Pod propulsor bearing health monitoring device and method

Abstract

The invention relates to a pod propulsor bearing health monitoring device and a pod propulsor bearing health monitoring method, a driving system in the device is in driving connection with a complete machine test bed and a rolling bearing, the complete machine test bed and the rolling bearing are driven by the driving system to work, and a plurality of vibration sensors are arranged on a bearing seat of the rolling bearing. And the vibration sensor is connected with the PC end computer through the accelerometer acquisition card and the data terminal memory card. The method comprises the following steps: compressing collected original vibration data of a rolling bearing to obtain sparse vibration data; and a compressed sensing and sparse reconstruction method is used to obtain vibration data after sparse reduction, index decomposition is carried out on the vibration data, indexes are classified by using a density clustering algorithm, and a rolling bearing fault diagnosis result is obtained. By adopting the pod propeller bearing health monitoring device and method, the reliability and the operation efficiency of the pod propeller can be improved, the maintenance strategy can be optimized, the maintenance cost can be reduced, and the intelligent development of the pod propeller can be promoted.

Description

Technical Field

[0001] The invention relates to a pod propeller bearing of a ship power system, and in particular to a pod propeller bearing health monitoring device and method. Background Art

[0002] Pod thruster is a new type of special thruster, which belongs to the high-tech field of ship electric propulsion. The design of the propulsion device adopts a pod-type structure, breaking through the "diesel engine plus open transmission shaft system" thruster design pattern, and bringing changes to electric propulsion. The use of pod-type electric propulsion reduces the water resource pollution problem caused by fuel emissions. The pod thruster is a combination of a propulsion motor and a propeller installed in a davit housing, and this assembly (propulsion motor, davit housing, propeller) is "suspended" at the stern with a rotating base, and the hull only needs to provide the power required by the propulsion motor. The direction of travel of the ship can be changed by rotating the pod thruster 360°.

[0003] The rolling bearing of the pod thruster is one of the key rotating parts of the propulsion motor. The temperature around the propulsion motor is high, and there are certain iron filings in the lubricating medium. The working environment of the rolling bearing is harsh and complex, and various faults are prone to occur, resulting in abnormal propulsion motors and even major safety accidents, causing serious casualties and economic losses. In order to avoid damage to the rolling bearings, vibration sensors, temperature sensors, liquid level sensors, etc. are installed at the rolling bearings at home and abroad to conduct empirical evaluation of the data collected by the sensors. This empirical evaluation method generally has the following problems:

[0004] 1. The amount of data collected by the sensors during the operation of the pod thruster is large, which requires a large storage space to meet. At the same time, the collected sensor signals cannot be well evaluated.

[0005] 2. The service conditions of the rolling bearings of pod thrusters are complex, the amount of data generated during operation is large, and the relationship between the data is complex. Traditional data processing methods often find it difficult to effectively extract useful information, resulting in inaccurate monitoring results.

[0006] 3. Traditional pod thruster bearing monitoring systems often fail to achieve real-time monitoring and early warning. When a bearing fails, it often causes significant losses or impacts.

[0007] The present invention is a pod thruster bearing health monitoring method and device, which can effectively solve the existing pod thruster rolling bearing service status monitoring problem. The original vibration signal of the rolling bearing to be monitored is obtained; compressed sensing is used to sparse the original vibration signal to reduce the local storage space; the compressed data is restored, and the vibration data is decomposed into indicators using a time-frequency domain analysis method, and finally the indicators are classified using a density clustering algorithm to obtain the rolling bearing fault diagnosis result. The device includes a pod thruster whole machine test bench, a vibration acceleration sensor at the rolling bearing, a local data acquisition card, a power adapter, a processor and a memory. Summary of the invention

[0008] The purpose of the present invention is to propose a pod thruster bearing health monitoring device and method, which can accurately judge the health status and service life of the rolling bearing by overcoming the shortcomings in the existing pod thruster rolling bearing monitoring, promote the development of intelligent operation and maintenance system, and avoid a large amount of resource waste caused by regular maintenance.

[0009] To achieve the above-mentioned purpose, the technical solution of the present invention is: a pod thruster bearing health monitoring device, including a whole machine test bench, a drive system, a rolling bearing, a vibration sensor, an accelerometer acquisition card, a data terminal storage card, a PC computer, and a power adapter. The drive system drives the whole machine test bench and the rolling bearing to work, and a plurality of vibration sensors are arranged on the bearing seat of the rolling bearing. The vibration sensor is connected to the PC computer through the accelerometer acquisition card and the data terminal storage card.

[0010] Furthermore, when the drive system is given different operating speeds, the whole machine test bench will have different operating conditions, and the rolling bearing will work together with the whole machine test bench.

[0011] Furthermore, when the whole machine test bench is running, the vibration sensor obtains the operation signal of the rolling bearing, and performs signal perception and transmission; the accelerometer acquisition card, the data terminal storage card and the PC computer are connected through a power adapter; the accelerometer acquisition card synchronously collects data as the rolling bearing runs; and the data terminal storage card stores data.

[0012] Furthermore, when the drive system drives the whole machine test bench, the data terminal storage card compresses and stores the data from the accelerometer acquisition card according to the compression sensing algorithm program in the PC computer, and restores the compressed data from the data terminal storage card.

[0013] A pod thruster bearing health monitoring method comprises: obtaining the original vibration signal of a rolling bearing to be monitored; compressing the collected original vibration data of the rolling bearing to obtain sparse vibration data; using the principles of compressed sensing and sparse reconstruction to obtain sparsely restored vibration data, performing frequency domain index decomposition on the vibration data, classifying the indexes using a density clustering algorithm, performing fault diagnosis on the rolling bearing to be monitored, and obtaining a rolling bearing fault diagnosis result.

[0014] Furthermore, the compressed sensing and sparse reconstruction method is used: first, the orthogonal matching pursuit algorithm is selected to restore the compressed data. Then, the time-frequency domain analysis method is used to select five characteristic symptom indicators, namely waveform indicator, pulse indicator, kurtosis indicator, margin indicator, and peak-to-peak value, which can reflect the overall state of the original vibration signal of the rolling bearing of the pod thruster, and the restored rolling bearing vibration signal is decomposed into indicators; then the density clustering algorithm is selected to classify the characteristic symptom indicators into different categories. Based on the above five characteristic symptom indicators, the rolling bearing fault diagnosis and analysis of the pod thruster is carried out, and four types of fault categories, namely rolling bearing outer ring peeling, rolling element peeling, cage peeling and bearing inner ring peeling, are obtained. The fault categories can separably reflect the distribution of different categories in the feature space, and characterize the importance of each characteristic component in classification and identification. Through feature formation, feature extraction, feature selection and fault pattern recognition, different rolling bearing faults of the pod thruster are identified, and the service status of the rolling bearing of the pod thruster is characterized.

[0015] Furthermore, the orthogonal matching pursuit algorithm is used to iteratively calculate the percentage of the original signal restored by the optimal sparse coefficient, and the reconstructed vibration signal is determined. The reconstructed vibration signal is used as data input and converted into a characteristic symptom index of the rolling bearing vibration signal. Monitoring and diagnosing the pod rolling bearing can timely discover the trend of fault changes over time, select five characteristic symptom indicators that can reflect the overall state of the pod rolling bearing, including waveform index, pulse index, kurtosis index, margin index, and peak-to-peak value, and determine the fault type and cause.

[0016] Furthermore, a density clustering algorithm is used. If the distances between the above four fault categories are small and basically concentrated in one place, it indicates that the current rolling bearing is in good service condition; when the distance between one of the above four categories is large with the other three categories, it indicates that the current rolling bearing has a fault corresponding to this category; when the distances between several of the above four categories and the other categories are large, it indicates that the current rolling bearing has faults corresponding to these categories; thus, the current service status of the rolling bearing of the pod thruster is obtained.

[0017] The beneficial effects of the present invention are mainly reflected in:

[0018] 1. Improve the reliability and operating efficiency of the rolling bearings of pod thrusters. Rolling bearings are key components of pod thrusters, and their service status directly affects the operation of pod thrusters. Through health monitoring, potential failures of bearings can be discovered in a timely manner to ensure the continuous and stable operation of pod thrusters.

[0019] 2. Optimize the maintenance strategy of pod thruster rolling bearings. Based on the results of bearing health monitoring, the shipyard can formulate a reasonable maintenance plan. In this way, over-maintenance and under-maintenance can be avoided, ensuring the effectiveness and economy of maintenance work.

[0020] 3. Reduce maintenance costs. Traditional bearing maintenance methods often rely on experience and regular replacement, which is not only inefficient but also costly. Bearing health monitoring can achieve real-time monitoring and analysis of bearing status, detect potential faults in a timely manner and take corresponding measures to deal with them. This can not only reduce downtime and maintenance costs caused by faults, but also extend the service life of the bearing, reducing the overall maintenance cost.

[0021] 4. Provide data support. The large amount of data collected during the bearing health monitoring process can provide strong support for the operation and maintenance of mechanical equipment. These data can be used to analyze the operating rules, failure modes, and maintenance effects of the rolling bearings of the pod thruster, providing a scientific basis for subsequent equipment management and maintenance.

[0022] Through the invention and design of the above-mentioned pod thruster bearing health monitoring device and method, the reliability and operating efficiency of the pod thruster can be improved, the maintenance strategy can be optimized, and the maintenance cost can be reduced, thereby promoting the intelligent development of the pod thruster. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 is a schematic diagram of the pod thruster bearing health monitoring device;

[0024] Figure 2 is a flow chart of a method for monitoring the health of a pod thruster bearing;

[0025] Figure 3 This is the installation diagram of the vibration sensor of the rolling bearing assembly

[0026] Figure 4 It is a schematic diagram of the whole machine test bench;

[0027] Figure 5 It is the overall description diagram of compressed sensing;

[0028] Figure 6 It is the flow chart of density clustering algorithm (DBSCAN). DETAILED DESCRIPTION

[0029] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0030] like Figure 1 As shown in Figures 3 and 4, a pod thruster bearing health monitoring device proposed in an embodiment of the present invention includes a whole machine test bench 1, a drive system 2, a rolling bearing 3, a vibration sensor 4, an accelerometer acquisition card 5, a data terminal storage card 6, a PC computer 7, and a power adapter 8.

[0031] The driving system 2 drives the whole machine test bench 1 and the rolling bearing 3 to connect. The driving system 2 drives the whole machine test bench 1 to work. When the driving system 2 is given different operating speeds, the whole machine test bench 1 has different operating conditions, and the rolling bearing 3 works together with the whole machine test bench 1.

[0032] The rolling bearing 3 is installed in the bearing seat 3-1, including the bearing inner ring 3-3, the bearing outer ring 3-2 and the rolling element 3-3. The bearing seat is provided with a plurality of vibration sensors 4. The vibration sensors 4 are connected to the PC computer 7 through the accelerometer acquisition card 5 and the data terminal storage card 6.

[0033] When the whole machine test bench 1 is running, the vibration sensor 4 obtains the running signal of the rolling bearing 3, and performs signal perception and transmission; the accelerometer acquisition card 5, the data terminal storage card 6 and the PC computer 7 are connected through the power adapter 8. The accelerometer acquisition card 5 synchronously collects data as the rolling bearing 3 runs; the data terminal storage card 6 stores data. Since the compressed sensing algorithm program is set on the PC computer 7, when the drive system 2 drives the whole machine test bench 1, the data terminal storage card 6 compresses and stores the data from the accelerometer acquisition card 5 according to the compressed sensing algorithm. The compressed data from the data terminal storage card 6 is then restored.

[0034] like Figure 2 As shown, the pod thruster rolling bearing health monitoring method of the present invention comprises: S01 acquiring the original vibration signal of the rolling bearing to be monitored; S02 compressing the collected original vibration data of the rolling bearing to obtain sparse vibration data; S03 using the principle of compressed sensing and sparse reconstruction, such as Figure 5 As shown, the vibration data after sparse restoration is obtained, and the frequency domain index decomposition is performed on the vibration data. S04 classifies the indexes by using the density clustering algorithm, performs fault diagnosis on the rolling bearing to be monitored, and obtains the rolling bearing fault diagnosis result.

[0035] like Figure 6As shown, the restoration algorithm used is the orthogonal matching pursuit algorithm, which iteratively calculates the percentage of the original signal restored by the optimal sparse coefficient to determine the reconstructed vibration signal. The reconstructed vibration signal is used as the data input and converted into the characteristic symptom index of the rolling bearing vibration signal. Monitoring and diagnosing the pod rolling bearing can timely discover the trend of fault changes over time. Five characteristic symptom indicators, namely waveform index, pulse index, kurtosis index, margin index, and peak-to-peak value, which can reflect the overall state of the pod rolling bearing, are selected to determine the type and cause of the fault. The density clustering algorithm is used to monitor the state of the pod thruster rolling bearing using the above five characteristic symptom indicators to identify different rolling bearing faults. Based on the above five characteristic symptom indicators, the fault diagnosis and analysis of the pod thruster rolling bearing is carried out, and four types of fault categories are obtained, namely, bearing outer ring peeling, rolling element peeling, cage peeling, and bearing inner ring peeling. The density clustering algorithm is based on the concepts of density reachability and density connectivity of data points. Starting from the selected core point, it continuously expands to the area of ​​the same data type, thereby obtaining a maximum area containing core points and boundary points. Any two points in the area are connected, and the division result is defined as a set of density-connected objects. When the spacing between the above four categories is small, it indicates that the current rolling bearing is in good condition; otherwise, it is in a faulty state. This can realize the health monitoring of the pod thruster bearing.

[0036] The pod thruster bearing health monitoring device and method provided by the present invention are used to solve the problem that the health status and service life of rolling bearings cannot be accurately judged at present, resulting in a large amount of resource waste. When the whole machine test bench is running, the control system of the pod thruster starts to work, and the original vibration signal of the rolling bearing to be tested under different working conditions is obtained, and the original vibration signal is stored in the local storage card; because the sensor is in a working state during the operation of the bench, all collected data will be stored in the local storage card, resulting in the need for a larger capacity storage card. However, most of the data can be compressed. Consider using the method of compressed sensing and signal sparse reconstruction to compress the original vibration data stored in the local storage card to reduce storage requirements; this method is based on the column coherence of the original vibration signal affecting the reconstruction accuracy of the signal. The smaller the column coherence, the greater the sparsity of the signal, and the fewer the number of measurements required to reconstruct the signal. The orthogonal matching pursuit algorithm (OMP algorithm) is selected to restore the compressed data. Then, the time-frequency domain analysis method is used to select five characteristic symptom indicators, namely waveform index, pulse index, kurtosis index, margin index and peak-to-peak value, which can reflect the overall state of the original vibration signal of the pod thruster rolling bearing. The restored rolling bearing vibration signal is decomposed into indicators; the density clustering algorithm (DBSCAN algorithm) is selected, through which the characteristic symptom indicators can be divided into different categories. Based on the above five characteristic symptom indicators, the fault diagnosis and analysis of the pod thruster rolling bearing is carried out, and four fault categories are obtained, namely, the outer ring peeling, rolling element peeling, cage peeling and inner ring peeling of the bearing. The category separability reflects the distribution of different classes in the feature space, and can characterize the importance of each feature component in classification and identification. Through feature formation, feature extraction, feature selection and fault pattern recognition, different pod thruster rolling bearing faults are identified, and the service status of the pod thruster rolling bearing is characterized. The density clustering algorithm is based on the concepts of density reachability and density connectivity of data points. Starting from the selected core point, it continuously expands to the area of ​​the same data type, thereby obtaining a maximum area containing core points and boundary points. Any two points in the area are connected, and the division result is defined as a set of density-connected objects. If the distances between the above four fault categories are small and basically concentrated in one place, it indicates that the current rolling bearing is in good service; when one of the above four categories is far away from the other three categories, it indicates that the current rolling bearing has a fault corresponding to this category; when several of the above four categories are far away from the other categories, it indicates that the current rolling bearing has a fault corresponding to these categories; thus, the current service status of the pod thruster rolling bearing is obtained.

[0037] The pod thruster bearing health monitoring device and method provided by the present invention obtains the original vibration signal of the rolling bearing to be monitored; compresses the original signal using compressed sensing and stores it in a local storage card; uses an orthogonal matching pursuit algorithm to restore the compressed data; and determines the health index data of the rolling bearing to be monitored using a time-frequency domain index decomposition method based on the original vibration signal; performs fault diagnosis on the rolling bearing to be monitored based on the health index data, and obtains the fault diagnosis result of the rolling bearing to be monitored. The method can accurately judge the health status and service life of the rolling bearing, promote the development of the intelligent operation and maintenance system, and avoid a large amount of resource waste caused by regular maintenance.

Claims

1. A pod thruster bearing health monitoring device, characterized in that: It includes a whole machine test bench, a drive system, a rolling bearing, a vibration sensor, an accelerometer acquisition card, a data terminal storage card, a PC computer, and a power adapter. The drive system drives the whole machine test bench and the rolling bearing to work. Multiple vibration sensors are arranged on the bearing seat of the rolling bearing. The vibration sensors are connected to the PC computer through the accelerometer acquisition card and the data terminal storage card.

2. The pod thruster bearing health monitoring device according to claim 1, characterized in that: When the drive system is given different operating speeds, the whole machine test bench will have different operating conditions, and the rolling bearing will work together with the whole machine test bench.

3. The pod thruster bearing health monitoring device according to claim 1, characterized in that: When the whole machine test bench is running, the vibration sensor obtains the running signal of the rolling bearing, senses and transmits the signal; the accelerometer acquisition card, the data terminal storage card and the PC are connected through the power adapter; the accelerometer acquisition card synchronously collects data as the rolling bearing runs; The data terminal memory card is used for data storage.

4. The pod thruster bearing health monitoring device according to claim 3, characterized in that: When the drive system drives the whole machine test bench, the data terminal storage card compresses and stores the data from the accelerometer acquisition card according to the compression sensing algorithm program in the PC computer, and restores the compressed data from the data terminal storage card.

5. A pod propeller bearing health monitoring method, using the pod propeller bearing health monitoring device according to any one of claims 1 to 4, characterized in that: include: Acquire the original vibration signal of the rolling bearing to be monitored; compress the collected original vibration data of the rolling bearing to obtain sparse vibration data; By using the principles of compressed sensing and sparse reconstruction, the sparsely restored vibration data is obtained, the frequency domain index decomposition of the vibration data is performed, the density clustering algorithm is used to classify the indexes, the fault diagnosis of the rolling bearing to be monitored is performed, and the rolling bearing fault diagnosis result is obtained.

6. The pod thruster bearing health monitoring method according to claim 5, characterized in that: Compressed sensing and sparse reconstruction method: First, the orthogonal matching pursuit algorithm is selected to restore the compressed data. Then, the time-frequency domain analysis method is used to select five characteristic symptom indicators, namely waveform indicator, pulse indicator, kurtosis indicator, margin indicator, and peak-to-peak value, which can reflect the overall state of the original vibration signal of the rolling bearing of the pod thruster, and the restored rolling bearing vibration signal is decomposed into indicators; then the density clustering algorithm is selected to classify the characteristic symptom indicators into different categories. Based on the above five characteristic symptom indicators, the rolling bearing fault diagnosis and analysis of the pod thruster is carried out, and four types of fault categories, namely rolling bearing outer ring peeling, rolling element peeling, cage peeling, and bearing inner ring peeling, are obtained. The fault categories can separably reflect the distribution of different categories in the feature space, and characterize the importance of each characteristic component in classification and identification. Through feature formation, feature extraction, feature selection and fault pattern recognition, different rolling bearing faults of the pod thruster are identified, and the service status of the rolling bearing of the pod thruster is characterized.

7. The pod thruster bearing health monitoring method according to claim 6, characterized in that: The orthogonal matching pursuit algorithm is used to iteratively calculate the percentage of the original signal restored by the optimal sparse coefficient, and the reconstructed vibration signal is determined. The reconstructed vibration signal is used as data input and converted into a characteristic symptom indicator of the rolling bearing vibration signal. Monitoring and diagnosis of the pod rolling bearing can timely discover the trend of fault changes over time, select five characteristic symptom indicators that can reflect the overall state of the pod rolling bearing, including waveform index, pulse index, kurtosis index, margin index, and peak-to-peak value, and determine the type and cause of the fault.

8. The pod thruster bearing health monitoring method according to claim 6, characterized in that: Using the density clustering algorithm, if the distances between the above four fault categories are small and basically concentrated in one place, it indicates that the current rolling bearing is in good service condition; when the distance between one of the above four categories is large with the other three categories, it indicates that the current rolling bearing has a fault corresponding to this category; when the distances between several of the above four categories and the other categories are large, it indicates that the current rolling bearing has faults corresponding to these categories; thus, the current service status of the rolling bearing of the pod thruster is obtained.

Images