Parametric Models for Optimizing Aircraft Performance

By dividing the flight trajectory into multiple segments and combining aerodynamics and engine models, using least squares method to iterate the parameters, the problem of poor convergence in the existing technology relying on unpublished data and nonlinear parameters is solved, and more accurate and reliable The construction of the aircraft performance model of the stick has improved the effect of aircraft operation optimization.

CN111353209BActive Publication Date: 2024-11-29THALES SA
View PDF 10 Cites 0 Cited by

Patent Information

Application Number
CN201911315109.3
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2019-12-19
Publication Date
2024-11-29
Estimated Expiration
2039-12-19

AI Technical Summary

Technical Problem

The prior art relies on performance data provided by undisclosed aircraft manufacturers when optimizing aircraft operations, and nonlinear parameter estimation methods tend to converge to local minimum values ​​during large-scale modeling, lack robustness, and it is difficult to independently determine aircraft. performance model.

Method used

By dividing the flight trajectory into multiple segments, each segment controlled by different systems of equations, combining aerodynamic models and engine speed models, the parameters are iteratively determined using least squares method to ensure convergence on the thrust and drag models and at different flight stages The aircraft performance model is independently constructed using the recursive least squares method and the two-stage least squares method.

Benefits of technology

A closer to reality thrust and drag model is achieved, ensuring convergence of least squares estimation, able to determine aircraft performance independently of the manufacturer, and improving the accuracy and robustness of aircraft operation optimization.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN111353209B_ABST
    Figure CN111353209B_ABST
Patent Text Reader

Abstract

A method and system for optimizing aircraft flight is disclosed. The trajectory is divided into multiple segments, each segment is controlled by a different set of equations, depending on the engine thrust mode and vertical guidance (climb, cruise or descent). Data is received from the flight record by assuming two models, an aerodynamic model and an engine speed model, and multiple parameters from a parameter optimization engine are iteratively determined by applying a least squares calculation until a predefined minimum criterion is met. The parameter optimization engine is then used to predict trajectory points after a given point. Software aspects and system aspects (e.g., FMS and / or EFB) are described.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates generally to the field of avionics. In particular, the present invention relates to methods and systems for optimizing all or some operations of an aircraft. Background Art

[0002] The methods known in the prior art (e.g., WO 2017042166 or US9290262) are generally based on data sets that model the performance of the aircraft. Although these performance data are known to aircraft manufacturers, they are generally not public, which poses a significant obstacle to the development of methods for optimizing aircraft operations.

[0003] Furthermore, when such data is made public, the data describes (generally) the behavior of aircraft types, not specific aircraft (i.e., the data is not specific). Finally, aircraft models may not be exempt from use tax.

[0004] Where appropriate, manufacturers of systems for optimizing aircraft operation must use performance data provided by the aircraft manufacturer (where appropriate) and may therefore find themselves in a disadvantageous position of dependence.

[0005] In the current state of the art, some tools such as "BADA" (acronym for "Aircraft Database") or "Safety-Line" of EUCASS ("European Conference of Aerospace Sciences") have limitations. The BADA model is limited in terms of thrust and drag. The EUCASS model is only applicable to aircraft propelled by turbojet engines (N1 drive).

[0006] Current optimization uses a variety of techniques. Specifically, some methods prefer nonlinear parameter estimation (e.g., least squares). These techniques are based on gradient descent algorithms, which sometimes have the defect of converging to local minima, which results in poor model parameter estimation.

[0007] Technical problems in aviation usually involve many different parameters and therefore nonlinear parameter methods converge little, poorly or not at all. This lack of convergence (or convergence towards local minima) is often observed especially when the process being modeled is large in scale (and follows different models depending on the viewpoint).

[0008] Methods based on models incorporating the aircraft's physical equations usually depend on the quality of the modeling and the knowledge of the aircraft's actual behavior. In fact, these methods are usually not robust to variations in the actual behavior of a given aircraft compared to the behavior of an "average" (modeled) aircraft.

[0009] There is a need in the industry for advanced methods and systems for optimizing all or some operations of an aircraft. Summary of the invention

[0010] Methods and systems for optimizing aircraft flight are disclosed. The trajectory is divided into multiple segments, each segment is controlled by a different set of equations, depending on the engine thrust mode and vertical guidance (climb, cruise or descent). Data from flight recordings are received by assuming two models, an aerodynamic model and an engine speed model, and multiple parameters from a parameter optimization engine are iteratively determined by applying least squares calculations until a predefined minimum criterion is met. The parameter optimization engine is then used to predict trajectory points after a given point. Software aspects and system (e.g., FMS and / or EFB) aspects are described.

[0011] Advantageously, the method according to the invention is based on a thrust (and drag) model that is closer to reality than the one implemented in the BADA model.

[0012] Advantageously, the invention makes it possible to determine the performance model of an aircraft independently of its manufacturer.

[0013] Advantageously, the present invention makes it possible to ensure the convergence of least squares estimation methods on large-scale mixed multi-model problems.

[0014] The present invention may preferably be implemented in various avionics industries, and in particular in systems for assisting aircraft piloting.

[0015] In particular, the invention may be deployed in a computer such as an FMS, or in a system assembly interconnecting an FMS with an EFB. More specifically, potential applications relate to calculating trajectories, assisting aircraft manufacturers in determining aircraft performance, optimizing airline flight operations in a broader sense, simulating flights, or tuning avionics systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Various aspects and advantages of the present invention will emerge with reference to the following drawings to support the description of a preferred but non-limiting mode of embodiment of the invention:

[0017] [ Figure 1 ] shows an example of steps of a method according to an embodiment of the present invention;

[0018] [ Figure 2 ] shows an example of parameter optimization. DETAILED DESCRIPTION

[0019] [ Figure 1 ] shows the influence of thrust mode and vertical guidance on parameters SEP, FF and N1.

[0020] The acronym SEP 111 for "Specific Excess Power" refers to the energy available to the aircraft for climb, i.e. the aircraft's climb capability divided by its weight (this parameter is not constant).

[0021] The acronym FF for "Fuel Flow" refers to the change in fuel weight.

[0022] The acronym N1 refers to the speed of the first stage of the engine, which is the speed that has the greatest impact on fuel consumption. The available power is strongly related to this speed N1.

[0023] The parameters SEP, FF and N1 are closely related.

[0024] One method of quantifying or defining the relationships between SEP, FF, and N1 involves formulating a model (equation) that models these relationships.

[0025] There are several relationships between the parameters FF, N1 and SEP, but they mainly depend on the flight mode of the aircraft, which is broken down into "thrust mode" (THR stands for "thrust" 111) on the one hand (whether adjusted 1111 or fixed 1111), and "vertical guidance" 112 on the other hand (for example, climb, cruise or descent).

[0026] These thrust and vertical guidance modes define N flight modes for the engine used in production (these are autopilot modes, e.g. modes called open-accel, IDLE, "energy sharing", engine idling, with or without airbrake deployment, etc.). For example, the "open-accel" mode corresponds to full thrust in order to reach and maintain predefined target parameters. These Nt modes correspond to the same number of models or systems of equations.

[0027] Therefore, each pattern leads to a way of predicting the next trajectory point (based on a pattern).

[0028] [ Figure 2 ] shows an example of the steps of the method according to the present invention.

[0029] In one embodiment, the method comprises the step of determining a parametric model comprising a plurality of flight phases describing the flight mechanics of the aircraft based on recordings obtained during (actual) commercial flights.

[0030] In one embodiment, the parameter model is determined using the method of least squares. The method of least squares makes it possible to compare experimental data, which are usually corrupted by measurement errors, with a mathematical model intended to describe these data. The error model is usually a Gaussian model (normal law). If the error model is a non-Gaussian model, the method can use the maximum likelihood method (of which the method of least squares is a special case).

[0031] In one embodiment, the parameter model is determined using recursive least squares (RLS), which is an adaptive filter (minimizing an error function or weighted least squares).

[0032] In one embodiment, a two-stage least squares method is used to determine the parametric model. This estimator makes it possible, in particular, to estimate a linear regression model with instrumental variables.

[0033] In one embodiment, the flight of an aircraft comprises different flight phases governed by different laws depending on the flight segment along the trajectory: the performance of the least squares method is advantageously improved if thrust and vertical guidance information are taken into account.

[0034] The flight phases include, in particular, the phases of take-off, climb, cruise, altitude change and landing.

[0035] In one embodiment, the method constitutes a step of determining, by the method of least squares (or according to a variant), an estimate of a model of the performance of the aircraft, this estimate being specifically made in a plurality of "passes" of increasing complexity, each pass using i) a data set enriched by the phase of flight relative to the previous pass; and ii) taking as a starting point the result of the previous pass (a first initial pass being carried out during the climb phase).

[0036] The present invention may use a predefined parameter model (ie, the use of equations).

[0037] There may be multiple models. In the following description, two models are described (an aerodynamic model and an engine model) which are necessary and sufficient to determine the performance model.

[0038] The model specified below further advantageously combines realism and fast computation speed.

[0039] Aerodynamic model

[0040] In one embodiment, the aerodynamics are modeled as follows:

[0041] [Mathematical formula 1]

[0042]

[0043] Among them, a i,j Represents coefficients or functions, C L (lift) and C D (Drag) establishes the relationship between lift, angle of attack and drag, and M (Mach) represents the Mach number (which expresses the ratio of the local velocity of the fluid to the speed of sound in the same fluid). This model is general.

[0044] Engine Model

[0045] In one embodiment, engine speed is modeled by an affine function of the following type:

[0046] [Mathematical formula 2]

[0047] TSP=a×z p +b×M+c

[0048] Among them, TSP, a, b and c represent coefficients, z p stands for pressure according to altitude, and M stands for Mach number.

[0049] More generally, in other embodiments, engine speed is modeled by a polynomial function (having real or complex coefficients).

[0050] Engine manufacturers have more complete equations at their disposal; drag and thrust in particular may be functions (aerodynamic configuration).

[0051] [ Figure 2 ] shows an example of parameter optimization according to the invention, in particular the step of improvement by successively adding flight phases.

[0052] In step 210 (which is a previous off-line initialization step), the method according to the invention consists in segmenting each flight into flight phases of number n_phase, which can be modeled by a predefined set of equations E depending on number p_exp of explanatory variables and number n_par of parameters, and providing an observation vector of size q_mes.

[0053] Explanatory variables use continuous values ​​or discrete variables to describe each sub-process, followed by the modeling process.

[0054] In step 220 , based on a set of files (for example, records from commercial flights, ie real data), an observation table is determined, comprising a number of rows m_obs and a number of columns p_exp+q_mes and containing points associated with the climb phase of the aircraft.

[0055] In step 230 , the method according to the invention comprises the steps of determining (for example, arbitrarily, but not always in order to speed up the calculations) a number n_par of initial parameters and calculating the least squares criterion from the table of step 220 .

[0056] The least squares criterion can be obtained as follows.

[0057] For each row of the table, the method determines the m_obs components (Y1…Ym_obs) of the vector Y=E(X1,..Xp_exp,P1,…Pn_par) using the system of equations E defined in step 1. Next, the method determines the Euclidean norm N of the vector [Y–Yobs], where the components of the vector Yobs are the m_obs last values ​​in the i-th row of the table. Finally, the average of the Euclidean norm N is calculated for all rows in the table.

[0058] In a next step, the result thus obtained is passed to the parameter optimization engine 240 which will deduce therefrom an iteration of n_par initial parameters and will repeat steps 230 and 240 until the minimum criterion 241 is met, at the end of which process a set of n_par parameters is provided, which are called final or optimized parameters.

[0059] In step 250, steps 220, 200 3240 are repeated according to predefined variations.

[0060] According to a variant, the observation table from step 220 contains the climb points and the descent points of the aircraft, and the initial parameters from step 230 are the end parameters obtained in step 240 .

[0061] According to another variant, steps 220 , 230 and 240 are repeated, wherein the observation table in step 220 contains the climb points, descent points and cruise points of the aircraft, and the initial parameters from step 230 are the end parameters obtained in step 240 .

[0062] In one embodiment, the watch table includes information related to thrust modes and vertical guidance.

[0063] The present invention can be implemented based on hardware and / or software elements. It can be used as a computer program product on a computer readable medium.

[0064] In a variant embodiment, one or more steps of the method may be implemented in a ground-based computer.

[0065] Advantageously, if the computer is located on board the aircraft and connected to a parameter recorder, the flight data can be used in a real-time architecture to gain real-time insights into aircraft performance.

[0066] In a variant embodiment, one or more steps of the method according to the invention are implemented in the form of a computer program hosted on an EFB (Electronic Flight Bag) portable computer. In a variant embodiment, one or more steps of the method can be implemented in an FMS (Flight Management System) computer or in the FM function of a flight computer.

[0067] Described is a method for optimizing the flight of an aircraft, the flight of the aircraft being divided into a number of flight phases n_phase, which can be modeled by a predefined set of equations E depending on a number of explanatory variables p_exp and on a number of parameters n_par, and providing an observation vector of size q_mes, the method comprising the following steps: receiving a plurality of flight records and determining a table comprising a number of rows m_obs and a number of columns p_exp+q_mes; determining a number of parameters n_par and calculating the least squares criterion of the table thus determined; determining n_par parameters (called end parameters) in a parameter optimization engine by iteration until a predefined minimum criterion is met.

[0068] Explanatory variables can be continuous numeric and / or discrete variables that describe various sub-processes behind the process being modeled.

[0069] In one embodiment, the least squares calculation step includes the following steps: - For each row of the table, use the defined equation group E to calculate the m_obs components (Y1...Ym_obs) of the vector Y=E(X1,..Xp_exp,P1,...Pn_par); - Determine the Euclidean norm N of the vector Y-Yobs, the components of the vector Yobs are the m_obs last values ​​in the Ith row of the table; and - Determine the average value of the Euclidean norm N in all rows of the table.

[0070] In one embodiment, the lookup table includes information related to thrust mode and vertical guidance. These data (full throttle, idle, climb, descent, etc.) are used as integrators (predictors).

[0071] In one embodiment, aerodynamics are modeled by linking models for lift, angle of attack, and drag with a description of the Mach number, M. The Mach number expresses the ratio of the local velocity of a fluid to the speed of sound in the same fluid.

[0072] In one embodiment, the engine speed is modeled by a polynomial function coupled to the aerodynamic model.

[0073] In one embodiment, the initial parameter is an end parameter.

[0074] In one embodiment, the least squares criterion is a recursive least squares criterion.

[0075] A description is given of a computer program product, the computer program comprising code instructions for performing one or more steps of a method when the program is executed on a computer.

[0076] Described is a system for implementing one or more steps of the method, the system comprising one or more avionics systems.

[0077] In one embodiment, the calculations are performed on the ground.

[0078] In one embodiment, the system also includes one or more non-avionics systems, such as an electronic flight bag (EFB).

[0079] In one embodiment, the system includes an onboard computer and storage system for processing aircraft performance in real time.

Claims

1. A method for determining a performance model of an aircraft, the flight of which is divided into a number n_phases of flight phases, wherein: Each stage can be modeled by a predefined set of equations E, which depends on a number of explanatory variables p_exp and a number of parameters n_par, and each stage provides an observation vector of size q_mes, the method comprising the following steps: (a) determining the contents of an observation table based on data derived from a plurality of flight records, the observation table comprising a number m_obs of rows and a number p_exp+q_mes of columns, the observation table containing points associated with a flight phase; (b) determining a plurality of initial parameters for n_par number of parameters; (c) calculating a least squares criterion for the observation table based on the plurality of initial parameters, wherein the least squares criterion is obtained by: - For each row of the observation table, the components (Y1 ... Ym_obs) of the observation vector Y = E (X1, .. Xp_exp, P1, ... Pn_par) are calculated using the defined system of equations E; - determining the Euclidean norm N of the vector Y-Yobs, the components of which are the m_obs last values ​​in the Ith row of the observation table; and - determining the average value of the Euclidean norm N among all rows in the observation table; and (d) using said least squares criterion obtained in the parameter optimization engine and iterating steps (b) and (c) until a predefined minimum criterion is met, said minimum criterion being met for a number n_par of parameters, said number n_par being called the end parameters, said end parameters determining the optimized aircraft performance model.

2. The method of claim 1, wherein the lookup table includes information related to thrust mode, vertical guidance and the minimum criteria.

3. The method according to claim 1, wherein: The aerodynamics of the aircraft are modeled using the following types of models: Among them, a i,j Represents coefficients or functions, C L "Lift" and C D "Drag" establishes the relationship between lift, angle of attack, and drag, and M represents the Mach number.

4. The method according to claim 1, wherein: The engine speed coupled with the aerodynamic model is modeled by a polynomial function. The method according to claim 1 , wherein the initial parameter is the end parameter. The method of claim 1 , wherein the least squares criterion is a recursive least squares criterion.

7. A computer program product, the computer program comprising code instructions for executing the steps of the method according to any one of claims 1 to 6 when the program is executed on a computer.

8. A system for implementing the steps of the method according to any one of claims 1 to 6, the system comprising one or more avionics systems.

9. The system of claim 8, further comprising one or more non-avionics systems of the electronic flight bag type.

10. A system according to claim 8 or 9, comprising an on-board computer and storage system for real-time processing of aircraft performance.

Citation Information

Patent Citations

  • Method for creating a vertical trajectory profile comprising multiple altitude levels

    US9290262B2

  • Optimising the trajectory of an aircraft

    WO2017042166A1

  • A fast vertical trajectory prediction method for air traffic management, and relevant atm system

    CN102610126A

  • Robustness identification method to suit helicopter's self-adaptive flight control

    CN106406092A

  • Method and system for predicting number of spare parts used in aircraft maintenance outfield

    CN108053079A