CN-117113172-B - Vibration analysis state index threshold value determining method, system, device and medium

CN117113172BCN 117113172 BCN117113172 BCN 117113172BCN-117113172-B

Abstract

The application discloses a vibration analysis state index threshold value determining method, a system, a device and a storage medium, wherein the method comprises the following steps of obtaining effective sample data sets of a plurality of aircraft state machines, wherein the effective data are used for representing data sets obtained by state machine measurement when the aircraft runs for a preset time under a preset condition, determining a single machine learning type threshold value and a cluster learning type threshold value according to the effective sample data sets, and determining a yellow attention threshold value and a red alarm threshold value according to the single machine learning type threshold value and the cluster learning type threshold value. The method can increase the alarm accuracy and reduce the maintenance cost. The application can be widely applied to the technical field of aircrafts.

Inventors

YE HONGKE
TAO ZHIYU
BAI YUNDONG
LIANG ZHONGZI
XIONG QIANG
Ma Ankang
CHEN SONGJING
MA SHUIQUAN

Assignees

广州航新航空科技股份有限公司

Dates

Publication Date: 20260512
Application Date: 20230804

Claims (5)

1. A vibration analysis status index threshold determination method, characterized by comprising: The method comprises the steps of obtaining effective sample data sets of a plurality of aircraft state machines, wherein the effective data are used for representing data sets obtained by measuring the state machines when the aircraft runs for a preset time under a preset condition, and samples of the effective data comprise an acquisition system, no faults of a sensor and no abnormality of acquisition signals, wherein the number of single machines is more than or equal to 100 when the effective data corresponding to single machines are required to meet the condition that the length of the flight time is more than or equal to 100h and the number of state indexes CI is more than or equal to 100, and the number of the single machines is more than or equal to 5 when the effective data corresponding to a cluster are required to meet a single machine learning threshold; The method comprises the steps of determining a single machine learning type threshold and a cluster learning type threshold according to an effective sample data set, wherein the determination of the cluster learning type threshold comprises the steps of extracting all effective sample data, carrying out cluster analysis, determining cluster analysis evaluation indexes when different classification numbers are determined, extracting the maximum indexes in the cluster analysis evaluation indexes when the different classification numbers are extracted, enabling a plurality of aircrafts to have no cluster learning type threshold if the classification number corresponding to the maximum indexes is not 1, enabling the aircrafts to be a single machine learning type threshold, re-sampling an effective sample of each aircrafts according to a quartile method if the classification number corresponding to the maximum indexes is 1, calculating the cluster analysis evaluation indexes corresponding to the different classification numbers again, and enabling the aircrafts to have the cluster learning type threshold if the classification number corresponding to the maximum indexes in the cluster analysis evaluation indexes of the different classification numbers is 1, otherwise, enabling the aircrafts to be the single machine learning type threshold; the average value of all the effective sample data and the variance of the effective sample data set are calculated, the average value and the variance are input into an alarm threshold value determining formula to obtain a yellow attention threshold value and a red alarm threshold value, wherein the alarm threshold value determining formula comprises the following steps: T = μ±N σ Wherein T is a yellow attention threshold or a red warning threshold, mu is a mean value, sigma is a variance, and N is a positive integer; Acquiring a state index of each part of the aircraft; When the state index is smaller than or equal to a yellow threshold value and a red threshold value, determining that the display part is normal; When the status indicator is greater than a yellow threshold and less than a red threshold, a yellow alert is given, the yellow alert being used to characterize an early alert for a potential problem; when the state index is greater than or equal to a red threshold, giving a red alarm, wherein the red alarm is used for representing that the monitored component has relatively clear fault symptoms; For normal or fault data sets with marks, the alarm accuracy is required to be more than 99%, the false alarm rate is less than 1%o and the false alarm rate is less than 1%, for normal data sets without marks, the false alarm rate is required to be less than 1%, and if the above criteria are not met, the effective sample data sets are collected again and the state index threshold is determined again.
2. The vibration analysis status index threshold value determining method according to claim 1, wherein the step of determining a stand-alone learning type threshold value from the valid sample data set comprises: Calculating the mean value of all the effective sample data and the variance of the effective sample data set; and inputting the mean value and the variance into a preset self-learning model to obtain a learning type single machine threshold value.
3. A vibration analysis status indicator threshold determination system for use in the method of any one of claims 1-2, comprising: the system comprises an acquisition unit, an acquisition unit and a state index CI (common element) acquisition unit, wherein the acquisition unit is used for acquiring effective sample data sets of a plurality of aircraft state machines, the effective data are used for characterizing a data set measured by the state machines when the aircraft runs for a preset time under a preset condition, and samples of the effective data comprise an acquisition system, no faults of a sensor and no abnormality of an acquisition signal, wherein the effective data corresponding to a single machine are required to meet the condition that the length of the flight is greater than or equal to 100h, the number of the state indexes CI is greater than or equal to 100, and the number of the single machines is greater than or equal to 5 under the condition that the effective data corresponding to a cluster is required to meet a single machine learning threshold; The first processing unit is used for determining a single machine learning type threshold value and a cluster learning type threshold value according to the effective sample data set; The second processing unit is used for determining a yellow attention threshold and a red alarm threshold according to the single machine learning type threshold and the cluster learning type threshold; Acquiring a state index of each part of the aircraft; When the state index is smaller than or equal to a yellow threshold value and a red threshold value, determining that the display part is normal; When the status indicator is greater than a yellow threshold and less than a red threshold, a yellow alert is given, the yellow alert being used to characterize an early alert for a potential problem; when the state index is greater than or equal to a red threshold, giving a red alarm, wherein the red alarm is used for representing that the monitored component has relatively clear fault symptoms; For normal or fault data sets with marks, the alarm accuracy is required to be more than 99%, the false alarm rate is less than 1%o and the false alarm rate is less than 1%, for normal data sets without marks, the false alarm rate is required to be less than 1%, and if the above criteria are not met, the effective sample data sets are collected again and the state index threshold is determined again.
4. A vibration analysis state index threshold value determining apparatus characterized by comprising: At least one processor; at least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor is caused to implement a vibration analysis status indicator threshold determination method according to any one of claims 1-2.
5. A computer readable storage medium having stored therein processor executable instructions which, when executed by a processor, are for performing a vibration analysis status indicator threshold determination method according to any of claims 1-2.

Description

Vibration analysis state index threshold value determining method, system, device and medium Technical Field The application relates to the technical field of aircrafts, in particular to a vibration analysis state index threshold value determining method, a vibration analysis state index threshold value determining system, a vibration analysis state index threshold value determining device and a vibration analysis state index threshold value storing medium. Background With the development of technology, a Health Use Monitoring System (HUMS) of a helicopter is widely applied to various helicopters and plays an increasingly important role. HUMS can realize health monitoring of three-big moving parts of the helicopter and the helicopter body structure based on vibration through methods such as data analysis, state monitoring, fault diagnosis, trend analysis and the like, improves safety and integrity, and reduces use and maintenance cost. In order to realize the health monitoring function, the reasonable setting of the health monitoring threshold is the key for realizing effective health monitoring, and the method for formulating, verifying and updating the HUMS threshold can improve the accuracy and the practicability of the health monitoring, so that the HUMS really plays the role of the HUMS, and has important significance. The foreign HUMS technology and products are very mature, and have a complete state index system and a threshold setting management method. The European helicopter HUMS system (such as an M' arms system provided with an EC225 helicopter and an EC155 helicopter) and the Western Coscose HUMS system (such as an IMD-HUMS system provided with an S92A) are provided with a complete set of threshold value determining and updating methods, the statistical characteristics of each state index are calculated through a statistical analysis method, and different threshold value types are set for each state index by combining the characteristics of the monitored component and the change rule of CI, so that the accuracy and the practicability of the threshold value are ensured. The domestic health monitoring and management technology starts later, the foundation is weak, especially the foundation data accumulation is insufficient, the health monitoring function, the vibration monitoring threshold value for judging whether the equipment is normal or not is mainly given by a design department, and the threshold value determining mode is mainly based on theoretical calculation and ground test results. The threshold type is single, only the threshold of the organic group is used, the individual difference is not considered, false alarm or false alarm is easy to generate, the vibration overrun state indexes are not differentiated in importance degree, maintenance suggestions are uniformly given, excessive maintenance is easy to cause, and the maintenance cost is increased. Therefore, a new vibration analysis state index threshold determination method is needed. Disclosure of Invention The present application aims to solve at least one of the technical problems existing in the prior art to a certain extent. Therefore, an object of the embodiments of the present application is to provide a method, a system, a device and a storage medium for determining a vibration analysis status index threshold, which can increase the accuracy of alarm and reduce maintenance cost. In order to achieve the technical aim, the technical scheme adopted by the embodiment of the application comprises a method for determining vibration analysis state index thresholds, wherein the method comprises the steps of obtaining effective sample data sets of a plurality of aircraft state machines, wherein the effective data are used for representing data sets measured by the state machines when the aircraft runs for a preset time under a preset condition, determining single-machine learning type thresholds and cluster learning type thresholds according to the effective sample data sets, and determining yellow attention thresholds and red alarm thresholds according to the single-machine learning type thresholds and the cluster learning type thresholds. In addition, according to the method for determining the vibration analysis state index threshold according to the above embodiment of the present invention, the following additional technical features may be provided: Further, in the embodiment of the application, the step of determining the single machine learning threshold according to the effective sample data set comprises the steps of calculating the mean value of all the effective sample data and the variance of the effective sample data set, and inputting the mean value and the variance into a preset self-learning model to obtain the learning single machine threshold. Further, the method for determining the cluster learning type threshold according to the effective sample data set comprises the steps of extracting all effective sample dat