Flight test data analysis method and system based on cloud platform

Abstract

The invention relates to the technical field of data analysis, in particular to a test flight data analysis method and system based on a cloud platform, and the method comprises the following steps: obtaining data, sorting logs according to sorties and timestamps, judging an undercarriage, speed and normal overload based on a time sequence event and a cumulative duration rule, generating a key value pair, and calculating the test flight data; the grouping event key value pair is input into the finite state automaton to output profile violation, the grouping duration key value pair calculates the total overrun duration to generate an accumulated overrun report, and the accumulated overrun report is combined to form a comprehensive report. According to the method, a time sequence log is constructed by collecting test flight sorties and timestamps, automatic data regulation is achieved, the undercarriage state, the flight speed and accumulated normal overload are judged by means of a time sequence event association rule, manual scripts are replaced, an event structure is input into a finite state automaton, violation logic state transition is matched, and the reliability of the undercarriage is improved. According to the method, the violation of the complex section is automatically identified, the total accumulated overrun duration is calculated, and the comprehensive violation verification report is combined and output, so that the efficiency and accuracy of data analysis are remarkably improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data analysis, in particular to a flight test data analysis method and system based on a cloud platform. BACKGROUND

[0002] The technical field of data analysis is a process of checking, cleaning, transforming and modeling data, aiming to discover useful information, inform conclusions and support decision-making. The core of this field includes extracting patterns, correlations and insights from massive raw data sets using statistical methods, computer science algorithms and domain-specific expertise to form a systematic knowledge system. Among them, the traditional flight test data analysis method refers to the post-processing of parameters collected by the on-board data recorder during the flight test of the aircraft. The commonly used way is to copy the data stored in the on-board device to the ground analysis workstation through physical medium, and then the analysis personnel write specific data analysis scripts, call state parameter calibration functions and time domain data alignment functions, and calculate and verify the dispersedly stored flight parameter files one by one.

[0003] The traditional flight test data analysis seriously depends on physical medium to copy data, resulting in lagging analysis process. The analysis personnel need to write specific data analysis scripts for the dispersedly stored flight parameter files, and manually call state parameter calibration and time domain data alignment functions. This way of calculating and verifying one by one not only consumes time and effort, but also is difficult to standardize. When facing complex time sequence event correlation, such as judging the violation flight profile under multiple parameter combinations, it is easy to miss and cannot guarantee the completeness and timeliness of the verification by manually writing scripts. SUMMARY

[0004] The purpose of the present application is to solve the shortcomings in the prior art, and to provide a flight test data analysis method and system based on a cloud platform.

[0005] In order to achieve the above purpose, the present application adopts the following technical scheme: a flight test data analysis method based on a cloud platform, comprising the following steps: S1: obtaining an original data packet, extracting a flight test session identifier, an absolute timestamp, a landing gear state, a flight speed and a normal overload, collecting and arranging in ascending order of the absolute timestamp based on the flight test session identifier to form an ordered time sequence data log; S2: based on the ordered time sequence data log, generating an event key-value pair set according to the time sequence event correlation rule of the landing gear state and the flight speed, and generating a time length key-value pair set according to the cumulative time length limit rule of the normal overload; S3: group the event key-value pair set according to the flight test flight number identifier, obtain an event structure list sorted according to the absolute time stamp, input the event structure list into a finite state automaton, match a time window and a rule violation logic state transition based on a time sequence event association rule, and output a profile violation record; S4: group the time length key-value pair set according to the flight test flight number identifier and calculate a total accumulated over-limit time length, judge whether the total accumulated over-limit time length exceeds a maximum allowed accumulated time length of an accumulated time length limit rule to generate an accumulated over-limit violation report, and combine the profile violation record and the accumulated over-limit violation report to form a comprehensive violation checking report.

[0006] As a further scheme of the present application, the ordered time sequence data log includes a time sequence data stream, a data frame structure, and a sorting index, the event key-value pair set includes an event identifier, an associated time point, and a state snapshot, the time length key-value pair set includes an over-limit type, a duration period, and an overload peak, the event structure list includes a flight phase division, an event sequence, and a logical association, the profile violation record includes a violation type code, a violation occurrence time, and a violation scenario reproduction, the accumulated over-limit violation report includes a total time length statistics, a judgment threshold, and an over-limit conclusion, and the comprehensive violation checking report includes a checking summary, profile violation details, and accumulated statistics results.

[0007] As a further scheme of the present application, the ordered time sequence data log includes a time sequence data stream, a data frame structure, and a sorting index, the event key-value pair set includes an event identifier, an associated time point, and a state snapshot, the time length key-value pair set includes an over-limit type, a duration period, and an overload peak, the event structure list includes a flight phase division, an event sequence, and a logical association, the profile violation record includes a violation type code, a violation occurrence time, and a violation scenario reproduction, the accumulated over-limit violation report includes a total time length statistics, a judgment threshold, and an over-limit conclusion, and the comprehensive violation checking report includes a checking summary, profile violation details, and accumulated statistics results. S101: collect bus interface transmission original data packets, perform frame header verification and data field analysis on the original data packets, locate a flight test flight number identifier field, an absolute time stamp field, a landing gear state field, a flight speed field, and a normal overload field according to a preset data dictionary, extract original values corresponding to the flight test flight number identifier field, the absolute time stamp field, the landing gear state field, the flight speed field, and the normal overload field from the original data packet data field, and generate instantaneous flight state data; S102: call the instantaneous flight state data, traverse the flight test flight number identifier in the instantaneous flight state data, set the flight test flight number identifier comparison reference, aggregate the instantaneous flight state data entries with the same flight test flight number identifier to the same data set, and attach the corresponding flight test flight number identifier as an index to each data set, and obtain a flight number collection data body; S103: according to the flight number collection data body, retrieve the flight number collection data body index one by one, extract multiple entry absolute time stamps inside the same flight test flight number identifier index, set an ascending order sorting rule of the absolute time stamp, call the ascending order sorting rule to perform rearrangement on the data entries inside the flight number collection data body, and establish an ordered time sequence data log.

[0008] As a further scheme of the present application, the step of S2 is specifically: S201: Separate the landing gear state sequence and the flight speed sequence from the multi-dimensional parameter data stream in the ordered time sequence data log, call the time sequence event association rule, perform time-by-time synchronous comparison and judgment on the landing gear state sequence and the flight speed sequence, filter time points whose numerical values meet the built-in logic conditions of the rule, and collect all time points to generate a flight event timestamp sequence; S202: Take the flight event timestamp sequence as an index to search for corresponding landing gear state values and flight speed values in the ordered time sequence data log, construct a record unit with a time point as a unique key name and a binary tuple composed of the landing gear state value and the flight speed value as a key value, aggregate all record units to establish an event key-value pair set; S203: According to the key names contained in the event key-value pair set, divide the corresponding normal overload data segments in the ordered time sequence data log, integrate and calculate the cumulative time length of sample points whose data values exceed the normal overload threshold in multiple data segments, call the cumulative time length limit rule, compare the obtained cumulative time length with the time length limit reference value, filter the time lengths that exceed the time length limit reference value, generate a new key-value pair with the event key as the index and the time length as the value, and obtain a time length key-value pair set.

[0009] As a further scheme of the present application, the step of obtaining the profile violation record is specifically: S301: Extract the test flight flight number identifier in each key-value pair record in the event key-value pair set, take the test flight flight number identifier as a unique grouping key, traverse and aggregate the event key-value pair set, merge all key-value pair records holding the same test flight flight number identifier into the same data subset, and establish a flight event grouping set; S302: For each data subset in the flight event grouping set, extract the absolute time stamps corresponding to all key-value pair records in the data subset, and take the absolute time stamps as the basis for sorting, perform ascending arrangement operation on the key-value pair records in the data subset, construct an event sequence, and obtain an event structure list; S303: Load the event structure list into a finite state automaton in sequence order, and obtain a preset time window threshold and a violation logic state transition matrix in the time sequence event association rule, calculate the absolute time stamp difference of adjacent two events when the state machine jumps between differentiated states, judge whether the difference value falls within the time window threshold interval, and compare whether the current state jump path is consistent with the path defined in the violation logic state transition matrix, if both conditions are met, record the complete event information of the violation transition, and output the profile violation record.

[0010] As a further scheme of the present application, the step of obtaining the comprehensive violation check report is specifically: S401: performing grouping operation on the time length key-value pair set according to the flight test flight number identifier, aggregating key-value pair records with the same flight test flight number identifier into flight data subsets, simultaneously traversing each flight data subset, performing arithmetic accumulation on the time length values of all key-value pairs in the flight data subset, calculating the total accumulated over-limit time length of the flight, and generating flight total over-limit time length statistics with the flight test flight number identifier as the key; S402: calling the flight total over-limit time length statistics, retrieving the maximum allowed accumulated time length from the accumulated time length limit rule, traversing all entries in the flight total over-limit time length statistics, performing numerical size judgment on the total accumulated over-limit time length of each entry record and the maximum allowed accumulated time length, screening all entries with the total accumulated over-limit time length greater than the maximum allowed accumulated time length, forming an accumulated over-limit violation report; S403: aligning the profile violation record and the accumulated over-limit violation report according to the flight test flight number identifier as the associated primary key, merging records holding the same flight test flight number identifier, and retaining all independent records in the two reports that are not aligned, then uniformly formatting all merged and independent records to form a comprehensive violation check report.

[0011] As a further scheme of the present application, the matching process of the finite state automaton is specifically: predefining a flight state set, the flight state set including a take-off phase, a climb phase, a cruise phase, a maneuver phase, a descent phase and a landing phase, and defining the violation logic state transition matrix, the violation logic state transition matrix storing illegal state transition triplets composed of a source state, a target state and a violation condition; loading the event structure list, reading key-value pair records in the event structure list one by one, analyzing state snapshots corresponding to the key-value pair records, mapping the state snapshots to target flight states in the flight state set as current states; obtaining the absolute timestamp of the key-value pair record, and calculating the difference value with the absolute timestamp of the previous event, judging whether the difference value falls within the time window threshold interval; If the difference value falls within the time window threshold interval, it is determined that the flight state of the previous event is the source state and the current state is the target state, the violation logic state transition matrix is retrieved, and it is compared whether the jump from the source state to the target state exists in the illegal state transition triplet, if it exists, it is determined as a violation transition, and the event identifier, associated time point and violation type code of the key-value pair record are extracted to generate the profile violation record.

[0012] As a further aspect of the present invention, the determination step of the cumulative duration limit rule is specifically as follows: Extract continuous normal overload sequences from the ordered time-series data log, and obtain the preset normal overload threshold in the cumulative duration limit rule. With minimum duration ; Traverse the normal overload sequence and identify all values ​​that continuously exceed the normal overload threshold. The data points are divided into multiple instantaneous over-limit data segments according to the continuity of absolute timestamps, and the duration of each instantaneous over-limit data segment is calculated. Filtering the duration to be greater than the minimum duration The instantaneous over-limit data segments are used to form a set of effective over-limit data segments, and the overload peak value in each effective over-limit data segment is extracted to construct the set of duration key-value pairs. The set of duration key-value pairs is grouped according to the test flight sortie identifier. The duration of all valid over-limit data segments under the same test flight sortie identifier is accumulated to calculate the total cumulative over-limit duration. The calculation logic is as follows: ; in, The total number of valid over-limit data segments corresponding to the test flight identifier. For the first The starting absolute timestamp of each of the aforementioned valid out-of-limit data segments. For the first The absolute timestamp of the end of each of the aforementioned valid over-limit data segments, and ; Call the maximum allowed cumulative duration and compare it with the With the maximum allowed cumulative duration, if the If the duration exceeds the maximum allowed cumulative duration, a cumulative violation report will be generated.

[0013] As a further aspect of the present invention, the determination step of the time-series event association rule is specifically as follows: The landing gear state sequence and flight speed sequence are synchronously extracted from the ordered time-series data log, and the preset landing gear state threshold in the time-series event association rules is obtained. Flight speed threshold and related weighting factors and ; The landing gear state sequence and the flight speed sequence are synchronously compared time-by-time at any absolute timestamp. At this point, calculate the event trigger score. The calculation formula is: ; in, For the Landing gear status values ​​at any given time. For the The flight speed value at any given moment. For the landing gear state matching function, when the Trigger the It returns 1 if the preset condition is met, otherwise it returns 0. For the flight speed matching function, when the Trigger the It returns 1 if the preset condition is met, otherwise it returns 0. and The associated weight factor and .

[0014] A cloud-based flight test data analysis system, wherein the cloud-based flight test data analysis system is used to implement the above-mentioned cloud-based flight test data analysis method, the system comprising: The time-series log generation module is used to acquire raw data packets, extract test flight sortie identifiers, absolute timestamps, landing gear status, flight speed and normal overload, collect them based on the test flight sortie identifiers and arrange them in ascending order according to the absolute timestamps to form an ordered time-series data log, and transmit the ordered time-series data log to the key-value pair parsing module; The key-value pair parsing module is used to determine the landing gear status and the flight speed according to the time-series event association rules through the ordered time-series data log, generate an event key-value pair set according to the cumulative duration limit rules, determine the normal overload and generate a duration key-value pair set, and pass the event key-value pair set to the profile violation matching module and the duration key-value pair set to the comprehensive verification report module. The profile violation matching module is used to group the event key-value pair set according to the test flight number identifier, obtain the event structure list sorted by the absolute timestamp, input the event structure list into the finite state automaton, match the time window of the time sequence event association rule with the violation logic state transition, output the profile violation record, and pass the profile violation record to the comprehensive verification report module. The comprehensive verification report module is used to group the set of duration key-value pairs according to the test flight identifier and calculate the total cumulative over-limit time, determine whether the total cumulative over-limit time exceeds the maximum allowable cumulative time of the cumulative time limit rule, generate a cumulative over-limit violation report, and merge the profile violation record and the cumulative over-limit violation report to form a comprehensive violation verification report.

[0015] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In the present application, by collecting key parameters such as flight test flight number identification and absolute time stamp, an ordered time sequence data log is constructed, the automatic regularization of the original data is realized, the subsequent time sequence event association rule is used to judge the landing gear state and flight speed, and the normal overload is judged according to the cumulative time limit rule. This method replaces the tedious manual script analysis, by inputting the event structure list to the finite state automaton, matching the illegal logic state transition, which can automatically identify complex profile violations, and calculate the total cumulative over-limit time, to ensure the comprehensiveness of the analysis, and finally output the comprehensive violation check report, to improve the efficiency and accuracy of data analysis. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1 The present application is a flight test data analysis method flow chart; Figure 2 The present application is an ordered time sequence data log acquisition flow chart; Figure 3 The present application is an event and time key-value pair set generation flow chart; Figure 4 The present application is a profile violation record acquisition flow chart; Figure 5 The present application is a comprehensive violation check report generation flow chart. DETAILED DESCRIPTION

[0017] In order to make the purpose, technical scheme and advantages of the present application clearer, the technical scheme realized by software is described in detail below combined with system architecture diagram and embodiment. It should be understood that the specific embodiments described herein are only used to explain the technical scheme of the present application, and do not constitute a limitation on the protection scope.

[0018] In the description of the present application, the system architecture relationship or data processing flow indicated by the terms "level", "module", "interface", "data flow", "client", "server" and the like are defined based on the corresponding architecture diagram or flow chart of the embodiment. This description method is only used to clearly explain the logical relationship of each element in the technical scheme, and is not limited to the physical deployment form. The "multiple" contains two or more technical units, including but not limited to multiple data nodes, processing threads, service instances or functional components, and other expandable elements, and the specific number is determined according to the actual business scenario.

[0019] Please refer to Figure 1 and Figure 2 The present application provides a technical scheme: a flight test data analysis method based on cloud platform, comprising the following steps: S1: Obtain the original data packet, extract the test flight frame number identifier, absolute timestamp, landing gear state, flight speed and normal overload, collect and arrange in ascending order of absolute timestamp based on the test flight frame number identifier to form an ordered time sequence data log; The ordered time sequence data log acquisition step is specifically: S101: Collect the bus interface transmission original data packet, perform frame header verification and data field analysis on the original data packet, locate the test flight frame number identifier field, absolute timestamp field, landing gear state field, flight speed field and normal overload field according to the preset data dictionary, extract the test flight frame number identifier field, absolute timestamp field, landing gear state field, flight speed field and normal overload field corresponding original value from the original data packet data field, and generate instantaneous flight state data; S102: Call the instantaneous flight state data, traverse the test flight frame number identifier in the instantaneous flight state data, set the test flight frame number identifier comparison reference, aggregate the instantaneous flight state data entries with the same test flight frame number identifier to the same data set, and add the corresponding test flight frame number identifier as the index for each data set, and obtain the frame number collection data body; S103: According to the frame number collection data body, retrieve the frame number collection data body index one by one, extract multiple entry absolute timestamps inside the same test flight frame number identifier index, set the ascending order sorting rule of the absolute timestamp, call the ascending order sorting rule to perform rearrangement on the data entries inside the frame number collection data body, and establish an ordered time sequence data log; The ordered time sequence data log includes time sequence data stream, data frame structure and sorting index.

[0020] S101: Real-time collect data frames from the bus interface (such as AFDX bus or ARINC429 bus), and the data frames are original data packets represented in hexadecimal. For example, an original data packet is received: [0x54462D32303235313032372D3031][0x00000181A1B2C3D4][0x02][0x10C0][0x0119][0xCRC16]. Frame header check is performed on this packet to confirm its origin and integrity, e.g. checking CRC-16 field [0xCRC16]. After passing the check, the data field [0x5446...][0x0000...][0x02][0x10C0][0x0119] is parsed. According to a pre-defined data dictionary (which defines the byte offset, length and parsing rule of a specific parameter in the data field), the key fields are located. The data dictionary is defined as follows: 1. Test flight cycle identification field: offset 8, length 12 bytes, ASCII encoding. 2. Absolute timestamp field: offset 20, length 8 bytes, 64-bit Unix timestamp (milliseconds). 3. Landing gear state field: offset 28, length 1 byte, Label 360. 4. Flight speed field: offset 29, length 2 bytes, Label 210, unit KCAS, scaling factor 0.1. 5. Normal acceleration field: offset 31, length 2 bytes, Label 380, unit g, scaling factor 0.01, offset -10. From the data field, the following are extracted: 1. Test flight cycle identification field (offset 8, 12 bytes): [0x54462D32303235313032372D3031], parsed as ASCII string "TF-20251027-01". 2. Absolute timestamp field (offset 20, 8 bytes): [0x00000181A1B2C3D4], parsed as Unix millisecond timestamp 1678886400000 milliseconds. 3. Landing gear state field (offset 28, 1 byte): [0x02]. According to Label 360 definition (0 = up, 1 = down, 2 = transition state), parsed as raw value 2. 4. Flight speed field (offset 29, 2 bytes): [0x10C0]. Parsed as decimal 4288. Scaling factor is applied: . Parsed as raw value 428.8 KCAS. 5. Normal acceleration field (offset 31, 2 bytes): [0x0119]. Parsed as decimal 281. Scaling and offset are applied: . Parses to raw value -7.19g. (This value is abnormal, subsequent packets will use 1.0g for example). Assume another packet (timestamp 1678886400100) parses as: Test Flight Gear ID = "TF-20251027-01", Absolute Timestamp = 1678886400100, Gear State = 0, Flight Speed = 430.2 KCAS, Normalized G-Force = 1.01g. Assume one more packet (timestamp 1678886400000) parses as: Test Flight Gear ID = "TF-20251027-02", Absolute Timestamp = 1678886400000, Gear State = 0, Flight Speed = 450.0 KCAS, Normalized G-Force = 1.00g. Combine these parsed raw values to generate multiple instantaneous flight state data.

[0021] S102: Invoke the set of instantaneous flight state data generated by S101. Entry 1: {ID: ,Time:1678886400000,Gear:2,Speed:428.8,G:-7.19} Entry 2: ,Time:1678886400100,Gear:0,Speed:430.2,G:1.01} Entry 3: ,Time:1678886400000,Gear:0,Speed:450.0,G:1.00} Entry 4: -01", Time: 1678886400050, Gear: 0, Speed: 429.5, G: 0.99} Traverse this collection, extract the first entry (entry 1) flight frame identifier "TF-20251027-01", set it as the first comparison reference. Create data collection 1, and aggregate entry 1 to data collection 1. Traverse entry 2, extract identifier "TF-20251027-01", compare with reference "TF-20251027-01", string completely matches. Aggregate entry 2 to data collection 1. Traverse entry 3, extract identifier "TF-20251027-02", compare with reference "TF-20251027-01", string does not match. Set new reference "TF-20251027-02". Create data collection 2, and aggregate entry 3 to data collection 2. Traverse entry 4, extract identifier "TF-20251027-01", compare with reference "TF-20251027-01", matches. Aggregate entry 4 to data collection 1. Traverse is completed. Attach index "TF-20251027-01" to data collection 1. Attach index "TF-20251027-02" to data collection 2. Get the flight frame aggregation data body. Data body [{Time: 1678886400000, …}, {Time: 1678886400100, …}, {Time: 1678886400050, …}], "TF-20251027-02": [{Time: 1678886400000, …}]}.

[0022] S103: According to the flight frame aggregation data body obtained in S102, retrieve its index one by one. Retrieve index "TF-20251027-01". Enter the data collection inside the index. Extract the absolute time stamp of multiple entries (3 entries in total) in this collection to form a list: [1678886400000, 1678886400100, 1678886400050]. Set the ascending order sorting rule for absolute time stamp, that is, for any two time stamps and , if , then is placed before . Call this rule (for example, use the quicksort algorithm) to perform rearrangement on data entries. Compare 1678886400100, 1678886400050. The sorting result is: 1678886400100. The order of entries in the rearranged data collection 1 (index "TF-20251027-01") is: 8886400050, Gear: 0, Speed: 429.5, G: 0.99}3. {Time: 1678886400100, Gear: 0, Speed: 430.2, G: 1.01}Retrieval index 1027-02”. The data set inside this index is entered. The absolute timestamps of the multiple entries (1 in total) in this set are extracted: [1678886400000]. The ascending order sorting rule is applied, and the number of entries is 1, so no rearrangement is needed. The order of the entries in the rearranged data set 2 (index “TF-20251027-02”) is: 00000, Gear: 0, Speed: 450.0, G: 1.00}An ordered time series data log is established: this log contains the time series data stream corresponding to the index “TF-20251027-01” (containing 3 sorted entries) and the time series data stream corresponding to the index “TF-20251027-02” (containing 1 entry), as well as the data frame structure defining the parameters and the sorting index (based on timestamps) for fast retrieval.

[0023] See Figure 1 and Figure 3 , S2: Based on the ordered time series data log, the landing gear state and flight speed are judged according to the time series event association rule to generate a set of event key-value pairs, and the normal overload is judged according to the cumulative duration limit rule to generate a set of duration key-value pairs; The steps of S2 are as follows: S201: Separate the landing gear state sequence and flight speed sequence from the multi-dimensional parameter data stream in the ordered time series data log, call the time series event association rule, and perform time-by-time synchronous comparison and judgment on the landing gear state sequence and flight speed sequence, filter the time points whose values meet the rule's built-in logical conditions, and collect all time points to generate a flight event timestamp sequence; S202: Take the flight event timestamp sequence as an index to retrieve the corresponding landing gear state value and flight speed value in the ordered time series data log, construct a record unit with the time point as the unique key name and the landing gear state value and flight speed value as the key-value binary tuple, and aggregate all record units to establish an event key-value pair set; S203: According to the key name contained in the event key-value pair set, divide the corresponding normal overload data segment in the ordered time series data log, integrate and calculate the cumulative duration of the sample points in multiple data segments whose data values exceed the normal overload threshold, call the cumulative duration limit rule, compare the obtained cumulative duration with the duration limit reference value, filter the durations that exceed the duration limit reference value, generate a new key-value pair with the event key as the index and the duration as the value, and obtain the duration key-value pair set; The judgment steps of the time series event association rule are as follows: Synchronize extraction of landing gear state sequence and flight speed sequence from ordered time series data log, obtain preset landing gear state threshold in time series event association rule , flight speed threshold and association weight factor and ; Synchronize comparison of landing gear state sequence and flight speed sequence at each time, at any absolute timestamp , calculate event trigger score , the calculation formula is: ; Wherein, is the landing gear state value at moment, is the flight speed value at moment, is the landing gear state matching function, which returns 1 when trigger preset condition, otherwise returns 0, is the flight speed matching function, which returns 1 when trigger preset condition, otherwise returns 0, and are association weight factors and ; The judgment steps of the cumulative duration limit rule are as follows: Extract the continuous normal overload sequence from the ordered time series data log, obtain the preset normal overload threshold in the cumulative duration limit rule and minimum duration ; Traverse the normal overload sequence, identify all data points whose values continuously exceed the normal overload threshold , divide the data points into multiple instantaneous over-limit data segments according to the absolute timestamp continuity, and calculate the duration of each instantaneous over-limit data segment; Screen the instantaneous over-limit data segments with duration greater than the minimum duration , form the effective over-limit data segment set, and extract the overload peak value in each effective over-limit data segment to construct the time length key-value pair set; The event key-value pair set includes event identifier, association time point and state snapshot; The time length key-value pair set includes over-limit type, duration period and overload peak value.

[0024] S201: Extract the data stream of flight "TF-20251027-01" from the ordered time series data log established in S1. Separate the landing gear state sequence and flight speed sequence . : [2, 0, 0, …, 1, 1, 1] (2 at timestamp 1678886400000, 0 at 1678886400050, …, 1 at some later time) . : [428.8, 429.5, 430.2, …, 162.0, 158.0, 155.0] Call the time series event association rule (see later). The built-in logical condition in this rule is . and sequences are compared and judged time by time. Suppose at KCAS is calculated (see later examples), and the result is . The judgment is , which is false. This time point is not filtered. Suppose at : KCAS is calculated (see later examples), and the result is . The judgment is , which is true. This time point 1678887000100 is filtered. Suppose at : KCAS is calculated , and the result is . The judgment is , which is true. This time point 1678887000200 is filtered. All filtered time points are collected to generate the flight event timestamp sequence: [1678887000100, 1678887000200, …].

[0025] S202: Take the flight event timestamp sequence [1678887000100, 1678887000200, …] generated in S201 as the index. Process the first index 1678887000100. Retrieve this timestamp in the ordered time series data log (flight “TF-20251027-01”). The entry is retrieved: Gear:1,Speed:158.0,G:1.05}. Extract the corresponding landing gear state value 1 and flight speed value 158.0. Construct a binary tuple with key value (1, 158.0). Construct a record unit: key name = 1678887000100, key value = (1, 158.0). Process the second index 1678887000200. Retrieve the entry: Gear:1,Speed:155.0,G:1.02}. Extract the corresponding landing gear state value 1 and flight speed value 155.0. Construct a binary tuple with key value (1, 155.0). Construct a record unit: key name = 1678887000200, key value = (1, 155.0). Aggregate all record units to establish the event key-value pair set: {1678887000100:(1, 158.0), 1678887000200:(1, 155.0),...}. This set contains event identifiers (i.e., key names, timestamps), associated time points (key names), and state snapshots (key values, binary tuples).

[0026] S203: According to the key names [1678887000100, 1678887000200,...] contained in the event key-value pair set established in S202, split the normal overload data segment in the ordered time series data log. Here, "split" means that these events (e.g., landing gear down and speed meeting) identify the start of a critical phase of flight (e.g., "approach"). We split the normal overload data segment from the first event (1678887000100) to the end of the flight (or the occurrence of another phase event). Assume that the split data segment is , from 1678887000100 to 1678887500000. Invoke the cumulative duration limit rule (see later) to obtain the normal overload threshold g and the duration limit reference value (i.e., the minimum duration ) = 0.5 seconds. Perform integral calculation (i.e., traversal) on this data segment to find the cumulative duration of sample points with data values exceeding 1.8g. Traverse the sequence: , g(<1.8g) , g(>1.8g)-> start of transient over-limit data segment 1 , g(>1.8g)... , g(>1.8g)-> peak g... , ​g (> 1.8g) -> end of transient over-limit data segment 1 , g (< 1.8g) calculate segment 1 cumulative duration: milliseconds = 0.9 seconds. Continue traversal... , g (> 1.8g) -> segment 2 begins , g (> 1.8g) -> peak , g (> 1.8g) -> end of segment 2 , g (< 1.8g) calculate segment 2 cumulative duration: milliseconds = 0.2 seconds. Invoke cumulative duration limit rule (i.e. compare ). Compare (0.9 seconds) with (0.5 seconds). , condition met. Filter . Compare (0.2 seconds) with (0.5 seconds). , condition not met. Generate new key-value pair with "event key" (here, the flight ID "TF-20251027-01" associated with this data segment) as key and the filtered duration (0.9 seconds) and its associated information (peak 1.91g, start and end times) as value. Get the set of duration key-value pairs (i.e. the output of the cumulative duration limit rule): { "TF-20251027-01": [{over-limit type: "G-Load", duration segment: , 1678887101000), peak over-load: 1.91}]} (if the event key of S202 is used for segmentation, the key here could be 1678887000100); Judgment steps of time-series event association rule: extract the landing gear state sequence and the flight speed sequence from the ordered time-series data log synchronously. Get the pre-set parameters in the rule. Landing gear state threshold : this parameter is non-numeric, which is a logical condition. The condition is set with reference to the configuration requirement in the "approach" phase of the aircraft operating manual (AFM). Set as "landing gear state value equals to 1 (indicating down) ". Flight speed threshold : this parameter is numeric. Its setting refers to the flight test outline for this type of aircraft (e.g. a certain type of twin-engine turbofan passenger aircraft). To determine Specialized flight tests were conducted. During the tests, the aircraft established a stable approach at different speeds (180, 175, 170, 165, 160, 155 KCAS) with the landing gear down and approach flaps in position.

[0027] Table 1. Flight Speed ​​Threshold Verification Test Data As shown in Table 1, the stability rating (1 being the best and 5 the worst) data shows that at speeds of 160 KCAS and below, the flight status is stable (rating 1) with minimal N1 fluctuations. At speeds of 165 KCAS, stability begins to decline (rating 2). At speeds of 170 KCAS and above, the stability rating is 3 or 4, indicating an unstable approach. Based on this test data, the flight speed matching condition is set to "Flight Speed..." "less than 160KCAS" means KCAS. Related weighting factor. and These two parameters are used to balance the importance of landing gear status and flight speed. In this scenario (approach), landing gear status (must be deployed) is a rigid condition, while speed (should be below a certain value) is a flexible condition. Referring to the Flight Safety Analysis (FHA) results, the risk level of an incorrect landing gear status is higher than the risk level of a slightly higher speed. Setting , ,satisfy Landing gear state matching function This is the quantization function. Input (Instantaneous value from S101: 0, 1, or 2) and (Condition "equals 1"). ,when hour. ,when Time (i.e.) or Flight speed matching function This is the quantization function. Input (Instantaneous value from S101) and (160KCAS). ,when hour. ,when hour.

[0028] formula Parameter description: yes The score is determined by the event trigger at a given moment. The weight of the landing gear status is 0.6. is a boolean matching function for gear status (returns 0 or 1) ; is a weight for airspeed (0.4) ; is a boolean matching function for airspeed (returns 0 or 1). This formula calculates the degree of coincidence between the airspeed at the moment and the preset event (speed below threshold) by weighted summation. is the degree of coincidence between the airspeed at the moment and the preset event (speed below threshold). The benefit of this formula is that it allows for a comprehensive evaluation of multiple parameters, rather than a simple "and" logic, allowing for future expansion to a more complex scoring mechanism. and the introduction of weights, which allows for a comprehensive evaluation of multiple parameters, rather than a simple "and" logic, allowing for future expansion to a more complex scoring mechanism.

[0029] Synchronize the gear status sequence and airspeed sequence by time, and compare them at : Obtain data: , KCAS. Quantize to meet conditions. . Quantize . (160). Substitute into the formula to calculate: Compare with the built-in logic condition in S201, if the moment does not trigger the event.

[0030] At : Obtain data: , KCAS. Quantize . . Quantize . (160). . Substitute into the formula to calculate: Compare with the built-in logic condition in S201, if the moment triggers the event. This result indicates that at , the gear and speed states simultaneously meet the definition of the approach event, and S201 will include (1678887000100) in the flight event timestamp sequence.

[0031] Judgment step of cumulative duration limit rule: Extract continuous normal acceleration sequence from ordered time series data log (flight “TF-20251027-01”) . Obtain preset parameters in the rule: Normal acceleration threshold : This parameter is set in reference to the aircraft design load limit (V-n diagram) and flight test subject requirements. For the approach and landing phase, the structural bearing capacity is relatively sensitive. To determine , a structural simulation analysis was performed. Using a finite element model (FEM), the stress levels at the main landing gear connection point and wing root under different normal accelerations (1.5g, 1.6g, 1.7g, 1.8g, 1.9g, 2.0g) in the approach configuration (landing gear and flaps down) were analyzed. The analysis results show that when g, the stress at the wing root reaches 70% of the design ultimate load, and the stress at the main landing gear connection point reaches 85%. When g, the stress exceeds 90%. Based on this analysis, to reserve a safety margin, the normal acceleration threshold that triggers monitoring is set to g. Minimum duration : This parameter is used to filter out invalid burrs caused by sensor noise or transient airflow disturbances. To determine , 100 known transient overload spikes in historical flight data were analyzed. Through spectral analysis and time domain analysis, 98% of the noise spike pulse widths are less than 300 milliseconds (0.3 seconds). At the same time, analysis of the structural dynamic response shows that loads with a duration shorter than 0.5 seconds have minimal contribution to structural fatigue. The minimum duration is set to seconds (i.e., 500 milliseconds). This setting value is higher than the 300 milliseconds noise spike pulse width obtained from analysis and corresponds to a time scale with minimal fatigue contribution in the structural dynamic response.

[0032] Traverse the normal acceleration sequence , identify all g data points. When , g, mark the starting point of the first transient overrun data segment . Initialize peak g. Traverse , g. When , g, update g. When , g, update g. When , g. When , g (<1.8g), mark the end point of data segment 1 (Previous timestamp). Calculate segment 1 duration: milliseconds = 0.9 seconds. Continue iteration… At , g, mark the start point of data segment 2 Initialize peak g. At , g. At , g. At , g (<1.8g), mark the end point of data segment 2 . Calculate segment 2 duration: milliseconds = 0.2 seconds. Filter duration: segment 1: seconds. Compare (0.5 seconds), condition holds. Segment 2: seconds. Compare (0.5 seconds), condition does not hold. Form valid over-limit data segment set, which contains only segment 1. Extract overload peak of segment 1 g. Construct duration key-value pair set (i.e. output of S203): {“TF-20251027-01”:[{over-limit type: “G-Load”, duration segment: 100, 1678887101000), overload peak: 1.91}]}.

[0033] See Figure 1 and Figure 4 , S3: Group event key-value pair set by flight test flight number, obtain event structure list sorted by absolute timestamp, input event structure list to finite state automaton, match time window and violation logic state transition based on time sequence event association rule, output profile violation record; The steps for obtaining the profile violation record are as follows: S301: Extract the flight test flight number in each key-value pair record in the event key-value pair set, use the flight test flight number as the unique grouping key, iterate and aggregate the event key-value pair set, merge all key-value pair records holding the same flight test flight number into the same data subset, and establish a flight event grouping set; S302: For each data subset in the flight event grouping set, extract the absolute timestamps corresponding to all key-value pair records in the data subset, and use the absolute timestamps as the sorting basis to perform ascending arrangement operation on the key-value pair records in the data subset, construct an event sequence, and obtain an event structure list; ​S303: Load the event structure list into the finite state automaton in sequence order, and obtain the preset time window threshold and violation logic state transition matrix in the time sequence event association rule. When the state machine jumps between differentiated states, calculate the absolute timestamp difference between the adjacent two events, judge whether the difference falls within the time window threshold interval, and compare whether the current state jump path is consistent with the path defined in the violation logic state transition matrix. If both conditions are met, record the complete event information of the violation transition, and output the profile violation record; The matching process of the finite state automaton is specifically: A set of flight states is defined in advance, including take-off phase, climbing phase, cruising phase, maneuvering phase, descending phase and landing phase, and a violation logic state transition matrix is defined, which stores illegal state transition triples composed of source state, target state and violation condition; Load the event structure list, read the key-value pair records in the event structure list one by one, parse the state snapshot corresponding to the key-value pair records, map the state snapshot to the target flight state in the flight state set as the current state; Get the absolute timestamp of the key-value pair record, and calculate the difference with the absolute timestamp of the previous event. Judge whether the difference falls within the time window threshold interval; If the difference falls within the time window threshold interval, it is determined that the flight state of the previous event is the source state and the current state is the target state. Retrieve the violation logic state transition matrix, compare whether the jump from the source state to the target state exists in the illegal state transition triple. If it exists, it is determined as a violation transition. Extract the event identifier, associated time point and violation type code of the key-value pair record to generate a profile violation record; The event structure list includes flight phase division, event sequence and logical association; The profile violation record includes violation type code, violation occurrence time and violation scenario reproduction.

[0034] S301: Extract the event key-value pair set generated by S202. Set , Time: 1678887000100, Val: (1, 158.0)), Evt2 (ID: , Time: 1678887000100, Val: (1, 159.0)), Evt3 (ID: , Time: 1678887050000, Val: (1, 150.0)), Evt4 (ID: "TF-01", Time: 1678887010000, Val: (1, 155.0))} (Note: S301 / S302 demonstration, data is expanded here, ID is simplified as TF-01, TF-02) Take flight session ID as the unique grouping key. Set the grouping key "TF-01". Create data subset 1. Traverse Evt1, ID = "TF-01", match. Merge Evt1 into data subset 1. Traverse Evt2, ID = "TF-02", not match. Set a new grouping key "TF-02". Create data subset 2. Traverse Evt2, ID = "TF-02", match. Merge Evt2 into data subset 2. Traverse Evt3, ID = "TF-01", match the grouping key "TF-01". Merge Evt3 into data subset 1. Traverse Evt4, ID = "TF-01", match the grouping key "TF-01". Merge Evt4 into data subset 1. Establish a flight session event grouping set: {"TF-01": [Evt1, Evt3, Evt4], "TF-02": [Evt2]}.

[0035] S302: Data subset "TF-01" in the flight session event grouping set established in S301. Extract the absolute timestamps corresponding to all key-value pair records (Evt1, Evt3, Evt4) in the subset "TF-01". Timestamp list: [1678887000100 (Evt1), 1678887050000 (Evt3), 1678887010000 (Evt4)]. Take the absolute timestamps as the sorting basis, and perform ascending arrangement operation on the key-value pair records (Evt1, Evt3, Evt4) in the data subset "TF-01". Compare the timestamps: 1678887000100 <1678887>010000 <1678887>050000. The corresponding event order is Evt1, Evt4, Evt3. Construct the event sequence "TF-01": [Evt1, Evt4, Evt3]. For data subset "TF-02", there is only Evt2 inside, and the sequence is [Evt2]. Get the event structure list (contains the event sequence of all flight sessions): ["TF-01": [Evt1, Evt4, Evt3], "TF-02": [Evt2]].

[0036] S303: Load the sequence "TF-01": [Evt1, Evt4, Evt3] in the event structure list obtained in S302 into the finite state automaton (FSM) in order. Get the preset time window threshold and violation logic state transition matrix in the time sequence event association rule (see FSM matching process for details). Time window threshold This parameter is used to define whether two consecutive events are logically related. To determine , 100 standard flight profile logs were analyzed. The time interval between consecutive flight phases (e.g. from the "start descent" event to the "establish approach" event) was counted. The statistics (histogram) showed that 95% of the valid consecutive event intervals were distributed between 5 seconds and 300 seconds. Events with intervals less than 5 seconds were usually considered as repeated confirmation of the same phase; events with intervals greater than 300 seconds (5 minutes) usually meant a phase interruption (e.g. level flight, go-around). Based on this analysis, the time window threshold interval was set to .

[0037] Violation logic state transition matrix: (see FSM procedure for details). FSM starts processing sequence "TF-01": 1. Load Evt1 (Time: 1678887000100, Val: (1, 158.0)). FSM (see later) maps state snapshot (1, 158.0) to flight state "approach" (state 4). FSM sets: current state = 4. previous state = (initial state). previous time = 1678887000100. 2. Load Evt4 (Time: 1678887010000, Val: (1, 155.0)). FSM maps state snapshot (1, 155.0) to flight state "approach" (state 4). FSM sets: current state = 4. Calculate the absolute timestamp difference of adjacent events: milliseconds. Determine whether the difference falls within the time window threshold interval [5000, 300000]. . Condition 1 is true. Compare state jump path: source state = 4 (previous state), target state = 4 (current state). Query violation logic state transition matrix for (Source: 4, Target: 4) violation entry. Query result: none. Condition 2 is false. No violation record. FSM updates: previous state = 4. previous time = 1678887010000. 3. Load Evt3 (Time: 1678887050000, Val: (1, 150.0)). FSM maps state snapshot (1, 150.0) to flight state "landing" (state 5). (Assume that in the mapping rules of FSM, speed and are mapped to landing). FSM sets: current state = 5. Calculate the timestamp difference: milliseconds. Determine whether the difference falls within the time window [5000, 300000]. Condition 1 is true. Compare state transition path: source state = 4 (previous state "approach"), target state = 5 (current state "landing"). Query violation logic state transition matrix for violation entry (Source: 4, Target: 5). Query result: no violation entry exists ((4, 5) is a legal transition). Condition 2 is false. No violation record. FSM update: previous state = 5. Previous time = 1678887050000.

[0038] Suppose a violation example: 1. Load EvtX (Time: 1678886500000, Val: (...)). FSM maps to "cruise" (state 3). Previous state = 3, previous time = 1678886500000. 2. Load EvtY (Time: 1678886550000, Val: (...)). FSM maps to "landing" (state 5). Current state = 5. 3. Calculate milliseconds. 4. Determine the time window: . Condition 1 is true. 5. Compare path: source state = 3 ("cruise"), target state = 5 ("landing"). 6. Query violation matrix (see below): violation entry (Source: 3, Target: 5) exists. Condition 2 is true. 7. Both conditions are true. Record the complete event information of the violation transition (EvtX, EvtY). Output profile violation record: {violation type code: "P-001" (cruise to landing), violation occurrence time: 1678886550000 (time of EvtY), violation scenario recurrence: (EvtX, EvtY details)}.

[0039] Matching process of finite state automaton: Define flight state set in advance: {1: takeoff, 2: climb, 3: cruise, 4: descent / approach, 5: landing}. Define violation logic state transition matrix (illegal state transition triplets) in advance: this matrix is determined based on standard operating procedures (SOP) and flight physical logic. milliseconds)-> missing "descent / approach" phase. (Source state: 1 (takeoff), target state: 4 (descent / approach), violation condition: milliseconds)-> missing "climb" and "cruise" phases. (Source state: 2 (climb), target state: 5 (landing), violation condition: milliseconds)-> missing "cruise" and "descent / approach" phases.

[0040] Load event structure list (sequence [EvtX, EvtY] for flight "TF-01"). Read EvtX (Time: 1678886500000, Val: (0, 430.0)). Parse state snapshot for EvtX (0, 430.0). Apply mapping logic: -IF (transition) -> State 1 (takeoff) -IF AND -> State 3 (cruise) -IF AND -> State 4 (approach) -IF AND -> State 5 (landing)... Apply mapping logic: (0, 430.0) satisfies AND . Map to target flight state "cruise" (State 3). Set: current state = 3. Set: previous event = EvtX (State 3). Previous time = 1678886500000.

[0041] Read EvtY (Time: 1678886550000, Val: (1, 150.0)). Parse state snapshot for EvtY (1, 150.0). Apply mapping logic: (1, 150.0) satisfies AND . Map to target flight state "landing" (State 5). Set: current state = 5. Get EvtY absolute timestamp 1678886550000. Calculate difference from previous event (EvtX) absolute timestamp 1678886500000: milliseconds. Determine if difference falls within time window threshold interval [5000, 300000]. . Condition holds. Determine flight state for EvtX as source state (3). Determine flight state for EvtY as target state (5). Retrieve illegal state transition matrix for violation logic, compare if a jump exists for (Source: 3, Target: 5). Retrieve triple (Source: 3, Target: 5, Condition: ) from illegal state transition matrix. Compare if violation condition is satisfied: . Condition holds. A jump exists in illegal state transition triple and time condition is satisfied. Determine as a violation transition. Extract event identifier for EvtY (1678886550000), associated time point (1678886550000), and violation type code ("P-001" - cruise to landing). Generate profile violation record: {Code: "P-001", Time: 1678886550000, Scenario: (EvtX, EvtY)}.

[0042] Please refer to Figure 1 and Figure 5 , S4: group the time length key-value pair set according to the flight test flight number identifier, calculate the total cumulative over-limit time length, judge whether the total cumulative over-limit time length exceeds the maximum allowed cumulative time length of the cumulative time length limit rule, generate a cumulative over-limit violation report, and merge the profile violation record and the cumulative over-limit violation report to form a comprehensive violation check report.

[0043] The obtaining step of the comprehensive violation check report is specifically: S401: group the time length key-value pair set according to the flight test flight number identifier, aggregate the key-value pair records with the same flight test flight number identifier into a flight number data subset, simultaneously traverse each flight number data subset, and perform arithmetic accumulation on the time length values of all key-value pairs in the flight number data subset to calculate the total cumulative over-limit time length of the flight number, and generate a flight number total over-limit time length statistic with the flight test flight number identifier as the key; S402: call the flight number total over-limit time length statistic, retrieve the maximum allowed cumulative time length from the cumulative time length limit rule, traverse all entries in the flight number total over-limit time length statistic, perform numerical size judgment on the total cumulative over-limit time length of each entry record and the maximum allowed cumulative time length, filter all entries with total cumulative over-limit time length greater than the maximum allowed cumulative time length, and form a cumulative over-limit violation report; S403: align the profile violation record and the cumulative over-limit violation report according to the flight test flight number identifier as the associated primary key, merge the records with the same flight test flight number identifier, and retain all independent records in the two reports that are not aligned, then format all the merged and independent records uniformly to form a comprehensive violation check report; Group the time length key-value pair set according to the flight test flight number identifier, accumulate the duration of all valid over-limit data segments under the same flight test flight number identifier, and calculate the total cumulative over-limit time length The calculation logic is: ; Among them, is the total number of valid over-limit data segments corresponding to the flight test flight number identifier, is the starting absolute timestamp of the th valid over-limit data segment, is the ending absolute timestamp of the th valid over-limit data segment, and ; Call the maximum allowed cumulative time length, compare with the maximum allowed cumulative time length, if is greater than the maximum allowed cumulative time length, then generate a cumulative over-limit violation report; The cumulative over-limit violation report includes total time length statistics, judgment threshold, and over-limit conclusion; The comprehensive violation verification report includes a verification summary, details of the violation profile, and cumulative statistical results.

[0044] S401: Perform grouping operations on the set of time-based key-value pairs generated by S203 (or the cumulative duration limit rule) according to the test flight sortie identifier. Input set = {Entry1(ID:“TF-01”,Type:“G-Load”,Duration:(t1,t2),Peak:1.91),(Note: t2-t1=0.9s)Entry2(ID:“TF-02”,Type:“G-Load”,Duration:(t3,t4),Peak:1.85),(Note: t4-t3=0.6s)Entry3(ID:“TF-01”,Type:“G-Load”,Duration:(t5,t6),Peak:1.88)(Note: t6-t5=0.7s)} Aggregate key-value pair records with the same test flight sortie identifier into a subset of sortie data. Subset "TF-01": [Entry1, Entry3] Subset "TF-02": [Entry2] Iterate through the flight data subset "TF-01". Perform an arithmetic sum of the duration values ​​for all key-value pairs within subset "TF-01". Duration value 1 (Entry1) = Seconds. Duration value 2 (Entry3) = Seconds. Calculate the total cumulative over-limit time of "TF-01". (TF-01)= Seconds. Iterate through the flight data subset "TF-02". Perform an arithmetic sum of the duration values ​​for all key-value pairs within subset "TF-02". Duration value 1 (Entry2) = Seconds. Calculate the total cumulative over-limit time of "TF-02". (TF-02)= Seconds. Generate a total overtime statistics for each test flight, with the test flight identifier as the key: {“TF-01”:1.6,“TF-02”:0.6}.

[0045] S402: Executes the total overrun time statistics generated by S401 {“TF-01”:1.6,“TF-02”:0.6}. Retrieves the maximum allowed cumulative time from the cumulative time limit rules. Maximum allowed cumulative duration This parameter is based on aircraft structural fatigue life analysis. It defines the acceptable limits of fatigue life during a single flight mission. (1.8g) total time. To determine Fatigue calculations based on the Miner cumulative damage criterion were performed. Analysis model: using SN curves (stress-life curves) and... The stress level corresponding to (1.8g) is (e.g., 85% of the design ultimate load). Calculation process: It is assumed that a certain percentage of fatigue life is consumed with each flight. Calculations show that when... cumulative time of g At 1.5 seconds, the fatigue life of critical components (such as the main landing gear connection points) reaches 90% of the permissible limit for a single flight. Based on this calculation, and with an additional 10% safety margin, the maximum permissible cumulative duration is set to... Seconds. Any sortie exceeding this value must undergo a mandatory special structural inspection.

[0046] Iterate through the entries in the total overtime statistics for all flights. Entry 1: {ID:“TF-01”, :1.6}. (This appears to be a typo and can be left as is.) (1.6 seconds) and (1.5 seconds) Perform numerical comparison. The condition is true. Filter item 1. Item 2: {ID:“TF-02”, :0.6}. (This likely refers to a value or value, but without further context, it's difficult to translate accurately. (0.6 seconds) and (1.5 seconds) Perform numerical comparison. . Determined as false. Item 2 is not filtered. A cumulative over-limit violation report is generated: [{Flight ID:“TF-01”, Total Duration: 1.6 seconds, Judgment Threshold: 1.5 seconds, Over-limit Conclusion: “Violation”}].

[0047] S403: Align the profile violation records (R_Profile) and the cumulative over-limit report (R_Cumulative) by the flight test serial ID. Process "TF-01": {ID: "TF-01", …} exists in R_Profile. {ID: "TF-01", …} exists in R_Cumulative. The primary key "TF-01" holds the same, and the data fields are merged. The merged record (M1): {ID: "TF-01", profile details: [{code: "P-001", …}], cumulative details: {total duration: 1.6, …}}. Process "TF-03": {ID: "TF-03", …} exists in R_Profile. "TF-03" does not exist in R_Cumulative. Keep the independent record in R_Profile. The independent record (M2): {ID: "TF-03", profile details: [{code: "P-002", …}], cumulative details: (empty)}. Process the remaining (if any) independent records in R_Cumulative (none in this example). Format all the merged records (M1) and independent records (M2) uniformly. Form the comprehensive violation check report: [{check summary: {serial: "TF-01", result: "profile violation, cumulative over-limit"}, profile violation details: [{code: "P-001", time: 1678886550000, …}], cumulative statistical results: {total duration: 1.6, threshold: 1.5, conclusion: "violation"}}, {check summary: {serial: "TF-03", result: "profile violation"}, profile violation details: [{code: "P-002", time: 1678889000000, …}], cumulative statistical results: {total duration: (N / A), threshold: 1.5, conclusion: "not over-limit"}}]. The calculation logic of the total cumulative over-limit duration: formula Parameter explanation: is the total cumulative over-limit duration for a single flight test serial ID. is the summation symbol, indicating traversal and accumulation of effective over-limit data segments. is the index of the data segment, from 1 to . is the total number of valid overrun data segments corresponding to the test flight session identifier. Valid overrun data segments are those identified as (1.8g) and the duration (0.5 seconds). is the start absolute timestamp (millisecond) of the th valid overrun data segment. is the end absolute timestamp (millisecond) of the th valid overrun data segment. is the duration of the th data segment. The logic of the whole formula is to find all the overload overrun ( ) time segments that satisfy the minimum duration ( ) and add up the duration of all these time segments to get the total overrun time. The benefit of this formula is that it accumulates multiple, non-continuous, instantaneous overrun events during the flight into one total quantitative index by using the operation (sum). This index is directly related to the cumulative fatigue damage of the structure, rather than relying only on the single peak or the duration of a single event.

[0048] The duration key-value pair set obtained in the grouping S401 according to the test flight session identifier "TF-01" is accumulated to add up the duration of all valid overrun data segments under the same test flight session identifier "TF-01". The obtained parameters are: valid overrun data segments identified by the session "TF-01" (which have passed the second screening): (2 valid data segments in total) millisecond (0.5 second) data segment (Entry 1): milliseconds milliseconds data segment (Entry 3): milliseconds (assuming value ) milliseconds (assuming value ).

[0049] Check (this check has been completed in S203 / Rule, and this is only for the formula constraint): Segment 1 duration = milliseconds. (500). Segment 2 duration = milliseconds. (500). is true.

[0050] Calculate the total cumulative overrun duration : milliseconds seconds.

[0051] call the maximum allowed cumulative duration set in S402 seconds. Compare (1.6 seconds) with (1.5 seconds). This result (greater than) indicates that the total cumulative overrun duration of the flight "TF-01" has exceeded the maximum allowed threshold. If greater than the maximum allowed cumulative duration, a cumulative overrun violation report (containing the entry of "TF-01") in S402 is generated.

[0052] The flight test data analysis system based on the cloud platform is used to execute the flight test data analysis method based on the cloud platform, and the system comprises: a time sequence log generation module, configured to acquire original data packets, extract flight test flight identifiers, absolute time stamps, landing gear states, flight speeds and normal overloads, collect and arrange in ascending order of absolute time stamps based on the flight test flight identifiers, form ordered time sequence data logs, and deliver the ordered time sequence data logs to a key-value pair analysis module; a key-value pair analysis module, configured to generate an event key-value pair set by judging the landing gear state and the flight speed according to a time sequence event association rule based on the ordered time sequence data logs, generate a duration key-value pair set by judging the normal overload according to a cumulative duration limit rule, deliver the event key-value pair set to a profile violation matching module, and deliver the duration key-value pair set to a comprehensive check report module; a profile violation matching module, configured to group the event key-value pair set according to the flight test flight identifiers, acquire an event structure list sorted according to the absolute time stamps, input the event structure list into a finite state automaton, match a time window and a violation logic state transition based on the time sequence event association rule, output a profile violation record, and deliver the profile violation record to the comprehensive check report module; a comprehensive check report module, configured to group the duration key-value pair set according to the flight test flight identifiers and calculate a total cumulative overrun duration, judge whether the total cumulative overrun duration exceeds a maximum allowed cumulative duration of the cumulative duration limit rule to generate a cumulative overrun violation report, and combine the profile violation record and the cumulative overrun violation report to form a comprehensive violation check report.

[0053] The above embodiments demonstrate the preferred implementation of the present application, and any equivalent adjustment of the technical solutions based on software engineering methods falls within the protection scope, including but not limited to: implementing algorithm logic in different programming languages, service reconstruction of functional modules, adjustment of data interaction protocols, optimization of resource scheduling strategies, etc. Any implementation derived through reasonable modification of the data processing flow, service calling link or system architecture level without deviating from the technical core of the present application shall be considered within the protection scope defined by the claims of the present application.

Claims

1. A flight test data analysis method based on a cloud platform, characterized in that, The method comprises the following steps: S1: obtaining original data packets, extracting flight test flight identification, absolute timestamp, landing gear state, flight speed and normal overload, collecting and arranging in ascending order of absolute timestamp based on the flight test flight identification to form an ordered time sequence data log; S2: based on the ordered time sequence data log, judging the landing gear state and the flight speed to generate an event key-value pair set according to the time sequence event correlation rule, and judging the normal overload to generate a time length key-value pair set according to the cumulative time length limit rule; S3: grouping the event key-value pair set according to the flight test flight identification, obtaining an event structure list sorted according to the absolute timestamp, inputting the event structure list into a finite state automaton, matching the time window and illegal state transition logic of the time sequence event correlation rule, and outputting a profile violation record; S4: grouping the time length key-value pair set according to the flight test flight identification and calculating the total cumulative over-limit time length, judging whether the total cumulative over-limit time length exceeds the maximum allowed cumulative time length of the cumulative time length limit rule to generate a cumulative over-limit violation report, merging the profile violation record and the cumulative over-limit violation report to form a comprehensive violation checking report.

2. The cloud platform based flight test data analysis method as claimed in claim 1, wherein, The ordered time sequence data log includes time sequence data stream, data frame structure and sorting index, the event key-value pair set includes event identifier, correlation time point and state snapshot, the time length key-value pair set includes over-limit type, duration period and overload peak, the event structure list includes flight phase division, event sequence and logical correlation, the profile violation record includes violation type code, violation occurrence time and violation scenario reproduction, the cumulative over-limit violation report includes total time length statistics, judgment threshold and over-limit conclusion, and the comprehensive violation checking report includes checking summary, profile violation details and cumulative statistical result. 3.The cloud platform-based flight test data analysis method of claim 2, wherein, The acquisition step of the ordered time sequence data log is specifically: S101: collecting bus interface transmission original data packets, performing frame header verification and data field analysis on the original data packets, locating flight test flight identification field, absolute timestamp field, landing gear state field, flight speed field and normal overload field according to a preset data dictionary, extracting the flight test flight identification field, the absolute timestamp field, the landing gear state field, the flight speed field and the normal overload field corresponding original values from the original data packet data field to generate instantaneous flight state data; S102: calling the instantaneous flight state data, traversing the flight test flight identification in the instantaneous flight state data, setting the flight test flight identification comparison reference, aggregating the instantaneous flight state data entries with the same flight test flight identification to the same data set, and attaching the corresponding flight test flight identification as an index to each data set, and obtaining a flight identification aggregation data body; S103: According to the flight cycle collection data body, retrieve the flight cycle collection data body index one by one, extract multiple entry absolute time stamps inside the same flight test flight cycle identification index, set the absolute time stamp ascending order sorting rule, call the ascending order sorting rule to rearrange the data entries inside the flight cycle collection data body, and establish an ordered time sequence data log.

4. The cloud platform based flight test data analysis method as claimed in claim 3, wherein, The steps of S2 are specifically: S201: Separate the landing gear state sequence and the flight speed sequence from the multi-dimensional parameter data stream in the ordered time sequence data log, call the time sequence event association rule, and perform time-by-time synchronization comparison and judgment on the landing gear state sequence and the flight speed sequence, filter the time points whose numerical values meet the rule built-in logic conditions, collect all time points to generate a flight event timestamp sequence; S202: Take the flight event timestamp sequence as an index to retrieve the corresponding landing gear state value and flight speed value in the ordered time sequence data log, construct a record unit with a time point as a unique key name and a binary tuple composed of the landing gear state value and the flight speed value as a key value, aggregate all record units to establish an event key-value pair set; S203: According to the key name contained in the event key-value pair set, divide the corresponding normal overload data segment in the ordered time sequence data log, integrate and calculate the cumulative time length of the sample points whose data values exceed the normal overload threshold value in multiple data segments, call the cumulative time length limit rule, compare the obtained cumulative time length with the time length limit reference value, filter the time lengths that exceed the time length limit reference value, generate a new key-value pair with the event key as the index and the time length as the value, and obtain a time length key-value pair set.

5. The cloud platform based flight test data analysis method as claimed in claim 4, wherein, The acquisition steps of the profile violation record are specifically: S301: Extract the flight test flight cycle identification in each key-value pair record in the event key-value pair set, take the flight test flight cycle identification as a unique grouping key, traverse and aggregate the event key-value pair set, merge all key-value pair records holding the same flight test flight cycle identification into the same data subset, and establish a flight cycle event grouping set; S302: For each data subset in the flight cycle event grouping set, extract the absolute time stamps corresponding to all key-value pair records in the data subset, and take the absolute time stamps as the basis for sorting, perform ascending arrangement operation on the key-value pair records in the data subset, construct an event sequence, and obtain an event structure list; S303: Load the event structure list into a finite state automaton according to the sequence order, obtain the preset time window threshold and violation logic state transition matrix in the time sequence event association rule, calculate the absolute time stamp difference of adjacent two events when the state machine jumps between differentiated states, judge whether the difference value falls within the time window threshold interval, and compare whether the current state jump path is consistent with the path defined in the violation logic state transition matrix, if both conditions are met, record the complete event information of the violation transition, and output the profile violation record.

6. The cloud platform based flight test data analysis method as claimed in claim 5, wherein, The acquisition steps of the comprehensive violation check report are specifically: S401: performing a grouping operation on the time length key-value pair set according to the test flight flight identification, aggregating key-value pair records with the same test flight flight identification into a flight data subset, simultaneously traversing each flight data subset, performing arithmetic accumulation on the time length values of all key-value pairs in the flight data subset, calculating the total accumulated over-limit time length of the flight, and generating a flight total over-limit time length statistic with the test flight flight identification as the key; S402: calling the flight total over-limit time length statistic and retrieving the maximum allowed accumulated time length from the accumulated time length limit rule, traversing all entries in the flight total over-limit time length statistic, performing numerical size judgment on the total accumulated over-limit time length of each entry record and the maximum allowed accumulated time length, filtering all entries with the total accumulated over-limit time length greater than the maximum allowed accumulated time length, and forming an accumulated over-limit violation report; S403: aligning the profile violation record and the accumulated over-limit violation report according to the test flight flight identification as the associated primary key, merging records holding the same test flight flight identification, and retaining all independent records in the two reports that are not aligned, then uniformly formatting all merged and independent records to form a comprehensive violation check report.

7. The cloud platform based flight test data analysis method as claimed in claim 6, wherein, The matching process of the finite state automaton is specifically: a flight state set is defined in advance, the flight state set includes a take-off phase, a climbing phase, a cruising phase, a maneuvering phase, a descending phase and a landing phase, and a violation logic state transition matrix is defined, the violation logic state transition matrix stores illegal state transition triples composed of a source state, a target state and a violation condition; the event structure list is loaded, key-value pair records in the event structure list are read one by one, state snapshots corresponding to the key-value pair records are parsed, and the state snapshots are mapped to target flight states in the flight state set as current states; the absolute timestamp of the key-value pair record is obtained, and the difference between the absolute timestamp and the absolute timestamp of the previous event is calculated to determine whether the difference falls within the time window threshold interval; if the difference falls within the time window threshold interval, it is determined that the flight state of the previous event is the source state and the current state is the target state, the violation logic state transition matrix is retrieved, and it is determined whether the jump from the source state to the target state exists in the illegal state transition triple, if it exists, it is determined as a violation transition, and the event identifier, associated time point and violation type code of the key-value pair record are extracted to generate the profile violation record.

8. The cloud platform based flight test data analysis method of claim 7, wherein, The judgment step of the accumulated time length limit rule is specifically: extracting a continuous normal overload sequence from the ordered time sequence data log, obtaining a preset normal overload threshold value in the cumulative time length limit rule with a minimum duration ; Traverse the normal overload sequence and identify all values ​​that continuously exceed the normal overload threshold. The data points are divided into multiple instantaneous over-limit data segments according to the continuity of absolute timestamps, and the duration of each instantaneous over-limit data segment is calculated. screening said transient over-limit data segments with a duration greater than said minimum duration to form a set of valid over-limit data segments, and extracting overload peaks in each of said valid over-limit data segments to construct said duration key-value pair set; According to the flight test flight number identifier, the time length key value pair set is grouped, the total accumulated over-limit time length is calculated by accumulating the continuous time length of all valid over-limit data segments under the same flight test flight number identifier The calculation logic is: ； wherein, is the total number of the valid over-limit data segments corresponding to the test flight number identification, is the starting absolute timestamp of the th valid over-limit data segment, is the ending absolute timestamp of the th valid over-limit data segment, and ; calling the maximum allowed cumulative duration, comparing the with the maximum allowed cumulative duration, if the greater than the maximum allowed cumulative duration, generating the cumulative overrun violation report. 9.The cloud platform-based flight test data analysis method of claim 8, wherein, The judgment step of the time sequence event association rule is specifically: synchronously extracting a landing gear state sequence and a flight speed sequence from the ordered time sequence data log, obtaining a preset landing gear state threshold, a flight speed threshold and an association weight factor in the time sequence event association rule , and and ; synchronizing the landing gear state sequence and the flight speed sequence on a time-by-time basis, at any absolute time stamp calculating an event trigger score , the calculation formula is: ； wherein, is the landing gear state value at the time instant, is the flight speed value at the time instant, is a landing gear state matching function, which returns 1 when the predefined condition is triggered, and 0 otherwise, is a flight speed matching function, which returns 1 when the predefined condition is triggered, and 0 otherwise, is a flight speed matching function, which returns 1 when the predefined condition is triggered, and 0 otherwise, and is the associated weight factor and .

10. A flight test data analysis system based on a cloud platform, characterized by, The system is used to implement the flight test data analysis method based on the cloud platform in any one of claims 1-9, and the system comprises: The time sequence log generation module is configured to acquire original data packets, extract flight test airframe identification, absolute time stamp, landing gear state, flight speed and normal overload, collect and arrange in ascending order of the absolute time stamp based on the flight test airframe identification to form an ordered time sequence data log, and deliver the ordered time sequence data log to the key-value pair analysis module; The key-value pair analysis module is configured to determine an event key-value pair set of the landing gear state and the flight speed according to a time sequence event association rule based on the ordered time sequence data log, determine a time length key-value pair set of the normal overload according to a cumulative time length limit rule, deliver the event key-value pair set to the profile violation matching module, and deliver the time length key-value pair set to the comprehensive check report module; The profile violation matching module is configured to group the event key-value pair set according to the flight test airframe identification, acquire an event structure list sorted according to the absolute time stamp, input the event structure list into a finite state automaton, match a time window and a violation logic state transition of the time sequence event association rule, output a profile violation record, and deliver the profile violation record to the comprehensive check report module; The comprehensive check report module is configured to group the time length key-value pair set according to the flight test airframe identification, calculate a total cumulative over-limit time length, determine whether the total cumulative over-limit time length exceeds a maximum allowable cumulative time length of the cumulative time length limit rule to generate a cumulative over-limit violation report, merge the profile violation record and the cumulative over-limit violation report to form a comprehensive violation check report.

Images