Small-sized lightweight rotor system fault simulation test bench and fault diagnosis method

Abstract

The invention discloses a small-sized lightweight rotor system fault simulation test bench and a fault diagnosis method, the test bench mainly comprises a mechanical fault simulation system and an embedded control system, various fault conditions are simulated through the mechanical fault simulation system, and data such as rotating speed, current and vibration are acquired in real time for analysis; the embedded control system adopts a fault diagnosis method based on cumulative spectrum distribution entropy and combines a multi-label nearest neighbor algorithm to realize accurate identification and classification of different fault types; the design focuses on miniaturization and light weight, is convenient to carry, is provided with a real-time display screen, and can display a fault diagnosis result and a classification condition in real time; through on-line demonstration of operation, maintenance and fault diagnosis processes of the rotor system, experimental research and classroom teaching in the field of fault diagnosis are assisted.

Description

Technical Field

[0001] The invention relates to the field of rotor fault simulation experiments, and in particular to a small-sized lightweight rotor system fault simulation test bench and a fault diagnosis method. Background Art

[0002] With the rapid development of modern aerospace, the working environment of aircraft engines is becoming increasingly severe. The mechanical parts of the engine operate under harsh conditions of high temperature, high pressure and high load, which often leads to engine damage and aviation safety accidents. Rotor structures and parallel shaft reducers are widely used in various types of aviation equipment. Blade and shaft friction, rotor imbalance and parallel shaft reducer failure are common mechanical failures of aircraft engines, which seriously affect the operating stability and safety of the engine.

[0003] During the manufacturing process of aircraft engine blades, cracks may occur due to process problems, resulting in asymmetric blades, which in turn cause rotor imbalance. When the engine is running, the high-temperature and high-pressure airflow in the flow channel causes the rotor to deform due to heat, aggravating the imbalance and generating vibration. The exciting force caused by the imbalance will increase the wear of the equipment, thereby changing its quality indicators such as structure, life and reliability. At the same time, when the engine is running at high speed, the fan blades are prone to deformation and rub against the shaft, resulting in strong self-excited vibration. There may be errors in the gear system during processing and installation. Coupled with the harsh working environment and complex working conditions, it will also affect the normal operation of the gear transmission system and generate large vibrations and shocks.

[0004] At present, the health management of equipment can be effectively achieved through vibration monitoring of aircraft engine rotor systems and parallel shaft reducers. In the teaching of fault prognosis and health management (PHM), it is difficult for students to understand the concept of fault monitoring and diagnosis due to the abstractness of fault mechanism theory and the lack of experimental teaching tools that demonstrate the entire process of fault diagnosis. Summary of the invention

[0005] The purpose of the present invention is to provide a small lightweight rotor system fault simulation test bench and a fault diagnosis method, which can clearly display the fault occurrence process, system protection after the fault occurs, and real-time status monitoring of the system, so as to facilitate teaching demonstration.

[0006] In order to achieve the above tasks, the present invention adopts the following technical solutions:

[0007] A small lightweight rotor system fault simulation test bench comprises a test bench base, a mechanical fault simulation system and an embedded system, wherein:

[0008] The mechanical fault simulation system is installed on the test bench base, including a DC motor, a motor base, a motor base, a rotor coupling, a rotating shaft bearing, a bearing seat, a balancing plate, a friction bracket, a sensor bracket, a parallel shaft reducer, a casing, a threaded plate, a screen display and a balancing plate rotating shaft;

[0009] The DC motor is installed on the motor base through the motor base, the motor base is fixed on the casing, and the casing is fixed on the test bench base; the motor shaft of the DC motor is connected to the balancing disk shaft through the rotor coupling, the balancing disk is installed on the bearing seat through the shaft bearing, and the bearing seat is fixed on the casing; the balancing disk shaft passes through the shaft bearing and is connected to the parallel shaft reducer fixed on the rear end of the casing; the balancing disk rotates under the drive of the DC motor; the friction bracket and the sensor bracket are arranged in sequence between the balancing disk and the parallel shaft reducer, and the screen display is installed at the front end of the casing; wherein the sensor bracket is installed with transverse and longitudinal vibration sensors for collecting vibration acceleration signals of the balancing disk shaft; the friction bracket is used to simulate friction faults;

[0010] The embedded system is installed at the bottom of the casing, and the embedded system includes a signal acquisition module, a data transmission module, a data processing module, an interactive display module and a control module;

[0011] The signal acquisition module is used to collect the current signal, speed signal and vibration acceleration signal of the DC motor and the balance disk shaft, and transmits the speed signal to the interactive display module in real time through the data transmission module, and transmits the current signal and vibration acceleration signal to the data processing module; the data processing module processes and analyzes the current signal and vibration acceleration signal from the signal acquisition module, on the one hand to determine whether to perform braking protection, and on the other hand to extract signal features and analyze its status information to determine the current system status; the system status includes fault diagnosis results, and the system status is transmitted to the interactive display module; the control module is used to set the operating parameters of the DC motor and adjust the operating status of the DC motor; the interactive display module displays the speed signal and system status of the DC motor based on the screen display.

[0012] Furthermore, a plurality of holes are evenly distributed on the circumference of the balancing disk. By screwing counterweights into some of the holes, the mass of the balancing disk can be asymmetric, thereby simulating a rotor imbalance fault. By adjusting the mass of the counterweights screwed in, simulation of rotor imbalance faults of varying degrees can be achieved.

[0013] Furthermore, the friction bracket is arranged on the outside of the balancing disk shaft and is a "door"-shaped structure; screw holes are arranged on the side and top of the friction bracket, which are used to screw in transverse screws and longitudinal screws respectively, so as to simulate the friction failure of the shaft; the transverse screws and longitudinal screws are both made of plastic, and the feed amount of the transverse screws and longitudinal screws is manually adjusted to make them have different degrees of contact with the balancing disk shaft, thereby adjusting the degree of friction.

[0014] Furthermore, the parallel shaft reducer has a built-in gear switching mechanism and a fault gear; wherein, the fault gear and multiple groups of normal gears are mounted on the output shaft of the parallel shaft reducer. When the gear is switched through the gear switching mechanism, different normal gears or faulty gears will cooperate with the gear mechanism inside the parallel shaft reducer, thereby realizing the normal deceleration function output of the parallel shaft reducer or the reducer fault simulation.

[0015] A small lightweight rotor system fault diagnosis method comprises the following steps:

[0016] Step 1: Start the small lightweight rotor system fault simulation test bench and initialize the test bench; the initialization content includes:

[0017] Set the speed and direction of the DC motor and the simulated fault. If a rotor imbalance fault is simulated, screw the counterweight into the hole of the balancing plate in advance. If a shaft friction fault is simulated, install the horizontal screws and vertical screws on the friction bracket. If a parallel shaft reducer fault is simulated, adjust the gear switching mechanism of the parallel shaft reducer to switch to the faulty gear or the normal gear according to the test requirements.

[0018] Setting an initial braking threshold, the initial braking threshold includes an initial threshold of a current signal of a DC motor and a vibration acceleration signal;

[0019] Step 2: After initialization, the DC motor starts to work, and the PID algorithm is used to stabilize the speed of the DC motor to the speed and direction set during initialization;

[0020] Step 3: The signal acquisition module collects the current signal of the DC motor and the vibration acceleration signal of the balancing disk shaft under the set simulated fault and sends them to the data processing module for signal processing and analysis, and judges the system status according to the fault diagnosis result; at the same time, the data processing module also judges whether to perform braking protection according to the initial braking threshold;

[0021] Step 4: If the braking protection does not occur, the fault simulation is performed; when simulating the shaft friction fault, the feed amount of the horizontal screw and the longitudinal screw is adjusted to control the degree of friction of different shafts; when simulating the fault of the parallel shaft reducer, the gear switching mechanism is used to switch the gear in time;

[0022] The system status is determined based on the data processing module; if a certain fault is determined to have occurred through the fault diagnosis results, it is transmitted to the interactive display module, and a control option that can stop the DC motor is given; when a combined fault occurs, emergency stop protection is directly performed;

[0023] Step 5, determining whether the fault simulation is completed, and stopping the DC motor when completed.

[0024] Furthermore, the data processing module further determines whether to perform a stop protection according to the initial stop threshold, including:

[0025] When it is detected that the current signal is greater than the initial threshold of the current signal, it is determined that a fault has occurred and the braking protection is performed; when it is detected that the vibration acceleration signal exceeds the initial threshold of the vibration acceleration signal, it is determined that a fault has occurred, and the interactive module displays whether to perform braking protection for the operator to choose.

[0026] Furthermore, the data processing module adopts a data processing method based on cumulative spectral distribution entropy to calculate the cumulative spectral distribution entropy of the vibration acceleration signal, and then inputs the cumulative spectral distribution entropy as a signal feature into a multi-label K nearest neighbor classifier for fault classification, thereby obtaining a fault diagnosis result.

[0027] Furthermore, the data processing method based on cumulative spectral distribution entropy is specifically as follows:

[0028] Step 4.1, perform discrete Fourier transform on the vibration acceleration signal time series x(n) of length N to obtain the signal X(k); n represents the nth data in the sequence, ;

[0029] Step 4.2: Determine the power spectrum distribution of the signal based on the signal X(k) :

[0030] ;

[0031] Where, k=0, 1, ..., N / 2-1, k represents the kth frequency component;

[0032] Step 4.3: Distribute the power spectrum The N / 2 points of the mid-spectral component are divided into M intervals, and then the power probability pdf of each interval is calculated m , expressed as:

[0033] ;

[0034] Among them, m=0, 1, ..., M, k represents the kth spectral component, Bin m Represents the set of frequency components in the mth interval;

[0035] Step 4.4: Determine the cumulative distribution function of the spectrum interval based on the probability density function :

[0036] ;

[0037] Step 4.5, cumulative distribution function based on spectrum interval , using the standardized cumulative residual function to determine the cumulative residual probability distribution :

[0038] ;

[0039] Step 4.6, the cumulative spectral distribution entropy CSDE is calculated as follows:

[0040] ;

[0041] In step 4.7, the calculated cumulative spectral distribution entropy value CSDE is used as the input of the multi-label K nearest neighbor classifier ML-KNN to identify the fault diagnosis result.

[0042] Furthermore, the training process of the multi-label K nearest neighbor classifier ML-KNN is as follows:

[0043] (1) Set different categories of fault diagnosis results; collect the vibration acceleration signals of the fault simulation test bench in healthy state and different fault types and calculate the corresponding cumulative spectral distribution entropy value CSDE, which are used as normal data and fault data respectively, and set corresponding labels for normal data and fault data respectively, so as to construct a data set; divide the data set into a training set and a test set;

[0044] (2) Build an ML-KNN classifier and put the divided training set into the classifier training;

[0045] (3) The trained ML-KNN classifier is input into the test set for testing. When the test effect meets the preset requirements, the trained ML-KNN classifier is saved;

[0046] After obtaining a set of new vibration acceleration signals, the cumulative spectral distribution entropy is extracted as the signal feature and input into the trained ML-KNN classifier to obtain the predicted fault diagnosis results, including whether the system is healthy or what type of fault has occurred.

[0047] Furthermore, the fault diagnosis result includes system health or system fault, and system fault is divided into different fault types, including rotor imbalance fault, shaft rubbing fault, parallel shaft reducer fault, and combination of these faults; each fault type is divided into mild and severe faults.

[0048] Compared with the prior art, the present invention has the following technical features:

[0049] The present invention effectively realizes the independent simulation of the rotor shaft friction, rotor imbalance and parallel shaft reducer fault during the operation of the rotor, realizes the health monitoring of the system by real-time monitoring the current, speed and acceleration of the system, realizes the real-time online fault diagnosis by cumulative spectral distribution entropy and multi-label nearest neighbor algorithm when the fault occurs, stops the rotor system and protects the test bench, which has high practical value. Moreover, the test bench is small in size and has an independent power supply, is easy to carry, simple to operate, and has a high degree of human-computer interaction of the screen display and control interface. It can also be used for teaching tasks, effectively improving the teaching quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 A top view of a small lightweight rotor system fault simulation test bench of the present invention;

[0051] Figure 2 A side view of a small lightweight rotor system fault simulation test bench of the present invention;

[0052] Figure 3 A schematic diagram of the dimensions of a small-sized lightweight rotor system fault simulation test bench according to an embodiment of the present invention;

[0053] Figure 4 It is a structural schematic diagram of the balancing disc part;

[0054] Figure 5 It is a structural schematic diagram of the friction support part;

[0055] Figure 6 It is the principle diagram of the embedded system part;

[0056] Figure 7 A fault classification confusion matrix established in one embodiment of the present invention;

[0057] Figure 8 It is a schematic diagram of the process of the present invention.

[0058] Explanation of the numbers in the figure: DC motor 1, motor base 2, motor base 3, rotor coupling 4, shaft bearing 5, bearing seat 6, balancing plate 7, friction bracket 8, sensor bracket 9, parallel shaft reducer 10, casing 11, threaded plate 12, screen display 13, balancing plate shaft 14, hole 7-1, horizontal screw 8-2, longitudinal screw 8-1. DETAILED DESCRIPTION

[0059] See attached Figure 1 and Figure 2The present invention provides a small lightweight rotor system fault simulation test bench, comprising a test bench base, a mechanical fault simulation system and an embedded system, wherein:

[0060] The mechanical fault simulation system is installed on the test bench base, including a DC motor 1, a motor base 2, a motor base 3, a rotor coupling 4, a shaft bearing 5, a bearing seat 6, a balancing disk 7, a friction bracket 8, a sensor bracket 9, a parallel shaft reducer 10, a casing 11, a threaded plate 12, a screen display 13 and a balancing disk shaft 14; the DC motor 1 is installed on the motor base 3 through the motor base 2, the motor base 3 is fixed on the casing 11, and the casing 11 is fixed on the test bench base; the motor shaft of the DC motor 1 is connected to the balancing disk shaft 14 through the rotor coupling 4, and the balancing disk shaft 14 is connected to the balancing disk shaft 14 through the rotor coupling 4. The disc 7 is mounted on the bearing seat 6 through the shaft bearing 5, and the bearing seat 6 is fixed on the casing 11; the balancing disc shaft 14 passes through the shaft bearing 5 and is connected to the parallel shaft reducer 10 fixed at the rear end of the casing 11; the balancing disc 7 rotates under the drive of the DC motor 1; the friction bracket 8 and the sensor bracket 9 are arranged in sequence between the balancing disc 7 and the parallel shaft reducer 10, and the screen display 13 is installed at the front end of the casing 11; wherein, the transverse and longitudinal vibration sensors are installed on the sensor bracket 9 for collecting the vibration acceleration signal of the balancing disc shaft 14; the friction bracket 8 is used to simulate the friction fault;

[0061] The embedded system is installed at the bottom of the casing 11. Figure 6 ,The embedded system includes a signal acquisition module, a data transmission module, a data processing module, an interactive display module and a control module;

[0062] The signal acquisition module is used to collect the current signal, speed signal and vibration acceleration signal of the DC motor 1 and the balance disk shaft 14, and transmit the speed signal to the interactive display module in real time through the data transmission module, and transmit the current signal and vibration acceleration signal to the data processing module; the data processing module processes and analyzes the current signal and vibration acceleration signal from the signal acquisition module, on the one hand, determines whether to perform the braking protection, and on the other hand, extracts the signal characteristics and analyzes its state information to determine the current system state; the system state includes the fault diagnosis result, and transmits the system state to the interactive display module; the control module is used to set the operating parameters of the DC motor 1 and adjust the operating state of the DC motor 1; the interactive display module displays the speed signal and system state of the DC motor 1 based on the screen display 13. The embedded system also includes a power supply and a switch module, which is used to provide a constant current source and a constant voltage source for the entire fault simulation test bench, perform power management, be responsible for power on and off, and control the start and stop of the fault simulation test bench.

[0063] The embedded system realizes visualization of control and fault demonstration results through a screen display 13 based on the main control chip; in one embodiment of the present invention, a control and fault demonstration interface program is constructed based on labview and burned into the main control chip, and the speed signal and system status of the DC motor 1 are displayed in real time on the display screen using an interactive display module through the main control chip in a SPI communication manner.

[0064] In one embodiment of the present invention, Figure 3 As shown, the fault simulation test bench adopts a lightweight structural design as a whole, and the relevant metal components are made of lightweight material aluminum-magnesium alloy, with a total weight not exceeding 1.5kg; in this embodiment, the casing 11 is designed as a rectangular parallelepiped of 40cm*16cm*6.45cm, and its small size makes it easy to carry.

[0065] See also Figure 1 and Figure 4 The balancing disk 7 has a plurality of holes 7-1 evenly distributed on its circumference. By screwing counterweights into some of the holes 7-1, the mass of the balancing disk 7 is asymmetric, thereby simulating a rotor imbalance fault. By adjusting the mass of the counterweights screwed in, simulations of rotor imbalance faults of varying degrees can be achieved.

[0066] See also Figure 5 The friction bracket 8 is arranged on the outside of the balancing disk shaft 14 and is a "door"-shaped structure; screw holes are arranged on the side and top of the friction bracket 8, which are respectively used to screw in the transverse screw 8-2 and the longitudinal screw 8-1, so as to simulate the shaft friction failure; the transverse screw 8-2 and the longitudinal screw 8-1 are both made of plastic. By manually adjusting the feed amount of the transverse screw 8-2 and the longitudinal screw 8-1, different degrees of contact are generated between them and the balancing disk shaft 14, thereby adjusting the degree of friction.

[0067] The parallel shaft reducer 10 has a built-in gear switching mechanism and a fault gear; wherein, the fault gear and multiple groups of normal gears are mounted on the output shaft of the parallel shaft reducer 10. When the gear is switched through the gear switching mechanism, different normal gears or faulty gears will cooperate with the gear mechanism inside the parallel shaft reducer 10, thereby realizing the normal deceleration function output of the parallel shaft reducer 10 or the reducer fault simulation.

[0068] like Figure 8 As shown, on the basis of the above technical solution, the present invention further provides a small lightweight rotor system fault diagnosis method, which is used to judge three faults of rotor imbalance, shaft friction and parallel shaft reducer fault, and comprises the following steps:

[0069] Step 1: Start the small lightweight rotor system fault simulation test bench through the power supply and switch module, and initialize the test bench; the initialization content includes:

[0070] Set the speed and direction of the DC motor 1, and set the simulated fault; if a rotor imbalance fault simulation is performed, screw the counterweight into the hole 7-1 of the balancing disk 7 in advance; if a shaft friction fault is performed, assemble the transverse screw 8-2 and the longitudinal screw 8-1 on the friction bracket 8; if a parallel shaft reducer 10 fault simulation is performed, adjust the gear switching mechanism of the parallel shaft reducer 10 to switch to the faulty gear or the normal gear according to the test requirements.

[0071] Setting an initial braking threshold, the initial braking threshold includes an initial threshold of a current signal and a vibration acceleration signal of the DC motor 1;

[0072] Step 2: After initialization, the DC motor 1 starts to work, and the PID algorithm is used to stabilize the speed of the DC motor 1 to reach the speed and direction set during initialization;

[0073] Step 3: The signal acquisition module collects the current signal of the DC motor 1 and the vibration acceleration signal of the balancing disk shaft 14 under the set simulated fault and sends them to the data processing module for signal processing and analysis, and judges the system status according to the fault diagnosis result; at the same time, the data processing module also judges whether to perform the braking protection according to the initial braking threshold, specifically:

[0074] When the rotor is unbalanced, the shaft rubs, or the parallel shaft reducer fails, the friction of the shaft rotation increases. To maintain a constant speed, the current signal of the DC motor 1 increases. When the current signal is detected to be greater than the initial threshold of the current signal, a fault is determined to occur and the brake is stopped for protection.

[0075] When the rotor system has rotor imbalance, shaft rubbing, or parallel shaft reducer failure, the shaft vibration intensifies. When the vibration acceleration signal is detected to exceed the initial threshold of the vibration acceleration signal, it is determined that a failure has occurred. The interactive module displays whether to perform braking protection for the operator to choose.

[0076] Step 4, if the braking protection does not occur, the fault simulation is performed; when simulating the shaft friction fault, the feed amount of the horizontal screw 8-2 and the longitudinal screw 8-1 is adjusted to control the degree of friction of different shafts; when simulating the fault of the parallel shaft reducer 10, the gear switching mechanism is used to switch the gear in time;

[0077] The system status is determined based on the data processing module; if a certain fault is determined to have occurred through the fault diagnosis result, it is transmitted to the interactive display module, and a control option for stopping the DC motor 1 is given; when a combined fault occurs, emergency stopping protection is directly performed.

[0078] Among them, the data processing module adopts a data processing method based on cumulative spectrum distribution entropy (CSDE) to calculate the cumulative spectrum distribution entropy of the vibration acceleration signal, and then inputs the cumulative spectrum distribution entropy as a signal feature into the multi-label K nearest neighbor classifier for fault classification, thereby obtaining the fault diagnosis result; the details are as follows:

[0079] For a given vibration acceleration signal time series of length N , n represents the nth data in the sequence, The process of obtaining the cumulative spectral distribution entropy of the vibration acceleration signal is as follows:

[0080] Step 4.1, perform discrete Fourier transform on the signal x(n) to obtain the signal X(k).

[0081] Step 4.2: Determine the power spectrum distribution of the signal based on the signal X(k) , as shown below:

[0082] ;

[0083] Wherein, k=0, 1, ..., N / 2-1, and k represents the kth frequency component.

[0084] It can be observed , which satisfies the definition of Shannon entropy; note that when using the discrete Fourier transform, the sum is from k=0 to k=N / 2-1.

[0085] Step 4.3: Distribute the power spectrum The N / 2 points of the mid-spectral component are divided into M intervals, and then the power probability pdf of each interval is calculated m , expressed as:

[0086] ;

[0087] Among them, m=0, 1, ..., M, k represents the kth spectral component, Bin m Represents the set of frequency components in the mth interval.

[0088] Step 4.4: Determine the cumulative distribution function of the spectrum interval based on the probability density function :

[0089] ;

[0090] Step 4.5, cumulative distribution function based on spectrum interval , using the Normalized cumulative residual function (NCRF) to determine the cumulative residual probability distribution , which can effectively preserve the distribution information and amplify the differences between frequency intervals. Mathematically, it can be expressed as:

[0091] ;

[0092] Step 4.6: The cumulative spectral distribution entropy proposed in this scheme is applied to the cumulative residual probability distribution To the Shannon entropy theoretical framework; the cumulative spectral distribution entropy value CSDE is calculated as follows:

[0093] ;

[0094] It can be observed , which satisfies the definition of Shannon entropy. Therefore, the obtained CSDE entropy value should be in the range [0, 1]; at the same time, it is common practice to only display and analyze the positive half of the spectrum (i.e., the first N / 2 points) because it contains all the information required for entropy calculation.

[0095] Through the above steps, the cumulative spectral distribution entropy can be extracted as a signal feature through the vibration acceleration signal reflecting the health status of the system.

[0096] In step 4.7, the calculated signal feature, i.e., the cumulative spectral distribution entropy value CSDE, is used as the input of the multi-label K nearest neighbor classifier ML-KNN to identify the fault diagnosis result; ML-KNN is a multi-label learning method proposed by extending KNN and combining Bayesian theorem; the training process of ML-KNN classifier is as follows:

[0097] (1) Set different categories of fault diagnosis results; collect the vibration acceleration signals of the fault simulation test bench in the healthy state and different fault types and calculate the corresponding cumulative spectral distribution entropy value CSDE, which are used as normal data and fault data respectively, and set corresponding labels for normal data and fault data respectively, so as to construct a data set; divide the data set into a training set and a test set with a ratio of 2:1;

[0098] (2) Use the sklearn library written in Python to build an ML-KNN classifier and put the divided training set into the classifier training;

[0099] (3) The trained ML-KNN classifier is input into the test set for testing. When the test effect meets the preset requirements, the trained ML-KNN classifier is saved;

[0100] Therefore, after obtaining a set of new vibration acceleration signals, the cumulative spectral distribution entropy is extracted as the signal feature and input into the trained ML-KNN classifier to obtain the predicted fault diagnosis results, including whether the system is healthy or what type of fault has occurred.

[0101] Among them, the fault diagnosis results include system health or system fault. System faults are divided into different fault types, including rotor imbalance fault, shaft friction fault, parallel shaft reducer fault, and combination of these faults; each fault type is divided into mild and severe faults.

[0102] like Figure 7 As shown, it is a fault classification confusion matrix established in one embodiment of the present invention, wherein the horizontal axis is the predicted label and the vertical axis is the true label.

[0103] Step 5, determining whether the fault simulation is completed, and stopping the DC motor 1 when completed.

[0104] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A small lightweight rotor system fault simulation test bench, characterized in that: It includes a test bench base, a mechanical fault simulation system and an embedded system, including: The mechanical fault simulation system is installed on a test bench base, and comprises a DC motor (1), a motor base (2), a motor base (3), a rotor coupling (4), a rotating shaft bearing (5), a bearing seat (6), a balancing plate (7), a friction bracket (8), a sensor bracket (9), a parallel shaft reducer (10), a casing (11), a threaded plate (12), a screen display (13) and a balancing plate rotating shaft (14); The DC motor (1) is mounted on a motor base (3) via a motor base (2); the motor base (3) is fixed on a casing (11); and the casing (11) is fixed on a test bench base; the motor shaft of the DC motor (1) is connected to a balancing disc shaft (14) via a rotor coupling (4); the balancing disc (7) is mounted on a bearing seat (6) via a shaft bearing (5); and the bearing seat (6) is fixed on the casing (11); the balancing disc shaft (14) passes through the shaft bearing (5) and is connected to a shaft bearing (14) fixed on the casing (11); The housing (11) is connected to a parallel shaft reducer (10) at the rear end of the housing (11); the balancing disc (7) rotates under the drive of the DC motor (1); a friction bracket (8) and a sensor bracket (9) are arranged in sequence between the balancing disc (7) and the parallel shaft reducer (10), and a screen display (13) is installed at the front end of the housing (11); wherein the sensor bracket (9) is installed with transverse and longitudinal vibration sensors for collecting vibration acceleration signals of the balancing disc rotating shaft (14); and the friction bracket (8) is used to simulate friction faults; The embedded system is installed at the bottom of the casing (11), and comprises a signal acquisition module, a data transmission module, a data processing module, an interactive display module and a control module; The signal acquisition module is used to acquire the current signal, speed signal and vibration acceleration signal of the DC motor (1) and the balance disk shaft (14), and transmits the speed signal to the interactive display module in real time through the data transmission module, and transmits the current signal and the vibration acceleration signal to the data processing module; the data processing module processes and analyzes the current signal and the vibration acceleration signal from the signal acquisition module, on the one hand to determine whether to perform braking protection, and on the other hand to extract signal features and analyze its status information to determine the current system status; the system status includes a fault diagnosis result, and the system status is transmitted to the interactive display module; the control module is used to set the operating parameters of the DC motor (1) and adjust the operating status of the DC motor (1); the interactive display module displays the speed signal and system status of the DC motor (1) based on the screen display (13).

2. The small lightweight rotor system fault simulation test bench according to claim 1 is characterized in that: The balancing disc (7) has a plurality of holes (7-1) evenly distributed on its circumference. By screwing counterweights into some of the holes (7-1), the mass of the balancing disc (7) is asymmetric, thereby simulating a rotor unbalance fault. By adjusting the mass of the counterweights screwed in, simulation of rotor unbalance faults of varying degrees is achieved.

3. The small lightweight rotor system fault simulation test bench according to claim 2 is characterized in that: The friction bracket (8) is arranged outside the balancing plate rotating shaft (14) and is a "door"-shaped structure; screw holes are arranged on the side and top of the friction bracket (8) for respectively screwing in a transverse screw (8-2) and a longitudinal screw (8-1) for simulating a rotating shaft friction fault; the transverse screw (8-2) and the longitudinal screw (8-1) are both made of plastic material, and the feed amount of the transverse screw (8-2) and the longitudinal screw (8-1) is manually adjusted to produce different degrees of contact with the balancing plate rotating shaft (14), thereby adjusting the degree of friction.

4. The small lightweight rotor system fault simulation test bench according to claim 3 is characterized in that: The parallel shaft reducer (10) has a built-in gear switching mechanism and a fault gear; wherein the fault gear and a plurality of groups of normal gears are mounted on the output shaft of the parallel shaft reducer (10); when the gear is switched by the gear switching mechanism, different normal gears or fault gears will cooperate with the gear mechanism inside the parallel shaft reducer (10), thereby realizing the normal deceleration function output of the parallel shaft reducer (10) or the simulation of reducer fault.

5. A small lightweight rotor system fault diagnosis method based on the fault simulation test bench according to claim 4, characterized in that: The following steps are involved: Step 1: Start the small lightweight rotor system fault simulation test bench and initialize the test bench; the initialization content includes: The speed and direction of the DC motor (1) are set, and the simulated fault is set; if a rotor imbalance fault is simulated, a counterweight is screwed into the hole (7-1) of the balancing plate (7) in advance; if a shaft friction fault is simulated, a transverse screw (8-2) and a longitudinal screw (8-1) are assembled on the friction bracket (8); if a parallel shaft reducer (10) fault is simulated, the gear switching mechanism of the parallel shaft reducer (10) is adjusted to switch to the faulty gear or the normal gear according to the test requirements; Setting an initial braking threshold, the initial braking threshold comprising initial thresholds of a current signal and a vibration acceleration signal of the DC motor (1); Step 2, after the initialization is completed, the DC motor (1) starts to work, and the PID algorithm is used to stabilize the speed of the DC motor (1) to reach the speed and direction set during initialization; Step 3, the signal acquisition module collects the current signal of the DC motor (1) and the vibration acceleration signal of the balancing disk shaft (14) under the set simulated fault and sends them to the data processing module for signal processing and analysis, and judges the system state according to the fault diagnosis result; at the same time, the data processing module also judges whether to perform braking protection according to the initial braking threshold; Step 4, if the braking protection does not occur, the fault simulation is performed; when simulating the shaft friction fault, the feed amount of the horizontal screw (8-2) and the longitudinal screw (8-1) is adjusted to control the degree of friction of different shafts; when simulating the fault of the parallel shaft reducer (10), the gear switching mechanism is used to switch the gear in time; The system state is determined based on the data processing module; if a certain fault is determined to have occurred through the fault diagnosis result, it is transmitted to the interactive display module, and a control option for stopping the DC motor (1) is given; when a combined fault occurs, emergency stopping protection is directly performed; Step 5, determining whether the fault simulation is completed, and stopping the DC motor (1) when completed.

6. The small lightweight rotor system fault diagnosis method according to claim 5, characterized in that: The data processing module further determines whether to perform a stop protection according to the initial stop threshold, including: When it is detected that the current signal is greater than the initial threshold of the current signal, it is determined that a fault has occurred and the braking protection is performed; when it is detected that the vibration acceleration signal exceeds the initial threshold of the vibration acceleration signal, it is determined that a fault has occurred, and the interactive module displays whether to perform braking protection for the operator to choose.

7. The small lightweight rotor system fault diagnosis method according to claim 5, characterized in that: The data processing module adopts a data processing method based on cumulative spectral distribution entropy to calculate the cumulative spectral distribution entropy of the vibration acceleration signal, and then inputs the cumulative spectral distribution entropy as a signal feature into the multi-label K nearest neighbor classifier for fault classification, thereby obtaining the fault diagnosis result.

8. The small lightweight rotor system fault diagnosis method according to claim 7, characterized in that: The data processing method based on cumulative spectral distribution entropy is specifically as follows: Step 4.1, perform discrete Fourier transform on the vibration acceleration signal time series x(n) of length N to obtain the signal X(k); n represents the nth data in the sequence, ; Step 4.2: Determine the power spectrum distribution of the signal based on the signal X(k) : ； Where, k=0, 1, ..., N / 2-1, k represents the kth frequency component; Step 4.3: Distribute the power spectrum The N / 2 points of the mid-spectral component are divided into M intervals, and then the power probability pdf of each interval is calculated m , expressed as: ； Among them, m=0, 1, ..., M, k represents the kth spectral component, Bin m Represents the set of frequency components in the mth interval; Step 4.4: Determine the cumulative distribution function of the spectrum interval based on the probability density function : ； Step 4.5, cumulative distribution function based on spectrum interval , using the standardized cumulative residual function to determine the cumulative residual probability distribution : ； Step 4.6, the cumulative spectral distribution entropy CSDE is calculated as follows: ； In step 4.7, the calculated cumulative spectral distribution entropy value CSDE is used as the input of the multi-label K nearest neighbor classifier ML-KNN to identify the fault diagnosis result.

9. The small lightweight rotor system fault diagnosis method according to claim 8, characterized in that: The training process of the multi-label K nearest neighbor classifier ML-KNN is as follows: (1) Set different categories of fault diagnosis results; collect the vibration acceleration signals of the fault simulation test bench in healthy state and different fault types and calculate the corresponding cumulative spectral distribution entropy value CSDE, which are used as normal data and fault data respectively, and set corresponding labels for normal data and fault data respectively, so as to construct a data set; divide the data set into a training set and a test set; (2) Build an ML-KNN classifier and put the divided training set into the classifier training; (3) The trained ML-KNN classifier is input into the test set for testing. When the test effect meets the preset requirements, the trained ML-KNN classifier is saved; After obtaining a set of new vibration acceleration signals, their cumulative spectral distribution entropy is extracted as signal features and input into the trained ML-KNN classifier to obtain predicted fault diagnosis results, including whether the system is healthy or what type of fault has occurred.

10. The small lightweight rotor system fault diagnosis method according to claim 9, characterized in that: The fault diagnosis result includes system health or system fault. System faults are divided into different fault types, including rotor imbalance fault, shaft rubbing fault, parallel shaft reducer fault, and combination of these faults; each fault type is divided into mild and severe faults.

Images