CN111259515B

CN111259515B - Aircraft health management method and system

Info

Publication number: CN111259515B
Application number: CN202010015332.2A
Authority: CN
Inventors: 芮云波; 颜军; 董文岳; 占连样
Original assignee: Zhuhai Orbita Aerospace Technology Co ltd
Current assignee: Zhuhai Aerospace Microchips Science & Technology Co ltd
Priority date: 2020-01-07
Filing date: 2020-01-07
Publication date: 2023-11-21
Anticipated expiration: 2040-01-07
Also published as: CN111259515A

Abstract

The invention discloses an aircraft health management method and system, which relate to the technical field of aerospace and are used for realizing: the fault monitoring, fault diagnosis, influence evaluation, fault prediction and the like of all subsystems of the aircraft and corresponding treatment measures, logistics guarantee arrangement and the like are integrated into a comprehensive management system for the health condition of the aircraft. The beneficial effects of the invention are as follows: and predicting the fault occurrence time, fault mode and the like according to certain fault symptoms, improving the maintenance guarantee of the aircraft, correctly implementing the optionally maintenance and reducing the fault occurrence rate of the aircraft.

Description

Aircraft health management method and system

Technical Field

The invention relates to the technical field of aerospace, in particular to an aircraft health management method and system.

Background

With the development of aerospace technology, the safety and reliability of an aircraft, the high efficiency and economy of an aircraft logistics guarantee system and the like are becoming unavoidable problems for developing the aerospace technology. Statistics from the federal aviation administration and the national transportation safety committee indicate that 24% of all world flight accidents have been caused by faults in aircraft subsystems and components, 26% have been caused by loss of flight, and a significant portion of the loss of flight is caused by faults in hardware and systems, in addition to the significant economic pressures faced by the aerospace industry, airlines spend 310 billions dollars on aircraft logistics each year, with an average of 12 hours of logistics time per hour.

In order to improve the reliability and safety of an aircraft and reduce the cost, how to establish a set of aircraft ground health management system is urgent.

Disclosure of Invention

In order to solve at least one of the technical problems existing in the prior art, the invention aims to provide an aircraft health management method and system, which synthesizes fault monitoring, fault diagnosis, influence evaluation, fault prediction and the like of all subsystems of an aircraft, corresponding treatment measures, arrangement of logistic guarantee and the like into a comprehensive management system for the health condition of the aircraft.

The first aspect of the technical scheme adopted by the invention for solving the problems is as follows: a method of aircraft health management comprising the steps of: a state monitoring step, namely establishing an anomaly monitoring library and setting a monitoring parameter range, and judging whether anomalies occur by collecting flight parameters of the aircraft; a health evaluation step of establishing a health behavior model and a health evaluation algorithm, and comparing real-time output parameters of a flight control system with the results output by the health behavior model based on the health evaluation algorithm to obtain the current flight control system health state evaluation result; a fault prediction step of acquiring aircraft equipment information, and obtaining the probability and time of occurrence of a predicted fault according to the aircraft equipment information for aircraft damage judgment, degradation state identification and residual service life prediction; and a maintenance management step, wherein a database is established for storing and managing historical data and state information of the aircraft, and analysis and research are carried out on the corresponding aircraft based on the output results of the three steps, so as to generate a maintenance scheme.

The beneficial effects are that: and predicting the fault occurrence time, fault mode and the like according to certain fault symptoms, improving the maintenance guarantee of the aircraft, correctly implementing the optionally maintenance and reducing the fault occurrence rate of the aircraft.

According to the first aspect of the invention, the establishment of the anomaly monitoring library specifically comprises: and establishing alarm monitoring conditions, and carrying out logic processing to obtain a decision tree based on the judgment conditions.

According to the first aspect of the present invention, the health evaluation step further includes: setting corresponding fault modes according to mutual influence factors of all components of the flight control system, and establishing a health mode table; simulating each health mode one by one, and obtaining response data under the corresponding health mode to form a neural network training sample space; training the neural network training sample space one by one to generate a health behavior model corresponding to the health mode; sorting the plurality of healthy behavior models to obtain a normal system model; and evaluating the data vector of the appointed test point of the flight control system based on a health evaluation algorithm, and carrying out comparison analysis by combining the result output by the normal system model to obtain the current flight control system health state evaluation result.

According to the first aspect of the present invention, the fault prediction step further includes: acquiring historical record data and corresponding state conditions of the aircraft, and modeling based on a neural network to acquire a mapping model between the historical record data and a predicted output state; and performing fault test on the aircraft, comparing, matching and evaluating the historical record data based on the mapping model according to the test information and the historical record data, setting the known state corresponding to the historical record data with the highest matching degree as the current state of the aircraft, and performing fault prediction according to the current state of the aircraft.

According to a first aspect of the present invention, the damage determination includes: collecting observation data of an aircraft in a specified time period, carrying out phase space reconstruction on the observation data and establishing a local linear model; estimating a tracking function according to the local linear model and constructing a tracking matrix, wherein the tracking matrix comprises tracking slow-change damage and working condition change; and separating the change trend of the slow-change damage from the tracking matrix by using a modal decomposition method to obtain a damage evolution process.

According to a first aspect of the invention, the degradation state identification comprises: an information acquisition step of acquiring working information of each time period for each component of the aircraft based on a plurality of kinds of sensors; an information processing step, namely extracting relevant characteristic vectors from the working state information by using a time domain analysis method and a time domain analysis method to obtain vector spaces constructed by different state characteristic vectors, and modeling corresponding state type spaces; an information identification step, namely constructing a nonlinear relation between a state feature vector space and a state type space, and training a model by adopting experimental sample data to obtain an information source state identification result; and a decision fusion step, namely comprehensively summarizing the state recognition results of the information sources, obtaining the total probability distribution of the state types based on fusion rules according to the basic confidence degrees of different recognition results, and further obtaining the final recognition result.

According to a first aspect of the invention, the remaining life prediction comprises: acquiring information of corresponding parts of the aircraft with fault characteristics, and acquiring observation data of the corresponding parts in a specified time period; selecting corresponding prediction features according to the components with fault features, preprocessing the prediction features, and performing noise smoothing on the observed data to obtain a preprocessed degradation feature sequence; performing regression fitting on the degradation characteristic sequence, and extracting a plurality of set data points corresponding to a non-zero basis function; establishing a degradation model according to the feature sequence, predicting and determining a priori degradation model based on a correlation vector machine method, and selecting the most suitable degradation model or determining to optimize and improve the model; performing congruent fitting on the observed data according to the degradation model to determine a model parameter value; and carrying out extrapolation prediction on the degradation model according to the parameter value to obtain estimation on the evolution trend of the prediction characteristic, wherein the estimation comprises a range estimation value of the residual service life of the corresponding part.

According to a first aspect of the present invention, the maintenance management step includes: a individuation step for designating corresponding maintenance items according to the maintenance manuals of the respective aircrafts; and a maintenance planning step, which is used for generating maintenance planning early warning according to the state parameters of the corresponding aircraft and issuing maintenance tasks.

The second aspect of the technical scheme adopted by the invention for solving the problems is as follows: an aircraft health management system, comprising: the state monitoring module is used for establishing an anomaly monitoring library and setting a monitoring parameter range, and judging whether anomaly occurs or not by collecting flight parameters of the aircraft; the health evaluation module is used for establishing a health behavior model and a health evaluation algorithm, and comparing the real-time output parameters of the flight control system with the results output by the health behavior model based on the health evaluation algorithm to obtain the current flight control system health state evaluation result; the fault prediction module is used for acquiring the information of the aircraft equipment, judging the damage of the aircraft, identifying the degradation state and predicting the residual service life of the aircraft according to the information of the aircraft equipment, and obtaining the probability and time of occurrence of the predicted fault; and the maintenance management module is used for establishing a database for storing and managing historical data and state information of the aircraft, analyzing and researching the corresponding aircraft based on the output results of the three modules, and generating a maintenance scheme.

According to a second aspect of the present invention, the failure prediction module further includes: the acquisition unit is used for acquiring the historical record data of the aircraft and the corresponding state conditions; the modeling unit is used for modeling based on the neural network according to the information acquired by the acquisition unit, and acquiring a mapping model between the historical record data and the predicted output state; the test unit is used for performing fault test on the aircraft to obtain test information; and the evaluation unit is used for comparing, matching and evaluating the historical record data based on the mapping model according to the test information and the historical record data, setting the known state corresponding to the historical record data with the highest matching degree as the current state of the aircraft, and carrying out fault prediction according to the current state of the aircraft.

Drawings

FIG. 1 is a schematic flow diagram of a method according to an embodiment of the invention;

FIG. 2 is a schematic diagram of a system architecture according to an embodiment of the invention;

FIG. 3 is a schematic diagram of a state monitoring decision according to an embodiment of the present invention;

FIG. 4 is a schematic diagram of a health assessment model according to an embodiment of the present invention;

FIG. 5 is a schematic diagram of a reconstructed phase space relationship according to an embodiment of the present invention;

FIG. 6 is a schematic diagram of remaining useful life prediction according to an embodiment of the present invention;

FIG. 7 is a schematic diagram of a management system main interface according to an embodiment of the invention;

FIG. 8 is a schematic diagram of an information gathering interface according to an embodiment of the invention;

FIG. 9 is a schematic diagram of a condition monitoring interface according to an embodiment of the present invention;

FIG. 10 is a health interface schematic according to an embodiment of the invention;

FIG. 11 is a schematic diagram of a fault prediction interface according to an embodiment of the present invention;

FIG. 12 is a schematic illustration of a maintenance interface according to an embodiment of the invention;

FIG. 13 is a schematic diagram of an integrated query interface according to an embodiment of the present invention.

Detailed Description

The conception, specific structure, and technical effects produced by the present invention will be clearly and completely described below with reference to the embodiments and the drawings to fully understand the objects, aspects, and effects of the present invention.

Referring to fig. 1, a method according to an embodiment of the present invention is shown schematically, comprising the steps of:

a state monitoring step, namely establishing an anomaly monitoring library and setting a monitoring parameter range, and judging whether anomalies occur by collecting flight parameters of the aircraft;

a health evaluation step of establishing a health behavior model and a health evaluation algorithm, and comparing real-time output parameters of a flight control system with the results output by the health behavior model based on the health evaluation algorithm to obtain the current flight control system health state evaluation result;

A fault prediction step of acquiring aircraft equipment information, and obtaining the probability and time of occurrence of a predicted fault according to the aircraft equipment information for aircraft damage judgment, degradation state identification and residual service life prediction;

and a maintenance management step, wherein a database is established for storing and managing historical data and state information of the aircraft, and analysis and research are carried out on the corresponding aircraft based on the output results of the three steps, so as to generate a maintenance scheme.

Referring to FIG. 2, a schematic diagram of a system architecture according to an embodiment of the present invention;

the aircraft ground health management system has an open architecture, and comprises a state monitoring module, a health evaluation module, a fault prediction module, a maintenance management module and an interface module, wherein the state monitoring module comprises two basic functions of abnormality monitoring and alarming. Firstly, an abnormality monitoring library and monitoring parameter range standards are established as the basis for judging whether equipment is abnormal or not, then data from an aircraft flight parameter collector are received, whether the abnormality exists or not is judged according to equipment alarm monitoring conditions stored in a fault library, and if the abnormality occurs, various modes such as voice, log, man-machine interaction interface and the like are adopted for alarming. An implementation of the status monitoring module is shown in fig. 3.

Referring to fig. 3, a schematic diagram of state monitoring and judging according to an embodiment of the present invention is shown;

abnormality monitoring

The anomaly monitoring library is used for storing all alarm monitoring conditions, various alarm monitoring conditions must be researched, and general monitoring logic is abstracted. The invention uses XML files to store alarm conditions, and forms decision tree from decision conditions to facilitate

And processing complex anomalies and giving reasonable suggestions.

The system analyzes according to the flying parameter data acquired by the flying parameter acquisition device to obtain parameter information such as the temperature of an engine of an aircraft, the rotation speed of the engine, the temperature of an auxiliary power device, the frequency of an alternating current power supply of a generator, the pressure of a hydraulic system, the pressure of left and right lubricating oil, the oil quantity, the pressure of a cabin, the existence of hanging points, the instantaneous consumption of fuel oil, the displacement of a steering column, the displacement of a steering wheel, the deflection angle of left and right ailerons, the heating of a airspeed tube, the state of a front landing gear and a rear landing gear and the like, then judges whether abnormal parameters exist according to alarm conditions, and if the abnormal parameters are found, the alarm is implemented.

Alarm device

When the system state monitors abnormality, the alarm function is started at the first time. The alarm will be generated in a versatile manner, ensuring that the current abnormal situation is notified to the logistics support personnel. The alarm system can send out alarm signals by adopting alarm sound, simultaneously display abnormality on a man-machine interaction interface, send the abnormality information to logistic support personnel through short messages and write the abnormality information into a system log file.

Referring to FIG. 4, a schematic diagram of a health assessment model according to an embodiment of the present invention is shown:

health directly affects the safety of an aircraft during flight, and therefore the need for health assessment techniques is particularly acute. Health assessment is largely divided into two parts: health behavior modeling and health assessment algorithm design.

Modeling of health behavior:

the health behavior model is based on analysis of a dynamic system, and is established to reflect the mapping relation between input excitation and behavior output under different health modes. Based on the full analysis of the dynamic characteristics of the system, a healthy behavior model is constructed by using a system identification method.

Health assessment algorithm

And a reasonable health evaluation algorithm is designed, and the actual output of the system and the predicted output of the health behavior model are correspondingly compared and analyzed, so that the current health state of the system is evaluated.

The overall scheme of the health evaluation module is designed as follows: firstly, establishing a clear health mode table, fully considering fault modes corresponding to various influencing factors in each component, secondly, applying computer simulation to acquire response data of the system in different health modes to form a neural network training sample space, respectively training to generate health behavior models in the corresponding health modes to form a system model in a normal state, and finally, analyzing test point data vectors based on health evaluation indexes to evaluate the health state of the system.

Fault prediction is an important link in aircraft ground health management systems. The main purpose of the system is to reduce the use and guarantee cost, improve the umbrella safety, the integrity and the task success of the equipment system, and realize maintenance and autonomous guarantee based on the state.

The fault prediction technology aims at the fault prediction of equipment, and estimates the residual service life at the moment when a pre-judging system breaks down, so as to guide task planning before the equipment breaks down, and formulate maintenance strategies of the equipment.

The fault prediction function in the present invention is defined as follows:

(1) The main application object of the fault prediction technology is the first level of the parts, so that the accuracy and the stability of the fault prediction of the parts are ensured to reach a higher level.

(2) The failure prediction behavior is the whole process from early damage occurrence to complete failure. No prediction was made until no damage occurred. When early damage is detected, estimating the evolution trend of specific damage according to known monitoring data, fault models or priori knowledge, and predicting the residual service life of the target until the target is completely invalid.

(3) And (3) determining the operation condition of the equipment in a period of time in the future after the current moment, and predicting the fault mode and the evolution process of the damage.

(4) Because the process from the failure of the target equipment to the complete failure is very unstable due to a plurality of factors such as environmental factors, working condition factors, damage of the material and the like, and the failure prediction result is highly uncertain, the failure growth probability model is established by means of simulation or experimental data verification, and a sufficient aging data statistics sample is collected and used for training, verifying and adjusting the failure prediction algorithm, so that the safety, reliability and accuracy of the failure prediction are ensured to the greatest extent.

1. The fault prediction method comprises the following steps:

the invention uses data driving to combine with neural network to model the obtained state monitoring data, and establishes a mapping model between the historical record data and the predicted output. Comparing, matching and evaluating the test information output by the current system with the historical test information under the condition of the known state, judging the known state which is best matched with the current test information as the current state of the system, and predicting the damage or fault of the current equipment in normal or different degrees.

2. The content of fault prediction is as follows:

The fault prediction is to predict damage, degradation state identification and residual service life of the aircraft equipment, and the probability and time of fault occurrence are predicted from quantification, description and evolution trend of the equipment health state, so that reasonable probability form expression of the target component is given.

(1) Damage tracking

The pattern of existing or impending lesions is tracked, evaluated and forecasted. The damage degradation of the component is a slow accumulation process, and a complex nonlinear relation exists between the damage state and the measured data, so that various nonlinear theory and methods are adopted for solving the problem.

Phase space curvature: in a nonlinear system, when a certain parameter changes (no matter how large the parameter changes), a phenomenon that the phase space of the system is curved is caused. This phenomenon illustrates that the slow accumulation of lesions will be manifested in the form of small curves in the reconstructed phase space of the system, based on which the system phase space can be reconstructed using observable fast-varying parameters, and further the characteristic parameters describing the slow-varying lesion accumulation process can be extracted from the reconstructed phase space.

Reconstructing phase space: the phase space of the reconstruction system, from the information contained in the phase space, has the space dimension including not only rapid-changing vibration and slow-changing damage, but also external time-varying factors (such as external excitation, environmental conditions, operation conditions and the like). Assuming that the slow-varying impairment is monotonically changing over time, an observation data segment is taken, the data length being on a medium time scale, i.e., the degree of slow-varying impairment can be considered to be substantially constant within the data segment. Therefore, the phase space trajectory corresponding to the data segment may be considered as a slice in the dimension of the slow-varying damage, as shown in fig. 5, referring to fig. 5, which is a schematic diagram of the reconstructed phase space relationship according to an embodiment of the present invention, where each slice represents a state of the slow-varying damage of the system, and the set of damage states of all slices represents the whole damage evolution process of the system. According to the principle of occurrence of phase space curvature transformation, if the tiny curvature of the phase space track caused by the accumulated change of the damage can be quantified among each data segment in the graph, the evolution process of the slow-change damage state can be described.

In order to quantitatively describe the phase space curvature, firstly reconstructing the phase space, then establishing a local linear model, estimating a tracking function and finally constructing a tracking matrix. The tracking matrix simultaneously reflects the change process of the slow-change damage, the working condition change and other possible factors, and then the change trend of the slow-change damage is separated by using a modal decomposition method, so that the damage evolution process is finally and reasonably represented.

(2) Degradation state identification

The degradation state of the device is from a normal health state to a functional degradation in its entirety

The base loses basic function and is currently in a state at some stage in its degradation process during complete failure.

The degradation state is an abstract concept, so that the degradation state of the device belongs to an unmeasurable "parameter" and needs to be estimated by some certain or empirical correspondence, using other measurable physical quantities, which are themselves calculated from the measurable time-frequency domain signal. The identification and estimation of which is entirely dependent on measurable information with the device.

Multiple information source state recognition algorithm: the invention adopts a multi-information source state recognition algorithm to realize degradation state recognition. When the performance degradation state of the equipment is just started, the state characteristics of the equipment are generally characterized by the following characteristics: the degraded state characteristic signal is weak, so that the state characteristic presents uncertain characteristics; because the internal structural difference is different from the transmission line, a weak degradation state characteristic signal usually appears at a certain local position of the device first, so that one sensor cannot comprehensively and accurately capture the state information of the device. The degradation state can be effectively identified by fusing the information of a plurality of sensors, the sensors positioned at a plurality of key positions are constructed into an information network, the measurement information of different positions is fully acquired, and the measurement information is fused by using an information fusion technology to obtain a final identification result.

The main steps of degradation state identification are as follows:

information processing and feature extraction

And aiming at the working state information of different time periods and different parts of equipment during working, respectively adopting different sensors to collect data during the working of the equipment, then using a time domain analysis method and a time domain analysis method to extract relevant characteristic vectors from the state information of the equipment, obtaining vector spaces constructed by different state characteristic vectors, and modeling corresponding state type spaces.

Information source status identification

And constructing a nonlinear relation between the state characteristic vector space and the state type space, and training the model by adopting experimental sample data.

State attribute decision fusion

And (3) summarizing different recognition results comprehensively, and applying a fusion rule according to the basic confidence coefficient of the different recognition results to obtain the total probability distribution of the state types, thereby obtaining the final recognition result.

(3) Remaining service life

The remaining service life refers to the remaining time that can also be effectively operated to realize the given function at a certain time in the operation process. Estimation and prediction of the remaining service life are central tasks of the ground health management system, and are needed to be realized by using available equipment running state information, health monitoring information, statistical information and the like.

The prediction of the remaining service life of the device is based on early damage detection and degradation state identification, and is based on prediction features, wherein the prediction features are associated with a life curve of the device, and the remaining service life of the device is predicted by estimating the change of the features, and referring to fig. 6, a schematic diagram of the prediction of the remaining service life according to an embodiment of the present invention is shown.

And according to the change condition of a certain selected health index capable of reflecting the performance of the equipment, the whole process of predicting the residual service life of the equipment is described. The initial state is considered to be a healthy state, and is also the life starting point of the device, irrespective of the failure of the target device (component) in the initial running-in stage. When the equipment fails thoroughly and the rescue cannot be carried out by maintenance means, the service life of the equipment is finished. The time span from the initial state to the end of the life of the device is the full life of the device. The overall life of a device can be divided into two parts, a healthy state and a faulty state, wherein for most devices the healthy state is usually maintained for a considerable period of time, and accordingly the health indicator is maintained at a relatively stable level. However, long-term operation inevitably causes fatigue and wear of the relevant equipment, which is initiated in the material interior, exists in the form of extremely minute internal cracks or the like, and is not significantly represented in various observable physical quantities of the exterior. With further increases in run time, microscopic damage within the material gradually increases and begins to manifest itself in observables. In this case, the existence of a fault can be found by various fault diagnosis methods, and the fault that can be found at the earliest is regarded as an initial fault, and a fault state is considered to be entered from this point.

Corresponding to the process, the health index also gradually decreases along with the health state of the equipment until reaching the level corresponding to the complete failure. However, the occurrence of a fault does not represent a loss of function of the device, and in fact, the intended task can be efficiently accomplished even though the operation of the device may be affected to some extent before the fault is reached. In the health management technology, in order to maximize the efficiency utilization rate of the equipment, after the initial fault is found, the degradation state identification and the fault prediction are performed on the equipment, and on the basis of ensuring the safety of the equipment, the equipment is enabled to continue to operate until the functional failure occurs.

The functional failure state refers to a state that the equipment cannot normally complete a given task, and the corresponding health index level is called a failure threshold. If forced operation is continued, the device will fail completely in a short period of time. Therefore, the initial failure state is the best timing for performing maintenance of the equipment, and the balance between the safety and efficiency utilization of the equipment can be achieved. By means of maintenance or replacement of parts and other measures, the system and the equipment can recover the health state and operate in service again, and the guarantee cost of the equipment is effectively reduced.

The period from the discovery of the initial fault to the failure of the device functionality is the development of the remaining life prediction phase.

Since the fault degradation phase of the device has strong nonlinearity, the rate of decline (severity of fault) of the health indicator, which is embodied as a health indicator, gradually increases with time. The predictable stage is divided into two stages of early degradation and accelerated degradation of faults, in the early degradation stage, the degradation is gradually started from weak faults, the aggravation speed of the fault degree is relatively slow, and the health index is not greatly reduced in a relatively long time; after entering the accelerated degradation stage, the aggravation speed of the fault degree is obviously increased, and the health index is quickly reduced to the failure threshold value in a short time.

After entering the predictable stage, the residual service life of the equipment can be predicted at any time according to the needs. Because predictions need to be modeled and calculated with the aid of operational state information, health monitoring information, and statistics, etc., before the current time of the device, available observed signals must be determined. Taking the initial fault moment as a starting point, intercepting observation data from the moment to the moment for prediction as known observation data, predicting the evolution trend of the health index after the current moment, estimating when the health index is reduced to a failure threshold value, and calculating the residual service life predicted value at the moment.

In the prediction process, attention should be paid to the following key point selection:

selecting a predicted point: when the prediction is performed after entering the predictable stage, a certain length of known observation data sample is required to be used as a modeling or calculating basis, and the required data length is different for different prediction methods. Therefore, when determining the initial observation point, it is necessary to determine whether the known observation data at that time is sufficient for the corresponding metering. Ensuring device security while also reducing unnecessary over-computation.

Selection of an observation data starting point: considering the problem of knowing the length of the observation, the starting point of the observation is selected by combining with a specifically adopted prediction algorithm. The starting point of the observation data is not limited to the point of time when the initial failure is found, and the known observation data may be taken from any point of time when the observation data is recorded in the health state stage, if necessary.

Determination of failure threshold: a failure threshold value in an approximate sense is set according to the specific situation or history experience of the equipment. According to different adopted health indexes, the numerical value of the failure threshold value can also have certain influence on the modeling steps in the prediction algorithm.

The implementation process of the residual service life can be divided into feature extraction and sparse data set construction, and degradation model determination and residual service life prediction are carried out.

Feature extraction and sparse data set construction: firstly, selecting proper prediction characteristics aiming at the characteristics of equipment to be predicted, preprocessing the prediction characteristics, and carrying out noise smoothing on an observation value sequence or eliminating the influence of factors such as working conditions as far as possible to obtain a preprocessed degradation characteristic sequence. And finally, carrying out regression fitting on the degradation characteristic sequences of the known observation data, extracting a plurality of most representative data points corresponding to the non-zero basis function, and giving prediction on occurrence of a future event in a probability form.

Determining a degradation model: after the degradation characteristics are determined, a degradation model can be determined according to the degradation characteristic sequence of the historical data, the prior degradation model is determined by prediction based on a correlation vector machine method, and the most suitable degradation model is selected or the model is determined to be optimized and improved.

Residual service life prediction: and fitting is carried out on the sparse data set by adopting the degradation model, and model parameter values of the degradation model are determined. And carrying out extrapolation prediction on the degradation model on the basis to obtain estimation on the evolution trend of the prediction characteristic. An estimate of the remaining useful life of the device and its upper and lower boundaries are calculated.

Maintenance management module

The maintenance man-hour and maintenance efficiency of the aircraft directly determine the maintenance guarantee level, and most of the maintenance modes of the air force in China adopt timing maintenance at present. The timing maintenance is a traditional maintenance mode, and embodies the maintenance thought centering on accident prevention. This maintenance method uses only time as a control parameter, and cannot effectively prevent a malfunction which has no direct relation with the use time. Theory and practice show that the failure of the engine has randomness, the failure rate is always 1 constant and is not in linear relation, and unnecessary work cannot be avoided in the mode, so that waste of manpower and material resources is caused. Moreover, the timing maintenance is not strong in predictability, and excessive additional maintenance and ineffective disassembly affect the working accuracy of the aircraft or the engine, so that the effective life of the aircraft or the engine is shortened.

The on-demand maintenance is based on the fact that a large number of faults occur with a 1-step progression and do not occur instantaneously, i.e. that most faults have some forenotice signal (called latent faults) when they are about to occur. If the status monitoring technique is used to monitor these signals, it can be found that the process is continuing, and measures can be taken to prevent the occurrence of a fault or to avoid the consequences of a fault.

After the steps of state monitoring, health evaluation and fault prediction are carried out, the aircraft ground health management system can accurately give out the parts with faults and the health conditions of the parts, and analyze the maintenance scheme.

The ground logistics support personnel can timely replace or overhaul the components, so that the reliability and usability of the aircraft system are improved and improved, the maintenance load is reduced, and the comprehensive maintenance efficiency is improved.

The maintenance management module improves the efficiency of maintenance management work, realizes informatization and intellectualization of the maintenance management work, realizes the input and monitoring of the state information of the aircraft and automatically generates maintenance plan early warning. The main work of the module is as follows:

1. an efficient database system is used for storing relevant data in maintenance management work and constructing an intelligent maintenance management system for the aircraft.

2. The state parameter information of the aircraft is stored through a database, and recorded and monitored in the whole life cycle of the aircraft, so that the digital management of the aircraft information is realized.

3. Maintenance project information is stored in a database, and the system can be used for making and modifying maintenance projects, so that on-line digital management of the maintenance projects is realized.

4. The system automatically generates the early warning of the maintenance plan according to the state information of the aircraft and the maintenance project information, reduces the burden of maintenance planning personnel and improves the efficiency of maintenance management work of the aircraft.

5. The system intuitively displays maintenance task information, and can perform overall flow operation of generating, issuing, submitting and ending the task on line, so that the task execution flow is more efficient.

The maintenance management module functions in the following parts:

aircraft information management: management of aircraft information, including factory information of the aircraft and state parameter information during service, has a decisive role in development of maintenance work. The intelligent maintenance management system of the aircraft has the functions of storing, checking and timely updating the status parameter information of the aircraft. After the aircraft performs the operation task, the aircraft needs to fill in a flight record table to record various parameter information of the aircraft and update the information into a system database, and after the maintenance task is performed, the state information of the aircraft needs to be changed in time.

Maintenance project management: in the daily maintenance work, the production department monitors the time control time of the time control project, and the time information of the time control project comes from a daily flight record sheet used by the airplane in flight and maintenance

Maintenance records of the squad. And then building according to the basic information managed by the time control unit on the aircraft maintenance manual

The time-controlled project management module stands up, so that the maintenance project management module firstly establishes all maintenance projects of each aircraft according to a maintenance manual of the aircraft, and mainly comprises the following steps: the aircraft timing and inspection items, the component inspection items, the time control items, the engine timing and inspection items and the engine life items. The timing inspection item refers to a regular inspection task to be performed by an aircraft according to an inspection outline at intervals, the component inspection item refers to a task in which internal components of the aircraft need to be inspected for component states at intervals according to a set, the time control item refers to a task in which components are replaced when the time control on the aircraft reaches the service life of the time control, the engine timing inspection item refers to a task in which inspection is required at intervals in an engine, and the engine time service task refers to a task in which components in the engine need to be replaced when the components reach the service life of the time control. The several item functions are mainly to write related maintenance items according to the examination outline, and can be checked, updated, modified and deleted. The main content of the task comprises main information such as an item number, a content description, a current state and the like. In addition to management of maintenance items, it is also necessary to manage work cards associated with maintenance items, which can be created, deleted, reviewed, and edited.

Maintenance planning management the maintenance planning needs to be developed according to the state parameters of the aircraft, such as the number of flight hours, cycles or number of landing and taking-off

When the manual or the upper limit of the maintenance outline requirement is approached, early warning of a maintenance plan can be automatically generated, a worker performs tasks such as issuing on the basis of the plan early warning, finally, the related worker executes the tasks and feeds back the tasks to report, the maintenance task is finished after the tasks are finished, and the finished maintenance plan has a history record for inquiry. The maintenance plan management can also pack a plurality of maintenance tasks according to the current situation, and then execute the operation flows of issuing and executing the packing tasks, and the like, so that the maintenance work is more flexibly executed. The maintenance schedule may be viewed with different time scales for the weekly schedule and the monthly schedule, respectively. The user can query the maintenance plan appointed by the history, and can query according to different conditions. Maintenance information including completion time, replacement information, etc. should be entered after each project is completed.

And (3) data file management: some file data, such as training plans, post qualification, maintenance outline and the like, are frequently used in aircraft maintenance management work, and the intelligent aircraft maintenance management system can have the function of digitally storing the data and can be conveniently checked. In the maintenance management workflow, there are places where uploading files is required, and for these files, the system can also store and manage efficiently and reliably.

Interface module

The interface module is used for communication and information exchange between each part of the airborne system of the aircraft and the ground health management system, and is mainly completed in a bus mode. The main function is to ensure that all parts in the whole ground health management system are communicated with each other correctly, smoothly, coordinately and safely, so that the informatization and integration of the whole ground health management system are realized.

System use

Mounting and connecting

The aircraft ground health management system mainly comprises an aircraft ground parameter collector, a power interface and aircraft ground health management system software.

Under normal temperature environment, the input 27V power supply of the aircraft ground parameter collector is connected to the direct current adjustable stabilized voltage power supply, the collector unloading line aircraft is used for connecting the circular aviation plug of the front panel of the collector with the PC network interface, 27V is electrified to the collector, under normal conditions, the current is 0.45A (+ -10%), the voltage is 27V (+ -1V), the front panel indicator lamp normally shows blue color at the instant of electrification, then the front panel indicator lamp turns into green light to flash all the time, and the computer shows that the connection network is successful. And observing whether the indicator lamp, the network, the current and the voltage are normal.

After logging in, the system main interface can be accessed, referring to fig. 7, which is a schematic diagram of a management system main interface according to an embodiment of the present invention, in the system main interface, an organic library, data acquisition, status monitoring, health evaluation, fault detection, maintenance management, and comprehensive query function buttons in a main control menu on the right side can be seen, and page switching can be performed by clicking the buttons.

And after the software is successfully logged in, entering a main interface. The default hangar function interface is a system main interface, and available airplane models of the airport can be checked in the hangar interface, and the airplane models are selected through button switching.

After one model is selected, the model states of all the models in the base are displayed in a lower model list, one model is selected, and the fault condition of the aircraft in the model can be seen in a lower fault icon.

Basic parameter information such as the aircraft number of the aircraft, such as the aircraft length, aircraft altitude, span, full aircraft empty weight, maximum height, maximum speed, maximum range, carrying capacity, and the like, is displayed in the display area on the right side.

3. And (3) data acquisition: clicking the button of the control menu on the left side can switch the function interface to the data acquisition interface, and can acquire the flight parameter information of the selected aircraft. FIG. 8 is a schematic diagram of an information collection interface according to an embodiment of the invention;

in the data acquisition interface, the model aircraft name, the aircraft number, the model number and the voyage information are displayed in the middle of the model aircraft library diagram displayed at the upper left.

From the resume selection drop-down frame, resume information to be downloaded is selected, the collection button is clicked to start collection, the collection duration and the collection progress can be seen from the lower side, the data collection can be suspended by clicking the suspension button, and the unloading of the resume information is terminated by clicking the stop button.

The characteristic information curve of the secondary history, such as the height, the speed, the atmospheric temperature, etc., can be checked in the characteristic curve interface, and the characteristic value curve to be checked is selected from the characteristic value selection frame, and the characteristic value curve is displayed in the graph.

In the left switching value signal list, the switching value parameter of the current resume information can be checked in real time, the indicator light is used for displaying the state of the switching value, red indicates that the switching value is not opened, and green indicates that the switching value is opened. More information is viewed by a scroll bar drag.

The analog signal table displaying the history information in the analog display area has signal parameters such as engine beat signal, ac analog signal, frequency signal, etc., and more information is checked by dragging with a scroll bar.

And (3) state monitoring: clicking the button of the control menu on the left side can switch the function interface to the state monitoring interface, so that the state of each system of the aircraft can be monitored, and as shown in fig. 9, the state monitoring interface is a schematic diagram according to the embodiment of the invention;

under the state monitoring page, state information of a plurality of modules of the aircraft can be monitored, and after the module to be detected is selected in the module selection drop-down frame, the states of subsystem functions of the module are displayed in a list below. The current monitoring module may also be switched by clicking a left switch or a right switch button. If the subsystem is found to have an abnormal or fault condition, the state early warning indicator lamp below the subsystem turns red, and meanwhile, an early warning message is sent to maintenance personnel.

Specific detailed status information of the module is listed in the information list on the right side, and detailed information of each subsystem of the module is displayed. The information report form of the module can be generated by clicking a button, and the module can be printed and is convenient to browse. And may click a button to save the current data as a history data file. If it is desired to view the monitoring data for a certain period before, a button can be clicked to load the previously saved data into a list for review.

Health assessment: clicking a button of a control menu on the left side, switching a functional interface to a health assessment interface, and checking the health status of each system of the aircraft, wherein the health status interface is shown in fig. 10;

in the on-board data structure, the composition of the overall system of the aircraft can be clearly seen. The health interface of the system can be popped up by directly clicking the health of the system or subsystem to be checked with a mouse.

In the underlying health list, the overall health status of the aircraft system may be viewed. The health condition and evaluation parameters of each subsystem under each system can be checked.

The health evaluation state report can be generated by clicking a button, and can be printed, so that the health evaluation state report is convenient to browse. And may click a button to save the current data as a history data file. If it is desired to view the monitoring data for a certain period before, a button can be clicked to load the previously saved data into a list for review.

And (3) fault prediction: clicking the button of the control menu on the left side can switch the functional interface to the fault prediction interface, so that possible faults of each functional module of the aircraft can be predicted, as shown in fig. 11, and the fault prediction interface schematic diagram according to the embodiment of the invention is shown.

In the fault prediction interface, fault prediction and pre-warning of sub-equipment of various systems of the aircraft can be seen. In the upper left corner an aircraft selection function box can be seen, and the displayed aircraft module is switched by clicking a button with a mouse. Basic information such as the total number of current devices, the total number of sensors, the total number of faults, downtime, utilization rate and the like can be checked and obtained in the right device information, and the displayed basic information is different according to the change of the selected devices.

In the equipment fault early warning list, the information such as the name, the equipment state, the fault occurrence probability, the estimated occurrence time, the early warning state, the last maintenance time and the like of each equipment of the current system are listed in detail, and a user can perform corresponding maintenance work according to prompts given in the fault early warning information.

The state curves of all the sensors of the system can be seen in the sensor list, and the display and the hiding of the sensor curves can be controlled by checking the boxes.

The running condition and the fault occurrence probability of each device of the system can be seen in the state of the device, and the device which may possibly have faults can be maintained in advance according to the results given by the chart.

Maintenance management: clicking the button of the control menu on the left side can switch the functional interface to the maintenance management interface, so that the maintenance conditions of all the systems of the aircraft can be recorded and checked, as shown in fig. 12, and the maintenance interface schematic diagram is provided according to the embodiment of the invention.

When the aircraft system generates maintenance requirements, ground logistics personnel should fill in a maintenance list first, record information such as time generated by faults of the aircraft system, maintenance reasons, maintenance starting time, consumed man-hours and the like, register and store the information, and can check historical maintenance information of the aircraft in a maintenance information list. The service ticket may click a print button to generate a preview and print.

The fitting information shows basic information for the replacement or repair of the fitting of the aircraft system. When each system needs to be replaced or newly added with an accessory, the functions of the accessory for being added and replaced are used, and if a certain accessory is to be maintained, a maintenance accessory button is clicked to enter a maintenance accessory interface for operation. If a certain accessory is to be removed, a delete accessory button is clicked to delete the accessory, and detailed accessory use, replacement and maintenance information can be seen in an accessory information list below.

Comprehensive query: clicking the button of the control menu on the left side can switch the functional interface to the comprehensive query interface, so that the historical data of each system of the aircraft can be queried, as shown in fig. 13, which is a schematic diagram of the comprehensive query interface according to the embodiment of the invention.

Detailed historical data such as health evaluation data, fault prediction data, maintenance management data and the like of each system can be checked in the comprehensive query page function. The method can accurately search according to the input time period, the search result is displayed in a list below, and the query result can be generated into a report form for convenient browsing.

It should be appreciated that embodiments of the invention may be implemented or realized by computer hardware, a combination of hardware and software, or by computer instructions stored in a non-transitory computer readable memory. The methods may be implemented in a computer program using standard programming techniques, including a non-transitory computer readable storage medium configured with a computer program, where the storage medium so configured causes a computer to operate in a specific and predefined manner, in accordance with the methods and drawings described in the specific embodiments. Each program may be implemented in a high level procedural or object oriented programming language to communicate with a computer system. However, the program(s) can be implemented in assembly or machine language, if desired. In any case, the language may be a compiled or interpreted language. Furthermore, the program can be run on a programmed application specific integrated circuit for this purpose.

Furthermore, the operations of the processes described herein may be performed in any suitable order unless otherwise indicated herein or otherwise clearly contradicted by context. The processes (or variations and/or combinations thereof) described herein may be performed under control of one or more computer systems configured with executable instructions, and may be implemented as code (e.g., executable instructions, one or more computer programs, or one or more applications), by hardware, or combinations thereof, collectively executing on one or more processors. The computer program includes a plurality of instructions executable by one or more processors.

Further, the method may be implemented in any type of computing platform operatively connected to a suitable computing platform, including, but not limited to, a personal computer, mini-computer, mainframe, workstation, network or distributed computing environment, separate or integrated computer platform, or in communication with a charged particle tool or other imaging device, and so forth. Aspects of the invention may be implemented in machine-readable code stored on a non-transitory storage medium or device, whether removable or integrated into a computing platform, such as a hard disk, optical read and/or write storage medium, RAM, ROM, etc., such that it is readable by a programmable computer, which when read by a computer, is operable to configure and operate the computer to perform the processes described herein. Further, the machine readable code, or portions thereof, may be transmitted over a wired or wireless network. When such media includes instructions or programs that, in conjunction with a microprocessor or other data processor, implement the steps described above, the invention described herein includes these and other different types of non-transitory computer-readable storage media. The invention also includes the computer itself when programmed according to the methods and techniques of the present invention.

The computer program can be applied to the input data to perform the functions described herein, thereby converting the input data to generate output data that is stored to the non-volatile memory. The output information may also be applied to one or more output devices such as a display. In a preferred embodiment of the invention, the transformed data represents physical and tangible objects, including specific visual depictions of physical and tangible objects produced on a display.

The present invention is not limited to the above embodiments, but can be modified, equivalent, improved, etc. by the same means to achieve the technical effects of the present invention, which are included in the spirit and principle of the present invention. Various modifications and variations are possible in the technical solution and/or in the embodiments within the scope of the invention.

Claims

1. A method of aircraft health management comprising the steps of:

a maintenance management step, namely establishing a database for storing and managing historical data and state information of the aircraft, and carrying out analysis and research on the corresponding aircraft based on the output results of the three steps to generate a maintenance scheme;

wherein the damage determination includes:

collecting observation data of an aircraft in a specified time period, carrying out phase space reconstruction on the observation data and establishing a local linear model;

Estimating a tracking function according to the local linear model and constructing a tracking matrix, wherein the tracking matrix comprises tracking slow-change damage and working condition change;

separating the change trend of the slow-change damage from the tracking matrix by using a modal decomposition method to obtain a damage evolution process;

wherein the degradation state identification comprises:

an information acquisition step of acquiring working state information of each time period for each component of the aircraft based on a plurality of kinds of sensors;

an information processing step, namely extracting relevant characteristic vectors from the working state information by using a time domain analysis method and a time domain analysis method to obtain vector spaces constructed by different state characteristic vectors, and modeling corresponding state type spaces;

an information identification step, namely constructing a nonlinear relation between a state feature vector space and a state type space, and training a model by adopting experimental sample data to obtain an information source state identification result;

a decision fusion step, namely comprehensively summarizing the state recognition results of the information sources, obtaining total probability distribution of the state types based on fusion rules according to the basic confidence degrees of different recognition results, and further obtaining a final recognition result;

Wherein the remaining life prediction comprises:

acquiring information of corresponding parts of the aircraft with fault characteristics, and acquiring observation data of the corresponding parts in a specified time period;

selecting corresponding prediction features according to the components with fault features, preprocessing the prediction features, and performing noise smoothing on the observed data to obtain a preprocessed degradation feature sequence;

performing regression fitting on the degradation characteristic sequence, and extracting a plurality of set data points corresponding to a non-zero basis function;

establishing a degradation model according to the feature sequence, predicting and determining a priori degradation model based on a correlation vector machine method, and selecting the most suitable degradation model or determining to optimize and improve the model;

performing congruent fitting on the observed data according to the degradation model to determine a model parameter value;

and carrying out extrapolation prediction on the degradation model according to the parameter value to obtain estimation on the evolution trend of the prediction characteristic, wherein the estimation comprises a range estimation value of the residual service life of the corresponding part.

2. The aircraft health management method according to claim 1, wherein the establishing an anomaly monitoring library specifically comprises:

And establishing alarm monitoring conditions, and carrying out logic processing to obtain a decision tree based on the judgment conditions.

3. The method of aircraft health management according to claim 1, wherein the health assessment step further comprises:

setting corresponding fault modes according to mutual influence factors of all components of the flight control system, and establishing a health mode table;

simulating each health mode one by one, and obtaining response data under the corresponding health mode to form a neural network training sample space;

training the neural network training sample space one by one to generate a health behavior model corresponding to the health mode;

sorting the plurality of healthy behavior models to obtain a normal system model;

and evaluating the data vector of the appointed test point of the flight control system based on a health evaluation algorithm, and carrying out comparison analysis by combining the result output by the normal system model to obtain the current flight control system health state evaluation result.

4. The aircraft health management method according to claim 1, wherein the fault prediction step further comprises:

acquiring historical record data and corresponding state conditions of the aircraft, and modeling based on a neural network to acquire a mapping model between the historical record data and a predicted output state;

And performing fault test on the aircraft, comparing, matching and evaluating the historical record data based on the mapping model according to the test information and the historical record data, setting the known state corresponding to the historical record data with the highest matching degree as the current state of the aircraft, and performing fault prediction according to the current state of the aircraft.

5. An aircraft health management system, comprising:

the state monitoring module is used for establishing an anomaly monitoring library and setting a monitoring parameter range, and judging whether anomaly occurs or not by collecting flight parameters of the aircraft;

the health evaluation module is used for establishing a health behavior model and a health evaluation algorithm, and comparing the real-time output parameters of the flight control system with the results output by the health behavior model based on the health evaluation algorithm to obtain the current flight control system health state evaluation result;

the fault prediction module is used for acquiring the information of the aircraft equipment, judging the damage of the aircraft, identifying the degradation state and predicting the residual service life of the aircraft according to the information of the aircraft equipment, and obtaining the probability and time of occurrence of the predicted fault;

the maintenance management module is used for establishing a database for storing and managing historical data and state information of the aircraft, analyzing and researching the corresponding aircraft based on the output results of the three modules, and generating a maintenance scheme;

Wherein the damage determination includes:

wherein the degradation state identification comprises:

Wherein the remaining life prediction comprises:

6. The aircraft health management system according to claim 5, wherein the fault prediction module further comprises:

the acquisition unit is used for acquiring the historical record data of the aircraft and the corresponding state conditions;

The modeling unit is used for modeling based on the neural network according to the information acquired by the acquisition unit, and acquiring a mapping model between the historical record data and the predicted output state;

the test unit is used for performing fault test on the aircraft to obtain test information;

and the evaluation unit is used for comparing, matching and evaluating the historical record data based on the mapping model according to the test information and the historical record data, setting the known state corresponding to the historical record data with the highest matching degree as the current state of the aircraft, and carrying out fault prediction according to the current state of the aircraft.