Fault gene-based health management monitoring parameter selection method

Abstract

The present application aims at the problems that the health management system of the rotating machinery exists in the selection of monitoring parameters, the structure of the rotating machinery rotor-supporting system is complex, the potential fault modes are numerous, and the experience of the selection of monitoring parameters is scattered; the present application provides a health management monitoring parameter selection method based on fault genes, analyzes the potential fault modes, establishes a monitoring parameter list, calculates the parameter weight, evaluates and compares the performance of each monitoring parameter, iteratively designs the monitoring parameters based on the above, deletes a large amount of redundant monitoring information, and finally determines a set of the best monitoring parameter list for monitoring and management, so that the monitoring parameters can be automatically evaluated, the efficiency of identifying and judging the fault characteristics of the health management system is improved, the test resources are saved, the performance of the monitoring system is optimized, and the accuracy of the fault monitoring is improved.

Description

Technical Field

[0001] The present invention belongs to the field of rotating machinery health management, and in particular relates to a health management monitoring parameter selection method based on fault genes. Background Art

[0002] Condition monitoring is currently the most pressing technical need in the field of equipment operation and maintenance. Selecting appropriate monitoring parameters is the fundamental guarantee for the stable operation of the health management system and improving the accuracy of condition monitoring. However, in actual engineering applications, it is found that the selection of traditional health management monitoring parameters has the following problems:

[0003] 1. In the application of traditional health management methods for rotating machinery rotor-support systems, a set of monitoring parameters is often selected to deal with all rotating machinery; however, the applicant has found in actual engineering that the information contained in a set of monitoring parameters is not applicable to all rotating machinery structures. The selection of parameters needs to vary according to the characteristics of the monitored object. Moreover, the information contained in a set of monitoring parameters often complements each other, and it is necessary to take into account both the accuracy and real-time performance of the system. If too few monitoring parameters are selected, the requirement for timely fault detection cannot be met, and if too many monitoring parameters are selected, the requirement for online / airborne real-time performance cannot be met.

[0004] 2. Traditional health management methods do not consider the potential fault characteristics of rotating machinery when selecting monitoring parameters, resulting in a large amount of redundant monitoring information, thereby wasting testing resources, and causing low efficiency in judging the health status of rotating machinery and reduced decision-making accuracy.

[0005] 3. The monitoring parameters of traditional health management systems are mainly selected based on expert experience, which leads to problems such as poor system monitoring accuracy and low parameter selection efficiency during the development of the health management system. In addition, the empirical parameter selection method has not formed a methodology and cannot break through the knowledge barriers between researchers, resulting in increased system R&D costs.

[0006] In response to the problems existing in the above-mentioned traditional health management methods, some studies have begun to consider how to select appropriate monitoring parameters. For example, in U.S. Patent US20200250584A1, based on the characteristics of the monitored object, a reference hypothesis model is selected based on similarity analysis, and on this basis, important parameters are quickly and effectively found from a small amount of monitoring data. Although this invention takes into account the mechanical structure characteristics of the object when selecting monitoring parameters, this method does not start from the perspective of structural failure. When selecting parameters, the parameter influence weight is not considered, and only qualitative elimination is performed, resulting in the selected parameters not reflecting the characteristics of the monitored object.

[0007] In a domestic patent application, the patent application with application number CN202210758045.X discloses a health management method and system for an aircraft engine. By collecting the target operating parameters of the engine in the working state, an abnormal operating feature set is constructed, and the abnormal operating feature set is weighted matched to generate the abnormal failure probability. However, this method only considers the monitoring parameter weight when generating a fault report, and redundant information is not eliminated in the monitoring stage, resulting in a waste of testing resources.

[0008] In addition, Yang Shuming et al., in their paper "Research on Monitoring Parameter Selection and Health Assessment Methods for Equipment Health Management," introduced a monitoring parameter selection method based on fuzzy hierarchical analysis. This method, combined with rough set attribute importance theory to determine monitoring parameter weights, provided a health assessment conclusion for rotating machinery. However, this method failed to consider the potential failure modes of rotating machinery, preventing the health management system from better understanding and identifying different failure characteristics, and parameter information was not fully utilized.

[0009] In summary, when selecting monitoring parameters, the current rotating machinery health management system and method have problems such as the complex structure of the rotating machinery rotor-support system, the numerous potential failure modes, and the scattered experience in selecting monitoring parameters. Summary of the Invention

[0010] In view of the problems existing in the prior art, the present invention proposes a health management monitoring parameter selection method based on fault genes.

[0011] The technical solution of the present invention is:

[0012] A health management monitoring parameter selection method based on fault genes is used to construct a rotating machinery monitoring and management system using a rotating machinery rotor-support tester. The special feature of the method is that it includes the following steps:

[0013] Step 1: Analyze the potential failure modes of the rotating machinery rotor-support system and establish a failure mode list;

[0014] The fault mode list includes several fault types and corresponding <fault characteristics>;

[0015] Step 2: Establish a fault gene library of the rotating machinery rotor-support system through a rotating machinery rotor-support experimenter;

[0016] Step 3: Supplement the fault gene library based on the fault mode list obtained in step 1;

[0017] The supplementing method is: checking whether the corresponding <fault feature> exists in the fault gene library, and when the fault gene library does not include the <fault feature> in the fault mode list, then adding the <fault feature> to the fault gene library;

[0018] Step 4: Read all characteristic parameters of the <fault characteristics> in the fault mode list obtained in step 1, and establish a list of candidate monitoring parameters for the rotating machinery rotor-support system;

[0019] The characteristic parameters of the <fault characteristics> are divided into three types of parameters, namely time domain characteristic parameters, frequency domain characteristic parameters or other characteristic parameters;

[0020] Step 5: Based on the list of candidate monitoring parameters of the rotating machinery rotor-support system, a fault experiment is performed on the rotating machinery rotor-support system using an experimenter to obtain a fault simulation data set;

[0021] Step 6: Calculate the weight of each parameter in the list of candidate monitoring parameters based on the fault simulation data set:

[0022] According to the fault simulation data set, calculate each monitoring parameter in the candidate monitoring parameters as the simulated fault characteristic parameter, and compare each simulated fault characteristic parameter with the fault characteristic parameter of <fault characteristic> in the fault gene library in step 2:

[0023] If a certain simulated fault feature parameter meets the <fault feature> of a certain fault mode in the fault gene library, the parameter count value of the fault feature can be indicated as +1, and the initial count value of each parameter is 0;

[0024] After all the parameters of the <fault feature> are counted, the counting results are normalized and the normalized results are used as parameter weights. The normalization process is:

[0025] X′=X / X MAX

[0026] Among them, X′ is the parameter weight, X is the parameter count value, and X MAX is the maximum parameter count value;

[0027] Step 7: Analyze all monitoring parameters in the list of selected monitoring parameters and estimate the time required for all health management algorithms to complete operation;

[0028] Step 8: Determine whether the total time obtained in step 7 meets the predetermined time requirement:

[0029] If the time requirement is not met, the feature with the lowest weight in the list of candidate monitoring parameters is deleted, and then go to step 7;

[0030] Otherwise, go to step 9;

[0031] Step 9: Check whether the engineering requirements indicators in the rotating machinery health management system meet the requirements:

[0032] If any indicator does not meet the requirements, reduce the monitoring characteristic parameters and return to step 4;

[0033] Otherwise, output the monitoring parameter list.

[0034] Furthermore, in step 1, the potential failure modes of the rotor-support system of the rotating machinery are analyzed using the FAMEC method in reliability theory to obtain a failure mode list.

[0035] Furthermore, the failure mode list is:

[0036] serial number Potential failure modes 1 Misalignment 2 unbalanced 3 Bearing outer ring failure 4 Bearing inner ring failure 5 Bearing rolling element failure .

[0037] Furthermore, the fault gene library of the rotating machinery rotor-support system established in step 2 is:

[0038]

[0039] Furthermore,

[0040] The time domain characteristic parameters include: mean, variance, standard deviation, peak-to-peak value, and single peak value;

[0041] The frequency domain characteristic parameters include: spectrum diagram, average power spectrum density, and spectrum kurtosis;

[0042] The other characteristic parameters include envelope characteristic parameters.

[0043] Furthermore, the list of monitoring parameters to be selected in step 4 is:

[0044] serial number Monitoring parameters 1 Vibration acceleration amplitude variance 2 Peak-to-peak vibration displacement 3 Vibration acceleration single peak 4 Vibration frequency domain mean 5 Signal power spectral density 6 Signal spectral kurtosis 7 Bearing inner ring characteristic frequency dominance 8 Bearing outer ring characteristic frequency dominance 9 Bearing rolling element characteristic frequency dominance .

[0045] Furthermore, the standard algorithm library includes all data processing and decision-making algorithms used in the health management system.

[0046] Concept and principle of the present invention:

[0047] like Figure 1 As shown in FIG, this method establishes a fault gene library for the rotating machinery system to indicate potential fault characteristics. Different monitoring parameters indicate different genes of the rotating machinery system. The process of assigning weights to monitoring parameters is similar to biological competition and evolution in nature.

[0048] This method first establishes a fault mode list and a fault gene library, and then establishes a list of candidate monitoring parameters by reading the characteristic parameters in the fault list. Then, the performance of each monitoring parameter is evaluated and compared through the simulation data set obtained from the rotating machinery rotor-support experiment device, and the inferior parameters are eliminated and the parameters with higher weights are retained. Finally, a set of the best monitoring parameter list is determined for monitoring and management.

[0049] Beneficial effects

[0050] The beneficial effects of the present invention are:

[0051] 1. Based on fault gene technology, the present invention selects the most important monitoring parameters by identifying and analyzing potential fault modes. By iteratively designing the monitoring parameters, a large amount of redundant monitoring information is deleted, and the selection of online / airborne monitoring parameters is realized. A set of parameter selection methodologies is formed, thereby improving the efficiency and accuracy of the monitoring system, breaking the knowledge barriers of empirical parameter selection methods, and saving the R&D costs of rotating machinery health management systems.

[0052] 2. The present invention targets potential fault modes, supplements the fault gene library through a fault mode list, and systematically represents the fault characteristics in the fault gene library, helping the rotating machinery health management system to establish the correlation between fault characteristics, further improving the system's perception of potential faults, enabling the health management system to more quickly understand and identify different fault characteristics, and providing more accurate data support for fault diagnosis and prediction.

[0053] 3. The present invention forms a set of evaluation methods for the selection results of monitoring parameters of the health management system. By establishing a monitoring parameter list for the rotor-support system of rotating machinery, the weight of each parameter is obtained to indicate the applicability of the parameter information to the monitored object; this method further optimizes the performance of the monitoring system and improves the accuracy of fault monitoring by evaluating whether the monitoring parameters can meet the engineering requirements of health management.

[0054] Additional aspects and advantages of the present invention will be set forth in part in the description which follows and, in part, will be obvious from the description which follows, or may be learned by practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the following description of the embodiments with reference to the accompanying drawings, in which:

[0056] Figure 1 This is a flow chart for selecting health management monitoring parameters of the present invention;

[0057] Figure 2 This is a schematic diagram of the structure of a rotating machinery rotor-support tester;

[0058] Figure 3 Flowchart of the health management algorithm.

[0059] Among them: 1-first fulcrum bearing (faulty bearing); 2-first fulcrum bearing seat; 3-first fulcrum load-bearing plate; 4-first fulcrum support; 5-disc; 6-second fulcrum support; 7-second fulcrum load-bearing plate; 8-second fulcrum bearing seat; 9-second fulcrum bearing (normal bearing); 10-coupling. DETAILED DESCRIPTION

[0060] The following describes embodiments of the present invention in detail, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and are not to be construed as limiting the present invention.

[0061] This embodiment proposes a method for selecting health management monitoring parameters based on fault genes, which is used to construct a rotating machinery monitoring and management system. The system uses a rotating machinery rotor-support tester to simulate imbalance faults, misalignment faults, and bearing faults that occur during aircraft engine operation, obtains rotating machinery fault signals, and uses the fault operation data of the tester as a simulation data set. By generating a qualified monitoring parameter list, the system selects health management monitoring parameters for the rotating machinery rotor-support system. The specific steps are:

[0062] Step 1: Analyze the potential failure modes of the rotating machinery rotor-support system and establish a failure mode list;

[0063] The fault mode list includes several fault types and corresponding <fault characteristics>;

[0064] In this embodiment, the rotor system of the rotating machinery rotor-support tester is designed through dynamics, adopts a rigid rotor, and the operating speed is set below the critical speed of 4217 rpm. It is powered by a three-phase asynchronous motor, and the single-disc rotor is supported on the bearing seat by ball bearings and roller bearings.

[0065] The potential failure modes of the rotating machinery rotor-support system are analyzed using the FAMEC method in reliability theory, and the failure mode list is shown in Table 1:

[0066] Table 1 Failure mode list

[0067] serial number Potential failure modes 1 Misalignment 2 unbalanced 3 Bearing outer ring failure 4 Bearing inner ring failure 5 Bearing rolling element failure

[0068] Step 2: Establish a fault gene library of the rotating machinery rotor-support system through a rotating machinery rotor-support experimenter;

[0069] This embodiment adopts the fault gene library construction method given on pages 272-281 of "Aero-Engine Fault Diagnosis", based on Figure 2 The structure and performance of the rotating machinery rotor-support tester shown in the figure list the fault types with high fault occurrence rate and easy to induce other types of faults as typical faults;

[0070] The rotating machinery rotor-support tester structure includes:

[0071] First fulcrum bearing (faulty bearing) 1; first fulcrum bearing seat 2; first fulcrum load-bearing plate 3; first fulcrum support 4; disk 5; second fulcrum support 6; second fulcrum load-bearing plate 7; second fulcrum bearing seat 8; second fulcrum bearing (normal bearing) 9; coupling 10.

[0072] The fault gene library established for the rotating machinery rotor-support tester established in this embodiment is shown in Table 2:

[0073] Table 2 Fault gene library

[0074]

[0075] Step 3: Supplement the fault gene library based on the fault mode list obtained in step 1;

[0076] The supplement method is: check whether the corresponding <fault feature> exists in the fault gene library shown in Table 2. If the fault gene library does not include the <fault feature> in the fault mode list, then add the <fault feature> to the fault gene library;

[0077] Step 4: As shown in Table 3, read the characteristic parameters of all <fault characteristics> in the fault mode list obtained in Step 1, and establish a list of candidate monitoring parameters for the rotating machinery rotor-support system;

[0078] Table 3 List of monitoring parameters to be selected

[0079] serial number Monitoring parameters 1 Vibration acceleration amplitude variance 2 Peak-to-peak vibration displacement 3 Vibration acceleration single peak 4 Vibration frequency domain mean 5 Signal power spectral density 6 Signal spectral kurtosis 7 Bearing inner ring characteristic frequency dominance 8 Bearing outer ring characteristic frequency dominance 9 Bearing rolling element characteristic frequency dominance

[0080] The characteristic parameters of <fault characteristics> are divided into three categories: time domain characteristic parameters, frequency domain characteristic parameters or other characteristic parameters;

[0081] Among them, the time domain characteristic parameters are a series of parameters that describe the change pattern of the signal on the time axis. The parameters indicate the dynamic characteristics, periodicity, and change trend of the signal. The time domain characteristic parameters of this embodiment are the vibration acceleration amplitude variance, the vibration displacement peak-to-peak value, and the vibration acceleration single peak value;

[0082] Frequency domain characteristic parameters are used to describe the properties and characteristics of a signal in the frequency domain. They are obtained by converting the signal from the time domain to the frequency domain using Fourier transform or other frequency domain analysis methods. The frequency domain characteristic parameters of this embodiment are the vibration frequency domain mean, signal power spectral density, and signal spectral kurtosis.

[0083] Other characteristic parameters are parameters that can reflect signal characteristics other than signal time domain characteristic parameters and frequency domain characteristic parameters; the other characteristic parameters of this embodiment are the characteristic frequency dominance of the bearing inner ring, the characteristic frequency dominance of the bearing outer ring, and the characteristic frequency dominance of the bearing rolling element.

[0084] The calculation method of each characteristic parameter according to the vibration signal is as follows:

[0085] Time domain characteristic parameter calculation process:

[0086] Signal peak value: the maximum value of time domain signal fluctuation in a specified time period: X p = max |x(t i ) |;

[0087] Signal peak-to-peak value (V p-p ): the difference between the maximum and minimum values of the time domain signal: V P-P = max (x(t) | - min x(t));

[0088] Signal mean value:

[0089] Signal variance:

[0090] Signal standard deviation:

[0091] Frequency domain characteristic parameter calculation process:

[0092] First, the vibration signal is subjected to FFT to obtain its components at different frequencies and their relative intensity;

[0093] Signal vibration frequency domain mean value: Where s i represents the frequency domain discrete point value;

[0094] Signal power spectral density: the average power P of a continuous time domain signal in a time range is defined as:

[0095] When a window function ω(t) is introduced to extract signal data in a limited range, the above formula can be rewritten as

[0096]

[0097] The time domain non-integrable energy infinite signal is converted into an integrable energy limited signal in an infinite time range, and at this time

[0098]

[0099] The signal power spectral density to be solved is:

[0100]

[0101] Signal spectral kurtosis: first calculate the expectation E, mean value μ, and standard deviation σ according to the signal data set X = {x1, x2,... x n}, and get the spectral kurtosis:

[0102] Kurtosis=E[(X-μ) 4 ] / σ 4

[0103] Calculation process of other characteristic parameters:

[0104] The envelope time-delay correlation demodulation method introduced on page 247 of "Aero Engine Fault Diagnosis" is used to obtain the signal envelope characteristics. Through envelope analysis, the fault characteristic frequency existing in the theoretical calculation of typical bearing faults is obtained, and dimensionless processing is performed to obtain the dominance parameter:

[0105] Bearing inner ring characteristic frequency dominance:

[0106] Bearing outer ring characteristic frequency dominance:

[0107] Bearing rolling element characteristic frequency dominance:

[0108] Where, is the characteristic frequency of the bearing inner ring; is the characteristic frequency of the bearing outer ring; is the characteristic frequency of the bearing rolling element; A T is the mean value of the vibration frequency domain.

[0109] Step 5: Based on the list of candidate monitoring parameters of the rotating machinery rotor-support system, a fault experiment is performed on the rotating machinery rotor-support system using an experimenter to obtain a fault simulation data set;

[0110] Step 6: Based on the fault simulation data set, calculate the weight of each parameter in the candidate monitoring parameter list to indicate the applicability of the parameter information to the monitoring object:

[0111] According to the fault simulation data set, calculate each monitoring parameter in the candidate monitoring parameters as the simulated fault characteristic parameter, and compare each simulated fault characteristic parameter with the fault characteristic parameter of <fault characteristic> in the fault gene library in step 2:

[0112] If a certain simulated fault feature parameter meets the <fault feature> of a certain fault mode in the fault gene library, the parameter count value of the fault feature can be indicated as +1, and the initial count value of each parameter is 0;

[0113] After all the parameters of the <fault feature> are counted, the counting results are normalized and the normalized results are used as parameter weights. The normalization process is:

[0114] X′=X / X MAX

[0115] Among them, X′ is the parameter weight, X is the parameter count value, and X MAX is the maximum parameter count value;

[0116] The monitoring parameter normalization parameter weight table in this embodiment is shown in Table 4:

[0117] Table 4 Normalized weights of monitoring parameters

[0118] Monitoring parameters Weight X′ Peak-to-peak vibration displacement 1 Bearing inner ring characteristic frequency dominance 0.91 Bearing outer ring characteristic frequency dominance 0.90 Bearing rolling element characteristic frequency dominance 0.89 Vibration frequency domain mean 0.80 Vibration acceleration single peak 0.72 Vibration acceleration amplitude variance 0.35 Signal power spectral density 0.12 Signal spectral kurtosis 0.09

[0119] Step 7: Analyze all monitoring parameters in the list of selected monitoring parameters and estimate the time required for all health management algorithms to complete operation;

[0120] The standard algorithm library contains all data processing and decision-making algorithms used in the health management system, including fast Fourier transform, time domain analysis, and envelope analysis;

[0121] like Figure 3 As shown, this embodiment uses three algorithms to construct a health management algorithm process, and the time taken to complete the operation is 320ms.

[0122] Step 8: Determine whether the total time obtained in step 7 meets the predetermined time requirement:

[0123] If the time requirement is not met, the feature with the lowest weight in the list of candidate monitoring parameters is deleted, and then go to step 7;

[0124] Otherwise, go to step 9;

[0125] This process is similar to biological competition and evolution in nature. By evaluating and comparing the performance of various monitoring parameters, inferior parameters are eliminated when they do not meet the system monitoring requirements, and parameters with excellent performance are retained. Finally, a set of the best parameters are selected for monitoring and management.

[0126] In this embodiment, the predetermined time requirement is 305ms. Therefore, in order to reduce the operation time of the health management system, the characteristic parameter "signal spectrum kurtosis" with the lowest weight in Table 4 is deleted, and step 7 is re-executed, with a timing result of 310ms; then the characteristic parameter "signal power spectral density" is further deleted, and step 7 is re-executed, with a timing result of 302ms, which meets the timing requirement.

[0127] Step 9: Check whether the engineering requirement indicators in the rotating machinery health management system meet the engineering requirements:

[0128] The engineering requirement indicators of this embodiment are the accuracy requirements of the fault location and the accuracy of the fault mode; specifically, the fault location and fault mode are identified by selecting monitoring parameters of a batch of rotating machinery, and then compared with the actual faults of the batch of rotating machinery to calculate the accuracy of error identification of this rotating machinery monitoring system.

[0129] If any indicator does not meet the requirements, reduce the monitoring characteristic parameters and return to step 4 to design a monitoring parameter list of the rotating machinery rotor-support system that meets the requirements through repeated iterations;

[0130] Otherwise, output the monitoring parameter list;

[0131] By leveraging the gene expression method in bionics, the present invention integrates the key information expression concept of "survival of the fittest" into the process of mechanical design and operation and maintenance, continuously updates the list of rotating machinery health monitoring parameters, and forms a methodology and process for monitoring parameter selection.

[0132] The above description is only a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with this technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed in the present invention, and these modifications or replacements should all be included in the scope of protection of the present invention.

Claims

1. A health management monitoring parameter selection method based on fault genes is used to construct a rotating machinery monitoring and management system using a rotating machinery rotor-support tester, which is characterized by: The following steps are involved: Step 1: Analyze the potential failure modes of the rotating machinery rotor-support system and establish a failure mode list; The fault mode list includes several fault types and corresponding <fault characteristics>; Step 2: Establish a fault gene library of the rotating machinery rotor-support system through a rotating machinery rotor-support experimenter; Step 3: Supplement the fault gene library based on the fault mode list obtained in step 1; The supplementing method is: checking whether the corresponding <fault feature> exists in the fault gene library, and when the fault gene library does not include the <fault feature> in the fault mode list, then adding the <fault feature> to the fault gene library; Step 4: Read all characteristic parameters of the <fault characteristics> in the fault mode list obtained in step 1, and establish a list of candidate monitoring parameters for the rotating machinery rotor-support system; The characteristic parameters of the <fault characteristics> are divided into three types of parameters, namely time domain characteristic parameters, frequency domain characteristic parameters or other characteristic parameters; Step 5: Based on the list of candidate monitoring parameters of the rotating machinery rotor-support system, a fault experiment is performed on the rotating machinery rotor-support system using an experimenter to obtain a fault simulation data set; Step 6: Calculate the weight of each parameter in the list of candidate monitoring parameters based on the fault simulation data set: According to the fault simulation data set, calculate each monitoring parameter in the candidate monitoring parameters as the simulated fault characteristic parameter, and compare each simulated fault characteristic parameter with the fault characteristic parameter of <fault characteristic> in the fault gene library in step 2: If a certain simulated fault feature parameter meets the <fault feature> of a certain fault mode in the fault gene library, the parameter count value of the fault feature can be indicated as +1, and the initial count value of each parameter is 0; After all the parameters of the <fault feature> are counted, the counting results are normalized and the normalized results are used as parameter weights. The normalization process is: X′=X / X MAX Among them, X′ is the parameter weight, X is the parameter count value, and X MAX is the maximum parameter count value; Step 7: Analyze all monitoring parameters in the list of selected monitoring parameters and estimate the time required for all health management algorithms to complete operation; Step 8: Determine whether the total time obtained in step 7 meets the predetermined time requirement: If the time requirement is not met, the feature with the lowest weight in the list of candidate monitoring parameters is deleted, and then go to step 7; Otherwise, go to step 9; Step 9: Check whether the engineering requirements indicators in the rotating machinery health management system meet the requirements: If any indicator does not meet the requirements, reduce the monitoring characteristic parameters and return to step 4; Otherwise, output the monitoring parameter list.

2. The method for selecting health management monitoring parameters based on fault genes according to claim 1, characterized in that: In the step 1, the potential failure modes of the rotor-support system of the rotating machinery are analyzed using the FAMEC method in reliability theory to obtain a failure mode list.

3. The method for selecting health management monitoring parameters based on fault genes according to claim 2, characterized in that: The failure mode list is: 。 4. The method for selecting health management monitoring parameters based on fault genes according to claim 1, characterized in that: The fault gene library of the rotating machinery rotor-support system established in step 2 is:

5. The method for selecting health management monitoring parameters based on fault genes according to claim 1, characterized in that: The time domain characteristic parameters include: mean, variance, standard deviation, peak-to-peak value, and single peak value; The frequency domain characteristic parameters include: spectrum diagram, average power spectrum density, and spectrum kurtosis; The other characteristic parameters include envelope characteristic parameters.

6. The method for selecting health management monitoring parameters based on fault genes according to claim 1, characterized in that: The list of monitoring parameters to be selected in step 4 is: 。 7. The method for selecting health management monitoring parameters based on fault genes according to claim 1, characterized in that: The standard algorithm library contains all data processing and decision-making algorithms used in the health management system.

Images