ (L - 1)), and τ is the time delay;

[0076] S24. Use the new "input" signal to repeat steps S21 - S23 until the remaining amount of the input signal is r[n] at this time;

[0077] S25. Calculate the OC component;

[0078]

[0079] Ω dominant : {ω: ω = ω dom}; (9).

[0080] Among them, the threshold P th and the variance threshold StDth The optimization range is set to (0, 1);

[0081] At this time, the optimization problem is described as:

[0082]

[0083] That is, under the premise of satisfying the value range, the threshold combination that can obtain the maximum eigenvalue is the optimal parameter combination.

[0084] S3. Process the vibration data of the ship lift reducer through the group decomposition algorithm to obtain the components of the vibration data;

[0085] S4. Calculate three indicators: kurtosis, envelope entropy, and correlation coefficient of the components, and combine the three indicators according to formula (1) as the eigenvalue of the signal component:

[0086]

[0087] Among them, kurtosis is used to measure the impulse impact. The larger the kurtosis, the more obvious the peak of the fault frequency. Envelope entropy is used to measure the uniformity of periodic impacts. The more pulses are detected, the cleaner the envelope spectrum, and the smaller the envelope entropy. The correlation coefficient reflects the correlation degree between variables. The eigenvalue comprehensively considers the advantages of kurtosis, envelope entropy, and correlation coefficient, making the eigenvalue more sensitive to periodic impact signals. The larger the eigenvalue, the more fault information contained in the component and the more reference value it has.

[0088] S5. Conduct spectrum analysis on the component with the largest eigenvalue, extract its characteristic frequency to determine whether the ship lift reducer has a fault.

[0089] Among them, the typical characteristics of gear wear are to excite the natural frequency of the gear and have sidebands at intervals of the rotational speed of the worn gear on both sides of the natural frequency of the gear. The amplitude of the gear meshing frequency may also remain unchanged. When the wear is obvious, high-amplitude sidebands appear on both sides of the gear meshing frequency, and these sidebands may be better wear indicators than the gear meshing frequency itself.

[0090] Among them, when a gear has a tooth breakage fault, its manifestation in the time domain is that the amplitude increases and has a certain periodic impact. The amplitude of the vibration energy increases significantly, and many side frequency bands with large amplitudes and wide distributions appear near the meshing frequency. The fault location and type can be judged through the side frequency band structure. The more serious the gear tooth breakage fault and the more broken teeth, the more obvious the impact phenomenon.

[0091] Furthermore, considering the actual working conditions of the ship lift reducer, the vibration of the reducer is the largest when the motor starts, stops, and makes an emergency stop. Based on this, collecting the vibration signals under these three working conditions can obtain more characteristic information.

[0092] Furthermore, the vibration sensor uses a uniaxial acceleration sensor. By using a uniaxial acceleration sensor, the vibration signal of the reducer under complex working conditions can be reliably achieved.

[0093] Furthermore, the data collector has built-in filtering and signal amplification functions. Through the above data collector, the data processing of the collected vibration signal can be realized.

[0094] Embodiment 2:

[0095] In this embodiment, in step S2 of the present invention, the particle swarm optimization algorithm is used to optimize the key parameters of the group decomposition algorithm. The key parameters are P th and StD th , in the present invention, the optimization range of the thresholds P th and StD th is set to (0, 1). Therefore, the optimization problem of this method can be described as:

[0096]

[0097] Here, the thresholds P th and StD th can take any value in [0.01, 0.99].

[0098] Specifically, in order to further reduce the computational amount of optimization, a preprocessing for testing the time length of each iteration is introduced. The value range of the threshold is the same, from 0.01 to 0.99, with an interval of 0.3. If the time of each iteration is less than 100 s, a larger value is used as the new boundary. For example, if the iteration time with 0.61 as the threshold is greater than 100 s and the iteration time with 0.31 as the threshold is less than 100 s, then [0.01, 0.31] is used as the new optimization range.

[0099] Specifically, in step S2 of the present invention, the number of particles m in the particle swarm optimization algorithm is 30, the maximum number of iterations L is 15, the eigenvalues of each component are calculated, and the velocity and position of each particle are updated and recorded. If the number of iterations i ≥ L, the iteration ends. Otherwise, continue the iteration and set i = i + 1.

[0100] Specifically, taking the broken tooth fault as an example, Figure 2 represents the vibration signal of the broken tooth gear.

[0101] Specifically, in the broken tooth fault, the number of teeth of the normal gear is 15, the number of teeth of the broken tooth gear is 110, the transmission ratio is 7.33, the rotational speed is 1450 rpm, its fault characteristic frequency is 25 Hz, and the meshing frequency is 365 Hz.

[0102] The optimization results are asFigure 3 As shown, the key parameters obtained at this time are (0.0677, 0.0337).

[0103] The optimized key parameters are brought into the group decomposition algorithm to process the broken tooth gear signal, and the signal components are obtained as Figure 4 shown.

[0104] Calculate the eigenvalues of each component, as shown in Table 1 below:

[0105] Table 1 Eigenvalues of Components

[0106]

[0107] By calculating the eigenvalues of each component, it can be known that Component 7 is the optimal component, and its spectrum is as Figure 4 shown. Observing its spectrum, it can be known that the fault characteristic frequency is 24.41 Hz, which is relatively close to the theoretical value of 25 Hz, and the second harmonic is also easy to identify. The amplitude at the meshing frequency (365 Hz) of the gear is very obvious, there are sideband phenomena on both sides, and there are also high-order harmonic components. The sideband bandwidth is basically the same as the fault characteristic frequency, and it can be judged that a broken tooth fault has occurred.

[0108] The above-described examples only represent certain embodiments of the present invention. The description is relatively specific and detailed, but it should not be construed as a limitation on the scope of the present invention; it should be noted that for those of ordinary skill in the art, without departing from the concept of the present invention, several deformations and improvements can still be made, and these all belong to the protection scope of the present invention; therefore, the protection scope of the present invention should be subject to the appended claims.

Claims

1. A vibration test fault diagnosis method for a vertical ship lift reducer, characterized in that, It includes the following steps: S1. Collect the vibration signals of the ship lift reducer, process the vibration signals, and then obtain the corresponding vibration data; S2. Use the particle swarm algorithm to optimize the key parameters in the group decomposition algorithm; S3. Process the vibration data of the ship lift reducer through the group decomposition algorithm to obtain the components of the vibration data; S4. Calculate three indexes of kurtosis, envelope entropy and correlation coefficient of the component, and combine the three indexes according to formula (1) as the eigenvalue of the signal component: S5. Conduct spectrum analysis on the component with the largest eigenvalue, extract its characteristic frequency, and judge whether the ship lift reducer fails.

2. The vibration test fault diagnosis method for a vertical ship lift reducer according to claim 1, wherein The specific steps for collecting the vibration data of the ship lift reducer in S1 are as follows: S11. Arrange corresponding vibration sensors at the monitoring points of each reducer, specifically set at 4 positions of the reducer, that is, arrange 4 vibration sensors in the directions perpendicular and parallel to the output shaft, capture the vibration signals of the reducer and emit voltage signals; S12. The data acquisition instrument receives the voltage signals, and after signal filtering and amplification conditioning, converts the voltage signals into digital signals through the analog-to-digital converter of the data acquisition instrument; S13. The continuous voltage signals are converted into discrete digital values through the analog-to-digital converter, and the computer analyzes and processes the digital values to obtain the final vibration data.

3. The vibration test fault diagnosis method for a vertical ship lift reducer according to claim 2, characterized in that, According to the actual working conditions of the ship lift reducer, when the motor starts, stops and makes an emergency stop, the vibration of the reducer is the largest. Collect the vibration signals under these three working conditions to obtain more characteristic information.

4. The vibration test fault diagnosis method for a vertical ship lift reducer according to claim 2, wherein The vibration sensor adopts a single-axis acceleration sensor.

5. The vibration test fault diagnosis method of a vertical ship lift speed reducer according to claim 2, characterized in that, The data acquisition instrument has built-in filtering and signal amplification functions.

6. The vibration test fault diagnosis method of a vertical ship lift reducer according to claim 2, characterized in that, The specific steps of the group decomposition algorithm in S2 are as follows: S21, determine the center frequency of the initial signal according to the peak value of the power spectrum Where, S x (ω) represents the Welch power spectrum, q represents the q-th time the frequency serves as the center frequency during the group decomposition process, and P th is the threshold; argmax ω S x (ω) represents the frequency corresponding to the maximum value taken each time by S x (ω); S22. Filter the input signal x[n] to obtain the output signal y[n], and calculate the variance StD value of the output signal and the input signal; When StD > StD th , filter with y[n] as the input signal repeatedly until StD < StD th , denote x i '[n] = y[n], where StD th is the variance threshold; S23. Update the input signal; In the formula, is the cross-correlation function (-(L - 1) ≤ τ ≤ (L - 1)), where τ is the time delay; S24, using the new "input" signal, repeat steps S21 - S23 until the input signal margin is r[n] at this time; S25. Calculate the OC component; Ω dominant : {ω: ω = ω dom}; (9).

7. The vibration test fault diagnosis method for a vertical ship lift reducer according to claim 2, characterized in that In S4, kurtosis is used to measure the impulse impact. The larger the kurtosis, the more obvious the peak of the fault frequency. Envelope entropy is used to measure the uniformity of periodic impacts. The more pulses are detected, the cleaner the envelope spectrum, and the smaller the envelope entropy. The correlation coefficient reflects the correlation degree between variables. The eigenvalue comprehensively considers the advantages of kurtosis, envelope entropy and correlation coefficient, making the eigenvalue more sensitive to periodic impact signals. The larger the eigenvalue, the more fault information contained in the component and the more reference value.

8. A vibration test fault diagnosis method for a vertical ship lift reducer according to claim 6, characterized in that The threshold P th and the variance threshold StD th are set to have an optimization range of (0, 1); At this time, the optimization problem is described as: That is, under the premise of satisfying the value range, the threshold combination that can obtain the largest eigenvalue is the optimal parameter combination.

9. The method for vibration test fault diagnosis of a vertical ship lift reducer according to claim 1, wherein The specific ship lift reducer fault diagnosis in S5 includes: The typical characteristics of gear wear are exciting the natural frequency of the gear and having sidebands at intervals of the rotational speed of the worn gear on both sides of the natural frequency of the gear. The amplitude of the gear meshing frequency may also remain unchanged. When the wear is obvious, high-amplitude sidebands appear on both sides of the gear meshing frequency. These sidebands may be better wear indicators than the gear meshing frequency itself.

10. The vibration test fault diagnosis method for a vertical ship lift reducer according to claim 9, characterized in that, The specific ship lift reducer fault diagnosis in S5 also includes: When the reducer gear has a tooth breakage fault, its manifestation in the time domain is that the amplitude increases and has a certain periodic impact. The amplitude of the vibration energy increases significantly, and many sidebands with large amplitudes and wide distributions appear near the meshing frequency. The fault location and type can be judged through the sideband structure. The more serious the tooth breakage fault of the gear and the more the number of broken teeth, the more obvious the impact phenomenon will be.

Images