Python language-based flight event detection system

Abstract

The invention provides a flight event detection system based on a Python language, which comprises an event detection model, and is characterized in that (1) in the event detection model of the system, specific parameters defined in the Python language are used as triggering conditions of an event, and the system provides a function of event modularization by looking up a parameter group; (2) an airport database capable of setting different airport parameters is created in the system, and information of the airport database can be quoted in the event detection model; and (3) an organic model parameter library is created in the system, and model parameters can be cited in the event detection model. According to the method, the flight event detection model is defined based on the Python language, event detection mechanisms of aircrafts of different types can be defined more accurately, and the QAR parameters are taken as triggering conditions of event detection to help a user more accurately position events to aircraft flight parameters.

Description

Technical Field

[0001] The present invention relates to the fields of civil aircraft flight event detection, flight quality monitoring and flight safety, and in particular to a flight event detection system based on Python (a computer programming language developed by Dutch scientist Guido van Rossum in 1989). Background Art

[0002] Flight Quality Assurance (FOQA), also known as Flight Data Analysis Program (FDAP) or Flight Data Monitoring (FDM), is a tool for systematic and proactive identification of hazard sources. Through the aircraft flight data collected by the China Civil Aviation Flight Quality Assurance Base Station and loaded on the Quick Access Recorder (QAR), airlines and the Civil Aviation Administration can use this data to monitor flight quality, analyze aircraft safety trends and study typical unsafe events, and provide data support for safety supervision in the civil aviation industry.

[0003] Event detection in flight quality monitoring includes both safety events and operational events such as fuel saving and maintenance. Detection refers to the use of tools such as editors, algorithms or models to calculate and analyze data, and to issue alarms for events that trigger event standards (conditional values). Simple event detection can be a pre-set threshold for a certain parameter. During the flight of the aircraft, the ground computer obtains the flight data of the aircraft. When the value exceeds the set threshold, the event is triggered, an alarm is issued, and the relevant monitoring personnel are prompted. Monitoring personnel can also use a large amount of historical alarm data and events for data analysis, statistical induction, and form creation.

[0004] The "Flight Quality Monitoring Information Management Measures (Draft for Comments)" stipulates mandatory flight quality monitoring events for different aircraft types, including monitoring items, monitoring parameters, monitoring points, monitoring standards (i.e. the triggering conditions for red events) and remarks information.

[0005] At present, the flight quality monitoring systems of domestic airlines have problems such as single event detection conditions resulting in low detection accuracy, inability to associate with differentiated information of airports and aircraft models, and difficulty in aligning data from different frames. Summary of the invention

[0006] The purpose of the present invention is to provide a flight event detection system based on Python language to address the problems of low detection accuracy, inability to associate with differentiated information of airports and aircraft models, and difficulty in aligning data of different frames caused by the single event detection conditions in current flight quality monitoring.

[0007] The object of the present invention is achieved through the following technical solutions: A flight event detection system based on Python language, including an event detection model, the system includes: (1) In the event detection model of the system, specific parameters defined in Python are used as trigger conditions for events, and the system provides the function of modularizing the parameter group events; (2) An airport database is created in the system to set different airport parameters, and the information in the airport database can be referenced in the event detection model; (3) A model parameter library is created in the system, and the model parameters can be referenced in the event detection model.

[0008] As a further technical solution, the specific steps of using the specific parameters defined in the Python language as trigger conditions for the event include: (1.1) An interactive interface is established in the system to enable the user to select the aircraft's QAR parameters as driving values ​​for the event detection model operation; (1.2) Use QAR parameters in the model in Python language.

[0009] As a further technical solution, the airport parameters include airport name, ICAO code (the code designated by the International Civil Aviation Organization for airports and airlines around the world), ITAT code (the three-letter code set by the International Air Transport Association IATA for airports), airport elevation, latitude, and accuracy.

[0010] As a further technical solution, users can define different parameter information for different models in the model parameter library.

[0011] As a further technical solution, when configuring parameters, the system will search for values ​​of parameters with different frames that are calculated in an event detection model according to set rules and logic.

[0012] As a further technical solution, the value-seeking strategy according to the set rules and logic is as follows: the minimum unit of the cursor is set to 1 / the greatest common multiple of the frame rate, and each frame data corresponding to the greatest common multiple of the frame rate is regarded as a frame group, with positive numbers pointing forward and negative numbers pointing backward; if the frame pointed to by the cursor is a blank frame, first search backward for the first valid value, and if an overflow or exception occurs, reverse the cursor to search forward for a valid value.

[0013] As a further technical solution, when a head overflow event occurs during the value search process, the first valid value is taken as the parameter value; when a tail overflow event occurs during the value search process, the last valid value is taken as the parameter value.

[0014] Compared with the prior art, the present invention defines the flight event detection model based on Python language, which can more accurately define the event detection mechanism of different types of aircraft, and use QAR parameters as the trigger condition for event detection to help users more accurately locate events to aircraft flight parameters. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 It is a logical case diagram for frame alignment; Figure 2 This is the main functional block diagram of this system; Figure 3 Screenshot of the relevant regulations document for large elevation angles off the ground; Figure 4 A screenshot of the system for filling in general information into the corresponding fields of a newly created event; Figure 5 System screenshots of the original code written in Python code; Figure 6 After writing the Python code, click "Code Verification" and take a screenshot of the dialog box that pops up; Figure 7 This is a system screenshot of the "Notes" page. DETAILED DESCRIPTION

[0016] The present invention is described in detail below with reference to the accompanying drawings and specific embodiments. Example

[0017] This embodiment provides a flight event detection system based on Python language, the functions of which include: (1) In the event detection model of the system, the specific parameters defined in the Python language are used as the triggering conditions of the event, and the system provides the function of modularizing the parameter group events; (2) An airport database is created in the system, and parameters of different airports can be set, such as the latitude and longitude of the airport, the airport elevation and other parameters, and the information of the airport database can be referenced in the event detection model; (3) A model parameter library is created in the system, and model parameters can be referenced in the event detection model; (4) When the system is configuring parameters, a frame alignment method is invented. The parameters with different frames that are calculated in an event detection model are searched according to the set rules and logic, so that the data of different frames can be "precisely synchronized".

[0018] As a further technical solution, the specific features of technical solution (1) are as follows: (1.1) Establish an interactive interface in the system to enable the user to select the aircraft's QAR parameters as driving values ​​for the event detection model operation; (1.2) The selectable QAR parameters are all the parameters specified in the parameter version corresponding to the aircraft model. The names of the parameters refer to the "Flight Quality Monitoring Parameter Name Specification" issued by the Civil Aviation Administration of China; (1.3) Support for using QAR parameters in Python models.

[0019] As a further technical solution, the specific features of the technical solution (2) are as follows: (2.1) Establish an airport database in the system, in which users can add and maintain airport parameter information. The airport parameter information includes airport name, ICAO code, ITAT code, airport elevation, dimension, accuracy and other parameters; (2.2) Airport information can be introduced into the event detection model. The airport parameters can participate in the model operation to improve the accuracy of event detection. The correlation between different airports (parameter information) and the same event can also be studied.

[0020] As a further technical solution, the specific features of the technical solution (3) are as follows: (3.1) Establish a model parameter library in the system, in which users can define different parameter information of different models, such as tail strike angle; (3.2) Aircraft model information can be introduced into the event detection model, and the model parameters can be used in the model calculation to improve the accuracy of event detection for different aircraft models.

[0021] As a further technical solution, the specific features of the technical solution (4) are as follows: If there are parameters with different acquisition frame rates in the model operator, the data needs to be aligned. Secondly, in the actual operation process, the value needs to be searched based on the padded value. The search strategy is: set the minimum unit of the cursor to 1 / the greatest common multiple of the frame rate, and each frame data corresponding to the greatest common multiple of the frame rate is regarded as a frame group, with positive numbers forward and negative numbers backward. If the frame pointed to by the cursor is a blank frame, according to the strategy, first search for the first valid value backward. If an overflow or exception occurs, reverse the cursor forward to search for a valid value. In special cases, when a head overflow event occurs during the search process, take the first valid value as the parameter value; when a tail overflow event occurs during the search process, take the last valid value as the parameter value.

[0022] Let's take a practical example to illustrate the logic of frame alignment. Figure 1 As shown, the data collection frequencies of the four parameters A, B, C, and D are set to be different. Parameter A is a data with one frame every two seconds, parameter B is a data with one frame every one second, parameter C is a data with two frames every one second, parameter D is a data with four frames every one second, and parameter E is a data with eight frames every one second. Figure 1As shown, the E data has eight frames per second, so eight frames are used as the maximum frame distribution, and the frame group size is recorded as 8. The other A, B, C, and D data need to be divided into eight frames for value search. The pseudo code for setting the model operation target is as follows: Goal 1:

[0023] Goal 2:

[0024] Goal 3:

[0025] Explain the expressions involved in the above objectives: A represents the current frame of the A parameter data set, which is regarded as abbreviation of; Indicates the value of the corresponding frame of the next frame group of the B parameter data set, that is, offset 8 frames forward from the current position; Indicates that the C parameter data set is offset 1 frame forward from the current position; Indicates that the D parameter data set is offset 1 frame backward from the current position; Indicates that the D parameter dataset is offset 10 frames forward from the current position.

[0026] by For example, the B parameter data set is offset 8 frames forward to a blank frame. According to the default value search strategy, the first valid value is searched backward, such as Figure 1 As shown in the figure, the value is a valid value if it is shifted backward by one frame, so this value is used as The actual value of .

[0027] Next, this embodiment describes a specific application scenario of the system of the present invention. This application scenario takes a pitch high at lift off event of a domestic aircraft as an example, and introduces in detail the process of creating a detection model for this event. Specific QAR parameters, airport parameters, and aircraft model parameters are introduced into the model. Figure 2 shown.

[0028] The Pitch High at Lift Off event is used to monitor whether the pitch angle of the aircraft exceeds the standard when it takes off. During the takeoff phase, the occurrence of this event may cause the aircraft to strike the tail and cause structural damage. This is a single event.

[0029] The process of using the model to calculate the parameters of the event includes the following steps: Step 1: Clarify the definition of the event. The specific steps are as follows: Step 1.1: Check relevant regulatory documents, such as Figure 3 As shown: It is clear from the document that this event is a single event, and the type is pitch; Step 1.2: It is clarified that the monitoring parameters of the event include: flight phase FLIGHT_PHASE, wheel-borne signal WOW and pitch angle PITCH. Among them, wheel-borne signal WOW and pitch angle PITCH are acquisition parameters, and flight phase FLIGHT_PHASE is a derived parameter; Step 1.3: The monitoring point of the event is determined to be the moment when the main wheel leaves the ground; Step 1.4: The main monitoring logic of the event is clarified, including: 1. The moment when the aircraft's main wheels leave the ground; 2. Monitoring the pitch angle parameter value at that moment. If it exceeds the threshold value, trigger different warnings; Step 1.5: The monitoring criteria (i.e. triggering conditions) for this event are clarified as follows: Early Warning Pitch angle (°) Blue Warning >80% of the models have tail strike angles Orange warning >90% of the models have tail strike angles Red Alert >Model tail-scratch angle Step 2: Create basic information of the event in the system. The specific steps are as follows: Step 2.1: Open the flight quality monitoring system, select the decoding version, create a new event in the event, and fill in the general information. Fill in the information specified in step 1 in the corresponding fields, such as Figure 4 shown.

[0030] Step 3: Create an event detection model. The specific steps are as follows: Step 3.1: In the event created in step 2, switch to the "Code Logic" page, select the parameters to be monitored on this page and write the original code in Python code, such as Figure 5 shown.

[0031] Step 3.2: In the Python code, according to the actual operation of the aircraft, the take-off airport and landing airport of the aircraft can be selected in the parameters as needed, and the information of the airport can be associated with the event; Step 3.3: In the Python code, introduce the tail angle (Tail) parameter created in the business parameter information module of this model.

[0032] Step 4: Python code verification, the specific steps are as follows: Step 4.1: After writing the Python code, click "Code Verification" and a dialog box will pop up as shown below: Figure 6 , select Flight Information from the list.

[0033] Step 3.3: Click the "Start Verification" button and wait for the calculation feedback. If the verification is successful, the event detection model will be saved and published; if the verification fails, return to the code logic definition window, check the defined model, and continue verification.

[0034] Step 4: Create trigger conditions and other content in the system. The specific steps are as follows: Step 4.1: Switch to the "Notes" page, such as Figure 7 , fill in the trigger conditions, event levels and corresponding values ​​in the corresponding fields on this page. This page is optional and can be filled in as needed.

[0035] When the collected flight parameters enter the published event detection model for calculation, for parameters of different frames, the system will automatically perform frame alignment "value search" processing during calculation, so that each parameter can be "precisely synchronized" at the same time.

[0036] When the flight data meets the monitoring criteria defined in the model, that is, the monitoring conditions are triggered, the system will issue relevant warnings to notify the monitoring personnel.

[0037] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. It should be pointed out that any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A flight event detection system based on Python language, including an event detection model, characterized in that: The system includes: (1) In the event detection model of the system, specific parameters defined in Python are used as trigger conditions for events, and the system provides the function of modularizing the parameter group events; (2) An airport database is created in the system to set different airport parameters, and the information in the airport database can be referenced in the event detection model; (3) A model parameter library is created in the system, and the model parameters can be referenced in the event detection model.

2. A flight event detection system based on Python language according to claim 1, characterized in that: The specific steps of using the specific parameters defined in Python language as the triggering conditions of the event include: (1.1) An interactive interface is established in the system to enable the user to select the aircraft's QAR parameters as driving values ​​for the event detection model operation; (1.2) Use QAR parameters in the model in Python language.

3. The flight event detection system based on Python language according to claim 1, characterized in that: The airport parameters include airport name, ICAO code, ITAT code, airport elevation, latitude, and accuracy.

4. The flight event detection system based on Python language according to claim 1, characterized in that: Users can define different parameter information for different models in the model parameter library.

5. The flight event detection system based on Python language according to claim 1, characterized in that: When configuring parameters, the system will search for values ​​of parameters with different frames that are calculated in an event detection model according to the set rules and logic.

6. The flight event detection system based on Python language according to claim 5, characterized in that: The value search strategy according to the set rules and logic is as follows: set the minimum unit of the cursor to 1 / the greatest common multiple of the frame rate, and each frame data corresponding to the greatest common multiple of the frame rate is regarded as a frame group, with positive numbers forward and negative numbers backward. If the frame pointed to by the cursor is a blank frame, first search backward for the first valid value. If an overflow or exception occurs, reverse the cursor to search forward for a valid value.

7. The Python-based flight event detection system according to claim 6, characterized in that: When a header overflow event occurs during the value search process, the first valid value is taken as the parameter value; when a tail overflow event occurs during the value search process, the last valid value is taken as the parameter value.

Images