A health status assessment method and system for a rescue helicopter

Abstract

The application provides a health state evaluation method and system of a rescue helicopter, and relates to the technical field of aeronautical engineering.The method comprises the following steps: step 1, an integrated data aggregation platform is established, flight operation parameters, maintenance archives, abnormal event records and surrounding environment monitoring indexes are synchronously collected, key dimensions of rotor devices, power transmission units, propulsion devices and sea climate conditions are comprehensively covered, and a comprehensive original data set is formed; and step 2, based on the comprehensive original data set, an adaptive parameter adjustment mechanism is used to comprehensively calculate cumulative operation time length indexes, periodic component residual efficiency values, core health scores, abnormal repeated occurrence probabilities, typical problem correlation strengths, sudden event severity and marine climate erosion elements, and generate comprehensive state indexes.The application can precisely evaluate the health state of a rescue helicopter, reduce flight risks and improve rescue decision-making and execution efficiency.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of aeronautics, in particular to a health state evaluation method and system for a rescue helicopter. BACKGROUND

[0002] Traditional technologies mostly do not fully consider the influence of the special environment of high salt mist and high humidity in the ocean on the helicopter, and it is difficult to accurately evaluate the damage of corrosion environmental factors to the health state of the helicopter. For example, when the rescue flight team performs a rescue mission at sea, the S-76C type helicopter is in the marine environment for a long time. The traditional evaluation technology cannot accurately quantify the problems such as corrosion of the fuselage skin and corrosion of the electronic circuit board because it does not take into account the corrosion factors in the marine environment, resulting in multiple unplanned stops due to corrosion and affecting the efficiency of rescue mission execution.

[0003] In addition, the traditional technology is relatively single in data fusion, and it is difficult to comprehensively process multi-dimensional data such as flight data, maintenance records, and fault information, and lacks a dynamic weight distribution mechanism, which cannot accurately evaluate the actual operating conditions of the helicopter, resulting in inaccurate evaluation of the health state of the helicopter and insufficient or excessive maintenance. SUMMARY

[0004] The technical problem to be solved by the present application is to provide a health state evaluation method and system for a rescue helicopter, which can accurately diagnose the health state of the helicopter and provide reliable basis for timely maintenance and safe rescue, and improve the reliability and success rate of rescue operations.

[0005] To solve the above technical problems, the technical solution of the present application is as follows:

[0006] In a first aspect, a health state evaluation method for a rescue helicopter, the method comprising:

[0007] Step 1, an integrated data aggregation platform is established to synchronously collect flight operation parameters, maintenance records, abnormal event records and surrounding environment monitoring indicators, comprehensively covering key dimensions of rotor devices, power transmission units, propulsion devices and sea climate conditions, forming a comprehensive raw data set;

[0008] Step 2, based on the comprehensive raw data set, an adaptive parameter adjustment mechanism is used to comprehensively calculate the cumulative running time index, the residual efficiency value of the periodic components, the core health score, the probability of repeated occurrence of abnormalities, the correlation strength of typical problems, the severity of sudden events and the marine climate erosion elements, and generate a comprehensive state index;

[0009] Step 3, based on the comprehensive state index, the number of surface oxidation points, the salt corrosion exposure period and the material stress microcrack characteristic quantity of the helicopter are standardized evaluated, and a dynamic quantitative correspondence between environmental erosion characteristics and equipment working conditions is established;

[0010] Step 4, using the quantitative relationship between environmental erosion and equipment working conditions, and combining the comprehensive state index of multiple machine groups, the diversified data of different machine groups are fused through a distributed collaborative learning architecture to construct a three-dimensional digital image of the helicopter, so as to deduce the performance change trend of the parts in the strong salt erosion and high moisture environment;

[0011] Step 5, based on the performance change trend of the parts and the comprehensive state index, the state classification reference value is adjusted in real time to obtain a complete evaluation report of visual warning signs, repair emergency degree and operation constraint suggestions.

[0012] Further, based on the comprehensive original data set, the cumulative running time index, the residual efficiency value of the periodic parts, the core health score, the abnormal repeated occurrence probability, the typical problem correlation strength, the severity of the sudden event and the marine climate erosion elements are calculated comprehensively by using an adaptive parameter adjustment mechanism to generate a comprehensive state index, including:

[0013] The vibration spectrum in the flight operation parameters and the torque fluctuation characteristics of the power transmission unit are analyzed in time and frequency domain, and the component replacement cycle and maintenance record time sequence data in the maintenance archives are combined to construct a component degradation feature set;

[0014] Based on the component degradation feature set, the fault mode space-time distribution characteristics in the abnormal event record are combined, and the correlation rule mining algorithm is used to calculate the typical problem correlation strength index to realize the correlation relationship between faults and construct a fault propagation path network;

[0015] Based on the fault correlation relationship framework, the marine climate erosion elements of real-time salt mist concentration and relative humidity are collected, and a factor library representing environmental stress intensity is constructed combined with corrosion damage data. At the same time, the climate erosion effects in different regions are spatially weighted and fused to form a comprehensive evaluation value of environmental erosion impact;

[0016] Based on the component degradation feature set, the fault propagation path network and the environmental erosion impact evaluation value, the core health score and the severity of the sudden event are spatio-temporally aligned, and a comprehensive state index with time sequence characteristics is generated through a multi-dimensional calculation framework, including mechanical health degree, environmental erosion degree and operation stability.

[0017] Further, based on the fault correlation relationship framework, the marine climate erosion elements of real-time salt mist concentration and relative humidity are collected, and a factor library representing environmental stress intensity is constructed combined with corrosion damage data. At the same time, the climate erosion effects in different regions are spatially weighted and fused to form a comprehensive evaluation value of environmental erosion impact, including:

[0018] According to the spatial distribution of key fault nodes in the fault propagation path network, the coordinates of the prominent erosion risk area are obtained;

[0019] Based on the highlighted erosion risk area coordinate set, real-time collection of salt fog concentration, relative humidity data and corrosion damage data in the corresponding area is performed;

[0020] The salt fog concentration, relative humidity data and corrosion damage data are input into a dynamic coupling algorithm to generate an environmental corrosion rate coefficient and a material degradation threshold, and a partitioned environmental stress intensity factor library is constructed;

[0021] According to the highlighted erosion risk area coordinate set, the rotor exposure area, the power cabin sealed area and the fuselage surface area are divided, and the spatial weighting coefficient is calculated in combination with the material corrosion resistance grade and exposure time of each area;

[0022] The partitioned environmental stress intensity factor library and the spatial weighting coefficient are calculated to obtain a comprehensive evaluation value of environmental erosion impact.

[0023] Further, based on the comprehensive state index, the number of oxidation points on the helicopter surface, the salt corrosion exposure period and the material stress microcrack characteristic quantity are standardized and evaluated to establish a dynamic quantitative correspondence between the environmental erosion characteristics and the equipment working conditions, including:

[0024] The sea salinity, air humidity, flight trajectory altitude and crew maintenance record data are separated from the comprehensive state index, and a corrosion environment feature vector including time stamp, geographic coordinates and working condition parameters is constructed by combining the corrosion case library, so as to associate the flight parameters with the environmental parameters in space and time, and obtain a multi-dimensional environmental stress factor matrix;

[0025] Based on the multi-dimensional environmental stress factor matrix, a three-dimensional point cloud is used to extract features from the helicopter surface structure data, and the oxidation point distribution and the salt corrosion exposure period are mapped to the fuselage structure coordinate system, and the time sequence change of the material stress microcrack characteristic quantity is combined to generate the distribution characteristics of the corrosion sensitive area through clustering analysis;

[0026] Based on the corrosion distribution characteristics, the correlation between the corrosion rate and the mechanical fatigue is analyzed, the material strength attenuation trend under different salt corrosion grades is fitted, and the key component corrosion sensitivity grading table including the residual life prediction value and the degradation rate confidence interval is generated in combination with the severity data of the sudden event;

[0027] Based on the key component corrosion sensitivity grading data, the corrosion damage information of multiple crews is integrated to construct a dynamic correlation between the environmental erosion intensity and the equipment health state, and a real-time updated corrosion evaluation system is formed.

[0028] Further, based on the corrosion distribution characteristics, the correlation between the corrosion rate and the mechanical fatigue is analyzed, the material strength attenuation trend under different salt corrosion grades is fitted, and the key component corrosion sensitivity grading table including the residual life prediction value and the degradation rate confidence interval is generated in combination with the severity data of the sudden event, including:

[0029] Based on the prominent corrosion risk area data in the corrosion distribution characteristics, the corrosion rate quantitative value and the mechanical fatigue strength index of the key components are extracted;

[0030] Using the corrosion rate quantitative value and the mechanical fatigue strength index, the dynamic correlation coefficient of the two is calculated through the correlation analysis algorithm, and the quantitative relationship between the environmental stress and the mechanical performance degradation is established;

[0031] Based on the quantitative relationship, and combined with the corrosion case library under different salt corrosion grades, the decay law reflecting the change of material strength with running time is fitted;

[0032] Using the severity data of the incident, the material strength decay law is dynamically calibrated, and based on the dynamically calibrated material strength decay law, the residual life prediction value and the degradation rate confidence interval of the key components are calculated;

[0033] Based on the residual life prediction value and the degradation rate confidence interval, the corrosion sensitivity of the key components is graded, and a grading table is generated.

[0034] Further, using the quantitative relationship between environmental erosion and equipment working conditions, and combining the comprehensive state indicators of multiple machine groups, the diversified data of different machine groups are fused through a distributed collaborative learning architecture to construct a three-dimensional digital image of the helicopter, so as to deduce the performance change trend of the parts in the strong salt corrosion and high moisture environment, including:

[0035] A multi-machine group data fusion platform based on digital twinning is established to collect real-time salt fog concentration, humidity fluctuation value and equipment running load parameters of rescue helicopters in different regions, and through a federal learning mechanism, the multi-source heterogeneous data is collaboratively encrypted and trained to generate a fused global environment-working condition feature matrix;

[0036] The global environment-working condition feature matrix is input into the environment degradation simulation processing unit, and the material corrosion resistance coefficient and structural stress distribution data of the parts in the three-dimensional digital image are simultaneously called, based on the salt corrosion exposure period and the humidity fluctuation range, the surface corrosion depth growth process and the internal crack extension trajectory of the parts are simulated, and a degradation data set including mechanical performance decay value is obtained;

[0037] The degradation data set is decoupled by the transfer learning processing unit to extract the corrosion rate correlation function of the high salt corrosion area machine group, which is adapted to the three-dimensional digital image of the low salt corrosion area machine group to generate the performance change trend of the parts in the low salt corrosion area.

[0038] Further, based on the performance change trend of the parts and the comprehensive state indicators, the state grading reference value is adjusted in real time to obtain a complete evaluation report of visual warning signs, repair emergency level and operation constraint suggestions, including:

[0039] The component performance change trend and the real-time collected flight mission urgency parameters are input into the dynamic classification processing unit. The predicted value of the component's remaining life and the environmental corrosion rate are weighted and calculated through a deep reinforcement learning mechanism to dynamically generate a health status classification threshold parameter set.

[0040] The health status classification threshold parameter set is input into the status identification processing unit, and the spatial positioning data of the parts in the three-dimensional digital image are called at the same time to perform the classification determination operation and generate an instruction data set bound to the three-dimensional space coordinates;

[0041] The instruction data set bound to three-dimensional spatial coordinates is input into the decision generation processing unit, and combined with the external input spare parts supply cycle, alternate airport location information and remaining flight time window to obtain a structured evaluation report, including visual identification mapping relationship, maintenance time series table and airspace restriction parameters.

[0042] Furthermore, the component performance change trend and the real-time collected flight mission urgency parameters are input into the dynamic classification processing unit. Through the deep reinforcement learning mechanism, the predicted value of the component's remaining life and the environmental corrosion rate are weighted and calculated to dynamically generate a health status classification threshold parameter set, including:

[0043] Based on the component failure time points in the fault records and the emergency maintenance frequency in the maintenance logs, the initial correlation weight matrix between the environmental corrosion rate and the remaining life prediction value is extracted;

[0044] The initial correlation weight matrix and the real-time flight mission urgency parameter are input into the dynamic calibration mechanism, and the weight matrix is ​​processed by the mission urgency coefficient to obtain the real-time weight coefficient set;

[0045] The real-time weight coefficient set, the predicted value of the remaining life of the components and the environmental erosion rate are input into the deep reinforcement learning mechanism to construct a multi-dimensional state feature vector in the state space and generate operation instructions in the action space;

[0046] The action space operation instructions are input into the pre-trained reward function mechanism, and a health status classification threshold parameter set including the first warning threshold and the second grounding threshold is generated based on the component failure risk cost and sudden maintenance loss.

[0047] In a second aspect, a health status assessment system for a rescue helicopter includes:

[0048] The data aggregation module is used to synchronously collect flight operation parameters, maintenance files, abnormal event records and surrounding environment monitoring indicators to form a comprehensive original data set;

[0049] The status assessment module is used to generate comprehensive status indicators in multiple dimensions, including cumulative operating time indicators, component remaining performance values, and core health scores, based on the comprehensive original data set;

[0050] An environmental erosion module is configured to normalize the number of oxidation points, the salt erosion exposure period and the material stress microcrack characteristics of the helicopter surface based on the comprehensive state index, and establish a quantitative relationship between the environmental erosion and the equipment working condition;

[0051] A digital image module is configured to use the quantitative relationship between the environmental erosion and the equipment working condition, combine the multi-helicopter group state index, construct a three-dimensional digital image of the helicopter to deduce the performance change trend;

[0052] An evaluation report module is configured to adjust the grading reference value in real time based on the performance change trend of the parts and the comprehensive state index, and generate a complete evaluation report including warning marks, maintenance suggestions and operation constraints.

[0053] In a third aspect, a computing device includes:

[0054] One or more processors;

[0055] A storage device for storing one or more programs, when the one or more programs are executed by the one or more processors, so that the one or more processors implement the method.

[0056] In a fourth aspect, a computer readable storage medium stores a program, which is executed by a processor to implement the method.

[0057] The above-mentioned scheme of the present application at least has the following beneficial effects:

[0058] Through the method, the health state information of the rescue helicopter can be quickly and accurately obtained, and in an emergency rescue scene, every second is crucial, so that the rescue commander can quickly understand the operation status of each key component of the helicopter, immediately judge whether the helicopter is suitable for performing a task, avoid delaying the rescue opportunity due to blindly deploying a faulty helicopter, improve the efficiency of rescue decision-making, and gain valuable time for saving lives; based on comprehensive health state evaluation, problems that may occur during the execution of the task can be predicted in advance, so that the rescue route and task arrangement can be reasonably planned, the performance indicators and component states of the helicopter can be monitored in real time, potential faults can be predicted through data analysis, which enables maintenance personnel to take targeted preventive measures before the fault occurs, perform advance maintenance or replace components, avoid fault expansion, reduce the number of unplanned maintenance, reduce maintenance cost, and prolong the service life of the helicopter.

[0059] Through comprehensive evaluation and data analysis of the health state of the rescue helicopter, the failure frequency and maintenance requirements of different components can be understood, which helps the maintenance department to reasonably plan the spare parts inventory, ensures the sufficient supply of key spare parts, avoids waste of resources caused by inventory accumulation, and further arranges the work tasks of maintenance personnel according to the health state of the helicopter, and improves the utilization efficiency of human resources; The rescue helicopter faces complex and changeable environment and various potential dangers during task execution, and the health state is directly related to flight safety. Timely discovery of safety hazards existing in the helicopter, such as engine failure and structural damage, enables the crew to take corresponding measures before takeoff to avoid flying with faults, thereby reducing the probability of flight accidents and protecting the lives and safety of rescue personnel and passengers. BRIEF DESCRIPTION OF DRAWINGS

[0060] Figure 1 is a flowchart of a health state evaluation method of a rescue helicopter provided by an embodiment of the present application.

[0061] Figure 2 is a schematic diagram of a health state evaluation system of a rescue helicopter provided by an embodiment of the present application. DETAILED DESCRIPTION

[0062] Exemplary embodiments of the present disclosure will be described in greater detail below with reference to the accompanying drawings. Although exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments described herein. On the contrary, these embodiments are provided so that the present disclosure can be more thoroughly understood and the scope of the present disclosure can be accurately conveyed to those skilled in the art.

[0063] As shown in Figure 1 , an embodiment of the present application proposes a health state evaluation method of a rescue helicopter, which comprises the following steps:

[0064] Step 1, an integrated data aggregation platform is established to synchronously collect flight operation parameters, maintenance records, abnormal event records and surrounding environment monitoring indicators, comprehensively cover key dimensions of rotor devices, power transmission units, propulsion devices and sea climate conditions, and form a comprehensive raw data set;

[0065] Step 2, based on the comprehensive raw data set, an adaptive parameter adjustment mechanism is used to comprehensively calculate the cumulative running time index, the residual efficiency value of the periodic component, the core health score, the abnormal repeated occurrence probability, the typical problem correlation strength, the severity of the sudden event and the marine climate erosion elements, and generate a comprehensive state index;

[0066] Step 3, based on the comprehensive state index, the number of helicopter surface oxidation points, salt corrosion exposure period and material stress microcrack characteristic quantity are standardized evaluated, and the dynamic quantitative corresponding relationship between environmental erosion characteristics and equipment working conditions is established;

[0067] Step 4, using the quantitative relationship between environmental erosion and equipment working conditions, and combining the comprehensive state index of multiple machine groups, the diversified data of different machine groups are fused through the distributed collaborative learning architecture, and the three-dimensional digital image of the helicopter is constructed to deduce the performance change trend of the parts in the strong salt corrosion and high moisture environment;

[0068] Step 5, based on the performance change trend of the parts and the comprehensive state index, the state classification reference value is adjusted in real time to obtain a complete evaluation report including visual warning mark, repair emergency degree and operation constraint suggestion.

[0069] In the embodiment of the application, by establishing an integrated data aggregation platform, the flight operation parameters, maintenance archives, abnormal event records and surrounding environment monitoring indexes are synchronously collected, the key dimensions of the rotor device and the power transmission unit are fully covered, the comprehensive raw data set is formed, the data island is eliminated, the data integrity and consistency are ensured, and the helicopter running state and environmental influence factors are fully mastered; based on the comprehensive raw data set, the adaptive parameter adjustment mechanism is used to generate the comprehensive state index by comprehensively calculating multiple indexes, the influence factors in the helicopter running process are fully considered, the dynamic and accurate evaluation of the helicopter health state is realized, the potential problems are found in time, and the scientificity and reliability of the evaluation are improved; the number of helicopter surface oxidation points is standardized evaluated, the dynamic quantitative corresponding relationship between environmental erosion characteristics and equipment working conditions is established, the specific influence degree and law of the marine environmental factors on the helicopter equipment are determined, the basis for formulating targeted protection and maintenance measures is provided, and the damage of the helicopter caused by the severe environment such as high corrosion on the sea is effectively coped with.

[0070] Using the quantitative relationship and combining the data of multiple machine groups, the three-dimensional digital image is constructed through the distributed collaborative learning architecture, the performance change trend of the parts in the strong salt corrosion and high moisture environment is deduced, the performance degradation of the parts can be predicted in advance, the reasonable maintenance period and replacement strategy are provided, and the forward-looking and effectiveness of the maintenance are improved; based on the performance change trend of the parts and the comprehensive state index, the state classification reference value is adjusted in real time, a complete evaluation report including visual warning mark, repair emergency degree and operation constraint suggestion is generated, the intuitive and clear decision basis for the operation and maintenance personnel is provided, the dynamic management of the helicopter state is realized, the maintenance decision is optimized, and the safe and efficient operation of the helicopter is ensured.

[0071] In a preferred embodiment of the present application, step 1, the integrated data aggregation platform is established, and flight operation parameters, maintenance records, abnormal event records and surrounding environment monitoring indicators are synchronously collected, comprehensively covering key dimensions of rotor devices, power transmission units, propulsion devices and sea climate conditions, forming a comprehensive raw data set, which can include:

[0072] In the embodiment of the present application, sensors (such as vibration sensors, temperature sensors, pressure sensors, etc.) are deployed at key positions of the core components of the rescue helicopter (rotor devices, power transmission units, propulsion devices) for real-time collection of physical parameters (such as rotation speed, temperature, vibration frequency, oil pressure, etc.); synchronous planning of sea environment monitoring equipment (such as weather stations, sea wave sensors, humidity sensors) is deployed around the helicopter take-off and landing base and the sea area where the helicopter often performs tasks for collecting climate data such as wind speed, wind direction, temperature, humidity, and sea wave height; a data collection terminal is installed inside the helicopter body, which is connected with the sensors through wired or wireless communication modules (such as CAN bus, WiFi, 4G / 5G) to receive flight operation parameters in real time; a server cluster (including database server, file storage server) is deployed on the ground to receive and store data transmitted from the helicopter and environment monitoring equipment; at the same time, a local area network is built to connect terminal devices of maintenance workshops, command centers and other departments to the network to upload maintenance records and abnormal event record data.

[0073] After the helicopter takes off, the data collection terminal reads the sensor data of each component in real time at a preset frequency (such as 100ms / time), for example:

[0074] Rotor device: rotation speed of rotor is collected by angular velocity sensor, and blade swing amplitude and abnormal vibration signal are monitored by vibration sensor;

[0075] Power transmission unit: temperature sensor records gearbox oil temperature, and pressure sensor monitors hydraulic system pressure change;

[0076] Propulsion device (engine): engine oil temperature, oil pressure, fuel flow and other parameters are collected;

[0077] The collection terminal denoises and filters (such as eliminating transient interference signals) the raw data, and adds time stamp (accurate to millisecond level) to form standardized real-time data stream, which is transmitted to the ground server in real time through the onboard communication module (such as 5G).

[0078] After each maintenance, the maintenance personnel manually enter the maintenance records (such as maintenance time, part replacement, maintenance process, and model and batch number of the replaced part) through the terminal or export data from the maintenance management system (such as the enterprise resource planning (ERP) system) and batch import the platform. The platform automatically associates the maintenance records with the corresponding helicopter number and flight hours to form a maintenance archive. During flight, if the helicopter's on-board fault diagnosis system (such as the engine warning system) triggers an abnormal alarm, the data acquisition terminal automatically captures the alarm signal, records the time, component location, alarm code, and current flight parameters (such as altitude and speed), and transmits them in real time to the ground platform. The ground command center personnel can manually supplement the abnormal event description (such as visual observation of oil leakage or abnormal noise) through the platform interface to form a complete abnormal event log.

[0079] The weather stations deployed around the sea area collect wind speed, wind direction, atmospheric temperature, humidity, and pressure at a fixed frequency (such as 1 minute per time). The wave sensor monitors wave height and direction through radar or ultrasonic technology, and the data is sent to the ground server through wireless transmission (such as LoRa, NB-IoT). The platform automatically associates the geographical location (such as latitude and longitude) and time of data collection to form an environmental data time series. The ground server receives the three types of data through different interfaces (such as API interface and file transfer protocol (FTP)).

[0080] Flight operation parameters: Real-time data tables are established according to helicopter number and component type and stored in a time series database (such as InfluxDB) for quick query of dynamic data.

[0081] Maintenance archives and abnormal events: Stored in a relational database (such as MySQL) in a structured data format (such as CSV and Excel) and indexed by key fields (such as helicopter number and time).

[0082] Environmental data: Stored by geographical location and time dimension for flight data correlation analysis.

[0083] The platform automatically associates flight operation parameters, environmental data (such as real-time wind speed in the flight area) of the corresponding time period, and the historical maintenance archives (such as the last engine maintenance time) of the helicopter using "helicopter unique identifier (such as tail number) + timestamp" as the primary key. When an engine T5 temperature anomaly occurs in a helicopter during flight, the platform can automatically retrieve the flight parameters (speed, load), environmental data (ambient temperature, wind speed), and engine maintenance records for the past six months, forming a multi-dimensional data combination. The platform packages the integrated data by time period (such as daily and weekly) to form a comprehensive raw data set indexed by helicopter number, containing:

[0084] Flight parameter sequence (including time and component parameters);

[0085] Maintenance record list (including time and maintenance content);

[0086] Abnormal event log (including time and abnormal type);

[0087] Environmental data series (including time and climate parameters).

[0088] Data management personnel ensure that there are no missing values ​​or logical contradictions in the original data set (such as temperature parameters are within a reasonable physical range) through manual sampling and automatic system verification (such as checking data integrity and timestamp continuity).

[0089] Specific implementation of key dimension coverage:

[0090] Rotor device: Vibration sensors and speed sensors cover parameters such as blade status, bearing wear, and dynamic balance;

[0091] Power transmission unit: covers the operating status of the gearbox, drive shaft, and hydraulic system through temperature, pressure, and oil sensors;

[0092] Propulsion device: Covering power output efficiency and potential fault hazards through full engine parameter collection (lubricating oil temperature and pressure);

[0093] Sea climate conditions: Weather stations and wave sensors cover environmental factors that have a direct impact on flight safety, such as wind speed, wave height, and temperature.

[0094] Through full-process operations, the integrated convergence of multi-source data is ultimately achieved, forming a comprehensive original data set covering the core components and operating environment of the helicopter.

[0095] In a preferred embodiment of the present invention, the above step 2, based on the comprehensive original data set, uses the adaptive parameter adjustment mechanism to comprehensively calculate the cumulative operating time index, the periodic component remaining efficiency value, the core health score, the probability of abnormal recurrence, the typical problem correlation strength, the severity of the emergency, and the marine climate erosion factor to generate a comprehensive status index, which may include:

[0096] Step 220 , performing time-frequency domain analysis on the vibration spectrum and torque fluctuation characteristics of the power transmission unit in the flight operation parameters, and combining the component replacement cycle and maintenance record time series data in the maintenance file to construct a component degradation feature set;

[0097] Step 221 , based on the component degradation feature set and in combination with the temporal and spatial distribution characteristics of the fault modes in the abnormal event records, an association rule mining algorithm is used to calculate the typical problem association strength index, thereby realizing the association relationship between faults and constructing a fault propagation path network;

[0098] Step 222, based on the fault correlation relationship framework, collect real-time marine climate corrosion elements such as salt fog concentration and relative humidity, and combine corrosion damage data to build a factor library representing environmental stress intensity; at the same time, the spatial weighted fusion of the climate erosion effect of different regions is carried out to form a comprehensive evaluation value of environmental erosion effect, which specifically includes: according to the spatial distribution of key fault nodes in the fault propagation path network, the coordinates of the prominent erosion risk area are obtained; based on the prominent erosion risk area coordinate set, the salt fog concentration, relative humidity data and corrosion damage data of the corresponding region are collected in real time; the salt fog concentration, relative humidity data and corrosion damage data are input into the dynamic coupling algorithm to generate the environmental corrosion rate coefficient and the material degradation threshold, and a partitioned environmental stress intensity factor library is built; according to the prominent erosion risk area coordinate set, the rotor exposure area, the power cabin sealed area and the fuselage surface area are divided, and the spatial weighting coefficient is calculated combined with the material corrosion resistance grade and exposure time of each region; the partitioned environmental stress intensity factor library and the spatial weighting coefficient are calculated to obtain the comprehensive evaluation value of the environmental erosion effect;

[0099] Step 223, based on the component degradation feature set, the fault propagation path network and the environmental erosion impact evaluation value, the core health score and the emergency severity are spatiotemporally aligned, and a comprehensive state index with time sequence characteristics is generated through a multi-dimensional calculation framework, including mechanical health degree, environmental erosion degree and operation stability.

[0100] In the embodiment of the application, the vibration spectrum and power transmission unit torque fluctuation data in the flight operation parameters are normalized, and are divided into N data segments of the same length in time sequence, each data segment is regarded as a sparrow, at this time an initial population containing N sparrows is formed, and the position vector of each sparrow is composed of the frequency component and amplitude of the vibration spectrum in the data segment, and the mean value and variance characteristic parameters of the torque fluctuation; according to the rules of the sparrow search algorithm, the top 20% sparrows with higher fitness (such as reflecting more data change characteristics) are determined as discoverers, and the discoverer sparrows explore in the data feature space with a larger step size, for example, in the vibration spectrum data, first pay attention to the high frequency area, and find abnormal peaks that may indicate component wear; in the torque fluctuation data, explore the time period where the fluctuation amplitude suddenly increases, and the discoverer constantly updates its own position (adjusts the feature parameter weight) to explore the feature area with high energy and high variation, and calculates the fitness of the new position (such as using mean square error to measure the fitting degree of the new feature combination and the original data) after each exploration, and retains the position with higher fitness.

[0101] The remaining 80% of the follower sparrows learn according to the position information of the discoverer, the follower evaluates the fitness of the discoverer, first approaches the discoverer with higher fitness, and in the approaching process, fine-tunes and excavates the characteristics with a small step, such as excavating periodic characteristics in the low-frequency component of the vibration spectrum or extracting rules from the trend change of the torque fluctuation. The follower continuously updates its own position to excavate more potential characteristics, extracts the component replacement period and maintenance record time sequence data from the maintenance archives, uses the maintenance time node as a boundary, corresponds the operation parameter characteristics to the maintenance data, makes each sparrow “remember” the changes of the characteristics before and after the maintenance of the component, for example, the disappearance of the abnormal peak value of the vibration spectrum after the maintenance or the trend of the torque fluctuation tends to be stable, and the change points of these characteristics are included in the component degradation characteristic set, and finally a complete and dynamic characteristic set reflecting the performance degradation of the component is constructed.

[0102] The fault occurrence time in the abnormal event record is divided according to different time scales such as hours and days, the fault occurrence position is divided according to the helicopter component area, different fault mode space-time units are formed, each space-time unit is regarded as a sparrow individual, and all individuals form a sparrow population. The individual position is composed of the time identifier, the position identifier and the fault type identifier of the space-time unit; the global search capability of the sparrow search algorithm is used, the sparrows in the population explore the potential relationship between fault modes in the space-time dimension, and the discoverer sparrow searches in a large-scale space-time range first, such as searching for different fault modes occurring in the same component area within a week; the follower sparrow further excavates in a small-scale space-time range based on the search result of the discoverer, such as exploring the appearance rules of different fault modes of the same component in adjacent two hours, and finding the aggregation phenomenon and the sequence rules of the fault modes in space-time by continuously adjusting the individual position (trying different space-time combinations).

[0103] The association rule mining algorithm is adopted, the fitness function of the sparrow search algorithm is introduced, the occurrence frequency of different fault mode combinations in the space-time unit is taken as the support, the close degree of the causal relationship between the fault modes is taken as the confidence, and the two are weighted to form the fitness evaluation standard. The discoverer sparrow continuously tries new fault mode combinations, calculates the fitness, and retains the combinations with high fitness; the follower sparrow adjusts the combinations locally according to the information of the discoverer to further improve the fitness, determines the fault combinations with high correlation strength of the typical problems through multiple rounds of adjustment, and calculates the correlation strength index; according to the calculated correlation strength index of the typical problems, the fault modes with correlation strength exceeding a set threshold are connected by lines, the fault modes are taken as nodes, and the correlation strength is taken as the weight of the edge, so as to construct a complete fault propagation path network and intuitively display the correlation relationship and propagation path between faults.

[0104] Real-time collection of marine climate erosion factor data such as salt spray concentration and relative humidity, each data sample is regarded as a sparrow. When the data fluctuates abnormally (such as a sudden and substantial increase in salt spray concentration), the "anti-predation" behavior in the sparrow search algorithm is simulated, and the abnormal data sample (sparrow) will quickly approach the surrounding normal data samples. By adjusting its own value (position) multiple times, the abnormal data is corrected. At the same time, according to the data and the normal fluctuation range, the missing data is supplemented to ensure the data is complete and accurate. The search capability of the sparrow search algorithm is used to find the key factors that can reflect the intensity of environmental stress in the processed data. The discoverer sparrow first explores the areas in the data with drastic changes and great impact on components, such as the time period with high salt spray concentration and high humidity at the same time; the follower sparrow, based on the discoverer, digs out the areas related to component corrosion. The relevant potential factors, such as humidity change rate, are continuously adjusted by factor combinations (sparrow positions), and the combination's ability to represent environmental stress intensity (fitness) is calculated. Factor combinations with high fitness are retained, and an environmental stress intensity factor library is constructed. Different regions are regarded as sparrow individuals, and their climate erosion data are used as individual positions. According to the geographical location of each region (such as the distance from the helicopter landing point), area size and other factors, the group collaboration characteristics of the sparrow search algorithm are used to assign corresponding weights. The discoverer sparrow is responsible for finding regional data that can represent the overall environmental characteristics. The follower sparrow performs weighted fusion of data from different regions based on the discoverer information and weight. For example, the area close to the landing point and with a large area has a higher weight, and its data accounts for a larger proportion in the fusion, ultimately forming a comprehensive assessment value that accurately reflects the impact of environmental erosion.

[0105] The core health score is aligned with the severity of the emergency according to the timestamp and the helicopter flight position. The dynamic adjustment mechanism of the sparrow search algorithm is used to evaluate the accuracy of the data alignment in real time as time passes and the helicopter flight position changes. If deviations are found in time or space, such as the core health score time record lags behind the actual event time, the data weight is adjusted (similar to the sparrow adjusting its position) to make the data consistent in time and space dimensions to ensure data comparability. Based on the component degradation feature set, the fault propagation path network and the environmental erosion impact assessment value, a multi-dimensional computing framework is constructed, and data of different dimensions are regarded as different sparrow populations. Each population The individuals in the dimensionality represent the specific data features under this dimension. During each adjustment process, each population interacts with the information of other populations based on its own data characteristics. For example, individuals in the component degradation feature set population (such as vibration spectrum characteristics) will refer to the relevant fault mode information in the fault propagation path network population to adjust their own weight (position) in the comprehensive calculation; individuals in the environmental erosion impact assessment value population will optimize their contribution to the overall health status based on the data of other populations. Through multiple adjustments, the weights and combinations of data in each dimension are continuously adjusted to ultimately generate comprehensive status indicators such as mechanical health, environmental erosion, and operational stability with time series characteristics.

[0106] The Sparrow Search algorithm's unique finder-follower mechanism and global search capabilities enable in-depth exploration of high-value areas and potential connections in the data when constructing component degradation feature sets and mining fault correlations, improving assessment accuracy compared to traditional methods. When calculating the correlation strength of typical problems and constructing a fault propagation path network, the Sparrow Search algorithm rapidly identifies high-fitness fault mode combinations. The constructed fault propagation path network more closely matches actual fault occurrence patterns, providing a scientific and accurate basis for fault prevention and maintenance decision-making, and reducing the risk of rescue mission interruptions caused by faults. When processing marine climate erosion factor data, the Sparrow Search algorithm simulates "anti-predation" behavior to correct abnormal data. Combined with group collaboration for spatially weighted fusion, the resulting environmental stress intensity factor library and comprehensive assessment value more realistically reflect the impact of the environment on helicopter components, effectively addressing the challenges posed by complex marine environments to helicopter health assessment. When generating comprehensive status indicators, the Sparrow Search algorithm's dynamic adjustment allows real-time adaptive adjustment of the indicator calculation process based on changes in helicopter operating data, ensuring that indicators such as mechanical health, environmental erosion, and operational stability accurately and timely reflect the helicopter's true health status, providing reliable, real-time decision support for scientific rescue mission scheduling and helicopter maintenance.

[0107] In a preferred embodiment of the present invention, the above step 3, based on the comprehensive status index, performs a standardized assessment of the number of oxidation points on the helicopter surface, the salt corrosion exposure cycle, and the material stress microcrack characteristic quantity, and establishes a dynamic quantitative correspondence between the environmental corrosion characteristics and the equipment operating conditions, which may include:

[0108] Step 330 , separating sea salinity, air humidity, flight trajectory altitude, and crew maintenance record data from the comprehensive status indicators, and combining them with the corrosion case library to construct a corrosion environment feature vector including timestamps, geographic coordinates, and operating condition parameters. The flight parameters are spatially and temporally correlated with the environmental parameters to obtain a multidimensional environmental stress factor matrix.

[0109] Step 331 , based on the multi-dimensional environmental stress factor matrix, feature extraction is performed on the helicopter surface structure data using a three-dimensional point cloud. The oxidation point distribution and salt corrosion exposure period are mapped to the fuselage structure coordinate system. In combination with the temporal variation of the material stress microcrack characteristic quantity, the distribution characteristics of the corrosion sensitive area are generated through cluster analysis.

[0110] Step 332: Analyze the correlation between corrosion rate and mechanical fatigue based on the corrosion distribution characteristics, fit the material strength attenuation trend under different salt corrosion levels, and generate a corrosion sensitivity grading table for key components including remaining life prediction values ​​and degradation rate confidence intervals in combination with the emergency severity data. Specifically, the table includes: extracting the corrosion rate quantification value and mechanical fatigue strength index of the key components based on the prominent corrosion risk area data in the corrosion distribution characteristics; using the corrosion rate quantification value and the mechanical fatigue strength index, calculate the dynamic correlation coefficient between the two through a correlation analysis algorithm, and establish a quantitative relationship that characterizes the interaction between environmental stress and mechanical performance degradation; based on the quantitative relationship and in combination with a corrosion case library under different salt corrosion levels, fit a decay law that reflects the change of material strength of key components with operating time; using the emergency severity data, dynamically calibrate the material strength attenuation law, and calculate the remaining life prediction value and degradation rate confidence interval of the key components based on the dynamically calibrated material strength attenuation law; based on the remaining life prediction value and degradation rate confidence interval, grade the corrosion sensitivity of the key components and generate a grading table;

[0111] Step 333 , based on the corrosion sensitivity classification data of key components, integrate the corrosion damage information of multiple units, build a dynamic correlation between environmental erosion intensity and equipment health status, and form a real-time updated corrosion assessment system.

[0112] In the embodiment of the present application, from the mass data of the comprehensive state index, the environmental parameters such as sea area salinity, air humidity and flight trajectory altitude are accurately determined by using data label and index technology. For the crew maintenance record data, not only the component maintenance time and maintenance content are extracted, but also the information such as maintenance personnel operation method and used maintenance tools which may affect the equipment state is refined. The determined data is arranged in time sequence to ensure that the parameter information of each time node is complete and has continuity. The corrosion case library stores a large number of corrosion cases of different types of helicopters in various environments. The environmental parameters of each time node are compared with the data in the case library, and the similarity calculation method is used to find multiple cases most similar to the current helicopter environmental parameters in the case library. In addition to matching the environmental parameters, the helicopter service life and flight frequency are also considered to ensure the high relevance of the reference cases.

[0113] A high-precision timestamp is added to the data of each time node, accurate to millisecond level, to accurately record the data collection time. At the same time, the precise geographic coordinates (latitude, longitude and altitude) of the helicopter at that time are obtained through the satellite positioning system. The timestamp, geographic coordinates, sea area salinity, air humidity, flight trajectory altitude and crew maintenance record data are deeply integrated to form a corrosion environment feature vector containing multi-dimensional information. For example, when integrating the crew maintenance record, the maintenance information of different components is associated with the environmental parameters of the corresponding time node, so that the feature vector can fully reflect the environmental and equipment maintenance state of the helicopter at a certain time. The flight parameters (such as rotor speed, engine temperature, power transmission unit torque, etc.) and environmental parameters are processed by space-time alignment. In the time dimension, the flight parameters and environmental parameters at the same time point are one-to-one corresponding based on the timestamp. In the spatial dimension, according to the flight trajectory of the helicopter and the position of the environmental monitoring point, the spatial position of the two is accurately corresponding by using the geographic information system (GIS) technology. All the feature vectors of the time nodes are arranged in time sequence to form a multi-dimensional environmental stress factor matrix. During the matrix generation process, the data is standardized to make different parameters comparable.

[0114] The helicopter surface is scanned in all directions using a high-precision laser scanner to obtain high-density three-dimensional point cloud data, ensuring that the fine structural features of the fuselage surface can be captured. During the collection process, the scanning parameters, such as scanning resolution and scanning angle, are adjusted according to different parts of the helicopter to ensure the integrity and accuracy of the data. After the collection is completed, the original point cloud data is denoised by using a filtering algorithm to remove abnormal points and noise points caused by equipment errors, environmental interference, etc. At the same time, the data is smoothed to make the point cloud data more uniform. The oxidation point distribution data and the salt corrosion exposure period data are mapped to the fuselage structure coordinate system constructed by the three-dimensional point cloud according to their actual positions on the helicopter surface through coordinate conversion algorithm. For the material stress microcrack characteristic quantity, the data of its change over time is extracted and associated with the corresponding position in the point cloud data. For example, for the oxidation points of a certain part of the fuselage, their accurate coordinate positions are found in the point cloud data, and the salt corrosion exposure period and stress microcrack characteristic quantity data at different time points are recorded. Through the establishment of a data association table, these information is integrated together to form a data set containing spatial position and time sequence information.

[0115] The integrated data is analyzed using advanced clustering algorithms. In the clustering process, appropriate clustering parameters such as the number of cluster centers and the distance measurement method are determined. Through multiple tests and adjustments, the clustering results accurately reflect the differences in corrosion characteristics of different regions. Points with similar oxidation point number, salt corrosion exposure period and stress microcrack characteristic quantity are divided into the same class. Each cluster represents a corrosion sensitive region. For each cluster, the center position, characteristic average value and other statistical quantities are calculated to quantitatively describe and analyze the corrosion sensitive region. At the same time, the clustering results are visualized to intuitively present the distribution of corrosion sensitive regions on the helicopter surface. For each corrosion sensitive region, a large amount of corrosion data and mechanical performance test data are collected to analyze the internal relationship between corrosion rate and mechanical fatigue. For example, different corrosion conditions are simulated to test the fatigue of the component, and the performance change data of the component during the corrosion process, such as strength, toughness and vibration frequency, are recorded. Through data analysis, the influence of corrosion rate on mechanical fatigue life is found out, and the quantitative relationship between the two is determined. At the same time, the reliability and applicability of the research results are verified by combining with the actual flight data.

[0116] According to the corrosion condition and material strength test data of the components under different salt corrosion grades, the trend fitting method such as least square method, polynomial fitting, etc. is used to fit the attenuation trend of material strength with the change of salt corrosion grade. In the fitting process, the data is preprocessed to remove abnormal data points, ensuring the accuracy of the fitting data. At the same time, considering the influence of various factors on the material strength, such as temperature, humidity, stress, etc., the fitting trend is modified and adjusted. Through multiple adjustment of fitting parameters, the fitting trend can accurately reflect the attenuation trend of material strength under different salt corrosion grades; the severity data of the emergency event is called, including the fault occurrence time, fault type, influence degree of the fault on the helicopter operation, etc. The influence difference of the emergency event on the key components of the helicopter under different corrosion sensitivity is analyzed, for example, for the components in the high corrosion sensitivity area, the same type of emergency event may be more likely to cause component failure. These influencing factors are taken into account, the material strength attenuation trend is comprehensively modified, the emergency event data is combined with the corrosion data, the parameters of the trend are adjusted, which can more truly reflect the actual situation.

[0117] Based on the modified material strength attenuation trend, combined with the current corrosion state and operating conditions of the components, the life prediction algorithm is used to calculate the residual life prediction value of the key components. In the calculation process, various uncertainty factors are considered, such as environmental changes, equipment usage frequency, maintenance quality, etc. Through methods such as Monte Carlo simulation, the uncertainty factors are quantitatively analyzed, and the confidence interval of the degradation rate is calculated. In specific operation, a large number of random simulation scenarios are set to simulate the operation of the components under different conditions, and the distribution of the residual life is counted to determine the confidence interval to represent the reliability of the prediction result. According to the residual life prediction value and the degradation rate confidence interval, the corrosion sensitivity of the key components is divided into multiple grades, such as low, lower, medium, higher, and high. Detailed grading standards are developed to clearly define the residual life range, degradation rate interval, and corrosion characteristic information corresponding to each grade. These information is arranged in table form to form a key component corrosion sensitivity grading table.

[0118] By establishing a unified data collection platform, the corrosion damage data of multiple helicopters are collected, including the corrosion conditions of different crews in the same or different environments, maintenance records, operation data, etc. During data collection, the integrity and accuracy of the data are ensured, and the data are strictly audited and checked. The collected data are summarized and arranged, and are stored in a classified manner according to the classification standards such as helicopter model, service life, flight area, etc., to form a comprehensive and orderly multi-crew corrosion damage information database. At the same time, a data index and query system is established to facilitate quick retrieval and analysis of data. The data in the multi-crew corrosion damage information database are analyzed in depth, and the relationship between environmental erosion intensity (such as the combined effects of factors such as salinity, humidity, and temperature) and equipment health status (such as component performance, residual life, and failure rate) is studied by using data mining and statistical analysis methods. By comparing the corrosion degree and health status changes of equipment under different environmental conditions, the internal relationship and rules between the two are found out. For example, the relationship between the corrosion rate and failure rate of key components of a helicopter in a high-salinity and high-humidity environment is analyzed to determine the influence degree and mechanism of environmental factors on equipment health status.

[0119] From the multi-crew corrosion damage information database, the corresponding data of different environmental conditions (salinity, humidity, temperature, etc.) and equipment health indicators (component wear degree, residual life estimate, and failure frequency) are extracted. The environmental parameter combinations are matched with the equipment health status one by one in a data mapping manner. For example, corrosion acceleration cases of components that occur multiple times in a high-salinity and high-humidity environment are sorted and classified. By analyzing these matching data, the rules between environmental factors and equipment health changes are summarized, such as finding that the corrosion rate of a certain key component increases by an average of 5% for every 1‰ increase in salinity. Based on these rules, a set of dynamic association rules is established, and whenever new environmental data is input, the trend of equipment health status change can be quickly judged according to these rules. The dynamic association rules are connected with the real-time data acquisition system. During the operation of the helicopter, sensors continuously collect environmental data such as sea salinity, air humidity, and flight duration, as well as equipment status data such as component vibration frequency and temperature. These data are transmitted in real time to the ground monitoring platform through a special communication link. After receiving the data, the monitoring platform analyzes and processes the data according to the preset association rules. For example, when it is detected that the salinity in a certain area exceeds the warning value and the duration is 2 hours, and the temperature of the corresponding component appears abnormal fluctuations, the corrosion risk level and health status score of the component are immediately calculated according to the association rules, and the evaluation results are presented through a visual interface. The corrosion risk level of each key component is displayed in the form of a dashboard. At the same time, built-in warning logic triggers an audible and visual alarm when the corrosion risk level of a component reaches "high risk" or the health status score is below the safety threshold, and sends prompt information containing the specific risk location and risk type to the terminal device of the maintenance personnel, so that the maintenance personnel can quickly respond and take targeted maintenance measures.

[0120] By deep screening data, fine matching with corrosion case library and multi-dimensional data integration, the multi-dimensional environmental stress factor matrix constructed comprehensively and accurately considers various environmental factors and equipment maintenance factors, can more accurately predict the erosion of the environment on the helicopter, and provide reliable basis for the protection and maintenance of the helicopter; Using high-precision three-dimensional point cloud data acquisition and advanced clustering analysis technology, the generated corrosion sensitive area distribution characteristics can clearly and intuitively present the corrosion risk difference of different parts of the helicopter surface, so that the maintenance personnel can quickly and accurately locate the high-risk area, thereby reducing the maintenance cost; By deeply studying the relationship between corrosion rate and mechanical fatigue, fine fitting the material strength attenuation trend, and combining with the overall correction of the emergency data, the calculated residual life prediction value and degradation rate confidence interval of the key components have high scientificity and reliability, which provides a solid basis for component replacement and maintenance planning, effectively avoids the risk of rescue task failure caused by accidental failure of components, and improves the success rate and safety of rescue tasks; The dynamic correlation between environmental erosion and equipment health state and the real-time updated corrosion evaluation system can quickly and accurately adjust the corrosion evaluation results according to the real-time changes of environmental and equipment operation data, realize efficient dynamic monitoring of the corrosion state of the helicopter, discover potential corrosion problems in time and take measures, and prolong the service life of the helicopter; Integrating multi-machine group corrosion damage information, the data is fully shared and deeply analyzed, the corrosion law and lessons are summarized as a whole, rich and comprehensive data support is provided for the design improvement, maintenance strategy optimization of the helicopter, the management level of the whole rescue helicopter fleet is improved, the overall reliability and safety of the fleet are improved, and the development of the industry is also provided with a beneficial reference.

[0121] In a preferred embodiment of the present application, step 4 utilizes the quantitative relationship between environmental erosion and equipment working conditions, and combines the comprehensive state indicators of multiple machine groups to fuse diversified data of different machine groups through a distributed collaborative learning architecture to construct a three-dimensional digital image of the helicopter to deduce the performance change trend of the parts in a strong salt corrosion and high moisture environment, which can include:

[0122] Step 440, a multi-machine group data fusion platform based on digital twinning is established, real-time salt fog concentration, humidity fluctuation value and equipment running load parameters of different regional rescue helicopters are collected, multi-source heterogeneous data is collaboratively encrypted and trained through a federal learning mechanism, and a fused global environment-working condition feature matrix is generated;

[0123] Step 441: Input the global environment-operating condition characteristic matrix into the environmental degradation simulation processing unit, synchronously call the material corrosion resistance coefficient and structural stress distribution data of the component in the three-dimensional digital image, and simulate the surface corrosion depth growth process and internal crack extension trajectory of the component based on the salt corrosion exposure cycle and humidity fluctuation range to obtain a degradation data set including the mechanical property attenuation value;

[0124] In step 442, the degradation dataset is decoupled from regional features through a transfer learning processing unit, the corrosion rate correlation function of the machine cluster in the high salt corrosion area is extracted, and the three-dimensional digital image of the machine cluster in the low salt corrosion area is adapted to generate the performance change trend of the components in the low salt corrosion area.

[0125] In an embodiment of the present invention, a data fusion platform is deployed on a central server, and data acquisition terminals are equipped for rescue helicopters in different regions. The acquisition terminals are connected to various sensors on the helicopters (such as salt spray concentration sensors, humidity sensors, and load sensors) to obtain in real time the salt spray concentration, humidity fluctuation values, and equipment operating load parameters (such as engine torque, rotor speed, etc.) of the helicopters during operation. At the same time, a corresponding digital twin is created for each helicopter. The digital twin pre-stores the basic parameters of the helicopter (such as model, production time, component material information) and operating data, so that it can map the status of the physical helicopter in real time. The data collected by helicopters in different regions have multi-source heterogeneous characteristics (such as different data formats and sampling frequencies). After uploading these data to the data fusion platform, the federated learning mechanism is activated. The data from each region does not need to leave the local area and is only trained locally (such as feature extraction of salt spray concentration data and smoothing of humidity data). Local parameters are generated and transmitted to the central server through a secure encrypted channel. The central server aggregates these parameters (such as taking the average and weighted sum) to obtain the global parameters, which are then sent to nodes in each region. Each node uses the global parameters to update the local area and continue with the next round of training. After multiple rounds of adjustment, each node is continuously optimized, and finally a fused global environment-operating condition feature matrix is ​​generated on the central server. This matrix integrates the environmental and operating condition feature data of different regions and different helicopters.

[0126] The global environment-working condition characteristic matrix is input into the environment degradation simulation processing unit, and in the three-dimensional digital image database, the material corrosion resistance coefficient (such as the corrosion resistance ability value of aluminum alloy in a salt spray environment) and the structural stress distribution data (such as the stress size and direction borne by each part of the component in different flight attitudes) of each component are called, the corrosion depth of the helicopter fuselage skin is set to 0 according to the salt corrosion exposure period and the humidity fluctuation range data in the global environment-working condition characteristic matrix, and the corrosion depth is gradually increased according to a certain corrosion rate (determined by the material corrosion resistance coefficient and the environment parameters) according to the salt spray concentration and the exposure time; for internal cracks, the crack extension trajectory in the stress concentration area is simulated according to the structural stress distribution data and the humidity change (humidity will affect the material toughness), and the mechanical property attenuation value (such as the strength reduction percentage and the toughness reduction degree) of the component is calculated in real time during the simulation process, and the corrosion depth, crack extension, mechanical property attenuation value and other data are recorded in time sequence, and finally a degradation data set containing multi-dimensional information is formed.

[0127] The degradation data set is input into the transfer learning processing unit, the data is analyzed, the key features (such as the corrosion mode unique to the high-salt corrosion area and the environment parameter combination in the rapid corrosion stage) in the high-salt corrosion area group data are identified, these features are separated from the whole data through comparative analysis, and a corrosion rate correlation function (such as a function expression describing the relationship between salinity and corrosion rate) of the high-salt corrosion area is formed, the corrosion rate correlation function of the high-salt corrosion area is adapted to the three-dimensional digital image of the low-salt corrosion area group, the environment parameters (such as average salinity and humidity range) and equipment working condition data of the low-salt corrosion area are adjusted (such as reducing the corrosion rate coefficient in the function), and the performance change process of the component in the low-salt corrosion environment is simulated by using the adjusted correlation function combined with the material and structure data of the component in the three-dimensional digital image of the low-salt corrosion area group, and the performance change trend (such as the predicted strength reduction amplitude and the possibility of crack occurrence) of the component in the future period of time is calculated, which provides a decision basis for the maintenance and use of the low-salt corrosion area helicopter.

[0128] The data fusion platform constructed based on digital twinning and federated learning realizes efficient fusion of different regions and multi-source heterogeneous data, avoids privacy leakage risk in data transmission process, guarantees the security of each fleet data, and can more realistically reflect the corrosion and performance degradation of parts in a high-salt corrosion and high-moisture environment by simulating the degradation process of parts combined with environmental parameters, helping maintenance personnel to master the health state changes of parts in advance, timely develop maintenance plans, and reduce the probability of sudden failure; the corrosion law of high-salt corrosion areas is adapted to low-salt corrosion areas by using transfer learning to process data of different regional fleets, breaking the geographical restrictions, realizing the deduction of the performance change trend of parts across regions, and enabling low-salt corrosion areas to learn from the experience of high-salt corrosion areas, take preventive measures in advance, reduce the risk of equipment damage caused by environmental factors, and prolong the overall service life of the helicopter; by constructing a three-dimensional digital image of the helicopter and deducing the performance change trend of the parts, more scientific maintenance strategies can be developed for the entire fleet, the maintenance cycle and content can be reasonably arranged according to the environmental characteristics of different regions, over-maintenance or insufficient maintenance can be avoided, and the overall reliability and rescue mission execution capability of the fleet can be improved.

[0129] In a preferred embodiment of the present application, step 5, based on the performance change trend of the parts and the comprehensive state index, the state classification reference value is adjusted in real time to obtain a complete evaluation report of the visual warning mark, the emergency degree of repair and the suggestion of operation restriction, which can include:

[0130] Step 550, input the performance change trend of the parts and the real-time collected flight task urgency parameter into the dynamic classification processing unit, perform weighted calculation on the residual life prediction value of the parts and the environmental erosion rate through the deep reinforcement learning mechanism, dynamically generate a set of health state classification threshold parameters, which specifically includes: based on the component failure time point in the fault record and the emergency repair frequency in the maintenance log, extract the initial correlation weight matrix of the environmental erosion rate and the residual life prediction value; input the initial correlation weight matrix and the real-time flight task urgency parameter into the dynamic calibration mechanism, process the weight matrix through the task urgency coefficient to obtain a set of real-time weight coefficients; input the set of real-time weight coefficients, the residual life prediction value of the parts and the environmental erosion rate into the deep reinforcement learning mechanism, construct a multi-dimensional state feature vector in the state space, and generate an action space operation instruction; input the action space operation instruction into the pre-trained reward function mechanism, generate a set of health state classification threshold parameters including a first warning threshold and a second shutdown threshold based on the component failure risk cost and the sudden maintenance loss;

[0131] Step 551, input the set of health state classification threshold parameters into the state identification processing unit, and simultaneously call the spatial positioning data of the parts in the three-dimensional digital image to perform classification determination operation, and generate instruction data set bound with three-dimensional space coordinates;

[0132] Step 552, input the three-dimensional spatial coordinate binding instruction data set into the decision generation processing unit, combine the externally input spare part supply cycle, spare landing site information and remaining flight time window to obtain a structured evaluation report including visual identification mapping relationship, maintenance time sequence table and airspace restriction parameter.

[0133] In the embodiment of the present application, from the analysis of the obtained part performance change trend data, core parameters such as the remaining life prediction value of the key component, the material strength attenuation ratio, the surface corrosion depth growth rate, etc. are extracted. These data are obtained by real-time monitoring and data deduction combined with sensors, for example, the remaining life of the engine turbine blade is predicted by referring to the current real-time sensing data such as vibration and temperature, and combining with similar failure cases under past similar working conditions; real-time collection of flight task related parameters, including task type (such as medical rescue, disaster survey), task priority (set by command center, 1-5 levels), remaining fuel quantity, target area weather severity (converted to a quantitative level through meteorological satellite data) and the like, integrate these parameters into a comprehensive task urgency index, for example, when the rescue task priority is 1 level and there is a strong storm in the target area, the urgency index is greatly improved, and all collected data are normalized, and different dimension parameters (such as remaining life in hours, corrosion depth in millimeters) are converted to a unified interval of 0-1, at the same time, abnormal values are removed, such as extreme data generated by instantaneous sensor failure.

[0134] With the core target of balancing flight safety and task completion efficiency, if a high-priority task can be completed under the premise of safety, a high reward is given; if a safety accident or task failure is caused by improper setting of the grading threshold, a penalty is given; the parameters such as the remaining life prediction value of the component, the environmental erosion rate and the task urgency index are combined into the "state" of the system; the "action" is the adjustment operation of the health state grading threshold (the boundary value of the five levels of normal, attention, warning, serious and danger); for example, the remaining life threshold of the "warning" level is adjusted from 50 hours to 40 hours; by simulating a large number of flight scenes (combining flight data and preset extreme situations), different grading threshold settings are tried, after each simulation, the reward value is calculated according to the actual result (such as whether it is safe to arrive, whether the task is completed in time), the reward feedback is obtained, the weighting strategy for each parameter is adjusted, for example, under high task urgency, the remaining life threshold of the "danger" level is appropriately reduced to trigger an early warning, after multiple learning, the optimal health state grading threshold parameter set suitable for different scenes is obtained.

[0135] The health state grading threshold parameter set generated in step 550 is transmitted to the state identification processing unit as the basis for judging the health level of the parts. From the three-dimensional digital image database of the helicopter, the precise spatial positioning data of each part is obtained, including its three-dimensional coordinates (x, y, z) in the body coordinate system, the structure area it belongs to (such as the engine compartment, the rotor system), and the connection relationship information with the surrounding parts, for example, the coordinate information of the main rotor transmission shaft, and the connection point position with the engine and rotor blades will be accurately retrieved; the actual performance parameters of each part (such as the current remaining life, the real-time corrosion rate) are compared one by one with the grading thresholds, for example, if the remaining life of a certain hydraulic pump is less than the threshold of the "dangerous" level, and the corrosion rate exceeds the corresponding standard, then it is determined that its health level is "dangerous"; for each part with a determined health level, its three-dimensional spatial coordinates are bound with the corresponding instruction code (such as the "dangerous" corresponding code 005), generating a data set containing coordinates (x, y, z) and instruction code, for example (1.2, 3.5, -0.8, 005) represents that the part at the specific coordinate position is in the "dangerous" state; the bound instruction data is classified according to the structure area of the helicopter (fuselage, power system, avionics system, etc.), and the parts in the same area are sorted according to importance (such as key flight control components first), and the sorted data is arranged in a unified format such as JSON or XML format, forming a structured instruction data set.

[0136] From the logistics management, the spare parts supply information of each part is obtained, including the inventory quantity, the required time for replenishment, the supplier's geographic location, etc., and the spare parts availability time of each part is calculated, for example, a certain type of bearing has sufficient inventory, and the replenishment requires 3 days, so its spare parts availability time is the current time + 3 days; the location coordinates of the emergency landing airport, the runway conditions (length, carrying capacity), the maintenance capacity (whether it has the qualification for maintenance of specific parts), etc. Data are collected, the flight distance and estimated flight time from the current position of the helicopter to each emergency landing airport are calculated, the flight time is corrected according to the real-time weather conditions, and the remaining flight time window is calculated according to the current fuel quantity of the helicopter, the flight speed, the fuel consumption rate and the task requirements, for example, if the remaining fuel is only enough for 2 hours of flight, and the task requires to arrive at the target area within 1.5 hours, then the remaining flight time window is 0-1.5 hours, each health level is assigned a corresponding visual warning mark, a mapping table of health level and visual mark is established, and it is clear how to display in the three-dimensional visualization interface.

[0137] According to the part health level, spare part available time and remaining flight time window, a maintenance plan is made, for the "dangerous" level parts, arrange for immediate maintenance at the nearest alternate airport; for the "serious" level parts, if the remaining flight time allows, maintenance can be arranged as soon as possible after completing the task; other level parts are included in the regular maintenance plan to form a detailed maintenance time sequence table; combined with the part health status and alternate airport location, airspace restrictions are set, for example, if the key navigation parts are in "dangerous" state, the helicopter is limited to fly within 50 kilometers around the alternate airport, and the flight height is not more than 3000 meters; if multiple parts have problems at the same time, further reduce the activity range and reduce the flight speed, integrate the visual identification mapping relationship, maintenance time sequence table, airspace restriction parameters and other contents into the evaluation report, set the report title, abstract, table of contents, main text and other blocks.

[0138] By dynamically adjusting the grading threshold through deep reinforcement learning, the health status evaluation can adapt to task and environmental changes in real time, reduce the risk of sudden failure caused by evaluation lag, and the three-dimensional space coordinate bound instruction data set and the structured maintenance time sequence table enable maintenance personnel to quickly locate the problem parts and determine the maintenance priority, combined with the operation constraint suggestions made by the spare parts supply, alternate airport and remaining flight time, to provide a scientific basis for airspace management and flight decision, to ensure the safety of pilots and rescue tasks; based on the evaluation report, the maintenance plan is reasonably arranged to avoid excessive maintenance and spare parts waste, and to improve resource utilization efficiency; the structured evaluation report presents information in a unified format and visual way, promotes efficient communication between pilots, maintenance teams and command centers, and ensures the smooth execution of rescue tasks.

[0139] As shown in Figure 2 The embodiment of the present application also provides a health status evaluation system of a rescue helicopter, comprising:

[0140] A data aggregation module is used for synchronously collecting flight operation parameters, maintenance archives, abnormal event records and surrounding environment monitoring indicators to form a comprehensive raw data set;

[0141] A state evaluation module is used for generating a comprehensive state index of multi-dimensions of cumulative running time length index, part remaining efficiency value and core health score based on the comprehensive raw data set;

[0142] An environmental erosion module is used for standardizing the assessment of the number of oxidation points on the surface of the helicopter, salt erosion exposure period and material stress microcrack characteristic quantity based on the comprehensive state index, and establishing a quantitative relationship between environmental erosion and equipment working condition;

[0143] A digital image module is used for constructing a three-dimensional digital image of the helicopter to deduce the performance change trend by using the quantitative relationship between environmental erosion and equipment working condition and combining the state indexes of multiple groups of machines;

[0144] The evaluation report module is used for adjusting the grading reference value in real time based on the performance change trend of the spare parts and the comprehensive state index, and generating a complete evaluation report including warning marks, maintenance suggestions and operation constraints.

[0145] It should be noted that the system is a system corresponding to the above method, and all the implementation manners in the above method embodiment are applicable to this embodiment and can also achieve the same technical effects.

[0146] Embodiments of the application also provide a computing device, comprising a processor, a memory storing a computer program, the computer program being executed by the processor to perform the method described above. All the implementation manners in the above method embodiments are applicable to this embodiment and can also achieve the same technical effects.

[0147] Embodiments of the application also provide a computer readable storage medium storing instructions, when the instructions are executed on a computer, the computer executes the method described above. All the implementation manners in the above method embodiments are applicable to this embodiment and can also achieve the same technical effects.

[0148] The above is the preferred embodiment of the application, it should be noted that for those skilled in the art, without departing from the principles of the application, can make a number of improvements and refinements, these improvements and refinements should also be considered as the protection scope of the application.

Claims

1. A method for assessing the health status of a rescue helicopter, characterized in that: The method comprises: Step 1: Establish an integrated data aggregation platform to simultaneously collect flight operation parameters, maintenance records, abnormal event records, and surrounding environmental monitoring indicators, comprehensively covering the key dimensions of the rotor assembly, power transmission unit, propulsion unit, and marine climate conditions to form a comprehensive original data set; Step 2: Based on the comprehensive original data set, an adaptive parameter adjustment mechanism is used to comprehensively calculate the cumulative operating time index, the remaining efficiency value of periodic components, the core health score, the probability of abnormal recurrence, the correlation strength of typical problems, the severity of emergencies, and the marine climate erosion factor to generate a comprehensive status index; Step 3: Based on the comprehensive status indicators, a standardized assessment is conducted on the number of oxidation points on the helicopter surface, the salt corrosion exposure cycle, and the material stress microcrack characteristics, establishing a dynamic quantitative correspondence between environmental corrosion characteristics and equipment operating conditions; Step 4: Leveraging the quantitative relationship between environmental corrosion and equipment operating conditions, combined with comprehensive status indicators from multiple fleets, a distributed collaborative learning architecture is used to fuse diverse data from different fleets to construct a three-dimensional digital image of the helicopters. This allows for the deduction of component performance trends in environments with strong salt corrosion and high moisture content. Step 5: Based on the performance change trend of components and the comprehensive status indicators, the status classification benchmark value is adjusted in real time to obtain a complete evaluation report of visual warning signs, maintenance urgency and operation constraint recommendations, which specifically includes: extracting the initial correlation weight matrix between environmental erosion rate and remaining life prediction value based on the component failure time point in the fault record and the emergency maintenance frequency in the maintenance log; inputting the initial correlation weight matrix and the real-time flight mission urgency parameter into the dynamic calibration mechanism, processing the weight matrix through the mission urgency coefficient, and obtaining the real-time weight coefficient set; inputting the real-time weight coefficient set, the component remaining life prediction value and the environmental erosion rate into the deep reinforcement learning mechanism, constructing a multi-dimensional state feature vector in the state space, and generating a dynamic The action space operation instructions are input into the pre-trained reward function mechanism, and a health status grading threshold parameter set including a first warning threshold and a second grounding threshold is generated based on the component failure risk cost and the sudden maintenance loss; the health status grading threshold parameter set is input into the status identification processing unit, and the spatial positioning data of the components in the three-dimensional digital image are called at the same time to perform a grading judgment operation to generate an instruction data set bound to the three-dimensional space coordinates; the instruction data set bound to the three-dimensional space coordinates is input into the decision generation processing unit, and combined with the external input spare parts supply cycle, alternate airport location information and remaining flight time window, a structured evaluation report is obtained, including a visual identification mapping relationship, a maintenance time series table and airspace restriction parameters.

2. The method for assessing the health status of a rescue helicopter according to claim 1, wherein: Based on the comprehensive original data set, the adaptive parameter adjustment mechanism is used to comprehensively calculate the cumulative operating time index, the remaining performance value of periodic components, the core health score, the probability of abnormal recurrence, the correlation strength of typical problems, the severity of emergencies and the marine climate erosion factors to generate comprehensive status indicators, including: The vibration spectrum and torque fluctuation characteristics of the power transmission unit in the flight operation parameters are analyzed in the time and frequency domain. At the same time, the component degradation feature set is constructed by combining the component replacement cycle and maintenance record time series data in the maintenance archive. Based on the component degradation feature set, combined with the temporal and spatial distribution characteristics of the fault mode in the abnormal event records, and using the association rule mining algorithm to calculate the typical problem association strength index, the association relationship between faults is realized to construct the fault propagation path network; Based on the fault correlation framework, we collect marine climate erosion factors such as real-time salt spray concentration and relative humidity, and combine them with corrosion damage data to build a factor library that characterizes environmental stress intensity. At the same time, we perform spatial weighted fusion of climate erosion effects in different regions to form a comprehensive assessment of environmental erosion impacts. Based on the component degradation feature set, fault propagation path network and environmental erosion impact assessment value, the core health score and the severity of the emergency are aligned in time and space, and a comprehensive status indicator with time series characteristics is generated through a multidimensional computing framework, including mechanical health, environmental erosion and operational stability.

3. The method for assessing the health status of a rescue helicopter according to claim 2, wherein: Based on the fault correlation framework, we collect marine climate erosion factors such as real-time salt spray concentration and relative humidity, and combine them with corrosion damage data to build a factor library that characterizes environmental stress intensity. At the same time, we perform spatial weighted fusion of climate erosion effects in different regions to form a comprehensive assessment of environmental erosion impacts, including: According to the spatial distribution of key fault nodes in the fault propagation path network, the coordinate set of the prominent erosion risk area is obtained; Based on the coordinate set of the prominent erosion risk area, the salt spray concentration, relative humidity data and corrosion damage data of the corresponding area are collected in real time; The salt spray concentration, relative humidity data and corrosion damage data are input into the dynamic coupling algorithm to generate the environmental corrosion rate coefficient and material degradation threshold, and build a partitioned environmental stress intensity factor library; Based on the coordinate set of the prominent erosion risk area, the rotor exposure area, the power cabin sealing area and the fuselage surface area are divided, and the spatial weighting coefficient is calculated based on the corrosion resistance level of the material in each area and the exposure time. The partitioned environmental stress intensity factor library and the spatial weighted coefficient are calculated to obtain a comprehensive assessment value of the environmental erosion impact.

4. The method for assessing the health status of a rescue helicopter according to claim 3, wherein: Based on comprehensive status indicators, a standardized assessment is conducted on the number of oxidation points on the helicopter surface, salt corrosion exposure cycles, and material stress microcrack characteristics. A dynamic quantitative correspondence between environmental corrosion characteristics and equipment operating conditions is established, including: The sea salinity, air humidity, flight trajectory altitude, and crew maintenance record data were separated from the comprehensive status indicators. Combined with the corrosion case library, a corrosion environment feature vector including timestamps, geographic coordinates, and operating parameters was constructed. The flight parameters were spatially and temporally correlated with the environmental parameters to obtain a multidimensional environmental stress factor matrix. Based on a multidimensional environmental stress factor matrix, a 3D point cloud was used to extract features from helicopter surface structure data. The distribution of oxidation points and salt corrosion exposure cycles were mapped to the fuselage structure coordinate system. Furthermore, cluster analysis was used to generate distribution characteristics of corrosion-sensitive areas, combining the temporal changes in the material stress microcrack characteristics. Based on the corrosion distribution characteristics, the relationship between corrosion rate and mechanical fatigue is analyzed, and the material strength attenuation trend under different salt corrosion levels is fitted. Combined with the emergency severity data, a corrosion sensitivity classification table for key components is generated, including the remaining life prediction value and the degradation rate confidence interval; Based on the corrosion sensitivity classification data of key components, the corrosion damage information of multiple units is integrated, and a dynamic correlation between environmental erosion intensity and equipment health status is constructed to form a real-time updated corrosion assessment system.

5. The method for assessing the health status of a rescue helicopter according to claim 4, wherein: Based on the corrosion distribution characteristics, the relationship between corrosion rate and mechanical fatigue is analyzed, and the material strength attenuation trend under different salt corrosion levels is fitted. Combined with the emergency severity data, a corrosion sensitivity grading table for key components is generated, including the remaining life prediction value and degradation rate confidence interval, including: Extract the corrosion rate quantification value and mechanical fatigue strength index of key components based on the prominent corrosion risk area data in the corrosion distribution characteristics; Using the corrosion rate quantification value and the mechanical fatigue strength index, the dynamic correlation coefficient between the two is calculated through the correlation analysis algorithm, and a quantitative relationship is established to characterize the interaction between environmental stress and mechanical performance degradation; Based on the quantitative relationship and combined with the corrosion case library under different salt corrosion levels, the attenuation law reflecting the change of key component material strength with operation time is fitted; Using emergency severity data, dynamically calibrate the material strength attenuation law, and calculate the remaining life prediction value and degradation rate confidence interval of key components based on the dynamically calibrated material strength attenuation law; Based on the remaining life prediction value and degradation rate confidence interval, the corrosion sensitivity of key components is graded and a grading table is generated.

6. The method for assessing the health status of a rescue helicopter according to claim 5, characterized in that: By leveraging the quantitative relationship between environmental erosion and equipment operating conditions, combined with comprehensive status indicators from multiple fleets, and integrating diverse data from different fleets through a distributed collaborative learning architecture, a three-dimensional digital image of the helicopter is constructed to deduce component performance trends in environments with strong salt corrosion and high moisture content, including: Establish a multi-fleet data fusion platform based on digital twins to collect real-time salt spray concentration, humidity fluctuation values, and equipment operating load parameters of rescue helicopters in different regions. Use a federated learning mechanism to conduct collaborative encryption training on multi-source heterogeneous data to generate a fused global environment-operating condition feature matrix. The global environment-operating condition characteristic matrix is ​​input into the environmental degradation simulation processing unit, and the material corrosion resistance coefficient and structural stress distribution data of the components in the three-dimensional digital image are synchronously called. Based on the salt corrosion exposure cycle and humidity fluctuation range, the surface corrosion depth growth process and internal crack extension trajectory of the components are simulated to obtain a degradation data set including the mechanical property attenuation value; The degradation dataset is decoupled from regional features through the transfer learning processing unit, and the corrosion rate correlation function of the machine cluster in the high salt corrosion area is extracted. It is adapted to the three-dimensional digital image of the machine cluster in the low salt corrosion area to generate the performance change trend of components in the low salt corrosion area.

7. A health status assessment system for a rescue helicopter, the system implementing the method according to any one of claims 1 to 6, characterized in that: include: The data aggregation module is used to synchronously collect flight operation parameters, maintenance files, abnormal event records and surrounding environment monitoring indicators to form a comprehensive original data set; The status assessment module is used to generate comprehensive status indicators in multiple dimensions, including cumulative operating time indicators, component remaining performance values, and core health scores, based on the comprehensive original data set; The environmental corrosion module is used to perform standardized assessments of the number of oxidation points on the helicopter surface, salt corrosion exposure cycles, and material stress microcrack characteristics based on comprehensive status indicators, establishing a quantitative relationship between environmental corrosion and equipment operating conditions; The digital image module is used to build a three-dimensional digital image of helicopters to deduce performance trends by utilizing the quantitative relationship between environmental erosion and equipment operating conditions, combined with multiple fleet status indicators; The evaluation report module is used to adjust the classification benchmark value in real time based on the performance change trend of components and comprehensive status indicators, and generate a complete evaluation report including warning signs, maintenance suggestions and operation constraints. Specifically, it includes: extracting the initial correlation weight matrix between environmental erosion rate and remaining life prediction value based on the component failure time point in the fault record and the emergency maintenance frequency in the maintenance log; inputting the initial correlation weight matrix and the real-time flight mission urgency parameter into the dynamic calibration mechanism, processing the weight matrix through the mission urgency coefficient, and obtaining the real-time weight coefficient set; inputting the real-time weight coefficient set, the component remaining life prediction value and the environmental erosion rate into the deep reinforcement learning mechanism, constructing a multi-dimensional state feature vector in the state space, and generating an action space. The action space operation instructions are input into the pre-trained reward function mechanism, and a health status grading threshold parameter set including a first warning threshold and a second grounding threshold is generated based on the component failure risk cost and the sudden maintenance loss; the health status grading threshold parameter set is input into the status identification processing unit, and the spatial positioning data of the components in the three-dimensional digital image are called at the same time to perform a grading judgment operation to generate an instruction data set bound to the three-dimensional space coordinates; the instruction data set bound to the three-dimensional space coordinates is input into the decision generation processing unit, and combined with the external input spare parts supply cycle, alternate airport location information and remaining flight time window, a structured evaluation report is obtained, including a visual identification mapping relationship, a maintenance time series table and airspace restriction parameters.

8. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a program, which, when executed by a processor, implements the method according to any one of claims 1 to 6.

Images