Intelligent flight decision assistance system and method based on multi-modal flight data

Abstract

The application provides an intelligent flight decision assistance system and method based on multi-modal flight data, and relates to the technical field of aviation flight. First, multi-source flight data containing multiple monitoring dimensions such as flight attitude, external environment and power parameters and having time correlation attributes are acquired. Then, the performance characteristics and cooperative change characteristics of different monitoring dimension data in each flight scene are extracted, a dynamic correlation model is established, the influence weight is determined based on the dynamic correlation model, the scene decision factor is generated, the association rules of the influence transmission characteristics and the flight operation response logic are established, the decision response link is constructed, and finally the logic in the decision response link is converted into natural language expression. The intelligent flight decision assistance instruction containing flight operation suggestions and data monitoring priorities is output, and the flight safety and efficiency can be effectively improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of aviation flight, in particular to an intelligent flight decision assistance system and method based on multi-modal flight data. BACKGROUND

[0002] In the field of aviation flight, it is crucial to ensure flight safety and efficiency. Traditional flight decisions mainly rely on the experience of pilots and relatively single flight data, such as making operation judgments based on only simple data of flight instruments. However, the modern flight environment is increasingly complex, and many monitoring dimensions are involved in the flight process, including flight attitude related dimensions, external environment related dimensions, and power parameter related dimensions.

[0003] Existing flight decision assistance methods often process data of each dimension in isolation, without fully considering the time correlation attributes and collaborative change characteristics between data of different monitoring dimensions. When facing complex and variable flight scenarios, due to the lack of comprehensive and in-depth correlation analysis of multi-source flight data, it is difficult to accurately grasp the actual influence of each monitoring dimension data on flight operations, resulting in insufficient accuracy and timeliness of decision assistance, which cannot effectively respond to various sudden flight conditions, and brings potential risks to flight safety. SUMMARY

[0004] In view of the above-mentioned problems, in combination with the first aspect of the present application, the embodiments of the present application provide an intelligent flight decision assistance method based on multi-modal flight data, the method comprising:

[0005] Obtaining multi-source flight data, the multi-source flight data containing continuous state data recorded by different monitoring dimensions in the flight process, the continuous state data of different monitoring dimensions having time correlation attributes, the different monitoring dimensions including flight attitude related dimensions recorded by flight instruments, external environment related dimensions collected by environment perception devices, and power parameter related dimensions recorded by power system operation;

[0006] Extracting performance characteristics of continuous state data of different monitoring dimensions under each flight scenario, and extracting collaborative change characteristics of monitoring dimension data under each flight scenario, forming a scene data correlation benchmark, and establishing a dynamic correlation model of the multi-source flight data and flight scenarios based on the scene data correlation benchmark, the performance characteristics including change laws, fluctuation characteristics of each monitoring dimension data under the corresponding flight scenario, and correlation performance between different monitoring dimension data;

[0007] Determine an influence weight of each monitoring dimension data based on a scene data association benchmark in the dynamic association model, extract an influence transmission characteristic of different monitoring dimension data under each flight scene, generate a scene decision factor by combining the influence weight and the influence transmission characteristic, and the influence transmission characteristic includes a transmission path, a duration and a form of influence transmission of monitoring dimension data change on associated affected monitoring dimension data.

[0008] Establish an association rule of the influence transmission characteristic and a flight operation response logic, filter a flight operation response logic suitable for the scene decision factor based on the association rule, concatenate the filtered flight operation response logic according to a logical relationship, and construct a decision response link corresponding to the scene decision factor;

[0009] Convert the flight operation response logic in the decision response link into a natural language expression conforming to a flight operation specification, determine a monitoring dimension and a monitoring implementation mode corresponding to each data monitoring focus, combine the expression result and the monitoring focus under each flight scene, and output an intelligent flight decision assistance instruction including a flight operation suggestion and a data monitoring focus matched under each flight scene.

[0010] In still another aspect, the embodiments of the present application also provide an intelligent flight decision assistance system based on multi-modal flight data, comprising:

[0011] A processor, a machine readable storage medium for storing machine executable instructions of the processor, wherein the processor is configured to execute the above-mentioned intelligent flight decision assistance method based on multi-modal flight data by executing the machine executable instructions.

[0012] In still another aspect, the embodiments of the present application also provide a computer program product, the computer program product comprising machine executable instructions stored in a computer readable storage medium, a processor of a computer device reading the machine executable instructions from the computer readable storage medium, and the processor executing the machine executable instructions, so that the computer device executes the above-mentioned intelligent flight decision assistance method based on multi-modal flight data.

[0013] Based on the above aspects, by acquiring multi-source flight data containing multiple monitoring dimensions and having time correlation attributes, the performance characteristics and cooperative change characteristics of different monitoring dimension data in each flight scene are extracted, and a dynamic correlation model is established, which reveals the internal relationship between the multi-source flight data and the flight scene. Based on the dynamic correlation model, the influence weight is determined and the scene decision factor is generated, the influence degree and transmission mode of each monitoring dimension data on flight operation are accurately quantified, the association rules of influence transmission characteristics and flight operation response logic are established, the decision response link is constructed, the complex data relationship is converted into clear flight operation suggestions, the operability and pertinence of decision are improved, and finally the output intelligent flight decision auxiliary instruction contains matching flight operation suggestions and data monitoring priorities, which can effectively assist the pilot to make correct decisions, and significantly improves the flight safety and efficiency. BRIEF DESCRIPTION OF DRAWINGS

[0014] Figure 1 is the execution flow schematic diagram of the intelligent flight decision assistance method based on multi-modal flight data provided by an embodiment of the application. DETAILED DESCRIPTION

[0015] The application will be described in detail below with reference to the accompanying drawings, Figure 1 is the flow schematic diagram of the intelligent flight decision assistance method based on multi-modal flight data provided by an embodiment of the application, and the intelligent flight decision assistance method based on multi-modal flight data will be described in detail below.

[0016] Step S110: acquiring multi-source flight data, the multi-source flight data containing continuous state data recorded in different monitoring dimensions during flight, the continuous state data of different monitoring dimensions having time correlation attributes, the different monitoring dimensions including flight attitude related dimensions recorded by flight instruments, external environment related dimensions collected by environmental perception devices, and power parameter related dimensions recorded by power system operation.

[0017] The embodiment takes the daily flight of a certain type of commercial jet as an example. The data of the related dimensions of the flight attitude are obtained through flight instruments such as attitude indicators and inertial navigation systems on the aircraft, such as the pitch angle, roll angle, yaw angle and other data of the aircraft. These data are continuously and dynamically recorded as the flight process advances, forming a continuous sequence of state data, and the sampling frequency of the data is several times per second to ensure the continuity and real-time of the data. The data of the related dimensions of the external environment are collected by environmental perception devices such as weather radars, air data computers and wind speed and direction sensors, such as atmospheric temperature, air pressure, wind speed, wind direction, cloud thickness and the like. These data also have time correlation and are recorded synchronously with the flight attitude data on the time axis to ensure that different dimensions of flight-related data can be obtained at the same time point. The power parameter related dimension data of the power system operation record come from the thrust sensor, fuel flow sensor, speed sensor and other devices of the engine, such as the thrust size, fuel flow rate and engine speed of the engine. These data are also updated continuously as the flight progresses, and maintain the same recording rhythm as other dimensions of data in time. When obtaining these multi-source flight data, the data needs to be read from each sensor and device through a data acquisition interface, and preliminary format conversion and storage are performed to ensure that the data of different monitoring dimensions remain consistent in time stamp, such as sampling and recording at a time interval of milliseconds, so that the time correlation attribute of the data can be utilized in subsequent processing.

[0018] Step S120: Extract the performance characteristics of the continuous state data of different monitoring dimensions under each flight scene, and extract the cooperative change characteristics of the monitoring dimension data under each flight scene to form a scene data correlation reference. A dynamic correlation model of the multi-source flight data and the flight scene is established based on the scene data correlation reference. The performance characteristics include the change law, fluctuation characteristics of each monitoring dimension data under the corresponding flight scene, and the correlation performance between different monitoring dimension data.

[0019] The embodiment continues the flight scene of the commercial jet described above. The flight scene includes the take-off stage, the cruising stage, the landing stage, and the normal weather condition and the complex weather condition (such as thunderstorm weather, strong wind weather, low-visibility weather, etc.).

[0020] Step S121: Divide the flight scene type to obtain a plurality of flight scene units. Each flight scene unit corresponds to a flight phase or flight environment condition that is pre-defined in the flight operation regulations and has a clear boundary. The flight phase includes the take-off stage, the cruising stage and the landing stage. The flight environment condition includes the normal weather condition and the complex weather condition.

[0021] In this embodiment, for the commercial jet aircraft, flight scenarios are divided into take-off phase scenario unit, cruising phase scenario unit, landing phase scenario unit, normal weather condition take-off scenario unit, normal weather condition cruising scenario unit, normal weather condition landing scenario unit, complex weather condition (such as thunderstorm) take-off scenario unit, complex weather condition (such as thunderstorm) cruising scenario unit, complex weather condition (such as thunderstorm) landing scenario unit, complex weather condition (such as strong wind) take-off scenario unit, complex weather condition (such as strong wind) cruising scenario unit, complex weather condition (such as strong wind) landing scenario unit, complex weather condition (such as low visibility) take-off scenario unit, complex weather condition (such as low visibility) cruising scenario unit, complex weather condition (such as low visibility) landing scenario unit, and the like. Each scenario unit has its specific flight phase or environmental condition, for example, the take-off phase scenario unit mainly corresponds to the process of the aircraft from the ground taxiing, leaving the ground to reaching the safe flight height (such as 10000 feet), in which the flight attitude, power parameters, and external environmental data will show specific change patterns, such as larger and continuously increasing thrust in the taxiing phase, gradually increasing pitch angle, and the like.

[0022] Step S122: For each flight scenario unit, collect the historical running records of multi-source flight data under the flight scenario unit, which contain the complete change process of the continuous state data of different monitoring dimensions under the flight scenario unit, and the time interval of the records remains consistent.

[0023] Taking the take-off phase scenario unit as an example, the historical flight data under this scenario are collected, including the continuous records of the flight attitude related dimensions (pitch angle, roll angle, yaw angle, etc.), external environment related dimensions (atmospheric temperature, air pressure, wind speed, wind direction, etc.), and power system running related dimensions (engine thrust, fuel flow, speed, etc.) in each take-off process. The time interval of these records needs to remain consistent, for example, all of them are sampled at a time interval of 500 milliseconds, to ensure the synchronization and comparability of the data. The historical running records can be obtained from the flight data recorder (FDR) of the aircraft, or retrieved from the flight data management system of the airline, and the integrity and accuracy of the data need to be ensured, and the records with damaged or abnormal data, for example, the engine thrust data in a record shows a cliff-like change without reasonable reasons, need to be excluded.

[0024] Step S123: Time synchronization is performed on the continuous state data of each monitoring dimension in the historical operation record, and performance characteristics of the monitoring dimension data are extracted from the synchronized historical operation record, the performance characteristics including change law, fluctuation characteristics and correlation performance, the change law including data change trend with flight progress and data state characteristics at key flight nodes, the fluctuation characteristics including data value distribution range in stable operation period and data fluctuation performance in state conversion period, and the correlation performance including change synchronization of different monitoring dimension data at the same time node.

[0025] Step S1231: For the continuous state data of each monitoring dimension, a data change curve is plotted in time sequence of flight progress, and a data change trend with flight progress is identified from the data change curve, the change trend including continuous rising trend, continuous falling trend, smooth keeping trend, fluctuation rising trend and fluctuation falling trend.

[0026] Taking the engine thrust data in the take-off stage scenario unit as an example, a data change curve is plotted in time sequence of flight progress (from the start of the taxiing to reaching the cruising altitude), in the taxiing stage, the engine thrust data usually presents a continuous rising trend, because sufficient thrust is needed to accelerate the aircraft taxiing and taking off; after taking off, the thrust is adjusted to a relatively stable level, presenting a smooth keeping trend, to maintain the climbing speed of the aircraft; when encountering air flow disturbance, the thrust data may present a fluctuation rising or fluctuation falling trend, depending on the direction and strength of the air flow. By observing the slope and fluctuation of the data change curve, the change trend of the monitoring dimension data with flight progress is identified.

[0027] Step S1232: Key flight nodes on the data change curve are marked, and monitoring dimension data values corresponding to each key flight node are recorded, to form data state characteristics at key flight nodes, the key flight nodes including nodes at which the flight state under the flight scenario unit changes.

[0028] On the engine thrust data change curve in the take-off stage scenario unit, the key flight nodes include the taxiing start node (the time point at which the aircraft starts moving on the runway), the taking-off node (the time point at which the aircraft tires leave the ground), the landing gear retracting node (the time point at which the aircraft retracts the landing gear after reaching a certain height and speed), the cruising altitude reaching node (the time point at which the aircraft climbs to the cruising altitude), etc. At these key flight nodes, the values of the engine thrust data are recorded, for example, the thrust is X kilonewtons at the start of taxiing, the thrust is Y kilonewtons at taking off, the thrust is Z kilonewtons at retracting the landing gear, and the thrust is A kilonewtons at reaching the cruising altitude, which constitute the state characteristics of the engine thrust data at the key flight nodes.

[0029] Step S1233: dividing the stable running period and the state transition period of the monitoring dimension data from the historical running records, counting the values of the monitoring dimension data in the stable running period, determining the value distribution range of the data in the stable running period, the stable running period being a period in which the data values have no significant fluctuations, the state transition period being a period in which the data values transition from one stable range to another stable range, and the value distribution range containing the common range of the data values in the period.

[0030] For the engine thrust data of the take-off phase scenario unit, the stable running period and the state transition period can be divided from the historical running records. The stable running period is, for example, the climb phase after taking off to reaching the cruising altitude, in which the thrust data values have no significant fluctuations and are in a relatively stable range. The state transition period is, for example, the taxiing phase, in which the thrust data values transition from the initial idle thrust range to the thrust range required for taking off. The values of the engine thrust data in the stable running period are counted, and the statistical quantities such as the average value, the maximum value, the minimum value, and the standard deviation are calculated to determine the value distribution range of the data in the stable running period, for example, the average value is B kN, the maximum value is C kN, the minimum value is D kN, and the standard deviation is E kN. Therefore, the value distribution range can be described as being between [D-E, C+E] kN, or other reasonable range representation methods are used according to actual conditions.

[0031] Step S1234: analyzing the change process of the monitoring dimension data in the state transition period, recording the fluctuation frequency and the fluctuation amplitude change of the data in the period, and forming the fluctuation performance of the data in the state transition period.

[0032] In the state transition period (taxiing phase) of the engine thrust data of the take-off phase scenario unit, the change process is analyzed, and the fluctuation frequency (for example, how many times per minute) and the fluctuation amplitude change (for example, the fluctuation amplitude gradually increases from F kN to G kN) are recorded. Through the analysis of multiple historical records, the typical fluctuation performance of the monitoring dimension data in the state transition period is summarized, for example, the fluctuation frequency gradually decreases, the fluctuation amplitude gradually increases, or both the fluctuation frequency and the amplitude remain relatively stable, etc.

[0033] Step S1235: selecting continuous state data of any two different monitoring dimensions, extracting the value change of the two at the same time node, judging whether the change direction is consistent and whether the change amplitude is related, and forming the change synchronization of the data of different monitoring dimensions at the same time node.

[0034] Taking the engine thrust data and the pitch angle data in the take-off stage scene unit as an example, the values at the same time nodes (such as 10 seconds, 20 seconds, 30 seconds after the start of the taxiing, etc.) are selected, the change amount of the engine thrust data (such as the thrust increased by H kilonewtons from the start of the taxiing to 10 seconds) and the change amount of the pitch angle data (such as the pitch angle increased by I degrees from the start of the taxiing to 10 seconds) are extracted, and it is judged whether the change direction is consistent (both increase, the change direction is consistent) and whether the change amplitude is related (for example, whether the ratio of the increase amount of the thrust to the increase amount of the pitch angle is within a relatively stable range, such as the pitch angle increases by J degrees for each increase of 1 kilonewton of the thrust). Through the analysis of multiple time nodes and multiple historical records, the description of the change synchronization of the two monitoring dimension data at the same time node is formed, such as the change direction is basically consistent, the change amplitude is positively correlated, and the correlation coefficient is K, etc.

[0035] Step S1236: The change trend, the key flight node state feature, the stable running period value distribution range, the state conversion period fluctuation performance and the change synchronization are integrated to form the complete performance feature of the monitoring dimension data under the corresponding flight scene unit.

[0036] Taking the engine thrust data of the take-off stage scene unit as an example, the change trend (continuously rising in the taxiing stage, stably maintaining in the climbing stage, and fluctuating rising or falling when encountering airflow), the key flight node state feature (the thrust value at the start of the taxiing, the take-off, the landing gear retraction, and the cruising altitude), the stable running period value distribution range (such as [D-E, C+E] kilonewtons), the state conversion period fluctuation performance (such as the fluctuation frequency of L times per minute in the taxiing stage, and the fluctuation amplitude increases from F kilonewtons to G kilonewtons), and the change synchronization with the pitch angle data (the change direction is consistent, the amplitude is positively correlated, and the correlation coefficient is K) are integrated to form the complete performance feature of the engine thrust data under the take-off stage scene unit.

[0037] Step S1237: The above operation is repeated until the complete performance features of all monitoring dimension data under the corresponding flight scene unit are extracted, and the complete performance features of all monitoring dimensions jointly constitute the performance feature set of the multi-source flight data under the flight scene unit.

[0038] In addition to the engine thrust data, the pitch angle, roll angle, yaw angle, atmospheric temperature, air pressure, wind speed, wind direction, fuel flow, engine speed, and other monitoring dimension data are also needed. According to the operations in steps S1231-S1236, the complete performance characteristics of the monitoring dimension data under the take-off stage scenario unit are extracted, and then the complete performance characteristics of all monitoring dimensions are integrated to form the performance characteristic set of the multi-source flight data under the take-off stage scenario unit. For the cruise stage, the landing stage, and the scenario units under different weather conditions, the complete performance characteristics of the monitoring dimension data are also extracted in the same way to form the corresponding performance characteristic set.

[0039] Step S124: Analyze the cooperative change characteristics between the monitoring dimension data, determine the correlation mode of the monitoring dimension data under each flight scenario unit based on the cooperative change characteristics, and the cooperative change characteristics include the echo relationship of the change direction of different monitoring dimension data and the associated performance of the change amplitude.

[0040] Taking the take-off stage scenario unit as an example, the cooperative change characteristics of the engine thrust data and the pitch angle data are analyzed. When the engine thrust increases, the pitch angle usually also increases, and the change direction presents an echo relationship. There is a certain associated performance between the amplitude of the thrust increase and the amplitude of the pitch angle increase, for example, the pitch angle increases by M degrees or so when the thrust increases by 1 kilonewton. The above-mentioned associated performance can be quantified by calculating the covariance, correlation coefficient, and other statistical quantities of the two. Similarly, the cooperative change characteristics of the fuel flow data and the engine speed data are analyzed. When the fuel flow increases, the engine speed usually also increases, the change direction echoes, and there is a certain associated performance between the change amplitudes, such as the engine speed increases by N revolutions per minute when the fuel flow increases by 1 kilogram per second. By analyzing the cooperative change characteristics between all monitoring dimension data, the correlation mode of the monitoring dimension data under each flight scenario unit is determined, such as the thrust-pitch angle correlation mode, the fuel flow-speed correlation mode, the atmospheric temperature-engine thrust correlation mode (when the atmospheric temperature rises, the engine thrust may decrease to a certain extent), and the like.

[0041] Step S125: Integrate the performance characteristics and the correlation mode to form the scenario data correlation reference corresponding to each flight scenario unit, and the scenario data correlation reference includes the performance characteristic parameters of each monitoring dimension data and the correlation mode description between the dimensions.

[0042] Taking the take-off stage scene unit as an example, the performance characteristic parameters (such as the trend description of the engine thrust, the key flight node value, the stable running period value distribution range, the state conversion period fluctuation performance statistics, and the change synchronization statistics with other dimensions) and the correlation mode description (such as the specific correlation relationship of the thrust-pitch angle correlation mode and the specific correlation relationship of the fuel flow-speed correlation mode) of each monitoring dimension data are integrated to form the scene data correlation benchmark corresponding to the take-off stage scene unit. The benchmark contains the typical characteristics and correlation relationships of the multi-source flight data under the scene unit, for example, the trend of the engine thrust in the taxiing stage is continuously rising, and the correlation mode with the pitch angle is that the change direction echoes and the amplitude is positively correlated, and the correlation coefficient is K. For other scene units, such as the cruise stage scene unit, its scene data correlation benchmark contains the performance characteristic parameters of each monitoring dimension data in the cruise stage and the correlation mode description, for example, the engine thrust in the cruise stage presents a steady trend, and the correlation mode with the fuel flow is that the change direction echoes (the fuel flow is stable, the thrust is stable, and the fuel flow changes, the thrust changes accordingly), and the amplitude correlation performance is that the fuel flow changes 1 kg / s, and the thrust changes about O kN.

[0043] Step S126: Establish the corresponding relationship between the scene data correlation benchmark of each flight scene unit and each monitoring dimension data in the multi-source flight data to form a preliminary correlation model, and the preliminary correlation model contains the matching conditions of the scene data correlation benchmark and each monitoring dimension data.

[0044] Taking the scene data correlation benchmark of the take-off stage scene unit as an example, the corresponding relationship between the scene data correlation benchmark and each monitoring dimension data in the multi-source flight data is established, for example, when the engine thrust data in the flight data presents a continuously rising trend in the taxiing stage, a steady maintaining trend in the climbing stage, and a fluctuating rising or falling trend when encountering air flow, and the change direction echoes and the amplitude correlation (the correlation coefficient is within a reasonable range) with the pitch angle data, the matching conditions of the scene unit are met. The preliminary correlation model is to clearly define the above matching conditions for judging whether the input multi-source flight data belongs to the flight scene unit. For the cruise stage scene unit, the matching conditions are that the engine thrust data presents a steady maintaining trend, the change direction echoes and the amplitude correlation (the correlation coefficient is within a reasonable range) with the fuel flow data, etc.

[0045] Step S127: Substitute the continuous state data in the multi-source flight data into the preliminary correlation model, check the matching of the continuous state data and the scene data correlation benchmark of each flight scene unit, and adjust the corresponding relationship between the scene data correlation benchmark and the monitoring dimension data to enable the continuous state data to match the corresponding flight scene unit.

[0046] Take the multi-source flight data of a certain take-off process as an example, and substitute it into the preliminary correlation model of the take-off stage scene unit. Check whether the change trend of the engine thrust data conforms to the description in the benchmark and whether the correlation performance with the pitch angle data matches. If it is found that the actual data has a large deviation in the rising speed of the thrust in the taxiing stage (such as the rising speed of the thrust in the taxiing stage in the benchmark is P kN / s, and the actual data is Q kN / s, the deviation of Q and P is large), or the correlation amplitude does not conform (such as the actual correlation coefficient is R, the deviation of R and K is large), the performance characteristic parameters or correlation mode description in the scene data correlation benchmark need to be adjusted, for example, the range of the thrust rising speed is adjusted (the rising speed range in the benchmark is adjusted from [P-S, P+S] kN / s to [Q-S, Q+S] kN / s), and the proportion relationship of the correlation amplitude is corrected (the range of the correlation coefficient is adjusted from [K-T, K+T] to [R-T, R+T]), so that the flight data can match the take-off stage scene unit. For the flight data of other scene units, the same way of substitution, checking and adjustment is needed.

[0047] Step S128: Repeat the adjustment process until any continuous state data in the multi-source flight data can be matched to a unique corresponding flight scene unit through the correlation model. Integrate all adjusted correlation models corresponding to the flight scene units to form a dynamic correlation model of the multi-source flight data and the flight scene.

[0048] The multi-source flight data of multiple take-offs, cruises, landings and other different flight processes are substituted and adjusted to ensure that any continuous state data can be uniquely matched to the corresponding flight scene unit. For example, for a certain segment of cruise stage flight data, the adjusted cruise stage scene unit correlation model can accurately identify its flight scene, that is, the engine thrust data presents a stable maintaining trend, and the correlation performance with the fuel flow data conforms to the benchmark description. After the correlation models of all flight scene units are adjusted, they are integrated to form a dynamic correlation model of the multi-source flight data and the flight scene. The model can automatically determine the flight scene unit to which the input multi-source flight data belongs.

[0049] Step S130: Determine the influence weight of each monitoring dimension data based on the scene data correlation benchmark in the dynamic correlation model, extract the influence transmission characteristics of different monitoring dimension data under each flight scene, and generate scene decision factors by combining the influence weight and the influence transmission characteristics. The influence transmission characteristics include the transmission path, duration and form of the influence of the change of the monitoring dimension data on the related affected monitoring dimension data.

[0050] Step S131: For each flight scene unit, a scene data correlation benchmark corresponding to the flight scene unit is extracted from the dynamic correlation model, and the scene data correlation benchmark includes the performance characteristics of each monitoring dimension data in the flight scene unit and the correlation mode between the dimensions.

[0051] Taking the take-off phase scene unit as an example, the scene data correlation benchmark thereof is extracted, wherein the performance characteristic of the engine thrust is that the thrust continuously increases during the taxiing phase, remains stable during the climbing phase, and fluctuates up or down when encountering air flow, the fluctuation characteristic is the value range of the stable running period (such as [D-E, C+E] kN), the fluctuation performance of the state transition period (such as the fluctuation frequency of the taxiing phase is L times / minute, and the fluctuation amplitude increases from F kN to G kN), the correlation mode with the pitch angle is that the change direction echoes and the amplitude is positively correlated, and the correlation coefficient is K; the performance characteristic of the pitch angle is that it gradually increases during the taxiing phase, remains stable during the climbing phase, and fluctuates when encountering air flow, and the correlation mode with the engine thrust is the same as above; the performance characteristic of the fuel flow is that it increases with the increase of the thrust during the taxiing phase, and remains stable during the climbing phase, and the correlation mode with the engine speed is that the change direction echoes and the amplitude is positively correlated.

[0052] Step S132: Determine the correlation strength between each monitoring dimension data based on the correlation mode, and determine the initial influence weight of each monitoring dimension data according to the correlation strength, the correlation strength includes the synchronization degree of data change and the mutual dependence degree of data value, and the mutual influence between monitoring dimension data is more significant when the correlation strength is high, and the initial influence weight is used to represent the importance of the change of the monitoring dimension data to the influence of the monitoring dimension data with which it has a correlation relationship.

[0053] In the take-off phase scene unit, the correlation mode between the engine thrust and the pitch angle is analyzed, the synchronization degree of data change between the two is high (when the thrust changes, the pitch angle almost changes at the same time), and the mutual dependence degree of data value is also high (the value of the thrust determines the value range of the pitch angle to a certain extent, and vice versa), so the correlation strength between them is high. According to the correlation strength, the initial influence weight of the engine thrust data is determined, for example, the initial influence weight of the engine thrust data is set to 0.8 because the change of the thrust has a significant influence on other monitoring dimension data such as the pitch angle; and for the atmospheric temperature data, the correlation strength between the atmospheric temperature data and the engine thrust data is relatively low (the change of the atmospheric temperature has a relatively indirect and weak influence on the thrust, and it usually takes a long time to reflect), so the initial influence weight of the atmospheric temperature data is set to 0.2. Similarly, the correlation strength between the fuel flow and the engine speed is analyzed, and the initial influence weight of the fuel flow data is determined to be 0.7, and the initial influence weight of the engine speed data is 0.7 because the correlation strength between them is also high; the correlation strength between the atmospheric pressure data and the engine thrust data is low, and the initial influence weight is set to 0.3.

[0054] Step S133: Select the monitoring dimension under the flight scene unit whose initial influence weight meets the preset importance requirement as the core influence dimension, and the rest of the monitoring dimensions as the affected dimensions, and analyze the influence of the state change of the core influence dimension data on the state of the affected dimension data.

[0055] In the take-off phase scene unit, the preset importance requirement is that the initial influence weight is higher than 0.5. According to the initial influence weight determined in step S132, the initial influence weights of the engine thrust (0.8), the pitch angle (0.7), the fuel flow (0.7), the engine speed (0.7), etc. meet the requirements, so they are taken as the core influence dimensions, and the atmospheric temperature (0.2), the atmospheric pressure (0.3), the wind speed (0.4), the wind direction (0.4), etc. are taken as the affected dimensions. Analyze the influence of the state change of the core influence dimension data on the state of the affected dimension data, for example, when the engine thrust suddenly increases in the taxiing phase, it will cause the acceleration of the aircraft to increase, thereby accelerating the rising speed of the pitch angle, and the fuel flow will also increase accordingly, because the increase of the thrust needs more fuel supply; the engine speed will increase with the increase of the fuel flow, because the increase of the fuel flow provides more energy source for the engine; and the change of the atmospheric temperature may affect the thrust output of the engine, but the above influence is indirect and relatively lagging, for example, when the atmospheric temperature rises, the air density decreases, and the engine thrust may decrease to a certain extent, but the above decrease usually needs to be observed after the temperature change lasts for a period of time.

[0056] Step S134: Track the process of the change of the state of the affected dimension data after the change of the state of the core influence dimension data, determine the transmission path of the influence, record the time length from the beginning of the change of the state of the core influence dimension data to the stable change of the state of the affected dimension data as the duration of the influence, and observe the state change performance of the affected dimension data in the influence process to determine the performance form of the influence. The transmission path includes a direct transmission path and an indirect transmission path. The direct transmission path is a path in which the core influence dimension data directly acts on the affected dimension data. The indirect transmission path is a path in which the core influence dimension data acts through the monitoring dimension data located between it and the affected dimension data. The duration distinguishes different durations corresponding to the direct transmission path and the indirect transmission path.

[0057] Step S1341: Select a core influence dimension data under the flight scene unit, record the initial stable state of the core influence dimension data, simulate a preset type of state change of the core influence dimension data, and the preset type of state change includes value increasing change, value decreasing change, fluctuation amplitude increasing change, and fluctuation amplitude decreasing change.

[0058] Taking the engine thrust data in the take-off stage scene unit as an example, record the initial stable state, for example, in the idle state before the taxiing, the engine thrust data keeps a relatively stable value and fluctuation amplitude. Then simulate the preset type of state change of the core influence dimension data, such as simulating the value rising change, so that the value of the engine thrust data gradually rises; or simulate the fluctuation amplitude increase change, so that the fluctuation amplitude of the engine thrust data increases.

[0059] Step S1342: While simulating the state change, record the state change process of all affected dimension data under the flight scene unit in real time, and record the time nodes at which each affected dimension data starts to change in time sequence.

[0060] While simulating the state change of the engine thrust data, record the state change process of the pitch angle, fuel flow, engine speed and other affected dimension data in real time. For example, record the time node at which the pitch angle data starts to change, the time node at which the fuel flow data starts to change, the time node at which the engine speed data starts to change, etc.

[0061] Step S1343: Compare the time nodes at which each affected dimension data starts to change with the time node at which the core influence dimension data starts to change, and determine the order of response of each affected dimension data.

[0062] Compare the time node at which the core influence dimension data (engine thrust) starts to change with the time node at which each affected dimension data starts to change. According to the order of time nodes, determine the order of response of each affected dimension data, for example, if the state change time node of a certain affected dimension data is earlier, then its response order is earlier.

[0063] Step S1344: Analyze the affected dimension data that responds to the influence first, check whether there is a direct physical or logical association between the affected dimension data and the core influence dimension data, if there is, determine that the path from the core influence dimension data to the affected dimension data is a direct transmission path.

[0064] Analyze the affected dimension data that responds to the influence first, check the association between it and the core influence dimension data. From the physical or logical principle, if there is a direct association between the two, for example, the change of engine thrust will directly cause the change of aircraft lift and in turn cause the change of pitch angle, then it is determined that the path from the core influence dimension data to the affected dimension data is a direct transmission path.

[0065] Step S1345: For the affected dimension data that is not the first to respond, analyze whether the time node when the state change begins has a sequential connection with the time node when the state change of other affected dimension data begins. If the time node when the state change of a certain affected dimension data begins is later than that of another affected dimension data, and there is a physical or logical connection between the two, it is determined that the path through which the core influence dimension data acts on the latter responding affected dimension data through the former responding affected dimension data is an indirect transmission path.

[0066] For the affected dimension data that is not the first to respond, analyze the time node relationship between the state change time node of the affected dimension data and that of other affected dimension data. If the state change time node of a certain affected dimension data is later than that of another affected dimension data, and there is a connection between the two, for example, the fuel flow change time node is later than the engine thrust change and the two are related, and at the same time, the engine speed change time node is later than the fuel flow change and the two are related, then it is determined that the path through which the core influence dimension data acts on the latter responding affected dimension data through the former responding affected dimension data is an indirect transmission path.

[0067] Step S1346: Record the affected dimension data that plays an intermediate transmission role in the indirect transmission path, and make clear the complete transmission link of the indirect transmission path.

[0068] After determining the indirect transmission path, record the affected dimension data that plays an intermediate transmission role, for example, in the indirect transmission path through which the engine thrust data acts on the engine speed data through the fuel flow data, the fuel flow data is the intermediate transmission data, and it is clear that the complete transmission link of the indirect transmission path is engine thrust data→fuel flow data→engine speed data.

[0069] Step S1347: Draw an influence transmission path diagram corresponding to the core influence dimension data, and clearly mark the core influence dimension, the affected dimension, the intermediate transmission dimension, and the influence transmission direction between the dimensions in the influence transmission path diagram.

[0070] Draw the influence transmission path diagram corresponding to the engine thrust data, and clearly mark the core influence dimension (engine thrust), the affected dimension (pitch angle, fuel flow, engine speed, etc.), the intermediate transmission dimension (such as fuel flow), and the influence transmission direction between the dimensions in the diagram, for example, engine thrust data→pitch angle data (direct transmission), engine thrust data→fuel flow data→engine speed data (indirect transmission).

[0071] Step S1348: Replace the core impact dimension data, and repeat the above operations of simulating state changes, recording changes in affected dimensions, determining response sequences, analyzing transmission paths, and drawing impact transmission path diagrams until all impact transmission paths corresponding to the core impact dimension data under the flight scene unit are determined.

[0072] Replace the core impact dimension data, for example, use the pitch angle data as the core impact dimension data, and repeat the above operations to simulate state changes in the pitch angle data, record changes in the affected dimension data, determine the response sequence, analyze the transmission path, and draw the impact transmission path diagram until all impact transmission paths corresponding to the core impact dimension data under the flight scene unit are determined.

[0073] Step S1349: Integrate all impact transmission path diagrams corresponding to the core impact dimension data to form a complete impact transmission path set under the flight scene unit.

[0074] Integrate all impact transmission path diagrams corresponding to the core impact dimension data under the flight scene unit to form a complete impact transmission path set, which contains impact transmission path information of each core impact dimension data on the affected dimension data under the flight scene unit.

[0075] Step S135: Integrate the transmission path, duration, form of expression, corresponding core impact dimension, affected dimension, and initial impact weight to form a structured impact feature item.

[0076] For the case of engine thrust as the core impact dimension, the direct transmission path (thrust → pitch angle), the indirect transmission path (thrust → fuel flow → engine speed → roll angle, yaw angle), the duration of the direct transmission path (T1 seconds), the duration of the indirect transmission path (T2 seconds), the form of expression of the impact (pitch angle rapidly rises and stabilizes, roll angle slightly fluctuates and stabilizes), and the corresponding core impact dimension (engine thrust), affected dimension (pitch angle, fuel flow, engine speed, roll angle, yaw angle), and initial impact weight (0.8) are integrated to form a structured impact feature item. This impact feature item contains information such as the impact mode, time, and performance of the core impact dimension on the affected dimension.

[0077] Step S136: Classify all impact feature items under the same flight scene unit, divide the impact feature items into multiple impact groups according to the different core impact dimensions, and each impact group contains all impact feature items corresponding to the same core impact dimension.

[0078] In the take-off phase scenario unit, all impact feature items are classified according to the core impact dimension, for example, the impact feature item of engine thrust as the core impact dimension is classified into an impact group, the impact feature item of pitch angle as the core impact dimension is classified into another impact group, the impact feature item of fuel flow as the core impact dimension is classified into a third impact group, and the impact feature item of engine speed as the core impact dimension is classified into a fourth impact group. Each impact group contains the impact feature items of the core impact dimension on different affected dimensions, for example, the engine thrust impact group contains the impact feature items of thrust on the pitch angle, fuel flow, engine speed, roll angle, yaw angle, etc. affected dimensions; the pitch angle impact group contains the impact feature items of the pitch angle on the roll angle, yaw angle, engine thrust, etc. affected dimensions (if any).

[0079] Step S137: analyze the impact coverage of each impact feature item in each impact group, adjust the initial impact weight of the impact feature item based on the impact coverage, and the impact coverage contains the number of affected dimensions involved in the impact feature item. When the impact coverage is wide, the impact weight is correspondingly increased.

[0080] Taking the engine thrust influence group as an example, the detailed information of all influence characteristic items in the influence group is extracted, and each influence characteristic item corresponds to a list of affected dimensions. The number of dimensions in the list of affected dimensions of each influence characteristic item is counted, for example, a certain influence characteristic item involves pitch angle, fuel flow, engine speed, roll angle, and yaw angle, which covers five affected dimensions, and the influence coverage size is 5; another influence characteristic item involves pitch angle, fuel flow, and engine speed, which covers three affected dimensions, and the influence coverage size is 3; and another influence characteristic item involves pitch angle and fuel flow, which covers two affected dimensions, and the influence coverage size is 2. The influence coverage size and the initial influence weight of all influence characteristic items in the influence group are recorded. The overall distribution of the influence coverage size in the influence group is analyzed to determine the benchmark size of the influence coverage, for example, the common level of the influence coverage size in the influence group is 3 (i.e. the number of affected dimensions of most influence characteristic items is 3). For the influence characteristic items with an influence coverage size greater than the benchmark size (such as a size of 5), the initial influence weight of the influence characteristic item is increased based on the relative degree of its exceeding the benchmark size, for example, the exceeding degree is (5-3) / 3=0.666..., the increasing amplitude can be determined according to the preset rules, such as increasing by 0.2, then the initial influence weight of the influence characteristic item is increased from 0.8 to 1.0 (but usually the weight is normalized to avoid exceeding 1, so the actual increasing amplitude may need to be adjusted, for example, the benchmark size is set to 3, the exceeding part is allocated weight in proportion to ensure that the sum of the weights of all influence characteristic items is 1 or a reasonable total); for the influence characteristic items with an influence coverage size equal to the benchmark size, the initial influence weight is kept unchanged; for the influence characteristic items with an influence coverage size less than the benchmark size, the initial influence weight is reduced based on the relative degree of its being lower than the benchmark size, for example, the size is 2, the relative degree of being lower than the benchmark size is (3-2) / 3=0.333..., the decreasing amplitude is 0.1, then the initial influence weight of the influence characteristic item is reduced from 0.8 to 0.7. After the adjustment is completed, the sum of the adjusted influence weights of all influence characteristic items in the influence group is calculated, and if the sum exceeds the preset reasonable range of the weight sum (for example, 0.9-1.1), the adjusted influence weights of all influence characteristic items are scaled in proportion to make the sum within the reasonable range. The final influence weight of each influence characteristic item is recorded and stored in association with other information of the influence characteristic item.

[0081] Step S138: integrate all adjusted influence characteristic items to form the initial scene decision factor corresponding to the flight scene unit.

[0082] The adjusted influence characteristic items in all influence groups, such as engine thrust influence group, pitch angle influence group, fuel flow influence group, engine speed influence group, etc., are integrated to form the initial scene decision factor corresponding to the take-off stage scene unit. The scene decision factor contains the influence transmission characteristics and influence weights of each core influence dimension on the affected dimension, and can reflect the influence relationship between the monitoring dimension data in the flight scene unit.

[0083] Step S139: Compare the initial scene decision factors of different flight scene units, identify the differences in influence characteristics between similar flight scene units, adjust the influence characteristic items and influence weights of similar flight scene units, so that the scene decision factor can distinguish the influence transmission characteristics of different flight scene units.

[0084] By comparing the initial scene decision factors of the take-off stage scene unit and the normal weather condition take-off scene unit, the differences in influence characteristics between them are identified. For example, in the normal weather condition take-off scene unit, the influence weight of engine thrust may be slightly lower, because it does not need to overcome the additional resistance of complex weather conditions, and the influence transmission path may be simpler, for example, the influence amplitude of thrust on pitch angle is smaller. According to these differences, the influence characteristic items and influence weights of similar flight scene units are adjusted, so that the scene decision factor can accurately distinguish different flight scene units. For example, the influence weight of engine thrust in the normal weather condition take-off scene unit is adjusted from 0.8 to 0.75, and the correlation proportion of the influence amplitude of thrust on pitch angle is adjusted, so that the scene decision factor can reflect the differences between take-off in normal weather conditions and complex weather conditions.

[0085] Step S1310: Integrate all adjusted initial scene decision factors of flight scene units to form a complete set of scene decision factors, each scene decision factor corresponding to a flight scene unit and the influence transmission characteristics in the flight scene unit.

[0086] The adjusted initial scene decision factors of the take-off stage, cruise stage, landing stage and flight scene units under different weather conditions are integrated to form a complete set of scene decision factors. Each scene decision factor corresponds to a specific flight scene unit and the influence transmission characteristics in that scene, for example, the take-off stage scene decision factor corresponds to the influence transmission characteristics in the take-off stage, the cruise stage scene decision factor corresponds to the influence transmission characteristics in the cruise stage, the complex weather condition (thunderstorm) landing stage scene decision factor corresponds to the influence transmission characteristics in the landing stage under complex weather conditions, etc.

[0087] Step S140: Establishing an association rule of the influence transmission characteristic and the flight operation response logic, screening the flight operation response logic of the adaptation scene decision factor based on the association rule, concatenating the screened flight operation response logic according to the logical relationship, and constructing a decision response link corresponding to the scene decision factor.

[0088] Step S141: Collecting the flight operation response logic, wherein the flight operation response logic comprises a coping operation process for the influence transmission characteristic of different monitoring dimension data, and each flight operation response logic corresponds to a group of influence coping requirements which are predefined in the flight operation standard and have explicit coping targets.

[0089] The flight operation response logic of the commercial jet is collected, for example, a coping operation process for the sudden increase of engine thrust causing the too fast increase of pitch angle: when the sudden increase of the thrust and the increase speed of the pitch angle exceeding a threshold value (such as 5 degrees / second) are monitored, the pilot should slowly pull the control column backward to reduce the increase speed of the pitch angle, and observe the engine instrument and the flight attitude instrument to ensure the stability of the aircraft attitude; a coping operation process for the abnormal increase of fuel flow: when the abnormal increase of the fuel flow (such as 20% of the normal cruising flow) is monitored, the pilot should check whether there is a leakage or other failure in the fuel system, and adjust the engine thrust and the fuel flow if necessary to ensure that the fuel consumption of the aircraft is within a reasonable range. The above flight operation response logic corresponds to specific influence coping requirements, such as the influence of the thrust change on the pitch angle, the influence of the fuel flow change on the engine, etc.

[0090] Step S142: Feature extraction is performed on each flight operation response logic to obtain operation logic features, wherein the operation logic features comprise an influence type to which the response logic is directed, a monitoring dimension involved, an operation execution process, and an operation target performance.

[0091] Taking the coping operation logic for the sudden increase of the engine thrust causing the too fast increase of the pitch angle as an example, the operation logic features are extracted: the influence type to which it is directed is the influence of the thrust change on the pitch angle (abnormal amplitude), the monitoring dimension involved is the engine thrust and the pitch angle, the operation execution process is that when the sudden increase of the thrust and the increase speed of the pitch angle exceeding a threshold value are monitored, the control column is slowly pulled backward to reduce the increase speed of the pitch angle, and the engine instrument and the flight attitude instrument are observed, and the operation target performance is that the increase speed of the pitch angle is returned to the normal range (such as 2-3 degrees / second), and the thrust is kept stable or adjusted as needed.

[0092] Step S143: Structurally describing the influence transmission characteristic of the scene decision factor to obtain influence characteristic description items, wherein the influence characteristic description items comprise an influence source dimension, an affected dimension, a transmission path type, an influence performance form, and an influence duration.

[0093] For one of the impact feature items in the scene decision factor of the take-off phase, the impact source dimension is engine thrust, the affected dimension is pitch angle, the transmission path type is a direct transmission path, the impact manifestation is that the pitch angle rises too fast (more than 5 degrees per second), and the impact duration is T1 seconds (for example, the time from the increase in thrust to the stabilization of the pitch angle is 10 seconds). The above information is structured and described to form an impact feature description item.

[0094] Step S144: Establish an association rule between the impact feature description item and the operation logic feature, and build an association mapping table between the impact transmission feature and the flight operation response logic based on the association rule. The association rule contains the matching requirements of each item of content in the impact feature description item and each item of content in the operation logic feature, including that the impact source dimension is consistent with the monitoring dimension involved in the operation logic feature, and the impact manifestation is adapted to the impact type targeted by the operation logic feature. Each impact feature description item in the association mapping table corresponds to at least one adapted flight operation response logic identifier.

[0095] The association rule is established, for example, when the impact source dimension in the impact feature description item is engine thrust, the affected dimension is pitch angle, and the impact manifestation is that the pitch angle rises too fast (more than 5 degrees per second), it is matched with the flight operation response logic in the operation logic feature, which involves the monitoring dimension of engine thrust and pitch angle, and the impact type targeted is the impact of thrust change on pitch angle (abnormal amplitude). Based on these association rules, an association mapping table is built, in which each impact feature description item corresponds to one or more adapted flight operation response logic identifiers, for example, the above impact feature description item corresponds to the flight operation response logic identifier of "adjusting the operation of the control column".

[0096] Step S145: For each scene decision factor, extract all impact feature description items in the scene decision factor, and query the corresponding adapted flight operation response logic identifier for each impact feature description item based on the association mapping table.

[0097] Taking the scene decision factor of the take-off phase as an example, all impact feature description items therein are extracted, such as the impact of engine thrust on pitch angle (abnormal amplitude), the impact of engine thrust on fuel flow (abnormal increase), the impact of fuel flow on engine speed (abnormal increase), etc. For each impact feature description item, the corresponding adapted flight operation response logic identifier is queried in the association mapping table, for example, the identifier corresponding to the impact of engine thrust on pitch angle (abnormal amplitude) is "adjusting the operation of the control column", the identifier corresponding to the impact of engine thrust on fuel flow (abnormal increase) is "checking the operation of the fuel system", and the identifier corresponding to the impact of fuel flow on engine speed (abnormal increase) is "adjusting the operation of the engine thrust".

[0098] Step S146: According to the influence weight of the influence characteristic item in the scene decision factor, the queried adaptive flight operation response logic identifiers are prioritized, and the flight operation response logic corresponding to the adaptive flight operation response logic identifier with high priority is selected as the candidate operation response logic of the scene decision factor. The flight operation response logic identifier corresponding to the influence characteristic item with high influence weight is prioritized.

[0099] In the take-off phase scene decision factor, the influence weight of the influence characteristic item of engine thrust on pitch angle is 0.8, the influence weight of the influence characteristic item of fuel flow is 0.7, and the influence weight of the influence characteristic item of fuel flow on engine speed is 0.6. According to the influence weight, the queried adaptive flight operation response logic identifiers are prioritized, and the identifier of "adjusting the operation of the control stick" is prioritized (because the corresponding influence characteristic item weight is 0.8), the identifier of "checking the operation of the fuel system" is second (the weight is 0.7), and the identifier of "adjusting the operation of the engine thrust" is last (the weight is 0.6). The flight operation response logic corresponding to the adaptive flight operation response logic identifier with high priority (such as the first two) is selected as the candidate operation response logic, for example, "adjusting the operation of the control stick" and "checking the operation of the fuel system".

[0100] Step S147: Analyze the logical association relationship between the candidate operation response logics, and determine whether there is an execution order requirement or a functional complementary relationship between different candidate operation response logics.

[0101] Step S1471: Extract the execution prerequisite condition and execution result performance of each candidate operation response logic, wherein the execution prerequisite condition includes the monitoring dimension data state that needs to be met before executing the candidate operation logic, and the execution result performance includes the expected change of the monitoring dimension data state after executing the candidate operation logic.

[0102] Taking "adjusting the operation of the control stick" as an example, the execution prerequisite condition is that the engine thrust suddenly increases and the pitch angle rising speed exceeds the threshold (5 degrees / second), and the execution result performance is that the pitch angle rising speed returns to the normal range (2-3 degrees / second) and the thrust remains stable; taking "checking the operation of the fuel system" as an example, the execution prerequisite condition is that the fuel flow abnormally increases (more than 20% of the normal cruising flow), and the execution result performance is that the fuel system returns to normal (the fuel flow returns to the normal range) or the fault is found and the corresponding measures are taken (such as closing the faulty fuel pipeline).

[0103] Step S1472: Compare the execution result performance of the first candidate operation response logic with the execution precondition of the second candidate operation response logic, to determine whether the execution result performance of the first candidate operation response logic can meet the execution precondition of the second candidate operation response logic. If yes, it is determined that the two candidate operation response logics have a sequence of execution, and the first candidate operation response logic needs to be executed before the second candidate operation response logic.

[0104] The execution result performance of the "adjusting the driving stick operation" (the pitch angle rising speed returns to the normal range, and the thrust remains stable) does not directly meet the execution precondition of the "checking the fuel system operation" (the fuel flow abnormally increases), because the "adjusting the driving stick operation" mainly affects the state of the pitch angle and the thrust, rather than the state of the fuel flow. Therefore, the two do not have a sequence of execution.

[0105] Step S1473: Compare the execution preconditions and the execution result performances between all candidate operation response logics in sequence in the above manner, to construct an execution sequence relationship table of the candidate operation response logics, in which the execution sequence of each operation logic is explicitly marked.

[0106] Since the "adjusting the driving stick operation" and the "checking the fuel system operation" do not have a sequence of execution, the execution sequence of the two operation logics is marked as "no sequence" in the execution sequence relationship table.

[0107] Step S1474: For the candidate operation response logics that do not have a sequence of execution, analyze whether the execution targets thereof have functional complementarity, which includes whether the affected dimensions to which the operation logics are directed are different but jointly serve the same core affected dimension in terms of impact response, or whether the impact performances to which the operation logics are directed are different but jointly cover all impact types of the core affected dimension.

[0108] The execution targets of the "adjusting the driving stick operation" and the "checking the fuel system operation" are analyzed. The "adjusting the driving stick operation" is directed to the pitch angle and mainly responds to the impact of the engine thrust change on the pitch angle, and the "checking the fuel system operation" is directed to the fuel flow and mainly responds to the impact of the engine thrust change on the fuel flow. The two are directed to different affected dimensions but jointly serve the impact response of the engine thrust change, and the execution processes do not interfere with each other (adjusting the driving stick will not affect the checking of the fuel system, and checking the fuel system will not affect the adjustment of the driving stick), so the two have a functional complementarity.

[0109] Step S1475: If the two candidate operation response logics are directed to different affected dimensions but serve the same core affected dimension's impact response and the execution processes do not interfere with each other, it is determined that the two have functional complementary relationship.

[0110] According to the analysis of step S1474, the "adjusting the driving stick operation" and the "checking the fuel system operation" are directed to different affected dimensions (pitch angle and fuel flow, respectively) but serve the same core affected dimension's impact response (change of engine thrust) and the execution processes do not interfere with each other, so it is determined that the two have functional complementary relationship.

[0111] Step S1476: If the two candidate operation response logics are directed to different impact manifestations but cover all impact types of the core affected dimension and the execution processes can be cooperatively linked, it is determined that the two have functional complementary relationship.

[0112] Suppose there is another candidate operation response logic "adjusting the engine thrust operation" directed to the impact manifestation of abnormal increase of engine thrust, and the "adjusting the driving stick operation" is directed to the impact manifestation of too fast pitch angle rising speed, the two are directed to different impact manifestations but cover all impact types of the core affected dimension (abnormal increase of engine thrust and the resulting change of pitch angle) and the execution processes can be cooperatively linked (adjusting the engine thrust can solve the problem of abnormal increase of engine thrust from the source, and adjusting the driving stick can solve the problem of the resulting change of pitch angle), so there is also functional complementary relationship.

[0113] Step S1477: The candidate operation response logics with functional complementary relationship are integrated to form a combined operation response logic, and the execution mode and cooperative requirement of each operation logic are specified in the integration process.

[0114] The "adjusting the driving stick operation" and the "checking the fuel system operation" are integrated into a combined operation response logic, and the execution mode and cooperative requirement are specified: when the sudden increase of engine thrust and the pitch angle rising speed exceeding the threshold are monitored, the pilot first performs the "adjusting the driving stick operation" to slowly pull the driving stick backward to reduce the pitch angle rising speed, and at the same time, the "checking the fuel system operation" is started to check whether there is leakage or other faults in the fuel pipeline, fuel pump, fuel nozzle and other components. The two are cooperatively executed to jointly respond to the impact of the change of engine thrust.

[0115] Step S1478: Repeat the above functional complementary relationship judgment and integration operation until all candidate operation response logics without execution sequence but with functional complementary relationship are integrated into combined operation response logics.

[0116] If there are other candidate operation response logics that have functional complementarity, for example, "adjust engine thrust operation" and "adjust steering column operation", then the operations in steps S1474-S1477 are performed to integrate them to form a new combined operation response logic.

[0117] Step S1479: Check whether there is a logic conflict between the integrated combined operation response logic and the candidate operation response logics that have execution sequence, the logic conflict including execution precondition contradiction and execution result interference.

[0118] Check whether there is a logic conflict between the combined operation response logic of "adjust steering column operation" and "check fuel system operation" and other candidate operation response logics that have execution sequence (if any). For example, assume that there is a candidate operation response logic of "adjust engine thrust operation", whose execution precondition is to monitor abnormal increase of engine thrust, and whose execution result is to return the engine thrust to the normal range. The execution precondition of "adjust steering column operation" is to monitor sudden increase of engine thrust and the pitch angle rising speed exceeding a threshold, and the execution result is to return the pitch angle rising speed to the normal range. After the execution of "adjust engine thrust operation", the engine thrust returns to the normal range, which may cause the execution precondition of "adjust steering column operation" not to be met, because the pitch angle rising speed may return to the normal range, which is a logic conflict of execution precondition contradiction.

[0119] Step S14710: If there is a logic conflict, re-analyze the execution precondition and execution result of the candidate operation response logic, adjust the execution parameters of the operation logic or select other suitable candidate operation response logics to eliminate the logic conflict.

[0120] For the above logic conflict, re-analyze the execution precondition and execution result of the candidate operation response logic, adjust the execution parameters of "adjust steering column operation", for example, reduce the pitch angle rising speed threshold in the execution precondition of "adjust steering column operation" from 5 degrees per second to 3 degrees per second, or adjust the execution result of "adjust engine thrust operation", for example, make the speed of returning the engine thrust to the normal range slower, to ensure that the execution precondition of "adjust steering column operation" is still met during the execution of "adjust engine thrust operation". Alternatively, select other suitable candidate operation response logics, for example, replace "adjust steering column operation" with "monitor pitch angle operation", to eliminate the logic conflict.

[0121] Step S14711: Finally form an operation logic set containing the candidate operation response logics arranged in execution sequence and the integrated combined operation response logic, and there is no logic conflict between the operation logics in the operation logic set and they can cooperatively cope with the influence transmission characteristics of the scene decision factor.

[0122] After adjustment and integration, a set of operation logics are formed, such as the combined operation response logic including "adjusting the operation of the driving rod" and "checking the operation of the fuel system", and (if there are) the candidate operation response logic of "adjusting the operation of the engine thrust" arranged in sequence, which do not have logical conflicts and can coordinate the influence transmission characteristics of the take-off phase scene decision factors.

[0123] Step S148: Determine the execution steps of the candidate operation response logic or the combined operation response logic, and specify the operation content, operation implementation time, and required multi-source flight data support conditions of each execution step, including the monitoring dimension data and data state requirements that need to be paid attention to when executing the execution step.

[0124] Taking the combined operation response logic of "adjusting the operation of the driving rod" and "checking the operation of the fuel system" as an example, determine its execution steps:

[0125] Step one: Monitor the engine thrust and pitch angle data, and when the engine thrust suddenly increases (such as more than 10% of the normal taxiing thrust) and the pitch angle rising speed exceeds the threshold (such as 3 degrees per second), execute "adjusting the operation of the driving rod", the operation content is to slowly pull the driving rod backward, the operation implementation time is the moment when the data anomaly is monitored, and the required support conditions are sudden increase of engine thrust data and pitch angle rising speed exceeding the threshold.

[0126] Step two: At the same time as step one, monitor the fuel flow data, and when the fuel flow abnormally increases (such as more than 20% of the normal taxiing fuel flow), execute "checking the operation of the fuel system", the operation content is to check whether there is leakage or other faults in the fuel pipeline, fuel pump, fuel nozzle and other components, the operation implementation time is the moment when the fuel flow anomaly is monitored, and the required support conditions are abnormal increase of fuel flow data.

[0127] Step S149: Associate the scene decision factors, candidate operation response logic or combined operation response logic, corresponding execution steps, and support conditions according to the logical relationship to form a single decision response sub-link, which clearly specifies the complete logical path from the scene decision factor to the operation response.

[0128] The combination operation response logic of the take-off phase scene decision factor, "adjusting the operation of the driving rod" and "checking the operation of the fuel system" and the corresponding execution steps and support conditions are associated according to a logical relationship to form a single decision response sub-link: when the influence transmission features of the sudden increase of engine thrust and the pitch angle rising speed exceeding the threshold value and the abnormal increase of fuel flow appear in the take-off phase scene decision factor, the combination operation response logic is triggered, and "adjusting the operation of the driving rod" is executed first (satisfying the support condition of the sudden increase of engine thrust and the pitch angle rising speed exceeding the threshold value), and at the same time, if the abnormal increase of fuel flow is monitored during the execution of "adjusting the operation of the driving rod" (satisfying the support condition of the abnormal increase of fuel flow data), "checking the operation of the fuel system" is executed. The sub-link clearly defines the complete logical path from the scene decision factor to the operation response.

[0129] Step S1410: Perform the above-mentioned operations of obtaining the influence feature description item to forming a single decision response sub-link for all scene decision factors (including sequentially obtaining the influence feature description item, querying the adapted operation logic identifier, sorting, selecting the candidate operation logic, analyzing the logical association, determining the execution step and the support condition, and forming a single decision response sub-link).

[0130] The scene decision factors in the cruise phase, the landing phase and different weather conditions are also subjected to the above-mentioned operations, for example, the influence transmission features of the steady maintenance of engine thrust but the abnormal increase of fuel flow in the cruise phase scene decision factor, the influence feature description item is extracted, the adapted flight operation response logic identifier (such as "checking the operation of the fuel system", "adjusting the operation of the engine thrust" and the like) is queried, the sorting, the selection of the candidate operation logic, the analysis of the logical association, the determination of the execution step and the support condition are performed, and the corresponding single decision response sub-link is formed.

[0131] Step S1411: Classify all single decision response sub-links according to the flight scene unit, group the single decision response sub-links under the same flight scene unit, analyze the trigger relationship between different single decision response sub-links under the same flight scene unit, identify the single decision response sub-links with conditional triggering or time sequence connection, and connect the single decision response sub-links according to the trigger relationship.

[0132] All single decision response sub-links are classified by flight scene units, and those in the take-off stage are classified as a group, those in the cruising stage are classified as a group, and those in the landing stage are classified as a group. The triggering relationship between different single decision response sub-links in the take-off stage is analyzed, for example, when the combined operation response logic of "adjusting the operation of the control column" and "checking the operation of the fuel system" is executed, if the abnormal increase of engine speed is monitored, another single decision response sub-link "adjusting the operation of the engine speed" may be triggered, which is the conditional triggering relationship. The single decision response sub-links with conditional triggering or time sequence connection are identified, and they are connected in series according to the triggering relationship.

[0133] Step S1412: The single decision response sub-links connected in series are integrated to form a decision response sub-link corresponding to the scene decision factor in the flight scene unit, and the decision response sub-links corresponding to all flight scene units are integrated to form a complete decision response link corresponding to the scene decision factor. The complete decision response link contains the complete association path from the decision factor to the operation response in each flight scene, and the switching trigger conditions between different flight scene sub-links are clear.

[0134] The single decision response sub-links connected in series in the take-off stage are integrated to form a decision response sub-link in the take-off stage. Similarly, a decision response sub-link in the cruising stage and a decision response sub-link in the landing stage are formed. The above decision response sub-links are integrated to form a complete decision response link, which contains the complete association path from the decision factor to the operation response in each flight scene, and the switching trigger conditions between different flight scene sub-links are clear, for example, when the aircraft enters the cruising stage from the take-off stage, the decision response link switches from the take-off stage sub-link to the cruising stage sub-link, and the trigger condition for switching is that the flight height of the aircraft reaches the cruising height (such as 10000 feet), the flight speed reaches the cruising speed (such as 450 knots), the engine thrust is adjusted to the cruising thrust, etc.

[0135] Step S150: The flight operation response logic in the decision response link is converted into a natural language expression conforming to the flight operation specification, the monitoring dimensions and monitoring implementation modes corresponding to each data monitoring focus are clear, and the intelligent flight decision assistance instructions containing the matching flight operation suggestions and data monitoring focuses in each flight scene are output by combining the expression results and monitoring focuses in each flight scene.

[0136] Step S151: Each single decision response sub-link in the decision response link is traversed, and the flight scene unit, operation response logic and support condition corresponding to each single decision response sub-link are extracted.

[0137] Extract the corresponding flight scene unit for the take-off phase, the operation response logic for the combination of "adjusting the operation of the control column" and "checking the operation of the fuel system", the support condition for the sudden increase in engine thrust and the pitch angle rising speed exceeding the threshold (3 degrees / second), and the abnormal increase in fuel flow (more than 20% of the normal taxiing fuel flow).

[0138] Step S152: For each operation response logic in each single decision response sub-link, disassemble its execution steps, convert each execution step into a natural language expression consistent with the flight operation specification, extract key operation points from the execution step expression of the operation response logic, integrate the key operation points with the corresponding execution step expression to form the flight operation suggestion corresponding to the operation response logic, and clearly express the specific implementation content of the operation, the starting time of the operation, the duration of the operation, and the flight parameters to be controlled during the operation. The key operation points include the implementation links that need to be focused on during the operation, the control requirements of the operation parameters, and the judgment basis of the operation effect. The flight operation suggestion includes the complete operation implementation process and the core control points.

[0139] For example, step S1521: obtain the complete execution flow description of the operation response logic, which includes all execution steps of the operation response logic and detailed implementation requirements of each step.

[0140] Taking the combination of "adjusting the operation of the control column" and "checking the operation of the fuel system" as an example, the complete execution flow description is as follows: when the sudden increase in engine thrust and the pitch angle rising speed exceeding 3 degrees / second are monitored, the pilot slowly pulls the control column backward to reduce the pitch angle rising speed, the speed of pulling the control column does not exceed 1 degree / second, and the amplitude does not exceed 5 degrees. At the same time, during the process of pulling the control column, the fuel flow data is monitored, and when the fuel flow exceeds 20% of the normal taxiing fuel flow, the pulling of the control column is stopped (or the current control column position is maintained), and the checking of whether there is leakage or other faults in the fuel pipeline, fuel pump, fuel nozzle, etc. is started. During the checking process, the readings of the fuel pressure gauge, fuel temperature gauge, etc. need to be observed to ensure that the fuel system is working normally.

[0141] Step S1522: disassemble the execution flow description according to the time sequence of operation implementation, and divide the operation response logic into multiple independent execution steps, each of which corresponds to a specific operation implementation behavior.

[0142] The above execution flow description is disassembled into the following independent execution steps according to the time sequence:

[0143] Step S15221 : Monitor engine thrust and pitch angle data, determine if the condition of sudden increase in engine thrust and pitch angle rising speed exceeding 3 degrees per second is met.

[0144] Step S15222: If the condition is met, slowly pull back the control column, the speed of pulling back the control column does not exceed 1 degree per second, and the amplitude does not exceed 5 degrees.

[0145] Step S15223: During the process of pulling back the control column, monitor the fuel flow data, determine if the condition of fuel flow exceeding 20% of the normal taxiing fuel flow is met.

[0146] Step S15224: If the condition is met, stop the operation of pulling back the control column (or keep the current control column position).

[0147] Step S15225: Start checking whether there is leakage or other faults in the fuel pipeline, fuel pump, fuel nozzle and other components.

[0148] Step S15226: During the checking process, observe the readings of fuel pressure gauge, fuel temperature gauge and other instruments, and determine whether the fuel system is working normally.

[0149] Step S1523: For each independent execution step, extract the operation implementation subject of the execution step, extract the operation implementation object of the execution step, determine the operation implementation mode of the execution step, analyze the operation starting time of the execution step, determine the operation duration of the execution step, extract the flight parameters to be controlled during the execution of the execution step, the operation implementation subject includes the device system or operation mechanism that performs the operation, the operation implementation object includes the components or parameter items that the operation behavior acts on, the operation implementation mode includes the specific action form of the operation and the implementation intensity requirement of the operation, the starting time includes the starting time node or starting condition that the execution step needs to meet, the duration includes the operation time length required from the start to the end of the execution step, or the execution step needs to be continuously executed until the end condition is met, the flight parameters include the specific name of the parameter and the control requirement of the parameter, the control requirement includes the state range that the parameter needs to reach or the stability degree that the parameter needs to maintain.

[0150] Step S15231 : The operation implementation subject is the pilot (monitoring data through the instrument system of the aircraft), the operation implementation object is the engine thrust data and the pitch angle data, the operation implementation mode is to observe the instrument readings and make a judgment, the operation starting time is after the start of the take-off phase (or any time as long as the condition is met), the operation duration is continuous monitoring until the condition is met or the flight phase is ended, and there is no flight parameter to be controlled (this step is a monitoring step).

[0151] Step S15232: The operation implementation subject is the pilot, the operation implementation object is the control stick, the operation implementation manner is to pull the control stick backward, the pulling speed is not more than 1 degree per second, the pulling amplitude is not more than 5 degrees, the operation starting time is after step S15221 determines that the condition is met, the operation duration period is from the start of pulling the control stick to step S15223 determines that the condition is met or the pitch angle rising speed returns to the normal range (2-3 degrees per second), and the flight parameter to be controlled is the pulling speed of the control stick (≤1 degree per second), the pulling amplitude (≤5 degrees), and the pitch angle rising speed (target range 2-3 degrees per second).

[0152] Step S15233: The operation implementation subject is the pilot (by monitoring the data of the fuel flow instrument of the aircraft), the operation implementation object is the fuel flow data, the operation implementation manner is to observe the instrument reading and make a judgment, the operation starting time is after step S15222 starts, the operation duration period is continuous monitoring until the condition is met or step S15222 ends, and the flight parameter to be controlled is none (this step is a monitoring step).

[0153] Step S15234: The operation implementation subject is the pilot, the operation implementation object is the control stick, the operation implementation manner is to stop pulling the control stick or keep the current position, the operation starting time is after step S15223 determines that the condition is met, the operation duration period is from the start of step S152224 to the end of step S152225, and the flight parameter to be controlled is the control stick position (keep the current position).

[0154] Step S15235: The operation implementation subject is the pilot, the operation implementation object is the fuel pipeline, fuel pump, fuel nozzle and other components, the operation implementation manner is visual inspection or judgment through instrument data, the operation starting time is after step S15224 starts, the operation duration period is from the start of step S15225 to step S15226 determines that the fuel system is normal or a fault is found, and the flight parameter to be controlled is none (this step is an inspection step).

[0155] Step S15236: The operation implementation subject is the pilot (by monitoring the data of the fuel pressure gauge, fuel temperature gauge and other instruments of the aircraft), the operation implementation object is the fuel pressure data, fuel temperature data and the like, the operation implementation manner is to observe the instrument reading and make a judgment, the operation starting time is after step S15225 starts, the operation duration period is continuous monitoring until the fuel system is determined to be normal or a fault is found, and the flight parameter to be controlled is none (this step is a monitoring step).

[0156] Step S1524: The operation implementation subject, operation implementation object, operation implementation manner, operation starting time, operation duration period and flight parameter to be controlled are integrated and expressed in natural language with professional terms in accordance with the flight operation specifications.

[0157] Step S15241: During takeoff phase, continuously monitor engine thrust and pitch angle data, when engine thrust suddenly increases (more than 10% of normal taxiing thrust) and pitch angle rises at a speed exceeding 3 degrees per second, proceed to the next step.

[0158] Step S15242: After meeting the conditions of step S15241, slowly pull back the control column, the speed of pulling back the control column does not exceed 1 degree per second, the amplitude does not exceed 5 degrees, continuously pull back the control column until the fuel flow exceeds 20% of the normal taxiing fuel flow or the pitch angle rise rate returns to the range of 2-3 degrees per second.

[0159] Step S15243: During the process of pulling back the control column, continuously monitor the fuel flow data, when the fuel flow exceeds 20% of the normal taxiing fuel flow, proceed to the next step.

[0160] Step S15244: After meeting the conditions of step S15243, stop pulling back the control column operation, maintain the current control column position.

[0161] Step S15245: While maintaining the control column position, start checking the fuel pipeline, fuel pump, fuel nozzle, etc. to see if there are leaks, blockages or other abnormalities.

[0162] Step S15246: During the process of checking the fuel components, continuously monitor the readings of fuel pressure gauge, fuel temperature gauge, etc. to determine if the fuel system is working properly. If the fuel pressure is within the normal range (e.g. 20-30 psi) and the fuel temperature is within the normal range (e.g. 40-60°C), it is determined that the fuel system is normal, otherwise it is determined that there is a fault and appropriate measures are taken (e.g. closing the faulty fuel pipeline, notifying the maintenance personnel, etc.).

[0163] Step S1525: Check if the expression accurately reflects the core implementation requirements of the step, whether it conforms to the professional expression habits of flight operations, if the expression is not accurate or does not conform to the habits, adjust it, and check if there is redundant information in the expression, if there is, delete the redundant content.

[0164] Check the above expression to ensure that each step accurately reflects the core implementation requirements of the step, such as the speed and amplitude requirements of pulling back the control column in step S15242, the normal range requirements of fuel pressure and temperature in step S15246, etc., which conform to the professional expression habits of flight operations. At the same time, delete redundant information, such as the redundant information "or any time" in step S15231 has been deleted, because the operation in the takeoff phase is mainly carried out in the takeoff phase.

[0165] Step S1526: Repeat the above operations of disassembling, extracting, determining, analyzing, integrating, expressing, checking and adjusting until all independent execution steps are converted into natural language expressions that meet the requirements.

[0166] Other steps of "checking fuel system operation" are also disassembled, extracted, determined, analyzed, integrated, expressed, checked and adjusted in the above manner to ensure that all independent execution steps are converted into natural language expressions that meet the requirements.

[0167] Step S1527: Arrange the natural language expressions of all independent execution steps in sequence to form a complete operation step expression sequence corresponding to the operation response logic of the operation, and the time connection relationship and logical association relationship between the step expressions in the complete operation step expression sequence.

[0168] Arrange the natural language expressions of steps S15241 to S15246 in sequence to form a complete operation step expression sequence corresponding to the combined operation response logic of "adjusting the steering column operation" and "checking the fuel system operation":

[0169] 1. During takeoff, continuously monitor engine thrust and pitch angle data. When engine thrust suddenly increases (more than 10% of normal taxiing thrust) and pitch angle rises at a speed exceeding 3 degrees per second, proceed to the next operation.

[0170] 2. After meeting the above conditions, slowly pull the steering column backward at a speed not exceeding 1 degree per second and an amplitude not exceeding 5 degrees. Continue to pull the steering column until the fuel flow exceeds 20% of the normal taxiing fuel flow or the pitch angle rise rate returns to the range of 2-3 degrees per second.

[0171] 3. During the steering column pulling process, continuously monitor the fuel flow data. When the fuel flow exceeds 20% of the normal taxiing fuel flow, proceed to the next operation.

[0172] 4. After meeting the above conditions, stop the steering column pulling operation and maintain the current steering column position.

[0173] 5. While maintaining the steering column position, start checking the fuel pipeline, fuel pump, fuel nozzle and other components to see if there are leaks, blockages or other abnormalities.

[0174] 6. During the inspection of fuel components, continuously monitor the readings of fuel pressure gauges, fuel temperature gauges and other instruments to determine whether the fuel system is working normally. If the fuel pressure is within the normal range (e.g. 20-30 psi) and the fuel temperature is within the normal range (e.g. 40-60°C), the fuel system is determined to be normal, otherwise a fault is determined and appropriate measures are taken (e.g. closing the faulty fuel pipeline, notifying the maintenance personnel, etc.).

[0175] Extract the key operation points from the sequence of expressions, such as the pull handle speed and amplitude control in step S15242, and the fuel pressure and temperature normal range in step S15246. Integrate the above key operation points with the corresponding execution step expressions to form flight operation suggestions:

[0176] Flight operation suggestions (take-off phase - combined operation):

[0177] Operation implementation process:

[0178] 1. During the take-off phase, continuously monitor engine thrust and pitch angle data. When the engine thrust suddenly increases (more than 10% of the normal taxiing thrust) and the pitch angle rises at a speed exceeding 3 degrees per second, slowly pull back the handle, with a speed not exceeding 1 degree per second and an amplitude not exceeding 5 degrees. Continue to pull back the handle until the fuel flow exceeds 20% of the normal taxiing fuel flow or the pitch angle rise rate returns to the range of 2-3 degrees per second.

[0179] 2. During the process of pulling back the handle, continuously monitor the fuel flow data. When the fuel flow exceeds 20% of the normal taxiing fuel flow, stop pulling back the handle and maintain the current handle position.

[0180] 3. While maintaining the handle position, start checking the fuel pipeline, fuel pump, fuel nozzle, and other components for leaks, blockages, or other abnormalities.

[0181] 4. During the process of checking the fuel components, continuously monitor the readings of the fuel pressure gauge, fuel temperature gauge, and other instruments to determine whether the fuel system is working normally. If the fuel pressure is within the normal range (e.g., 20-30 psi) and the fuel temperature is within the normal range (e.g., 40-60°C), the fuel system is considered normal. Otherwise, a fault is detected and appropriate measures are taken (e.g., closing the faulty fuel pipeline, notifying the maintenance personnel, etc.).

[0182] Core control points:

[0183] When pulling back the handle, strictly control the speed (≤1 degree per second) and amplitude (≤5 degrees) to ensure that the pitch angle rise rate smoothly decreases to the range of 2-3 degrees per second.

[0184] When checking the fuel components, carefully observe to ensure there are no leaks, blockages, or other abnormalities, and pay attention to the readings of the fuel pressure gauge and temperature gauge to ensure that the fuel system is working normally.

[0185] Step S153: Extracting the support condition in each single decision response sub-link, determining the monitoring dimension data that needs to be focused on from the support condition, clarifying the focus content of each monitoring dimension data, determining the monitoring implementation of each monitoring dimension data, integrating the monitoring dimension data, focus content and monitoring implementation to form the data monitoring focus corresponding to the single decision response sub-link, the focus content including the state change range of the data, the change trend characteristics of the data, the association state requirement of the data and the determined associated monitoring dimension data, the monitoring implementation including the device type used for monitoring, the collection method of monitoring data, and the recording form of monitoring data, the monitoring implementation being adapted to the focus content and being able to accurately obtain the data required by the focus content.

[0186] From the support condition of the single decision response sub-link in the take-off phase (sudden increase of engine thrust and pitch angle rising speed exceeding 3 degrees per second, abnormal increase of fuel flow exceeding 20% of normal taxiing fuel flow), the monitoring dimension data that needs to be focused on is determined as engine thrust, pitch angle and fuel flow.

[0187] Engine thrust data:

[0188] Focus content: state change range of data (±10% of normal taxiing thrust), change trend characteristics of data (sudden increase, i.e. increase of thrust exceeding 10% of normal taxiing thrust in a short time (such as 5 seconds)), and association state requirement of data and pitch angle data (when the thrust increases, the pitch angle rising speed exceeds 3 degrees per second).

[0189] Monitoring implementation: the device type used for monitoring is engine thrust sensor, the collection method of monitoring data is real-time collection (sampling frequency is 10 times per second), and the recording form of monitoring data is time series data (including time stamp and thrust value).

[0190] Pitch angle data:

[0191] Focus content: state change range of data (0-20 degrees in take-off phase), change trend characteristics of data (rising speed exceeding 3 degrees per second), and association state requirement of data and engine thrust data (when the thrust increases, the pitch angle rising speed exceeds 3 degrees per second).

[0192] Monitoring implementation: the device type used for monitoring is attitude sensor (such as inertial measurement unit), the collection method of monitoring data is real-time collection (sampling frequency is 10 times per second), and the recording form of monitoring data is time series data (including time stamp and pitch angle value).

[0193] Fuel flow data:

[0194] Focus: the range of state changes of data (±20% of normal taxiing fuel flow), the trend characteristics of data changes (abnormal increase, i.e. fuel flow increases more than 20% of normal taxiing fuel flow within a short time (such as 5 seconds)), and the correlation state requirements of data and engine thrust data (when the thrust increases, the fuel flow abnormally increases).

[0195] Monitoring implementation: the type of device used for monitoring is a fuel flow sensor, the collection method of monitoring data is real-time collection (the sampling frequency is 10 times per second), and the recording form of monitoring data is time series data (including time stamp and fuel flow value).

[0196] Integrating the above monitoring dimension data, focus and monitoring implementation, the data monitoring focus corresponding to the single decision response sub-link is formed:

[0197] Data monitoring focus (take-off phase-combination operation sub-link):

[0198] Engine thrust data:

[0199] Monitoring device: engine thrust sensor.

[0200] Collection method: real-time collection, sampling frequency 10 times per second.

[0201] Recording form: time series data (including time stamp and thrust value).

[0202] Focus: the thrust value changes within the range of ±10% of normal taxiing thrust, if the thrust increases more than 10% of normal taxiing thrust within a short time (5 seconds), and the pitch angle rises at a speed of more than 3 degrees per second at this time, it needs to be focused on.

[0203] Pitch angle data:

[0204] Monitoring device: attitude sensor (inertial measurement unit).

[0205] Collection method: real-time collection, sampling frequency 10 times per second.

[0206] Recording form: time series data (including time stamp and pitch angle value).

[0207] Focus: the pitch angle value changes within the range of 0-20 degrees, if the pitch angle rises at a speed of more than 3 degrees per second within a short time (5 seconds), and the engine thrust increases more than 10% of normal taxiing thrust at this time, it needs to be focused on.

[0208] Fuel flow data:

[0209] Monitoring device: fuel flow sensor.

[0210] Acquisition mode: real-time acquisition, sampling frequency 10 times / second.

[0211] Record form: time series data (including timestamp and fuel flow value).

[0212] Focus: if the fuel flow value changes within ±20% of the normal taxiing fuel flow, and the engine thrust increases more than 10% of the normal taxiing thrust within a short period of time (5 seconds), it needs to be paid attention to.

[0213] Step S154: Add the flight scene unit identifier corresponding to each single decision response sub-link to the flight operation suggestion and data monitoring focus corresponding to the single decision response sub-link. The flight scene unit identifier clearly indicates the flight scene to which the operation suggestion and monitoring focus apply.

[0214] For the flight operation suggestion and data monitoring focus corresponding to the single decision response sub-link in the take-off phase, add the flight scene unit identifier, and the identifier content is "take-off phase", which clearly indicates that the operation suggestion and monitoring focus apply to the take-off phase of the aircraft, i.e. the process from taxiing on the ground to reaching a safe flight altitude (10000 feet).

[0215] Step S155: Add time identifiers to the flight operation suggestion and data monitoring focus. The time identifiers include the time node for starting to execute the operation, the time node for ending to execute the operation, and the time node for starting to monitor, and the time node for ending to monitor.

[0216] Time identifier of flight operation suggestion:

[0217] Time node for starting to execute the operation: immediately start to execute "adjust the control stick operation" when the engine thrust suddenly increases (more than 10% of the normal taxiing thrust) and the pitch angle rises at a speed of more than 3 degrees / second; immediately start to execute "check fuel system operation" when the fuel flow abnormally increases (more than 20% of the normal taxiing fuel flow).

[0218] Time node for ending to execute the operation: end "adjust the control stick operation" when the pitch angle rises at a speed of 2-3 degrees / second; end "check fuel system operation" when the fuel system is checked and judged to be normal (or after taking appropriate fault handling measures).

[0219] Time identifier of data monitoring focus:

[0220] Time node for starting to monitor: start to monitor the engine thrust, pitch angle, and fuel flow data at the beginning of the take-off phase (e.g. when the aircraft starts to taxi).

[0221] The time node for suggesting the end of monitoring: the end of the take-off phase (e.g., when the aircraft reaches a safe flight altitude of 10,000 feet), and the end of monitoring engine thrust, pitch angle, and fuel flow data.

[0222] Step S156: Classify the flight operation suggestions and data monitoring priorities with flight scene unit identification and time identification by flight scene unit, and group the contents under the same flight scene unit.

[0223] Group the flight operation suggestions and data monitoring priorities with "take-off phase" identification and time identification, and classify them separately from the contents of the cruise phase and the landing phase. For example:

[0224] Take-off phase group:

[0225] Flight operation suggestion: combined operation suggestion of "adjusting the steering wheel operation" and "checking the fuel system operation", containing time identification and operation implementation process, core control points.

[0226] Data monitoring priority: monitoring priority of engine thrust, pitch angle, and fuel flow data, containing time identification and monitoring implementation method.

[0227] Cruise phase group:

[0228] Flight operation suggestion: operation suggestion for cruise phase scene decision factor, such as combined operation suggestion of "adjusting engine thrust operation" and "monitoring fuel consumption operation", containing time identification and operation implementation process, core control points.

[0229] Data monitoring priority: monitoring priority of engine thrust, fuel flow, and atmospheric temperature data, containing time identification and monitoring implementation method.

[0230] Landing phase group:

[0231] Flight operation suggestion: operation suggestion for landing phase scene decision factor, such as combined operation suggestion of "adjusting pitch angle operation" and "checking landing gear operation", containing time identification and operation implementation process, core control points.

[0232] Data monitoring priority: monitoring priority of pitch angle, landing gear, and aircraft altitude data, containing time identification and monitoring implementation method.

[0233] Step S157: Sort the flight operation suggestions in each group by execution priority, and the execution priority is determined based on the influence weight of the corresponding influence feature item of the operation suggestion. The operation suggestion with high influence weight of influence feature item is sorted in the front.

[0234] In the take-off phase group, the influence weight of the corresponding influence feature item of "adjust the operation of the control column" is 0.8, and the influence weight of the corresponding influence feature item of "check the operation of the fuel system" is 0.7. According to the influence weight, the flight operation suggestions are prioritized, and the suggestion of "adjust the operation of the control column" is ranked first, and the suggestion of "check the operation of the fuel system" is ranked last.

[0235] Step S158: The data monitoring priorities in each group of content are sorted according to the urgency of attention, and the urgency of attention is determined based on the position of the monitoring dimension data in the influence transmission path. The monitoring dimension data at the front end of the influence transmission path corresponds to the monitoring priority ranked first.

[0236] In the data monitoring priorities of the take-off phase group, the engine thrust data is at the front end of the influence transmission path (thrust changes will directly affect the changes of pitch angle and fuel flow), so the monitoring priority corresponding to the engine thrust data is ranked first; the pitch angle data is in the middle of the influence transmission path (thrust changes cause changes in pitch angle, and changes in pitch angle affect the attitude of the aircraft), and the sorting is second; the fuel flow data is at the back end of the influence transmission path (thrust changes cause changes in fuel flow, and changes in fuel flow affect engine performance), and the sorting is last.

[0237] Step S159: Integrate all sorted flight operation suggestions and data monitoring priorities, arrange them in the order of flight scene units, and present them in the order of operation suggestion priority and monitoring priority urgency under each flight scene unit, to form a structured intelligent flight decision assistance instruction. The content of the intelligent flight decision assistance instruction includes flight scene unit identification, time identification, sorted operation suggestions and monitoring priorities.

[0238] Integrate the flight operation suggestions and data monitoring priorities of the take-off phase, the cruise phase and the landing phase, and arrange them in the order of flight scene units:

[0239] 1. Take-off phase:

[0240] Flight operation suggestions (sorted by priority):

[0241] Priority 1: "Adjust the operation of the control column" suggestion, including time identification (start: thrust mutation and pitch angle speed exceeds threshold; end: pitch angle speed returns to normal), operation implementation process (pull control column speed ≤1 degree / second, amplitude ≤5 degrees, etc.), core control points (control the pull rod speed amplitude, ensure the pitch angle speed to be stable and reduce).

[0242] Priority 2: "Check fuel system operation" suggestion, including time identification (start: fuel flow anomaly; end: fuel system check completed and normal or after handling failure), operation implementation process (check fuel components, monitor fuel instruments, etc.), core control points (carefully check components, pay attention to fuel instrument readings).

[0243] Data monitoring focus (sorted by emergency level):

[0244] Emergency level 1: Engine thrust data monitoring, including time identification (start: takeoff roll; end: reach safe altitude), monitoring implementation (thrust sensor, 10 times / second sampling, time series recording), and attention content (thrust anomaly more than 10% and pitch angle rate more than 3 degrees / second).

[0245] Emergency level 2: Pitch angle data monitoring, including time identification (start: takeoff roll; end: reach safe altitude), monitoring implementation (attitude sensor, 10 times / second sampling, time series recording), and attention content (pitch angle rate more than 3 degrees / second and thrust anomaly more than 10%).

[0246] Emergency level 3: Fuel flow data monitoring, including time identification (start: takeoff roll; end: reach safe altitude), monitoring implementation (fuel flow sensor, 10 times / second sampling, time series recording), and attention content (fuel flow anomaly more than 20% and thrust anomaly more than 10%).

[0247] 2. Cruise phase:

[0248] Flight operation suggestions (sorted by priority):

[0249] Priority 1: "Adjust engine thrust operation" suggestion, including time identification (start: thrust deviates from cruise thrust; end: thrust returns to cruise thrust range), operation implementation process (adjust thrust lever, control thrust change rate, etc.), and core control points (control thrust change rate, ensure aircraft speed stability).

[0250] Priority 2: "Monitor fuel consumption operation" suggestion, including time identification (start: cruise phase begins; end: cruise phase ends), operation implementation process (monitor fuel flow and remaining fuel, calculate fuel consumption rate, etc.), and core control points (pay attention to fuel consumption rate, ensure sufficient fuel to reach destination).

[0251] Data monitoring focus (sorted by emergency level):

[0252] Emergency level 1: Engine thrust data monitoring, including time identification (start: cruise phase begins; end: cruise phase ends), monitoring implementation (thrust sensor, 5 times / second sampling, time series recording), and attention content (thrust deviates from cruise thrust range).

[0253] Urgency Level 2: Fuel Flow Data Monitoring, including Time Identification (Start: Cruise Phase Start; End: Cruise Phase End), Monitoring Implementation (Fuel Flow Sensor, 5 samples / second, Time Series Recording), Concerns (Abnormal Changes in Fuel Flow, Fuel Consumption Rate Deviation from Normal Range).

[0254] Urgency Level 3: Atmospheric Temperature Data Monitoring, including Time Identification (Start: Cruise Phase Start; End: Cruise Phase End), Monitoring Implementation (Atmospheric Temperature Sensor, 5 samples / second, Time Series Recording), Concerns (Abrupt Changes in Atmospheric Temperature, Impact on Engine Thrust and Fuel Consumption).

[0255] 3. Descent Phase:

[0256] Flight Operation Recommendations (Sorted by Priority):

[0257] Priority 1: "Adjust Pitch Angle Operation" Recommendation, including Time Identification (Start: Approach Phase Start; End: Aircraft Touchdown), Operation Implementation Process (Adjust the Control Stick, Control the Pitch Angle within the Approach Range, etc.), Core Control Points (Stable Control of Pitch Angle, Ensure Correct Approach Trajectory).

[0258] Priority 2: "Check Landing Gear Operation" Recommendation, including Time Identification (Start: Approach Phase Start; End: After Aircraft Touchdown), Operation Implementation Process (Check Landing Gear Indicator Lights, Visually Check Landing Gear Deployment, etc.), Core Control Points (Ensure Normal Deployment and Locking of Landing Gear).

[0259] Data Monitoring Focus (Sorted by Urgency):

[0260] Urgency Level 1: Pitch Angle Data Monitoring, including Time Identification (Start: Approach Phase Start; End: Aircraft Touchdown), Monitoring Implementation (Attitude Sensor, 10 samples / second, Time Series Recording), Concerns (Pitch Angle Deviation from Approach Range, Approach Trajectory Abnormality).

[0261] Urgency Level 2: Landing Gear Data Monitoring, including Time Identification (Start: Approach Phase Start; End: After Aircraft Touchdown), Monitoring Implementation (Landing Gear Position Sensor, 5 samples / second, State Recording), Concerns (Landing Gear Not Normally Deployed and Locked).

[0262] Urgency Level 3: Aircraft Height Data Monitoring, including Time Identification (Start: Approach Phase Start; End: Aircraft Touchdown), Monitoring Implementation (Altitude Sensor, 10 samples / second, Time Series Recording), Concerns (Height Deviation from Approach Profile, Abnormal Descent Speed).

[0263] The intelligent flight decision assistance instruction contains each flight scene unit identifier, time identifier, sorted operation suggestion and monitoring focus, assists the pilot to make correct flight decision, and ensures flight safety and high efficiency.

[0264] Based on the same inventive concept, the embodiments of the present application provide a multi-modal flight data based intelligent flight decision assistance system for executing the above-mentioned multi-modal flight data based intelligent flight decision assistance method. The multi-modal flight data based intelligent flight decision assistance system can include a communication unit, a machine readable storage medium and a processor.

[0265] In the embodiments, the machine readable storage medium can also be integrated into the processor and can communicate and interact with external systems through the communication unit. The machine readable storage medium is used to store machine executable instructions for executing the scheme of the present application, and the processor is used to execute the machine executable instructions stored in the machine readable storage medium to realize the multi-modal flight data based intelligent flight decision assistance method provided by the foregoing method embodiments.

[0266] It should be noted that, in order to simplify the description of the present application and to help the understanding of one or more embodiments of the present application, in the foregoing description of the embodiments of the present application, various features are sometimes combined into one embodiment, figure or description thereof.

Claims

1. An intelligent flight decision support method based on multimodal flight data, characterized in that, The method includes: Acquire multi-source flight data, which includes continuous state data recorded by different monitoring dimensions during flight. The continuous state data of different monitoring dimensions have time correlation attributes. The different monitoring dimensions include flight attitude related dimensions recorded by flight instruments, external environment related dimensions collected by environmental sensing devices, and power parameter related dimensions recorded by the power system operation. Extract the performance characteristics of continuous state data of different monitoring dimensions under each flight scenario, and extract the collaborative change characteristics of monitoring dimension data under each flight scenario to form a scenario data association benchmark. Based on the scenario data association benchmark, establish a dynamic association model between the multi-source flight data and the flight scenario. The performance characteristics include the change pattern and fluctuation characteristics of each monitoring dimension data under the corresponding flight scenario, as well as the association performance between different monitoring dimension data. Based on the scenario data association benchmark in the dynamic association model, the influence weight of each monitoring dimension data is determined, the influence transmission characteristics of different monitoring dimension data under each flight scenario are extracted, and the scenario decision factor is generated by combining the influence weight and the influence transmission characteristics. The influence transmission characteristics include the transmission path, duration and manifestation of the influence of changes in monitoring dimension data on the associated affected monitoring dimension data. Establish association rules between influence transmission characteristics and flight operation response logic, filter flight operation response logic that is suitable for scenario decision factors based on the association rules, and connect the filtered flight operation response logics in series according to logical relationships to construct the decision response link corresponding to the scenario decision factors. The flight operation response logic in the decision response chain is transformed into a natural language expression that conforms to the flight operation specifications. The monitoring dimensions and monitoring implementation methods corresponding to each data monitoring focus are clarified. Combining the expression results and monitoring focus under each flight scenario, intelligent flight decision support instructions containing flight operation suggestions and data monitoring focus matched under each flight scenario are output. The process of determining the influence weights of each monitoring dimension based on the scenario data association benchmark in the dynamic association model, extracting the influence transmission characteristics of different monitoring dimension data under each flight scenario, and generating scenario decision factors by combining the influence weights and influence transmission characteristics includes: For each flight scenario unit, the scene data association benchmark corresponding to the flight scenario unit is extracted from the dynamic association model. The scene data association benchmark includes the performance characteristics of each monitoring dimension data under the flight scenario unit and the association pattern between dimensions. Based on the correlation pattern, the correlation strength between the data of each monitoring dimension is determined, and the initial influence weight of each monitoring dimension is determined according to the correlation strength. The correlation strength includes the degree of synchronization of data changes and the degree of mutual dependence of data values. When the correlation strength is high, the mutual influence between the monitoring dimension data is more significant. The initial influence weight is used to characterize the importance of the change of the monitoring dimension data on the monitoring dimension data that are related to it. The monitoring dimensions whose initial influence weights meet the preset importance requirements under the flight scenario unit are selected as the core influence dimensions, and the other monitoring dimensions are selected as the affected dimensions. The influence of the state changes of the core influence dimension data on the state of the affected dimension data is analyzed. After tracking the changes in the state of the core impact dimension data, the process of changes in the state of the affected dimension data is determined. The transmission path of the impact is identified, and the duration of the impact is recorded as the time from the start of the change in the state of the core impact dimension data to the stabilization of the change in the state of the affected dimension data. The state changes of the affected dimension data during the impact process are observed to determine the manifestation of the impact. The transmission path includes a direct transmission path and an indirect transmission path. The direct transmission path is the path in which the core impact dimension data directly affects the affected dimension data. The indirect transmission path is the path in which the core impact dimension data affects the monitored dimension data located between it and the affected dimension data in the transmission path. The duration distinguishes the different durations corresponding to the direct and indirect transmission paths. The transmission path, duration, and form of manifestation are integrated with the corresponding core impact dimensions, affected dimensions, and initial impact weights to form structured impact feature items; All impact features under the same flight scenario unit are classified, and the impact features are divided into multiple impact groups according to the different core impact dimensions. Each impact group contains all impact features corresponding to the same core impact dimension. Analyze the impact coverage of each impact feature item within each impact group, and adjust the initial impact weight of the impact feature item based on the impact coverage. The impact coverage includes the number of affected dimensions involved in the impact feature item. All the adjusted influencing features are integrated to form the initial scenario decision factors corresponding to the flight scenario unit; By comparing the initial scenario decision factors of different flight scenario units, the differences in influence characteristics between similar flight scenario units are identified, and the influence characteristic items and influence weights of similar flight scenario units are adjusted so that the scenario decision factors can distinguish the influence transmission characteristics of different flight scenario units. The initial scenario decision factors after the adjustment of all flight scenario units are integrated to form a complete set of scenario decision factors. Each scenario decision factor corresponds to a flight scenario unit and the influence transmission characteristics under that flight scenario unit.

2. The intelligent flight decision support method based on multimodal flight data according to claim 1, characterized in that, The process involves extracting the performance characteristics of continuous state data from different monitoring dimensions under various flight scenarios, extracting the collaborative change characteristics of monitoring dimension data under various flight scenarios, forming a scenario data association benchmark, and establishing a dynamic association model between the multi-source flight data and flight scenarios based on the scenario data association benchmark, including: The flight scenario types are divided into multiple flight scenario units. Each flight scenario unit corresponds to a flight phase or flight environment condition that is predefined in the flight operation procedures and has clear boundaries. The flight phase includes takeoff phase, cruise phase, and landing phase. The flight environment condition includes normal weather conditions and complex weather conditions. For each flight scenario unit, historical operation records of multi-source flight data under that flight scenario unit are collected. The historical operation records contain the complete change process of continuous state data of different monitoring dimensions under that flight scenario unit, and the time interval of the records is kept consistent. The continuous state data of each monitoring dimension in the historical operation record are time-synchronized. The performance characteristics of each monitoring dimension data are extracted from the synchronized historical operation record. The performance characteristics include change patterns, fluctuation characteristics and correlation characteristics. The change patterns include the data change trend with the flight process and the state characteristics of the data at key flight nodes. The fluctuation characteristics include the value distribution range of the data during the stable operation period and the fluctuation performance of the data during the state transition period. The correlation characteristics include the change synchronization of data of different monitoring dimensions at the same time node. The collaborative change characteristics among the data of each monitoring dimension are analyzed, and the correlation pattern of the monitoring dimension data under each flight scenario unit is determined based on the collaborative change characteristics. The collaborative change characteristics include the correspondence relationship of the change direction of different monitoring dimension data and the correlation performance of the change magnitude. The performance characteristics and correlation patterns are integrated to form a scene data correlation benchmark for each flight scenario unit. The scene data correlation benchmark includes the performance characteristic parameters of each monitoring dimension data and the correlation pattern description between dimensions. Establish the correspondence between the scenario data association benchmark of each flight scenario unit and the monitoring dimension data in the multi-source flight data to form a preliminary association model. The preliminary association model includes the matching conditions between the scenario data association benchmark and the monitoring dimension data. Substitute the continuous state data from the multi-source flight data into the preliminary correlation model, check the matching of the continuous state data with the correlation benchmark of the scene data of each flight scenario unit, and adjust the correspondence between the scene data correlation benchmark and the monitoring dimension data so that the continuous state data can match the corresponding flight scenario unit. The adjustment process is repeated until any continuous state data segment in the multi-source flight data can be matched with a unique corresponding flight scenario unit through the association model. The adjusted association models corresponding to all flight scenario units are integrated to form a dynamic association model between multi-source flight data and flight scenarios.

3. The intelligent flight decision support method based on multimodal flight data according to claim 2, characterized in that, The extraction of performance characteristics of each monitoring dimension data from the synchronized historical operation records includes: For continuous state data of each monitoring dimension, data change curves are plotted in chronological order of flight process. The data change trend with flight process is identified from the data change curves. The change trend includes continuous upward trend, continuous downward trend, stable trend, fluctuating upward trend, and fluctuating downward trend. Mark the key flight nodes on the data change curve, record the monitoring dimension data values ​​corresponding to each key flight node, and form the state characteristics of the data at the key flight nodes. The key flight nodes include the nodes where the flight state changes under the flight scenario unit. The monitoring dimension data is divided into stable operating periods and state transition periods from historical operation records. The values ​​of the monitoring dimension data within the stable operating periods are statistically analyzed to determine the value distribution range of the data within the stable operating periods. The stable operating period is the period in which the data values ​​do not fluctuate significantly. The state transition period is the period in which the data values ​​transition from one stable range to another. The value distribution range includes the common range of data values ​​within the period. Analyze the changes in the monitored dimension data during the state transition period, record the changes in the fluctuation frequency and amplitude of the data during this period, and form the fluctuation performance of the data during the state transition period; Select any two continuous state data from different monitoring dimensions, extract the changes in their values ​​at the same time point, determine whether the direction of change is consistent and whether there is a correlation in the magnitude of change, and form a synchronization of changes in data from different monitoring dimensions at the same time point; The changes, key flight node status characteristics, stable operating period value distribution range, fluctuation performance during state transition periods, and change synchronization are integrated to form the complete performance characteristics of the monitoring dimension data under the corresponding flight scenario unit. Repeat the above steps until the complete performance features of all monitoring dimensions under the corresponding flight scenario unit are extracted. The complete performance features of all monitoring dimensions together constitute the performance feature set of multi-source flight data under that flight scenario unit.

4. The intelligent flight decision support method based on multimodal flight data according to claim 1, characterized in that, The process of tracking changes in the state of data in the core impact dimension, and then determining the transmission path of the impact, includes: Select the next core impact dimension data of the flight scenario unit, record the initial stable state of the core impact dimension data, and simulate the state change of the core impact dimension data of a preset type. The preset type of state change includes value increase, value decrease, fluctuation amplitude increase, and fluctuation amplitude decrease. While simulating state changes, the state change process of all affected dimension data under the flight scenario unit is recorded in real time, and the time node when each affected dimension data begins to change state is recorded in chronological order. By comparing the time points when the data of each affected dimension began to change state with the time points when the data of the core affected dimension began to change state, the order in which the data of each affected dimension responded to the impact can be determined. Analyze the data of the affected dimension that responds first to the impact, and check whether there is a direct physical or logical relationship between the affected dimension data and the core impact dimension data. If there is, determine that the path from the core impact dimension data to the affected dimension data is a direct transmission path. For affected dimension data that is not the first to respond to the impact, analyze whether the time node when it begins to change state is sequentially connected with the time node when other affected dimension data begins to change state. If the time node when a certain affected dimension data begins to change state is later than that of another affected dimension data, and there is a physical or logical connection between the two, then the path through which the core affected dimension data acts on the affected dimension data that responds first to respond later is determined to be an indirect transmission path. Record the affected dimension data that plays an intermediate role in the indirect transmission path, and clarify the complete transmission links of the indirect transmission path; Draw the impact propagation path diagram corresponding to the core impact dimension data. The impact propagation path diagram should clearly mark the core impact dimension, the affected dimension, the intermediate propagation dimension, and the direction of impact propagation between each dimension. Replace the core impact dimension data, repeat the above operations of simulating state changes, recording changes in affected dimensions, determining response order, analyzing transmission paths, and drawing impact transmission path diagrams, until the impact transmission paths corresponding to all core impact dimension data under this flight scenario unit are determined. Integrate the impact transmission path diagrams corresponding to all core impact dimensions to form a complete set of impact transmission paths for this flight scenario unit.

5. The intelligent flight decision support method based on multimodal flight data according to claim 1, characterized in that, The process of establishing association rules between influence transmission characteristics and flight operation response logic, selecting flight operation response logics suitable for scenario decision factors based on the association rules, and connecting the selected flight operation response logics in series according to logical relationships to construct the decision response chain corresponding to the scenario decision factors includes: Collect flight operation response logic, which includes response operation procedures for the impact transmission characteristics of data from different monitoring dimensions. Each flight operation response logic corresponds to a set of impact response requirements that are predefined in the flight operation standards and have clear response objectives. For each flight operation response logic, feature extraction is performed to obtain operation logic features. The operation logic features include the type of impact targeted by the response logic, the monitoring dimensions involved, the operation execution process, and the performance of the operation target. The influence transmission characteristics of scenario decision factors are described in a structured manner to obtain influence feature description items, which include the influence source dimension, the affected dimension, the transmission path type, the influence manifestation form, and the influence duration. Establish association rules between impact feature description items and operational logic features. Based on the association rules, construct an association mapping table between impact transmission features and flight operation response logic. The association rules include matching requirements between each item in the impact feature description item and each item in the operational logic feature. The matching requirements include consistency between the impact source dimension and the monitoring dimension involved in the operational logic feature, and adaptation between the impact manifestation form and the impact type targeted by the operational logic feature. Each impact feature description item in the association mapping table corresponds to at least one adapted flight operation response logic identifier. For each scenario decision factor, extract all impact feature description items of that scenario decision factor, and query the corresponding adaptive flight operation response logic identifier for each impact feature description item based on the association mapping table; Based on the influence weight of the influencing feature items in the scenario decision factor, the queried adaptive flight operation response logic identifiers are prioritized and sorted. The flight operation response logic corresponding to the top-ranked adaptive flight operation response logic identifier is selected as the candidate operation response logic for the scenario decision factor. The flight operation response logic identifiers corresponding to the influencing feature items with high influence weights are ranked higher. Analyze the logical relationships between candidate operation response logics to determine whether there are execution order requirements or functional complementarity relationships between different candidate operation response logics; If there is an execution order requirement, the execution order is determined based on the execution prerequisites and results of the operation logic, and the candidate operation response logic is arranged in order. If there is a complementary relationship between functions and the execution processes do not interfere with each other, the complementary candidate operation response logic will be integrated into a combined operation response logic, and the collaborative execution method of the two will be clarified during the integration process. Determine the execution steps of candidate operation response logic or combined operation response logic, and clarify the operation content, operation timing and required multi-source flight data support conditions corresponding to each execution step. The support conditions include the monitoring dimension data and data status requirements that need to be paid attention to when executing the execution step. The scenario decision factors, candidate operation response logic or combined operation response logic, and corresponding execution steps and supporting conditions are logically associated to form a single decision response sub-link. The single decision response sub-link clearly defines the complete logical path from the scenario decision factors to the operation response. The above-described operation is performed on all scenario decision factors to obtain the influencing feature descriptions and form a single decision response sub-link; All single decision-response sub-links are classified according to flight scenario units. Single decision-response sub-links under the same flight scenario unit are grouped together. The triggering relationship between different single decision-response sub-links under the same flight scenario unit is analyzed. Single decision-response sub-links with conditional triggering or temporal connection are identified. Single decision-response sub-links are connected in series according to the triggering relationship. The integrated and connected single decision response sub-links form the decision response sub-links corresponding to the scenario decision factors under the flight scenario unit. The decision response sub-links corresponding to all flight scenario units are integrated to form the complete decision response link corresponding to the scenario decision factors. The complete decision response link contains the complete association path from decision factors to operation responses under each flight scenario, and the switching triggering conditions between different flight scenario sub-links are clearly defined.

6. The intelligent flight decision support method based on multimodal flight data according to claim 5, characterized in that, The analysis of the logical relationships between candidate operation response logics, and the determination of whether different candidate operation response logics have execution order requirements or functional complementarity, includes: Extract the execution preconditions and execution result performance of each candidate operation response logic. The execution preconditions include the monitoring dimension data status that must be met before executing the candidate operation logic, and the execution result performance includes the expected changes in the monitoring dimension data status after executing the candidate operation logic. Compare the execution result of the first candidate operation response logic with the execution preconditions of the second candidate operation response logic, and determine whether the execution result of the first candidate operation response logic can satisfy the execution preconditions of the second candidate operation response logic. If they are satisfied, it is determined that the two candidate operation response logics have a sequential execution order, and the first candidate operation response logic must be executed before the second candidate operation response logic. By comparing the execution preconditions and execution results of all candidate operation response logics in the above manner, an execution order relationship table of candidate operation response logics is constructed, and the execution order relationship table clearly marks the execution order of each operation logic; For candidate operation response logics that do not have a sequential execution order, analyze whether their execution objectives have functional complementarity. The functional complementarity includes whether the affected dimensions targeted by the operation logic are different but jointly serve the response to the same core affected dimension, or whether the affected manifestations targeted by the operation logic are different but jointly cover all types of the core affected dimension. If two candidate operation response logics target different affected dimensions but jointly serve the impact response of the same core affected dimension, and their execution processes do not interfere with each other, then it is determined that the two have a complementary functional relationship. If two candidate operation response logics target different forms of impact but jointly cover all impact types of the core impact dimension, and their execution processes can be coordinated and connected, then it is determined that the two have a complementary functional relationship. Candidate operation response logics with complementary functions are integrated to form combined operation response logic. During the integration process, the execution method and coordination requirements of each operation logic are clarified. Repeat the above functional complementarity judgment and integration operations until all candidate operation response logics that do not have a sequential execution order but have functional complementarity are integrated into combined operation response logic; Check whether there is a logical conflict between the integrated combined operation response logic and the candidate operation response logic that has a sequential execution order. The logical conflict includes contradictions in the execution preconditions and mutual interference in the execution results. If a logical conflict exists, re-analyze the execution preconditions and execution result performance of the candidate operation response logic, adjust the execution parameters of the operation logic, or select other suitable candidate operation response logic to eliminate the logical conflict. Ultimately, an operation logic set is formed, which includes candidate operation response logic arranged in execution order and integrated combined operation response logic. There are no logical conflicts between the operation logics in this operation logic set, and they can work together to cope with the influence transmission characteristics of scenario decision factors.

7. The intelligent flight decision support method based on multimodal flight data according to claim 1, characterized in that, The process involves transforming the flight operation response logic in the decision-making response chain into natural language expressions that conform to flight operation specifications, clarifying the monitoring dimensions and implementation methods corresponding to each data monitoring focus, and combining the expression results and monitoring focuses under each flight scenario to output intelligent flight decision-making assistance instructions that include matching flight operation suggestions and data monitoring focuses for each flight scenario, including: Traverse each individual decision response sub-link in the decision response chain and extract the flight scenario unit, operation response logic and supporting conditions corresponding to each individual decision response sub-link; For each single decision response sub-link, the operation response logic is broken down into its execution steps. Each execution step is transformed into a natural language expression that conforms to flight operation specifications. Key operation points are extracted from the execution step expressions of the operation response logic. These key operation points are integrated with the corresponding execution step expressions to form flight operation suggestions for that operation response logic. The expression clearly defines the specific implementation content of the operation, the timing of the operation, the duration of the operation, and the flight parameters that need to be controlled during the operation. The key operation points include the implementation links that need to be focused on during the operation, the control requirements of the operation parameters, and the criteria for judging the operation effect. The flight operation suggestions include a complete operation implementation process and core control points. Extract supporting conditions from each individual decision response sub-link, determine the key monitoring dimensions from these conditions, clarify the focus of each monitoring dimension, and determine the monitoring implementation method for each monitoring dimension. Integrate the monitoring dimension data, focus, and monitoring implementation method to form the data monitoring focus corresponding to that individual decision response sub-link. The focus includes the range of data state changes, the trend characteristics of data changes, and the association state requirements between the data and the determined associated monitoring dimension data. The monitoring implementation method includes the type of equipment used for monitoring, the method of data acquisition, and the recording format of the monitoring data. The monitoring implementation method is adapted to the focus and can accurately obtain the data required for the focus. For each single decision response sub-link, add a flight scenario unit identifier corresponding to that single decision response sub-link for the flight operation suggestions and data monitoring focus. The content of the flight scenario unit identifier clearly indicates the flight scenario to which the operation suggestions and monitoring focus are applicable. Add time markers to flight operation recommendations and data monitoring priorities. The time markers include the time nodes when the recommended operation starts and when the recommended operation ends, as well as the time nodes when the recommended monitoring starts and when the recommended monitoring ends. Flight operation suggestions and data monitoring priorities with flight scenario unit identifiers and time identifiers are classified according to flight scenario units, and the content under the same flight scenario unit is grouped together. For each group of content, flight operation suggestions are sorted by execution priority. The execution priority is determined based on the influence weight of the influence feature item corresponding to the operation suggestion. Operation suggestions with higher influence weight of the influence feature item are sorted first. The data monitoring priorities in each group are sorted by their urgency. The urgency is determined based on the position of the monitoring dimension data in the impact transmission path. The monitoring priorities corresponding to the monitoring dimension data at the front of the impact transmission path are sorted first. The system integrates all categorized and sorted flight operation suggestions and data monitoring priorities, arranging them in the order of flight scenario units. The content under each flight scenario unit is presented in order of priority of operation suggestions and urgency of monitoring priorities, forming a structured intelligent flight decision support instruction. The intelligent flight decision support instruction includes flight scenario unit identifier, time identifier, sorted operation suggestions and monitoring priorities.

8. An intelligent flight decision support system based on multimodal flight data, characterized in that, include: processor; A machine-readable storage medium for storing machine-executable instructions of the processor; The processor is configured to execute the intelligent flight decision support method based on multimodal flight data according to any one of claims 1 to 7 by executing the machine-executable instructions.

9. A computer program product, characterized in that, The computer program product includes machine-executable instructions stored in a computer-readable storage medium. A processor of a computer device reads the machine-executable instructions from the computer-readable storage medium and executes the machine-executable instructions, causing the computer device to perform the intelligent flight decision support method based on multimodal flight data as described in any one of claims 1 to 7.

Images