A method and system for judging the stable state during takeoff and landing on a helicopter deck

Abstract

The present invention belongs to the technical field of ship and ocean engineering, and discloses a method and system for judging the stable period state of helicopter deck takeoff and landing. The method includes: constructing a multi-layer long short-term neural network prediction model Multi-LSTM-Aero based on dynamic constraints, reading and preprocessing ship motion data and environmental wind field data; predicting the three-degree-of-freedom motion envelope of the ship; predicting environmental wind field data; calculating the helicopter dynamic model with the prediction results of the ship's three-degree-of-freedom motion envelope and the prediction results of environmental wind field data; based on the obtained prediction results of helicopter control quantities and helicopter attitude quantities, judging and calculating the stable period of deck takeoff and landing through the set corresponding helicopter control limits and attitude limits. The present invention can effectively improve the ability of operation personnel and command personnel to grasp the takeoff and landing state, and provide technical support for improving the efficiency and safety of takeoff and landing operations.

Description

Technical Field

[0001] The invention belongs to the technical field of ship and ocean engineering, and in particular relates to a method and system for determining the state of a helicopter deck during take-off and landing safety period. Background Art

[0002] With the increase in the demand for offshore operations, the coordinated operation of ship-borne helicopters and ships has become an indispensable mode of operation. The diversified mission requirements have put forward higher requirements for the take-off and landing capabilities of ship-borne helicopters in complex marine environments. However, due to the swaying motion of the ship, taking off and landing operations on the ship's helicopter deck has become one of the difficulties in ship-aircraft collaborative operations. The take-off and landing process of ship-borne helicopters on the moving deck is significantly different from the land-based take-off and landing operations. The movement of the ship itself and the complex external environment pose a great threat to the maneuverability of the helicopter. These environmental factors not only directly affect the take-off and landing capabilities of the helicopter, but also weaken the response speed and execution efficiency of the ship-aircraft system in emergency tasks to a certain extent. For ships sailing at sea, as the wind and wave environment changes at sea, the hull posture will also change accordingly, especially in high sea conditions, the movement amplitude of the ship is very violent. However, during the violent ship movement, there is often a period of stable state, which is usually called the ship movement stability period. For the take-off and landing operations of ship-borne helicopters, the stable period is the most suitable time segment for take-off and landing operations in high sea conditions, so these segments can also be called the take-off and landing stability period of ship-borne helicopter decks.

[0003] At present, the perception of the deck take-off and landing safety period mostly relies on the shipborne helicopter deck take-off and landing indication system, but this system can only indicate the ship's motion state at the current moment, and cannot predict the safety period state in the future, let alone the helicopter attitude in the future. Therefore, how to predict the environmental factors in the take-off and landing process through historical information, and to judge the take-off and landing attitude and control characteristics of the target helicopter through the predicted information has become a difficult problem in solving the research on the shipborne helicopter take-off and landing safety period.

[0004] Through the above analysis, the problems and defects of the existing technology are as follows: the current traditional method of judging the safe period of shipborne helicopter take-off and landing is mainly based on the light indication of the deck take-off and landing auxiliary system in conjunction with the command of the deck take-off and landing guide. This method cannot meet the needs of judging the ship's motion trend in the future and predicting the surrounding wind field environment. Therefore, it can only provide an experience-based operation safety judgment. The judgment standard is difficult to quantify, and the judgment result lacks a description of the helicopter take-off and landing characteristics, which causes many uncertainties in the shipborne helicopter take-off and landing operations. Summary of the invention

[0005] To overcome the problems existing in the related technologies, the disclosed embodiments of the present invention provide a method and system for judging the stable landing and takeoff period state of a helicopter deck, specifically related to a method for judging the stable landing and takeoff period state of a helicopter deck based on a Multi-LSTM-Aero prediction model.

[0006] The technical solution is as follows: A method for judging the stable landing and takeoff period state of a helicopter deck, including:

[0007] S1, constructing a multi-layer long short-term neural network prediction model Multi-LSTM-Aero based on dynamic constraints;

[0008] S2, the model realizes the reading and preprocessing of ship motion data and environmental wind field data;

[0009] S3, the model realizes the prediction of the three-degree-of-freedom motion envelope of the ship;

[0010] S4, the model realizes the prediction of environmental wind field data;

[0011] S5, inside the model, the prediction results of the three-degree-of-freedom motion envelope of the ship and the prediction results of environmental wind field data are used for helicopter dynamics model calculation;

[0012] S6, based on the obtained helicopter control quantity prediction results and helicopter attitude quantity prediction results, through the set corresponding helicopter control limits and attitude limits, the judgment and calculation of the stable landing and takeoff period on the deck are carried out.

[0013] In step S2, when the model realizes the reading and preprocessing of ship motion data and environmental wind field data, it includes:

[0014] Before preprocessing the input data, according to requirements, set the basic parameters of the multi-layer long short-term neural network prediction model Multi-LSTM-Aero based on dynamic constraints. The basic parameters specifically include the total duration of reading historical data , the total duration of predicting data ; after setting the basic parameters of the model, read and preprocess the input data; using ship motion data and environmental wind field data as inputs, which are represented in turn as: ship motion data , environmental wind field data ; where is the timestamp of the read data, ; clip and correct the data at equal time intervals to make the data frequencies of all time history data the same; the required types of ship motion data include ship roll and ship pitch data; the type of wind field data only includes vector wind speed data.

[0015] Further, the input data is sequentially transmitted to the motion envelope prediction module for predicting the three-degree-of-freedom motion envelope of the ship, the environmental wind prediction module for predicting environmental wind field data, and the helicopter dynamics constraint module for calculating the helicopter dynamics model in the Multi-LSTM-Aero prediction model based on dynamic constraints. According to the set total duration of the predicted data an output of a specified length is obtained, with the helicopter control amount and the helicopter attitude amount as the outputs, which are sequentially expressed as: helicopter control amount and helicopter attitude amount , where is the time change amount of the prediction result compared to the time stamp of the historical data, , is the total duration of the predicted data.

[0016] In step S3, the model realizes the prediction of the three-degree-of-freedom motion envelope of the ship, including:

[0017] The data stream enters the motion envelope prediction module, and the processed ship motion data is subjected to envelope extraction, and the ship motion data is transformed into the upper envelope of the motion data and the lower envelope of the motion data , which are simultaneously used as the input of the LSTM model, so as to obtain the ship motion envelope prediction result, which is respectively expressed as the upper envelope prediction result of the motion data and the lower envelope prediction result of the motion data .

[0018] In step S4, the model realizes the prediction of environmental wind field data, including:

[0019] After the environmental wind field data enters the environmental wind prediction module, FFT filtering is performed to remove the high-frequency noise of the environmental wind field data. This denoising process smooths the wind field data to obtain , and this is used as the input of the Multi-LSTM-Aero model to obtain the smoothed environmental wind prediction result .

[0020] In step S5, the helicopter dynamics model is calculated, including:

[0021] After aligning the time stamps of the upper envelope prediction result of the motion data, the lower envelope prediction result of the motion data, and the smoothed environmental wind prediction result obtained in the intermediate process, they are input into the dynamics constraint module;

[0022] The rigid body dynamic equation of the helicopter is used to solve the force condition of the helicopter's rigid body, and then the moment condition of the helicopter is solved through the rotational motion equation; the aerodynamic lift is solved through the lift equation in the aerodynamic equation, and the resistance of the rotor power is solved through the resistance equation;

[0023] Based on the calculation results of the rigid body dynamics, aerodynamics, and rotor dynamics in the helicopter dynamics system, the pilot's control amount is calculated through the force control equation; meanwhile, the change in the helicopter's attitude is calculated through the Euler angle differential equation.

[0024] Furthermore, the use of the rigid body dynamic equation to solve the force condition of the helicopter's rigid body includes:

[0025] Using the rigid body dynamic equation to solve the force condition of the helicopter, where is the mass of the helicopter, is the velocity vector, is the derivative of the velocity vector with respect to time, is the aerodynamic force, is the gravity;

[0026] The aerodynamic force is a function of the environmental wind forecast result, that is , is the smoothed environmental wind forecast result, ;

[0027] The use of the rotational motion equation to solve the moment condition of the helicopter includes:

[0028] Using the rotational motion equation to solve the moment condition of the helicopter, where is the inertia matrix, is the angular velocity vector, is the derivative of the angular velocity vector with respect to time, is the aerodynamic moment, is the remaining moment; the aerodynamic moment is a function of the environmental wind forecast result.

[0029] Furthermore, the use of the lift equation in the aerodynamic equation to solve the aerodynamic lift includes: using the lift equation in the aerodynamic equation to solve the lift, where is the lift, is the lift coefficient, is the air density, is the relative velocity of the rotor, is the force area of the rotor;

[0030] The use of the resistance equation to solve the resistance of the rotor power includes:

[0031] Through the drag equation Solve for the drag force, where is the drag force is the drag coefficient;

[0032] Use the rotor dynamic equation Solve for the forces acting on the rotor, where is the rotor torque is the rotor thrust is the rotor radius;

[0033] The Euler angle differential equation is expressed as , respectively represent the angular velocity components of the helicopter in the body coordinate system.

[0034] In step S6, perform the judgment and calculation of the stable period of deck takeoff and landing, including:

[0035] After obtaining the predicted results of the helicopter control quantity and the predicted results of the helicopter attitude quantity, through the set helicopter control limit and the attitude limit , successively compare the predicted results of the helicopter control quantity , the predicted results of the helicopter attitude quantity . If and , then it is considered that is the safe time, so as to screen out the time segments that meet the takeoff and landing conditions, and further obtain more comprehensive prediction and discrimination information on the stable period of takeoff and landing.

[0036] Another object of the present invention is to provide a system for judging the stable state of the helicopter deck takeoff and landing period, which implements the method for judging the stable state of the helicopter deck takeoff and landing period. The system includes:

[0037] A data reading and preprocessing module for reading and preprocessing ship motion data and environmental wind field data;

[0038] A motion envelope prediction module for predicting the three-degree-of-freedom motion envelope of the ship;

[0039] An environmental wind field prediction module for predicting environmental wind field data;

[0040] A helicopter dynamics calculation module for performing helicopter dynamics model calculations on the predicted results of the three-degree-of-freedom motion envelope of the ship and the predicted results of environmental wind field data;

[0041] The deck takeoff and landing stable period judgment module is used to judge and calculate the deck takeoff and landing stable period based on the obtained helicopter control quantity prediction results and helicopter attitude quantity prediction results, through the set corresponding helicopter control limits and attitude limits.

[0042] Combining all the above technical solutions, the beneficial effects of the present invention are as follows: The present invention uses the Multi-LSTM-Aero prediction model, with the three-degree-of-freedom ship motion data and environmental wind field data as the model inputs, and the helicopter attitude and control quantity within a future period of time as the model outputs. At the same time, the prediction model can output the three-degree-of-freedom motion envelope prediction results and environmental wind speed prediction results through the intermediate process, and then give the time segments that meet the takeoff and landing safety requirements through the takeoff and landing attitude limits and control limits of shipborne helicopters.

[0043] The present invention integrates ship motion information, environmental wind field information, and helicopter dynamics calculation models, integrates ships, shipborne helicopters, and environmental information, and constructs a relatively comprehensive helicopter takeoff and landing stable period discrimination model, further increasing the comprehensiveness and credibility of discrimination information.

[0044] The present invention uses a multi-layer neural network structure to integrate two types of multivariate information of ship motion and environmental wind field, and uses historical information to predict the ship motion and environmental wind field information within a future period of time, so as to provide effective prediction data for the subsequent dynamics model calculation, and then obtain relatively rich takeoff and landing state information.

[0045] Compared with the prior art, the advantages of the present invention further include: The traditional method for judging the stable period of shipborne helicopter takeoff and landing mainly relies on the light indication of the deck takeoff and landing assistance system and the command of the deck takeoff and landing guide. This method cannot meet the requirements of judging the ship motion trend within a future period of time and predicting the surrounding wind field environment, and at the same time cannot provide additional helicopter attitude and control information for the deck takeoff and landing operation, which makes helicopter pilots face greater uncertain factors, thus affecting the operation efficiency and safety. This method comprehensively considers the three-degree-of-freedom motion data of the ship and the environmental wind field data. On the basis of considering ship and environmental factors, the takeoff and landing dynamics calculation formula is incorporated into the prediction model, so that the model output results include the prediction of the helicopter takeoff and landing attitude and control quantity. By comparing with various limits, the discrimination result of the takeoff and landing stable period is obtained, so as to describe the takeoff and landing state of shipborne helicopters in more detail. This method can effectively improve the ability of operation executors and commanders to grasp the takeoff and landing state, and provide technical support for improving the takeoff and landing operation efficiency and takeoff and landing operation safety. Description of the Drawings

[0046] The drawings here are incorporated into the specification and constitute a part of this specification, showing the embodiments consistent with the present disclosure, and are used together with the specification to explain the principles of the present disclosure;

[0047] Figure 1 It is a flow chart of a method for judging the stable landing and take-off period state of a helicopter deck provided by an embodiment of the present invention;

[0048] Figure 2 It is a structural diagram of a Multi-LSTM-Aero prediction model provided by an embodiment of the present invention;

[0049] Figure 3 It is a diagram of the output result of the motion envelope prediction module provided by an embodiment of the present invention;

[0050] Figure 4 It is an effect diagram of the filtering process of the environmental wind prediction module provided by an embodiment of the present invention;

[0051] Figure 5 It is a diagram of the output result of the environmental wind prediction module provided by an embodiment of the present invention;

[0052] Figure 6 It is a diagram of the output result of the helicopter dynamics constraint module provided by an embodiment of the present invention;

[0053] Figure 7 It is a diagram of the input data of ship motion data provided by an embodiment of the present invention;

[0054] Figure 8 It is a diagram of the input data of environmental wind field data provided by an embodiment of the present invention;

[0055] Figure 9 It is a diagram of the output result of the motion envelope prediction module provided by an embodiment of the present invention;

[0056] Figure 10 It is a diagram of the output result of the environmental wind prediction module provided by an embodiment of the present invention;

[0057] Figure 11 It is a diagram of the predicted output result of the Multi-LSTM-Aero prediction model provided by an embodiment of the present invention;

[0058] Figure 12 It is a diagram of the discrimination result of the Multi-LSTM-Aero prediction model provided by an embodiment of the present invention. Specific embodiments

[0059] In order to make the above objects, features and advantages of the present invention more obvious and understandable, the following detailed description of the specific embodiments of the present invention will be made in conjunction with the accompanying drawings. Many specific details are set forth in the following description in order to fully understand the present invention. However, the present invention can be implemented in many other ways different from those described herein, and those skilled in the art can make similar improvements without departing from the connotation of the present invention. Therefore, the present invention is not limited by the specific embodiments disclosed below.

[0060] The innovation of the present invention lies in: in view of the problems of single discriminant information and neglect of the attitude of shipborne helicopters existing in the traditional means of judging the stable period when predicting and judging the takeoff and landing window period of shipborne helicopters in a complex environment, a multi-layer long short-term neural network prediction model Multi-LSTM-Aero based on dynamic constraints constructed by an environmental motion information prediction module and a helicopter dynamics constraint module is proposed, which optimizes the method and process for judging the takeoff and landing window period and enriches the means for obtaining takeoff and landing discriminant information. The model uses the stacking idea to combine the ship motion envelope prediction module based on LSTM and the wind speed time history prediction module to form an environment-motion information fusion prediction module, and then calculates the relative attitude between the ship and the aircraft and the helicopter control state in the target environment through the helicopter dynamics constraint module, so as to provide more abundant reference for the stable period judgment.

[0061] Example 1, Technical terms in the field related to the present invention:

[0062] Six-degree-of-freedom motion of a ship:

[0063] (1) Three degrees of freedom of a ship: roll, pitch, heave.

[0064] The bow-stern (front-rear) direction of the ship is called the longitudinal direction, the port-starboard (left-right) direction is called the transverse direction, and the upper deck-bottom of the cabin (up-down) direction is called the vertical direction.

[0065] The translational motion in the front-rear direction is called surge, the translational motion in the left-right direction is called sway, and the translational motion in the up-down direction is called heave.

[0066] The angular change in the left-right direction is called roll, the angular change in the front-rear direction is called pitch, and the angular change of the bow in the left-right direction is called yaw.

[0067] (2) Ship motion stable period: When a ship is operating at sea, a period of time during which the ship's motion is relatively stable, and the probability of risk occurring during this period of operation is relatively small.

[0068] The method for judging the stable period state of helicopter deck takeoff and landing provided by the embodiment of the present invention inputs the ship motion time history data and environmental wind field information measured by the sensors carried on the ship, and uses a multi-layer long short-term neural network prediction model based on dynamic constraints (Multi-LSTM-Aero prediction model) incorporating the shipborne helicopter takeoff and landing dynamics model to predict the predicted values of the fuselage attitude and control performance during helicopter takeoff and landing, and at the same time can output the prediction results of the ship motion envelope and environmental wind field information. Relying on the model prediction results, according to the helicopter operation restrictions and attitude restriction requirements, the time segments of the stable period of deck takeoff and landing are screened, and then more abundant information reference is provided for the shipborne helicopter takeoff and landing operation.

[0069] Such asFigure 1 As shown in the figure, a method for judging the stable period state of helicopter deck take-off and landing includes:

[0070] S1, construct a multi-layer long short-term neural network prediction model Multi-LSTM-Aero based on dynamic constraints;

[0071] S2, the model realizes the reading and preprocessing of ship motion data and environmental wind field data;

[0072] The present invention aims at the problem of predicting and judging the stable period of shipborne helicopter take-off and landing, and proposes a prediction model based on Multi-LSTM-Aero. Its structure mainly includes three modules, namely, the motion envelope prediction module (ship motion envelope prediction module), the environmental wind prediction module (environmental wind speed time history prediction module), and the dynamic constraint module (helicopter dynamic constraint module). The structure of the Multi-LSTM-Aero prediction model is as Figure 2 shown;

[0073] Before preprocessing the input data, the basic parameters of the model need to be set according to requirements. The basic parameters of the model specifically include the total duration of reading historical data , the total duration of predicted data . After setting the basic parameters of the model, the input data is read and preprocessed. The model takes ship motion data and environmental wind field data as inputs, which are represented in turn as: ship motion data , including ship roll, ship pitch, and ship heave data; environmental wind field data , where represents the timestamp of the read data, , represents the total duration of reading historical data. The data is cropped and corrected at equal time intervals to ensure that the data frequencies of all time history data are the same. The types of ship motion data required by this model include ship roll and ship pitch data; the types of wind field data only include vector wind speed data.

[0074] The input data is sequentially transmitted to the motion envelope prediction module, the environmental wind prediction module, and the helicopter dynamic constraint module. According to the set total duration of predicted data of the model the model output of a specified length is obtained. The model takes helicopter control quantities and helicopter attitude quantities as outputs, which are represented in turn as: helicopter control quantity , helicopter attitude quantity ; where represents the total duration of reading historical data, represents the time change amount of the model prediction result compared with the timestamp of historical data, , represents the total duration of predicted data of the model.

[0075] It can be understood that the existing technology calculates the fuselage swing generated by the helicopter's rocking with the ship when the helicopter is static on the ship's deck, which belongs to the category of inertial force. The calculation of the helicopter's control amount and attitude amount in the present invention is for the attitude of the helicopter in the hovering state and the pilot's control amount, which belongs to the categories of aerodynamics and control.

[0076] S3. The model realizes the prediction of the ship's three-degree-of-freedom motion envelope;

[0077] In the model prediction stage, the data stream first enters the motion envelope prediction module. In this module, the present invention innovatively proposes to extract the envelope of the processed ship motion data, so as to convert the ship motion data into the upper envelope of the motion data and the lower envelope of the motion data , and at the same time use them as the input of the LSTM model, so as to obtain the ship motion envelope prediction results, which are respectively expressed as the upper envelope prediction result of the motion data and the lower envelope prediction result of the motion data . This result is the intermediate data stream of the model and provides data support for the subsequent dynamic constraint calculation. Among them, the output result of the motion envelope prediction module is as Figure 3 shown.

[0078] S4. The model realizes the prediction of the environmental wind field data;

[0079] After the environmental wind field data enters the environmental wind prediction module, it first undergoes an FFT filtering process to remove the high-frequency noise of the environmental wind field data. This denoising process smooths the wind field data, and then obtains , and uses this as the input of the Multi-LSTM-Aero model to obtain the smoothed environmental wind prediction result . Among them, represents the total duration of reading historical data, represents the time change amount of the model prediction result compared with the time stamp of the historical data, , represents the total duration of the model prediction data. This result is also the intermediate data stream of the model and provides data support for the subsequent dynamic constraint calculation.

[0080] Among them, the filtering effect of the environmental wind prediction module is as Figure 4 shown. The output result of the environmental wind prediction module is as Figure 5 shown.

[0081] S5. Inside the model, the ship's three-degree-of-freedom motion envelope prediction result and the environmental wind field data prediction result are used for the calculation of the helicopter dynamics model;

[0082] The upper envelope prediction result of the motion data obtained in the intermediate process , the lower envelope prediction result of the motion data , and the smoothed environmental wind prediction result After aligning the timestamps, input them into the dynamic constraint module (helicopter dynamic constraint module). In the formula represents the total duration of reading historical data, represents the time change amount of the model prediction result compared to the historical data timestamp, , represents the total duration of the model prediction data;

[0083] The present invention innovatively proposes to use the rigid body dynamics equation to solve the force situation of the helicopter. The technical effect of this formula is: incorporating the helicopter dynamic constraints into the prediction model to provide a calculation basis for the final output of the dynamic results. In the formula, is the mass of the helicopter, is the velocity vector, is the derivative of the velocity vector with respect to time, is the aerodynamic force, is the gravity.

[0084] The present invention innovatively proposes that the aerodynamic force is a function of the environmental wind prediction result, that is , represents the smoothed environmental wind prediction result , represents the total duration of reading historical data, represents the time change amount of the model prediction result compared to the historical data timestamp, , represents the total duration of the model prediction data.

[0085] Furthermore, solve the torque situation of the helicopter through the rotational motion equation . Among them, is the inertia matrix, is the angular velocity vector, is the derivative of the angular velocity vector with respect to time, is the aerodynamic torque, is the remaining torque; the aerodynamic torque is a function of the environmental wind prediction result, that is , represents the smoothed environmental wind prediction result , represents the total duration of reading historical data, represents the time change amount of the model prediction result compared to the historical data timestamp, , represents the total duration of the model prediction data.

[0086] Solve for the lift force through the lift equation in the aerodynamic equations, where, is the lift force, is the lift coefficient, is the air density, is the relative velocity of the rotor, is the force-bearing area of the rotor; Solve for the resistance force through the drag equation where, is the drag force, is the drag coefficient. Use the rotor power equation to solve for the force condition of the rotor, is the rotor torque, is the rotor thrust, is the rotor radius.

[0087] Integrate the calculation results of the rigid body dynamics, aerodynamics, and rotor dynamics in the helicopter dynamics system, and calculate the pilot's control amount through the force control equation Calculate the pilot's control amount, represents the control force calculation equation, which is the aerodynamic force the lift force the drag force the rotor torque function of, represents the helicopter pilot's control amount calculation equation, where, represents the helicopter pilot's control amount. At the same time, calculate the change in the helicopter's attitude amount through the Euler angle differential equation, and the Euler angle differential equation is expressed as respectively represent the angular velocity components of the helicopter in the body coordinate system.

[0088] Obtain the helicopter control amount prediction result the helicopter attitude amount prediction result through the above formulas, where, represents the total duration of reading historical data, represents the time change amount of the model prediction result compared to the historical data timestamp, represents the total duration of the model prediction data.

[0089] Figure 6 Among them, the output result of the helicopter dynamics constraint module is as shown in

[0090] S6. Based on the obtained helicopter control amount prediction result and helicopter attitude amount prediction result, perform the judgment and calculation of the stable landing and takeoff period on the deck through the set corresponding helicopter control limit and attitude limit;

[0091] ​​​​Specifically, after obtaining the helicopter control input prediction result and the helicopter attitude prediction result, through the set helicopter control limits and attitude limits , the helicopter control input prediction result and the helicopter attitude prediction result are compared in sequence. If and , then the moment is considered a safe moment. represents the total duration of reading historical data, and

[0092] represents the time change of the model prediction result compared to the time stamp of the historical data, thereby screening out the time segments that meet the takeoff and landing conditions, and further obtaining more comprehensive prediction and discrimination information on the stable takeoff and landing period.

[0093] Embodiment 2. The present invention provides a system for discriminating the state of the stable takeoff and landing period on a helicopter deck, including:

[0094] A data reading and preprocessing module, configured to read and preprocess ship motion data and environmental wind field data;

[0095] A motion envelope prediction module, configured to predict the three-degree-of-freedom motion envelope of the ship;

[0096] An environmental wind field prediction module, configured to predict environmental wind field data;

[0097] A helicopter dynamics calculation module, configured to perform helicopter dynamics model calculations on the predicted results of the three-degree-of-freedom motion envelope of the ship and the predicted results of environmental wind field data;

[0098] To further illustrate the related effects of the embodiments of the present invention, the following experiments are carried out.

[0099] Through on-board sensors or by calling offline data, ship motion data and environmental wind field data are obtained, where represents the time stamp of the read data, , and Figure 7 represents the total duration of reading historical data. After detecting outliers in the data, it is used as the input to the Multi-LSTM-Aero prediction model. The input data of the ship motion data is as shown in Figure 8 . The input data of the environmental wind field data is as shown in

[0100] Set the total duration for the prediction model to read historical data , and the total duration of the model's predicted data . Input the ship motion data , where represents the timestamp of the read data , represents the total duration of reading historical data. Input it into the motion envelope prediction module, and the model automatically extracts the ship motion envelope. The envelope consists of two upper and lower lines, corresponding to and transformed into the upper envelope of the motion data , the lower envelope of the motion data . Input the envelope data into the LSTM network structure embedded in the model for motion envelope prediction. The output result of the motion envelope prediction module is as shown in Figure 9 .

[0101] Input the environmental wind field data , where represents the timestamp of the read data , represents the total duration of reading historical data. Input it into the environmental wind prediction module, and after internal filtering processing of the model, obtain the smoothed wind field environmental data , and then input it into the LSTM prediction module. The output result of the environmental wind prediction module is as shown in Figure 10 .

[0102] After aligning the timestamps of the prediction results of the upper envelope of the motion data , the prediction results of the lower envelope of the motion data , and the smoothed prediction results of the environmental wind , input them into the dynamic constraint module. In the formula, represents the total duration of reading historical data represents the time change of the model's prediction result compared to the timestamp of the historical data , represents the total duration of the model's predicted data. Output the prediction results of the helicopter control amount and the prediction results of the helicopter attitude amount . The prediction output result of the Multi-LSTM-Aero prediction model is as shown in Figure 11 .

[0103] Set helicopter control limits and attitude limits . Then, the safety state of the helicopter at each moment can be judged from the output result of the prediction model. The discrimination result of the Multi-LSTM-Aero prediction model is as shown in Figure 12 .

[0104] As described above, it is only a relatively optimal specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention, any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be covered within the protection scope of the present invention.

Claims

1. A method for determining the state of a helicopter deck during take-off and landing stability period, characterized in that: The method includes: S1, construct a multi-layer long-term and short-term neural network prediction model Multi-LSTM-Aero based on dynamic constraints; S2, the model realizes the reading and preprocessing of ship motion data and environmental wind field data; S3, the model realizes the prediction of the three-degree-of-freedom motion envelope of the ship; S4, the model realizes the prediction of environmental wind field data; S5, using the prediction results of the three-degree-of-freedom motion envelope of the ship and the prediction results of the environmental wind field data to calculate the helicopter dynamics model within the model; S6, based on the obtained helicopter control amount prediction results and helicopter attitude amount prediction results, by setting corresponding helicopter control restrictions and attitude restrictions, performing deck take-off and landing safety period judgment and calculation; In step S2, the model implements the reading and preprocessing of ship motion data and environmental wind field data, including: Before preprocessing the input data, set the basic parameters of the Multi-LSTM-Aero multi-layer long-term and short-term neural network prediction model based on dynamic constraints according to the needs. The basic parameters specifically include the total time of reading historical data , total duration of forecast data ; After setting the basic parameters of the model, read and preprocess the input data; take the ship motion data and environmental wind field data as input, and express them as follows: ship motion data , Environmental wind field data ;in, The timestamp of the read data. ; The data is clipped and corrected at equal time intervals to make the data frequency of all time history data the same; the required ship motion data type includes ship roll and ship pitch data; the wind field data type only includes vector wind speed data; The input data is sequentially transmitted to the motion envelope prediction module for ship three-degree-of-freedom motion envelope prediction of the multi-layer long-term and short-term neural network prediction model Multi-LSTM-Aero based on dynamic constraints, the environmental wind prediction module for environmental wind field data prediction, and the helicopter dynamic constraint module for helicopter dynamic model calculation. The output of the specified length is obtained, with the helicopter control amount and the helicopter attitude amount as the output, which are expressed as: helicopter control amount 、Helicopter attitude ,in, is the time change of the predicted result compared to the historical data timestamp, , is the total duration of the predicted data.

2. The method for determining the state of the helicopter deck during take-off and landing stability period according to claim 1 is characterized in that: In step S3, the model implements the prediction of the three-degree-of-freedom motion envelope of the ship, including: The data stream enters the motion envelope prediction module, extracts the envelope of the processed ship motion data, and converts the ship motion data into Transformed into motion data upper envelope and motion data envelope , and used as the input of the LSTM model at the same time, so as to obtain the ship motion envelope prediction results, which are respectively expressed as the envelope prediction results on the motion data Envelope forecast results under motion data .

3. The method for determining the state of the helicopter deck during take-off and landing stability period according to claim 1, characterized in that: In step S4, the model implements environmental wind field data prediction, including: After entering the environmental wind forecast module, the environmental wind field data is subjected to FFT filtering to remove the high-frequency noise of the environmental wind field data. The denoising process smoothes the wind field data to obtain , and use this as the input of the Multi-LSTM-Aero model to obtain the smoothed ambient wind forecast result .

4. The method for determining the state of the helicopter deck during take-off and landing stability period according to claim 1, characterized in that: In step S5, helicopter dynamics model calculation is performed, including: The upper envelope forecast result of the motion data, the lower envelope forecast result of the motion data, and the smoothed ambient wind forecast result obtained in the intermediate process are aligned with the timestamp and input into the dynamic constraint module; The rigid body dynamics equation is used to solve the rigid body dynamic force of the helicopter, and then the torque of the helicopter is solved by the rotational motion equation; the aerodynamic lift is solved by the lift equation in the aerodynamic equation, and the drag of the rotor power is solved by the drag equation; The calculation results of rigid body dynamics, aerodynamics and rotor dynamics in the helicopter dynamics system are integrated to obtain the pilot control variable through the force control equation; at the same time, the helicopter attitude change is calculated through the Euler angle differential equation.

5. The method for determining the state of the helicopter deck during take-off and landing stability period according to claim 4 is characterized in that: The method of solving the rigid body dynamics equation to obtain the dynamic force of the helicopter rigid body includes: Using the rigid body dynamics equation Solve the helicopter force condition, where: is the mass of the helicopter, is the velocity vector, Take the time derivative of the velocity vector, For aerodynamics, is gravity; The aerodynamic force is a function of the ambient wind forecast, i.e. , is the smoothed ambient wind forecast result, ; The method of solving the moment condition of the helicopter by the rotational motion equation includes: By rotating the equation of motion Solve for the moment acting on the helicopter, where is the inertia matrix, is the angular velocity vector, is the time derivative of the angular velocity vector, is the aerodynamic torque, are the remaining moments; the aerodynamic moment is a function of the ambient wind forecast result.

6. The method for determining the state of the helicopter deck during take-off and landing stability period according to claim 4, characterized in that: The method of solving the aerodynamic lift by using the lift equation in the aerodynamic equation comprises: solving the aerodynamic lift by using the lift equation in the aerodynamic equation Solve for the lift force, where For lift, is the lift coefficient, is the air density, is the relative speed of the rotor, is the force bearing area of ​​the rotor; The method of solving the drag force on the rotor power by using the drag equation includes: Through the resistance equation Solve for the resistance, where For resistance, is the drag coefficient; Using the rotor dynamics equation Solve the rotor force condition, where is the rotor torque, is the rotor thrust, is the rotor radius; The Euler angle differential equation is expressed as , They respectively represent the angular velocity components of the helicopter in the body coordinate system.

7. The method for determining the state of the helicopter deck during take-off and landing stability period according to claim 1, characterized in that: In step S6, the deck take-off and landing safety period is judged and calculated, including: After obtaining the prediction results of the helicopter control amount and the helicopter attitude amount, the helicopter control limit is set. and posture restrictions , and the helicopter control prediction results are , Helicopter attitude prediction results For comparison, if and , then it is believed that The moment is a safe moment, so as to screen out the time segments that meet the take-off and landing conditions, and then obtain more comprehensive prediction and judgment information of the take-off and landing safe period.

8. A helicopter deck take-off and landing safety period state judgment system, characterized in that: The system implements the method for determining the state of the helicopter deck during take-off and landing stability period according to any one of claims 1 to 7, and the system comprises: Data reading and preprocessing module, used for reading and preprocessing ship motion data and environmental wind field data; Motion envelope prediction module, used for predicting the three-degree-of-freedom motion envelope of ships; Environmental wind field prediction module, used for environmental wind field data prediction; The helicopter dynamics calculation module is used to calculate the helicopter dynamics model based on the prediction results of the ship's three-degree-of-freedom motion envelope and the environmental wind field data; The deck take-off and landing safety period judgment module is used to judge and calculate the deck take-off and landing safety period based on the obtained helicopter control quantity prediction results and helicopter attitude quantity prediction results, by setting the corresponding helicopter control restrictions and attitude restrictions.

Images