METHOD AND DEVICE FOR MONITORING AND ESTIMATING PARAMETERS RELATING TO THE FLIGHT OF AN AIRCRAFT

Abstract

A device for monitoring and estimating parameters relating to the flight of an aircraft includes an estimation module 4 for determining an estimation of the values of the parameters relating to the flight of the aircraft and for generating residues, a detection module for determining the statuses associated with each of said sensors C1, C2, . . . , CN and with a parameter P1 corresponding to the weight of the aircraft, a transmission module for transmitting the statuses associated with each of said sensors C1, C2, . . . , CN to a user device and, on the next iteration, to the estimation module.

Description

FIELD OF THE INVENTION

The present invention relates to a method and an associated device for monitoring and estimating parameters relating to the flight of an aircraft in real time.

BACKGROUND OF THE INVENTION

The correct operation of an aircraft is partly guaranteed by that of the sensors. They supply information on its attitude, its trajectory, its speed, etc., and make it possible to ensure the navigation, the guidance and the control thereof via control laws and automatic control algorithms. The high levels of automation that are possible these days on some aircraft depend on the availability of these sensors. These days, any loss of information following sensor failures involves the pilots more in the control of the aircraft, which increases their workload. In order to simplify the piloting task, it is therefore necessary to extend the availability of the flight parameters and do so throughout the duration of the flight. This means using a step of validation of the quality of the information supplied by the sensors, called “monitoring”.

In order to ensure a better monitoring of the sensors and to increase the availability of the flight parameters in case of failure, two types of approaches have been tackled.

A first type of approach corresponds to the use of hardware redundancies, that is to say the use of several sensors of the same type in order to allow the identification of failure via in particular a majority-based vote (for example an average calculation or the choice of the median value). This first type of approach entails multiplying the number of embedded sensors, which in the case of aircraft with little instrumentation is unrealistic. In addition, in case of coherent and perfectly simultaneous failures of different sensors, the majority-based vote does not allow a fault to be identified. This is then called a common failure mode.

A second type of approach corresponds to the use of analytical redundancy. The sensor measurements can be linked via kinematic equations and the mechanics of the flight. These interdependency relationships between the measurements and the parameters relating to the flight have the advantage of being able to increase the dissimilarity of the information via the use of virtual sensors. These virtual sensors deliver estimations of parameters and of measurements computed from combinations of sensors measuring quantities of different types. This approach has the advantage of not increasing the number of sensors necessary for the monitoring since it uses only the information already present on board the aircraft. Nevertheless, it may require the availability of sufficient embedded computation means. Analytical redundancy is used in the development of estimators. These estimators also have the advantage of being able to deliver an estimation of the quantity invalidated after failure in order to guarantee a greater availability of the flight parameters.

The problems which arise these days concern the two types of approaches. First of all, the use of the majority-based vote does not allow complete monitoring of the common failure modes. Then, the analytical redundancy is based on assumptions of validity of certain quantities, or of other sensors. It is necessary for these assumptions to be verified and for the algorithm developed to be validated.

One of the problems that is thus sought to be resolved is that of the dissociation of a failure on the incidence, of a failure on the speed and of an error on the weight in addition to the estimation issues. Many estimators have been developed to date, but none makes it possible to broach the problem with assumptions that allow this distinction to be made.

BRIEF SUMMARY OF THE INVENTION

Aspects of the present invention may mitigate this drawback by proposing a method and a device for monitoring and estimating parameters relating to a flight of an aircraft.

The parameters relating to a flight comprise at least one of the following elements: flight parameters, atmospheric parameters, sensor biases or modelling biases.

To this end, the invention relates to a method for monitoring and estimating:

• • parameters relating to the flight of an aircraft;
  • statuses of sensors, these statuses being representative of an operation of said sensors; and
  • a status of a parameter corresponding to the current weight of the aircraft, this status being representative of the validity of said parameter.

According to an aspect of the invention, the method comprises the following steps:

• • an initialization step, implemented by an initialization module, consisting in initializing the statuses of sensors configured to determine flight parameters of the aircraft and the status of the parameter corresponding to the current weight of the aircraft and in initializing parameters used in the implementation of the monitoring and estimation device;

the method further comprising the following steps, implemented iteratively:

• • an estimation step, implemented by an estimation module, consisting in determining an estimation of the values of the parameters relating to the flight of the aircraft and an estimation of an error of the current weight parameter, from:
    • measurements of the parameters relating to the flight supplied by the sensors,
    • parameters relating to the flight initialized in the initialization step or estimated on the preceding iteration of the estimation step and
    • statuses associated with each of said sensors,

the estimation step consisting also in generating residues which are functions of the measured and estimated values of the parameters relating to the flight and of innovation terms which correspond to the difference between a measured flight parameter value and said estimated value;

• • a first transmission step, implemented by a first transmission module, consisting:
    • in transmitting to a user device and to a detection module a signal representative of the estimation of the values of the parameters relating to the flight of the aircraft and of the estimation of the error of the current weight parameter, determined in the estimation step,
    • in sending to said detection module a signal representative of the residues generated in the estimation step;
  • a detection step, implemented by a detection module, consisting in determining the different statuses associated with each of said sensors and with the parameter corresponding to the current weight of the aircraft, from:
    • the estimation of the values of the residues determined in the estimation step,
    • the estimation of the values of the parameters relating to the flight of the aircraft determined in the estimation step,
    • the measurements of the parameters relating to the flight supplied by the sensors,
    • the estimation of the error of the current weight parameter determined in the estimation step, and
    • the statuses determined on the preceding iteration of the detection step or initialized in the initialization step,
  • a second transmission step, implemented by a second transmission module, consisting in transmitting to the user device and, on the next iteration, to the estimation module the different statuses associated with each of said sensors and the status associated with said parameter corresponding to the current weight.

By virtue of an aspect of the invention, it is possible to identify the failures of sensors and a weight error in an embedded system of an aircraft. It also makes it possible to continuously supply an estimation of the parameters relating to the flight of the aircraft in real time and a status of the different sensors and of the parameter corresponding to the weight.

Furthermore, the estimation step comprises the following substeps:

• • an adaptation substep, implemented by an adaptation submodule, consisting in determining a variance and/or a validity Boolean associated with each of the measurements of the parameters relating to the flight supplied by the sensors and of the setting parameters associated with the estimation algorithm used in an estimation substep, from:
    • said measurements of the parameters relating to the flight, and
    • the statuses associated with each of said sensors;the adaptation substep consisting also in correcting the current weight from a weight error estimated on the preceding iteration or initialized in the initialization step, and from a status associated with the parameter corresponding to the weight,
  • the estimation substep, implemented by an estimation submodule, consisting in determining the estimation of the values of the parameters relating to the flight and an estimation of the error of said weight, from:
    • the measurements of the parameters relating to the flight supplied by said sensors,
    • the parameters relating to the flight estimated on the preceding iteration or initialized in the initialization step and
    • from the variance and/or the validity Boolean of each of the measurements of the parameters relating to the flight and of the setting parameters determined in the adaptation substep,the estimation substep consisting also in generating the residues from the estimated and measured parameters relating to the flight and from the innovation terms.

Furthermore, the detection step comprises the following substeps:

• • a substep of detection of a failing sensor and of incorrect parameters relating to the flight, implemented by a detection submodule, consisting in determining:
    • the common status associated with the sensor configured to measure the angle of incidence of the aircraft and with the parameter corresponding to the current weight and
    • the status associated with the other sensors,
  •  from:
    • the measurements of the parameters relating to the flight supplied by said sensors,
    • the estimation of the values of the parameters relating to the flight and the weight error,
    • the statuses associated with each of said sensors and with the parameter corresponding to a current weight of the aircraft determined on the preceding iteration or initialized in the initialization step and
    • the residues;
  • a substep of validation of the angle of incidence and of the current weight, implemented by a validation submodule, consisting in determining the status associated with the parameter corresponding to the current weight and the status associated with the sensor configured to measure the angle of incidence (a), from:
    • the common status associated with the sensor configured to measure the angle of incidence of the aircraft and with the parameter corresponding to the current weight,
    • the statuses associated with the other sensors,
    • the estimated parameters relating to the flight,
    • the estimated weight error,
    • the measured parameters relating to the flight,
    • the residues generated in the estimation substep and
    • a lift coefficient supplied from an embedded modelling fed by the parameters relating to the flight estimated and measured by the sensors.

Thus, it is possible to distinguish a fault between the measurement of angle of incidence and the parameter corresponding to the weight, including in the case of common failure modes.

According to a first embodiment, the estimation substep corresponds to an extended Kalman filter associated with a state vector, an observation vector and an auxiliary measurement vector,

the auxiliary measurement vector having for expression:

$Z=\left(i_{H_m},{\delta }_{q_{i_m}},{\delta }_{{\mathrm{sp}}_{i_m}},{\psi }_m,{\varphi }_m,{\theta }_m,n_{X_{1_m}},m,\mathrm{conf},V_{g_{x_{0_m}}},V_{g_{y_{0_m}}},V_{g_{z_{0_m}}},z_{g_m}\right),$

in which:

• • i_Hm corresponds to a measurement of deflection of the horizontal plane,

${\delta }_{q_{i_m}}$

corresponds to a measurement of deflection of elevators of the aircraft,

${\delta }_{{\mathrm{sp}}_{i_m}}$

corresponds to a measurement of deflection of spoilers of the aircraft,

• • ψ_m corresponds to a heading measurement,
  • φ_m corresponds to a list angle measurement,
  • θ_m corresponds to an inclination measurement,

$n_{X_{1_m}}$

corresponds to a measurement of longitudinal load factor in the reference frame associated with the aircraft,

• • m corresponds to the current weight parameter of the aircraft,
  • conf corresponds to a measurement of aerodynamic configuration of the aircraft,

$V_{g_{x_{0_m}}},V_{g_{y_{0_m}}},V_{g_{z_{0_m}}}$

• • correspond to measurements of the ground speed components in the Earth's reference frame, and
  • z_gm corresponds to a geometrical altitude measurement;the state vector having for expression:

$X=\left(\begin{array}{c}{W}_{x_0}\\ W_{y_0}\\ W_{z_0}\\ \Delta \mathrm{ISA}\\ b_{C_L}\\ C_{b_x}\end{array}\right),$

in which:

• • (W_x0, W_y0, W_x0) correspond to the three components of the speed of the wind in the Earth's reference frame,
  • ΔISA corresponds to a temperature deviation between a current static temperature and a temperature determined from a standard atmosphere model at a current geometrical altitude,
  • b_CL corresponds to a lift coefficient modelling bias, and
  • C_bx corresponds to a barometric correction term;the derivative of the state vector having for expression:

$\stackrel{.}{x}=\left(\begin{array}{c}{\stackrel{.}{W}}_{x_0}=0\\ {\stackrel{.}{W}}_{y_0}=0\\ {\stackrel{.}{W}}_{z_0}=0\\ \Delta \stackrel{.}{\mathrm{ISA}}=0\\ {\stackrel{.}{b}}_{C_L}=-\frac{b_{C_L}}{{\tau }_b}\\ {\stackrel{.}{C}}_{b_x}=0\end{array}\right),$

in which τ_b corresponds to a characteristic time associated with a dynamic range of the lift coefficient modelling bias b_CL,

• • the observation vector having for expression:

$Y=\left(\begin{array}{c}{\alpha }_m=\mathrm{atan}\left(\frac{w}{u}\right)\\ {\beta }_m=\mathrm{atan}\left(\frac{v}{\sqrt{u^2+w^2}}\right)\\ P_{S_m}=\zeta \left(z_P\right)\\ n_{Z_{1_m}}=\frac{\gamma {\mathrm{SP}}_SM^2\left(C_L+b_{C_L}\right)}{2\mathrm{mg}\cos \alpha }+n_{X_{1_m}}\tan \alpha \\ P_{T_m}={P_S\left(1+\frac{\gamma -1}{2}M^2\right)}^{\frac{\gamma }{\gamma -1}}\\ T_{T_m}=T_S\left(1+\frac{\gamma -1}{2}M^2\right)\end{array}\right),$

in which:

• • α_m corresponds to an incidence measurement,
  • β_m corresponds to a slide slip measurement,
  • P_Sm corresponds to a static pressure measurement,

$n_{Z_{1_m}}$

• • corresponds to a measurement of vertical load factor in a reference frame associated with the aircraft,
  • P_Tm corresponds to a total pressure measurement,
  • T_Tm corresponds to a total temperature measurement, and
  • z_p corresponds to the pressure altitude and is expressed according to the equation

$z_p=\frac{z_g}{1+\frac{\Delta \mathrm{ISA}}{T_{15}}}-C_{b_x},$

• •  with T₁₅=288.15 K,
  • ζ corresponds to a function linking the measurement of the pressure altitude z_P to the static pressure with the following expression:

$\{\begin{array}{c}\zeta \left(z_{P_m}\ge z_{P_{11}}\right)={10}^{{\log }_{10}\left(P_{11}\right)+\frac{g\left(z_{P_{11}}-z_{P_m}\right)}{{\mathrm{RT}}_{11}\log \left(10\right)}}\\ \zeta \left(z_{p_m}<z_{P_{11}}\right)={P_0\left(1-\frac{G_{T_{z_0}}}{T_{15}}z_{P_m}\right)}^{\frac{g}{{\mathrm{rG}}_{T_{z_0}}}}\end{array},$

in which:

• • z_P11=11 km corresponds to the standard altitude of the tropopause,
  • P₁₁=226.321 mbar and T₁₁=216.65 K correspond to the standard static pressure and static temperature at the tropopause,
  • G_Tz0=0.0065 K/m corresponds to the standard temperature gradient for z_Pm<z_P11,
  • G corresponds to the acceleration of gravity,
  • R corresponds to the specific constant of the air,
  • T_S corresponds to the statistical temperature parameter and is expressed according to the equation T_S=T₀+G_Tz0z_g+ΔISA, with T₀=273.15 K and G_Tz0=0.0065 K/m,
  • V corresponds to the air speed flight parameter and is expressed according to the equation V=√{square root over (u²+v²+w²)}, with each component of the air speed defined in the aircraft reference frame (u, v, w) expressed according to the following expression:

$\left[\begin{array}{c}u\\ v\\ w\end{array}\right]=\left[\begin{array}{ccc}\cos \theta \cos \psi & \cos \theta \sin \psi & \sin \theta \\ \sin \varphi \sin \theta \cos \psi -\cos \varphi \sin \psi & \sin \varphi \sin \theta \sin \psi +\cos \varphi \cos \psi & -\sin \varphi \cos \theta \\ \cos \varphi \sin \theta \cos \psi +\sin \varphi \sin \psi & \cos \varphi \sin \theta \cos \psi +\sin \varphi \sin \psi & -\cos \varphi \cos \theta \end{array}\right]\left[\begin{array}{c}{V}_{g_{x_0}}-W_{x_0}\\ V_{g_{y_0}}-W_{y_0}\\ V_{g_{z_0}}-W_{z_0}\end{array}\right]$

• • M corresponds to the Mach number flight parameter computed as a function of the norm of the air speed V according to the equation

$M=\frac{V}{\sqrt{\gamma {\mathrm{rT}}_s}}$

• •  with r corresponding to the specific constant of the air,the estimation substep generating, on an iteration k, the following residues:
  • a first residue having for expression at a time

$t_kr_1\left(t_k\right)={\varepsilon }_{\alpha }\left(t_k\right)+\frac{{\hat{b}}_{c_z}\left(t_k\right)}{C_{z_{\alpha }}\left(\mathrm{conf}\left(t_k\right)\right)}$

• •  in which:
    • ε_α(t_k) corresponds to the innovation term associated with the measurement of the angle of incidence α at the time t_k,
    • {circumflex over (b)}_Cz corresponds to an estimation of the modelling bias of the lift coefficient C_z at the time t_k, and
    • C_zα(conf(t_k)) corresponds to a tabulated value of the lift coefficient depending on a value of a configuration parameter of the aircraft at the time t_k,
  • a second residue having for expression at the time t_k

$r_2\left(t_k\right)=\sqrt{\frac{2}{\gamma -1}\left(\frac{T_{T_m}}{\widehat{T_s}\left(z_{P_m},t_k\right)}-1\right)}-\sqrt{\frac{2}{\gamma -1}\left({\left(\frac{P_{T_m}}{\zeta \left(z_{P_m}\right)}\right)}^{\frac{\gamma -1}{\gamma }}-1\right)}$

• •  in which:
    • γ corresponds to the adiabatic coefficient of the air,
    • T_Tm corresponds to a total temperature measured by one of the sensors used as input for the estimation module,
    • P_Tm corresponds to a total pressure measured by one of the sensors used as input for the estimation module,
    • z_Pm corresponds to a pressure altitude measured by one of the sensors used as input for the estimation module,
    • ζ corresponds to the function linking the measurement of the pressure altitude to the static pressure and
    • {circumflex over (T)}_s(z_Pm, t_k) corresponds to an estimation of the static temperature computed from parameters relating to the flight estimated at the time t_k on the current iteration and at a time t_k-1 on a preceding iteration as a function of the first residue according to the formula:

${\hat{T}}_s\left(z_{P_m},t_k\right)=T_{15}+G_{T_{z0}}\left(z_{P_m}\left(t_k\right)+\left(s\right)\left[{\hat{C}}_{b_x}\left(t_k\right){\Pi }_{r_1^{\pm }}\left(r_1\left(t_k\right)\right)+{\hat{C}}_{b_x}\left(t_{k-1}\right)\left(1-{\Pi }_{r_1^{\pm }}\left(r_1\left(t_k\right)\right)\right)\right]\right)+\left(s\right)\left[\left(t_k\right){\Pi }_{r_1^{\pm }}\left(r_1\left(t_k\right)\right)+\left(t_{k-1}\right)\left(1-{\Pi }_{r_1^{\pm }}\left(r_1\left(t_k\right)\right)\right)\right],$

• • •  in which is the transfer function of a low-pass filter and

${\Pi }_{r_1^{\pm }}\left(r\right)=H\left(r+\frac{r_1^{\pm }}{2}\right)-H\left(r-\frac{r_1^{\pm }}{2}\right)$

where H is the Heaviside function, and

$r_1^{\pm }$

• • • corresponds to the limits associated with the residue r₁ defined hereinafter in the description;
  • a third residue having for expression at the time

$t_kr_3\left(t_k\right)=\sqrt{\frac{2}{\gamma -1}\left(\frac{T_{T_m}}{\widehat{T_s}\left(z_{P_m},t_k\right)}-1\right)}-\sqrt{\frac{2}{\gamma -1}\left(\frac{2\mathrm{mg}n_{Z_{1_m}}\cos {\alpha }_m}{\gamma S\zeta \left(z_{P_m}\right)\widehat{C_Z}}-n_{X_{1_m}}\tan {\alpha }_m\right)}$

in which:

• • • m corresponds to the weight of the aircraft,
    • S corresponds to the reference surface of the aircraft,
    • g corresponds to the acceleration of gravity,

$n_{Z_{1_m}}$

• • • corresponds to a vertical load factor,

$n_{X_{1_m}}$

• • • corresponds to a longitudinal load factor,
    • corresponds to the estimated lift coefficient, obtained from the estimated and measured parameters relating to the flight and from the configuration of the aircraft.

According to a first variant, the observation vector has for expression:

$Y=\left(\begin{array}{c}{\alpha }_m=\mathrm{atan}\left(\frac{w}{u}\right)\\ {\beta }_m=\mathrm{atan}\left(\frac{v}{\sqrt{u^2+w^2}}\right)\\ P_{S_m}=\zeta \left(z_p\right)\\ n_{Z_m}=\frac{{\mathrm{SP}}_d\left(C_L+b_{C_L}\right)}{\mathrm{mg}\cos \alpha }+n_{X_{1_m}}\tan \alpha \\ P_{T_m}={P_S\left(1+\frac{\gamma -1}{2}M^2\right)}^{\frac{\gamma }{\gamma -1}}\\ T_{T_m}=T_S\left(1+\frac{\gamma -1}{2}M^2\right)\\ P_{T_{\mathrm{eng}}}={P_S\left(1+\frac{\gamma -1}{2}M^2\right)}^{\frac{\gamma }{\gamma -1}}\\ P_{S_{\mathrm{eng}}}=\zeta \left(z_p\right)\\ T_{T_{{\mathrm{eng}}_m}}=T_S\left(1+\frac{\gamma -1}{2}M^2\right)\end{array}\right)$

in which:

• • P_Teng corresponds to a total pressure value derived from an engine-nacelle modelling,
  • P_Seng corresponds to a static pressure value derived from the engine-nacelle modelling, and

$T_{T_{{\mathrm{eng}}_m}}$

corresponds to a measured total temperature value derived from an engine-nacelle total temperature sensor;

the estimation substep also generating the following residues:

• • a fourth residue having for expression

$r_4\left(t_k\right)={\varepsilon }_{P_{T_{\mathrm{eng}}}}\left(t_k\right)$

• •  in which

${\varepsilon }_{T_{P_{\mathrm{eng}}}}\left(t_k\right)$

• •  corresponds to the difference between the measured total pressure value and said total pressure value derived from the modelling at a time (t_k),
  • a fifth residue having for expression

$r_5\left(t_k\right)={\varepsilon }_{P_{s_{\mathrm{eng}}}}\left(t_k\right)$

• •  in which

${\varepsilon }_{P_{s_{\mathrm{eng}}}}\left(t_k\right)$

• •  corresponds to the difference between the measured static pressure value and said static pressure value derived from the modelling at a time (t_k), and
  • a sixth residue having for expression r₆(t_k)=ε_TATeng (t_k) in which ε_TATeng(t_k) corresponds to the difference between the measured total temperature value and said total temperature value derived from the engine-nacelle measurement at a time (t_k).

The engine-nacelle modelling is described in the patent FR 2 977 942. It makes it possible to deduce therefrom the two virtual measurements P_Seng and

$P_{T_{\mathrm{eng}}}$

that are useful to the present patent. In summary, the static pressure P_Seng obtained is made up of digital data corresponding to the measurements of static pressures of the ambient air in the nacelle P_nac whereas the total pressure P_Teng obtained is deduced from the input of data corresponding to the measurements of static pressures of the ambient air in the nacelle P_nac, of “engine” static pressure P_mot, of the speed of rotation of the fan of the engine and from the measurement of total temperature of the air.

According to a second variant, the estimation substep corresponds to an extended Kalman filter associated with a state vector and an observation vector and an auxiliary measurement vector,

the auxiliary measurement vector having for expression:

$Z=\left(i_{H_m},{\delta }_{q_{i_m}},{\delta }_{{\mathrm{sp}}_{i_m}},{\psi }_m,{\varphi }_m,{\theta }_m,n_{X_{1_m}},n_{Y_{1_m}},m,\mathrm{conf},V_{g_{x_{0_m}}},V_{g_{y_{0_m}}},V_{g_{z_{0_m}}},z_{g_m},\right),$

in which

$n_{Y_{1_m}}$

corresponds to a measurement of a lateral load factor in the reference frame associated with the aircraft,the state vector having for expression:

$X=\left(\begin{array}{c}{W}_{x_0}\\ W_{y_0}\\ W_{z_0}\\ \Delta \mathrm{ISA}\\ b_{C_L}\\ C_{b_x}\\ v_{g_{x_0}}\\ v_{g_{y_0}}\\ v_{g_{z_0}}\\ b_{n_{x_1}}\\ b_{n_{y_1}}\\ b_{n_{z_1}}\\ z_G\end{array}\right),$

in which:

$V_{g_{x_0}},V_{g_{y_0}},V_{g_{z_0}}$

correspond to the three components of the ground speed in the Earth's reference frame,

$b_{n_{x_1}},b_{n_{y_1}},b_{n_{z_1}}$

• • correspond to the three bias components of accelerometers in the reference frame associated with the aircraft,
  • z_G corresponds to a geometrical altitude;the derivative of the state vector having for expression:

$\stackrel{.}{X}=\left(\begin{array}{c}{\stackrel{.}{W}}_{x_0}=0\\ {\stackrel{.}{W}}_{y_0}=0\\ {\stackrel{.}{W}}_{z_0}=0\\ \Delta \stackrel{.}{\mathrm{ISA}}=0\\ {\stackrel{.}{b}}_{C_L}=-\frac{b_{C_L}}{{\tau }_b}\\ {\stackrel{.}{C}}_{b_x}=0\\ \begin{array}{c}{\stackrel{.}{v}}_{g_{x_0}}\\ {\stackrel{.}{v}}_{g_{y_0}}\\ {\stackrel{.}{v}}_{g_{z_0}}\end{array}=-g*M_{\mathrm{rot}}\left(\begin{array}{c}\left(n_{x_{1_m}}+b_{\mathrm{nx}}\right)\sin \theta \\ \left(n_{y_{1_m}}+b_{\mathrm{ny}}\right)\cos \theta \sin \phi \\ \left(n_{z_{1_m}}+b_{\mathrm{nz}}\right)\cos \theta \cos \phi \end{array}\right)\\ {\stackrel{.}{b}}_{n_{x_1}}=0\\ {\stackrel{.}{b}}_{n_{y_1}}=0\\ {\stackrel{.}{b}}_{n_{z_1}}=0\\ {\stackrel{.}{z}}_G=V_{g_{z_0}}\end{array}\right),$

in which:

• • M_rot corresponds to a standard matrix of rotation of the Earth's reference frame to the reference frame associated with the aircraft and has for expression

$M_{\mathrm{rot}}=\left(\begin{array}{ccc}\cos \theta \cos \psi & \sin \phi \sin \theta \cos \psi -\cos \phi \sin \psi & \cos \phi \sin \theta \cos \psi +\sin \phi \sin \psi \\ \cos \theta \sin \psi & \sin \phi \sin \theta \sin \psi +\cos \phi \cos \psi & \cos \phi \sin \theta \sin \psi -\sin \phi \cos \psi \\ \sin \theta & -\sin \phi \cos \theta & -\cos \phi \cos \theta \end{array}\right),$

the observation vector having for expression:

$Y=\left(\begin{array}{c}{\alpha }_m=\mathrm{atan}\left(\frac{w}{u}\right)\\ {\beta }_m=\mathrm{atan}\left(\frac{v}{\sqrt{u^2+w^2}}\right)\\ P_{S_m}=\zeta \left(z_p\right)\\ n_{Z_m}=\frac{{\mathrm{SP}}_d\left(C_L+b_{C_L}\right)}{\mathrm{mg}\cos \alpha }+n_{X_{1_m}}\tan \alpha \\ P_{T_m}={P_s\left(1+\frac{\gamma -1}{2}M^2\right)}^{\frac{\gamma }{\gamma -1}}\\ T_{T_m}=T_S\left(1+\frac{\gamma -1}{2}M^2\right)\\ M_{\mathrm{eng}}=\frac{V}{\sqrt{\gamma {\mathrm{rT}}_S}}\\ P_{S_{\mathrm{eng}}}=\zeta \left(z_p\right)\\ T_{T_{{\mathrm{eng}}_m}}=T_S\left(1+\frac{\gamma -1}{2}M^2\right)\\ V_{g_{x_{0_m}}}=V_{g_{x_0}}\\ V_{g_{y_{0_m}}}=V_{g_{y_0}}\\ V_{g_{z_{0_m}}}=V_{g_{z_0}}\\ z_{G_m}=z_G\end{array}\right),$

in which:

$V_{g_{x_{o_m}}},V_{g_{y_{o_m}}},V_{g_{z_{o_m}}}$

• • correspond to measurements of three components of the ground speed, and
  • z_Gm corresponds to a measurement of the geometrical altitude.

According to a third variant, the lateral component v of the air speed V is assumed nil.

That means that the side slip p is assumed nil.

Moreover, the determination of the common status and of the status associated with the other sensors of the failing sensor detection substep comprises the following substeps:

• • a substep of assignment to each residue of a maximum limit and a minimum limit from the estimated flight parameters;
  • the substep being also a substep of construction of a first validity indicator associated with the static pressure sensor and of a second validity indicator associated with the total pressure sensor from the estimated parameters and the measurements derived from the sensors of the aircraft:
    • the first validity indicator having the value 1 if the relationship |_h(s)(z_Pm−z_Gm)|≤l_PS is borne out,
      • the first validity indicator having the value 0 if the relationship |_h(s)(z_Pm−z_Gm)|≤l_PS is not borne out,
      • in which l_PS corresponds to a limit determined experimentally as a function of the dynamics of the aircraft, _h corresponding to the transfer function of a high-pass filter,
    • the second validity indicator having the value 0 by default and 1 if the residue r₁ crosses its respective limits, at an instant t_k, determined later and that n ϵ such that t_nϵ[t_k−τ_PT, t_k] bearing out |(1−_τ) (M_PT(t_n)−M_PT(t_n-j))|≤l_PT, with τ_PT a time constant, _τ a transfer function of a low-pass filter having τ for time constant, jϵ a constant, and l_PT a limit determined experimentally as a function of the dynamics of the aeroplane;residue indicators being computed and associated with each residue and validity indicators;
  • a substep of identification of the occurrence of failures when the sum of the residue indicators is strictly greater than zero and identification, in case of failure, of the faulty sensor by comparing the current values of the residue indicators to those itemized in a table, identified offline or online, itemizing the failure cases as a function of the different combinations of values of the residue indicators.

According to a second embodiment, the statuses associated with each of said sensors are also determined from auxiliary statuses associated with each of said sensors that can be sent to the detection module by a monitoring module.

According to a third embodiment, for a measured consolidated flight parameter from a plurality of sensors, the method comprises the following substeps implemented by a unitary verification module, when a failure is detected for a sensor measuring said flight parameter for which said measurement is used as input for the estimation module:

• • a substep of reconfiguration of the estimation submodule to not take account of the incorrect measurement of the flight parameter used hitherto in the estimation substep,
  • a substep of computation of the difference between the estimation of said flight parameter and of the measurement of one of said sensors derived from the plurality of sensors, for each of the measurements derived from the plurality of sensors,if the absolute value of the difference is less than a predetermined value, the measurement being retained in the estimation substep.

The invention relates also to a device for monitoring and estimating:

• • parameters relating to the flight of an aircraft;
  • statuses of sensors, these statuses being representative of an operation of said sensors; and
  • a status of a parameter corresponding to the current weight of the aircraft, this status being representative of the validity of said parameter.

According to an embodiment of the invention, the monitoring and estimation device comprises:

• • an initialization module, configured to initialize the statuses of sensors configured to determine parameters relating to the flight of the aircraft and the status of the parameter corresponding to the current weight of the aircraft and to initialize the parameters used in the implementation of the monitoring and estimation device; the monitoring and estimation device further comprises the following modules, implemented iteratively:
  • an estimation module, configured to determine an estimation of the values of the parameters relating to the flight of the aircraft and an estimation of an error of the current weight parameter, from:
    • measurements of the parameters relating to the flight supplied by the sensors,
    • parameters relating to the flight initialized in the initialization step or estimated on the preceding iteration of the estimation step and
    • statuses associated with each of said sensors,the estimation module being also configured to generate residues which are functions of the measured and estimated values of the parameters relating to the flight and of innovation terms which correspond to the difference between a measured flight parameter value and said estimated value;
  • a first transmission module configured to:
    • transmit to a user device and to a detection module a signal representative of the estimation of the values of the parameters relating to the flight of the aircraft and of the estimation of the error of the current weight parameter, determined in the estimation step,
    • send a signal representative of the residues generated in the estimation step to said detection module;
  • a detection module configured to determine the different statuses associated with each of said sensors and with the parameter corresponding to the current weight of the aircraft, from:
    • the estimation of the values of the residues determined in the estimation step,
    • the estimation of the values of the parameters relating to the flight of the aircraft determined in the estimation step,
    • the measurements of the parameters relating to the flight supplied by the sensors,
    • the estimation of the error of the current weight parameter determined by the estimation module,
    • said statuses determined on the preceding iteration or initialized in the initialization step,
  • a second transmission module configured to transmit to the user device and, on the next iteration, to the estimation module the different statuses associated with each of said sensors and the status associated with said parameter corresponding to the current weight.

Furthermore, the estimation module comprises:

• • an adaptation submodule configured to determine a variance and/or a validity Boolean associated with each of the measurements of the parameters relating to the flight supplied by the sensors and of the setting parameters associated with the estimation algorithm used in an estimation submodule, from:
    • said measurements of the parameters relating to the flight and
    • from the statuses associated with each of said sensors;the adaptation submodule being also configured to correct the current weight from an estimation of a weight error obtained on the preceding iteration or initialized by the initialization module, and from a status associated with the parameter corresponding to the weight,
  • the estimation submodule configured to determine the estimation of the values of the parameters relating to the flight and an estimation of the error of said weight, from:
    • the measurements of the parameters relating to the flight supplied by said sensors,
    • the parameters relating to the flight estimated on the preceding iteration or initialized in the initialization step and
    • from the variance and/or the validity Boolean of each of the measurements of the parameters relating to the flight and of the setting parameters determined in the adaptation substep,the estimation submodule being also configured to generate residues from the estimated and measured parameters relating to the flight and from the innovation terms.

Furthermore, the detection module comprises:

• • a detection submodule configured to determine:
    • the common status associated with the sensor configured to measure the angle of incidence of the aircraft and with the parameter corresponding to the current weight and
    • the status associated with the other sensors,
  • from:
    • the measurements of the parameters relating to the flight supplied by said sensors,
    • the estimation of the values of the parameters relating to the flight and the weight error,
    • the statuses associated with each of said sensors and with a parameter corresponding to a current weight of the aircraft determined on the preceding iteration or initialized in the initialization step, and
    • the residues;
  • a validation submodule configured to determine the status associated with the parameter corresponding to the current weight and the status associated with the sensor configured to measure the angle, from:
    • the common status associated with the sensor configured to measure the angle of incidence of the aircraft and with the parameter corresponding to the current weight,
    • the statuses associated with the other sensors,
    • the estimated parameters relating to the flight,
    • the estimated weight error,
    • the measured parameters relating to the flight,
    • the residues generated in the estimation substep and
    • a lift coefficient supplied from an embedded modelling fed by the parameters relating to the flight estimated and measured by said sensors.

The invention also relates to an aircraft, in particular a transport aeroplane, which comprises a device for monitoring and estimating flight parameters of an aircraft such as described above.

BRIEF DESCRIPTION OF THE DRAWINGS

The invention, with its features and advantages, will emerge more clearly on reading the description given with reference to the attached drawings in which:

FIG. 1 represents a block diagram of an embodiment of the monitoring and estimation device;

FIG. 2 represents steps of the monitoring and estimation method;

FIG. 3 represents substeps implemented in another embodiment in which the aircraft uses hardware redundancy;

FIG. 4 represents a block diagram of the embodiment of the monitoring and estimation device in which the aircraft uses hardware redundancy;

FIG. 5 represents an aircraft with the monitoring and estimation device embedded.

DETAILED DESCRIPTION

The description below will refer to the figures cited above.

FIG. 1 illustrates an embodiment of the device 1 for monitoring and estimating parameters relating to the flight of an aircraft AC and statuses associated with the operation of the sensors C1, C2, . . . , CN and with the validity of a parameter P1 corresponding to the weight of said aircraft, that can be embedded on an aircraft AC (FIG. 5). Said device is called “monitoring and estimation device” hereinafter in the description. The monitoring and estimation device is configured to implement a monitoring and estimation method.

The parameters relating to the flight correspond to at least one of the following parameters: flight parameters, atmospheric parameters, sensor bias parameters, modelling bias parameters.

The flight parameters correspond to the flight parameters measured directly by sensors and/or to the flight parameters recomputed from measured flight parameters and/or estimated parameters relating to the flight.

The measured flight parameters comprise, for example, the angle of incidence α of the aircraft AC, the static pressure P_S, the total pressure P_T, the load factor n_Z1, etc. They are directly derived from sensor measurements.

The recomputed flight parameters comprise, for example, the weight of the aircraft AC, the Mach number, the air speed, the calibrated speed, etc.

The estimated parameters relating to the flight refer to any parameter relating to the flight derived from an estimation result.

The atmospheric parameters correspond to the parameters associated with the atmospheric environment. The atmospheric parameters comprise, for example, the wind speed, the local pressure or temperature gradients, the temperature deviations ΔISA between a recomputed static temperature and a temperature modelled by the international standard atmosphere model.

The sensor bias parameters correspond to parameters that make it possible to know the bias of a measurement derived from a sensor.

The modelling bias parameters correspond to deviations due to digital models that can comprise sequences of more or less approximated equations or interpolation tables.

The monitoring and estimation device 1 comprises an initialization module COMP13 configured to initialize the statuses of sensors C1, C2, . . . CN of the parameter P1 corresponding to the weight of the aircraft AC and of the parameters used in the implementation of the monitoring and estimation device 1. For example, the initialization can correspond at least to the fact that the statuses of all the sensors C1, C2, . . . CN are considered to be statuses representative of the correct operation of said sensors C1, C2, . . . CN. The parameters used in the implementation of the device 1 can comprise the parameters relating to the flight or intermediate parameters relating to the operation of the embedded algorithms during the implementation of the device 1.

The monitoring and estimation device 1 further comprises at least the following modules which are implemented iteratively:

• • an estimation module COMP24 (COMP for “computational module”); and
  • a detection module COMP35.

The estimation module 4 is configured to determine an estimation of the values of the parameters relating to the flight of the aircraft AC from:

• • measurements of the parameters relating to the flight supplied by the sensors C1, C2, . . . CN,
  • parameters relating to the flight estimated on the preceding iteration or initialized by the initialization module 3 and
  • statuses associated with each of said sensors C1, C2, . . . CN.

The estimation module 4 is also configured to generate residues which are functions of the measured and estimated values of the parameters relating to the flight and of innovation terms which correspond to the difference between a measured flight parameter value and said estimated value.

FIG. 1 represents a set 2 of N sensors C1, C2, . . . CN.

The detection module 5 is configured to determine the different statuses associated with each of said sensors C1, C2, . . . CN, with a parameter P1 corresponding to a current weight of the aircraft AC, and an estimation of an error of said current weight. The different statuses, the parameter P1 and the error estimation are determined from:

• • said residues determined by the estimation module 4,
  • estimated and measured parameters relating to the flight of the aircraft and the estimation of the weight error, determined by the estimation module 4,
  • said statuses determined on the preceding iteration or initialized by the initialization module 3.

The monitoring and estimation device 1 also comprises a first transmission module TRANS17 (TRANS for “transmission module”) configured to:

• • transmit to a user device 6 and to the detection module 5 a signal representative of the estimation of the values of the parameters relating to the flight of the aircraft AC and of the weight error, determined by the estimation module 4, and
  • send to the detection module 5 a signal representative of the residues generated by the estimation module 4.

The monitoring and estimation device 1 also comprises a second transmission module TRANS28 configured to transmit to the user device 6 and to the estimation module 4 the different statuses associated with each of said sensors C1, C2, . . . CN, with said parameter P1 corresponding to the weight. The transmission of the statuses to the estimation module 4 is performed on the next iteration.

This monitoring and estimation device 1 tackles resolving the problem of being capable of distinguishing a fault between the different sensors C1, C2, . . . CN and the parameter P1 corresponding to the current weight used as input for the monitoring and estimation device 1, and, more specifically, a fault between the incidence, the weight and the speed, including in the case of common failure modes. It makes it possible to continuously deliver an estimation of the parameters relating to the flight of the aircraft AC in real time, including its weight, as well as a status of the different anemometric and clinometric sensors.

The estimation module 4 can comprise the following submodules:

• • an adaptation submodule COMP2141; and
  • an estimation submodule COMP2242.

The adaptation submodule 41 is configured to determine a variance and/or a validity Boolean associated with each of the measurements of the parameters relating to the flight supplied by the sensors C1, C2, . . . CN and of the setting parameters associated with the estimation algorithm used in the estimation submodule 42. The determination is performed from said measurements of the parameters relating to the flight and from the statuses associated with each of said sensors C1, C2, . . . CN. The adaptation submodule 41 is also configured to correct the current weight as a function of the weight error estimated on the preceding iteration or initialized by the initialization module 3, and of the status associated with the parameter P1 corresponding to the weight.

The variances and/or the validity Booleans are determined for flight parameter measurements supplied by sensors C1, C2, . . . CN whose statuses are representative of an operation of the sensors C1, C2, . . . CN.

The adaptation submodule 41 allows the monitoring and estimation device 1 to be configured as a function of the failure cases detected, thus ensuring that said monitoring and estimation device 1 is never fed with a measurement from a sensor C1, C2, . . . CN exhibiting a failure.

The determined variance depends on the sensor C1, C2, . . . CN and reports its accuracy. In a case of failure of a sensor C1, C2, . . . CN, the variance of the measurement of the sensor C1, C2, . . . CN exhibiting a failure is significantly increased so that the latter no longer has an impact in the estimation module. With regard to the validity Booleans, their value is modified in case of failure and in case of revalidation. Only the use of estimation modules based on probability densities requires the association of variances with each of the measurements. In the other cases, the validity Booleans will be used in order to not update the monitoring and estimation method with incorrect measurements.

The adaptation submodule 41 also ensures the observability of the system. Consequently, in case of failure of a sensor C1, C2, . . . CN, it acts on the setting of the estimation module in order to freeze certain states to ensure the convergence of the estimation module. This procedure can be set offline by a user as a function of the failure cases that can be encountered.

Finally, the adaptation submodule 41 also makes it possible to correct the current weight as a function of an estimation of a weight error obtained on the preceding iteration or initialized by the initialization module 3 and of a status associated with the parameter corresponding to the weight.

The estimation submodule 42 is configured to determine the estimation of the values of the parameters relating to the flight and an estimation of the error of said weight, from:

• • the measurements of the parameters relating to the flight supplied by said sensors C1, C2, . . . CN,
  • the parameters relating to the flight estimated on the preceding iteration or initialized by the initialization module 3 and
  • the variance and/or the validity Boolean of each of the measurements of the parameters relating to the flight and of the setting parameters determined by the adaptation submodule 41.

The estimation submodule 42 is also configured to determine the innovations associated with each of the parameters relating to the flight. An innovation is equal to the difference between a measurement of a parameter relating to the flight and an estimation value of said parameter relating to the flight. Finally, it is configured to generate the residues from the innovation terms and the measured and estimated parameters relating to the flight.

The detection module 5 can comprise the following submodules:

• • a detection submodule COMP3151; and
  • a validation submodule COMP3252.

The detection submodule 51 is configured to determine:

• • the common status associated with the sensor C1 configured to measure the angle of incidence α of the aircraft AC and with the parameter P1 corresponding to the current weight and
  • the status associated with the other sensors C2, C3, . . . , CN.

These statuses are determined from:

• • the measurements of the parameters relating to the flight supplied by said sensors C1, C2, . . . , CN,
  • the estimation of the values of the parameters relating to the flight and the weight error,
  • the statuses associated with each of said sensors C1, C2, . . . , CN and with a parameter P1 corresponding to the current weight of the aircraft AC determined on the preceding iteration or initialized by the initialization module 3,
  • the residues.

The submodule for the validation of the angle of incidence and of the current weight 52 is configured to determine the status associated with the parameter P1 corresponding to the current weight and the status associated with the sensor C1 configured to measure the angle of incidence α.

The determination is performed from:

• • the common status associated with the sensors C1 configured to measure the angle of incidence of the aircraft AC and with the parameter P1 corresponding to the current weight,
  • the statuses associated with the other sensors (C2, C3, CN),
  • the estimated parameters relating to the flight,
  • the estimated weight error,
  • the measured parameters relating to the flight,
  • the residues generated by the estimation submodule 42, and
  • a lift coefficient C_L supplied by an embedded modelling and fed by the parameters relating to the flight measured by the sensors C1, C2, . . . , CN and by the estimated parameters relating to the flight.

Hereinafter in the description, the index m refers to the measurements derived from the sensors of the aircraft used as input for the estimation submodule 42.

According to a first embodiment, the estimation submodule 42 corresponds to a Bayesian estimation module of extended Kalman filter type associated with the following system:

$\{\begin{array}{c}\stackrel{.}{X}\left(t\right)=ℱ\left(X\left(t\right)\right)\\ Y\left(t\right)=\left(X\left(t\right),Z\left(t\right),C_L\left(X\left(t\right),Z\left(t\right)\right)\right)\end{array},$

• • in which X is the state vector, Y is the observation vector, Z is an auxiliary measurement vector, is the function associated with the state equation and is the function associated with the observation equation.

The auxiliary measurement vector Z has for expression:

$Z=\left(i_{H_m},{\delta }_{q_{i_m}},{\delta }_{{\mathrm{sp}}_{i_m}},\psi ,\varphi ,\theta ,n_{X_{1_m}},,m,\mathrm{conf},V_{g_{x_{0_m}}},V_{g_{y_{0_m}}},V_{g_{z_{0_m}}},z_{g_m}\right),$

in which:

• • i_Hm corresponds to a measurement of deflection of the horizontal plane,

${\delta }_{q_{i_m}}$

corresponds to a measurement of deflection of elevators of the aircraft (AC),

${\delta }_{{\mathrm{sp}}_{i_m}}$

corresponds to a measurement of deflection of spoilers of the aircraft (AC),

• • ψ corresponds to a heading measurement,
  • φ corresponds to a list angle measurement,
  • θ corresponds to an inclination measurement,

$n_{X_{1_m}}$

• • corresponds to a measurement of the longitudinal load factor in the reference frame associated with the aircraft (AC),
  • m corresponds to the current weight parameter of the aircraft (AC),
  • conf corresponds to a measurement of configuration of the aircraft (AC),

$V_{g_{x_{0_m}}},V_{g_{y_{0_m}}},V_{g_{z_{0_m}}}$

• • correspond to the measurements of the ground speed components in the Earth's reference frame, and
  • z_gm corresponds to a measurement of geometrical altitude.

The configuration measurement corresponds, for example, to the position of the leading edge and of the foil flaps of the aircraft AC.

The Kalman filter is associated with the state vector X and with the observation vector Y.

The Kalman filter does not consider the classic aircraft AC states but considers the atmospheric states as well as biases.

The state vector X has for expression:

$X=\left(\begin{array}{c}{W}_{x_0}\\ W_{y_0}\\ W_{z_0}\\ \Delta \mathrm{ISA}\\ b_{C_L}\\ C_{b_x}\end{array}\right),$

in which:

• • (W_x0, W_y0, W_z0) correspond to the three components of the speed of the wind in the Earth's reference frame,
  • ΔISA corresponds to a temperature deviation between a current static temperature and a temperature determined from a standard atmosphere model at a current geometrical altitude,
  • b_CL corresponds to a lift coefficient modelling bias, and
  • C_bx corresponds to a barometric correction term.

The derivative of the state vector X has for expression:

$\stackrel{.}{x}=\left(\begin{array}{c}{\stackrel{.}{W}}_{x_0}=0\\ {\stackrel{.}{W}}_{y_0}=0\\ {\stackrel{.}{W}}_{z_0}=0\\ \Delta \stackrel{.}{\mathrm{ISA}}=0\\ {\stackrel{.}{b}}_{C_L}=-\frac{b_{C_L}}{{\tau }_b}\\ {\stackrel{.}{C}}_{b_x}=0\end{array}\right),$

• • in which τ_b corresponds to a characteristic time associated with the Markov process used to characterize the trend of the modelling bias. Without being limiting, this characteristic time is set at a few seconds, for example 30 seconds. A similar modelling can be used to describe the trend of the components of the wind.

The observation vector Y has for expression:

$Y=\left(\begin{array}{c}{\alpha }_m\\ {\beta }_m\\ P_{S_m}\\ n_{Z_{1_m}}\\ P_{T_m}\\ T_{T_m}\end{array}\right),$

in which:

• • α_m corresponds to a measurement of angle of incidence,
  • β_m corresponds to a side slip measurement,
  • P_Sm corresponds to a static pressure measurement,

$n_{Z_{1_m}}$

• • corresponds to a measurement of vertical load factor in a reference frame associated with the aircraft,
  • P_Tm corresponds to a total pressure measurement, and
  • T_Tm corresponds to a total temperature measurement.

The incidence has for expression:

${\alpha }_m=\mathrm{atan}\left(\frac{w}{u}\right),$

• • in which u and w are, respectively, the longitudinal and vertical components of an air speed having for norm V=√{square root over (u²+v²+w²)} where v is the lateral component.

The components of the air speed V have for expression:

$\left[\begin{array}{c}u\\ v\\ w\end{array}\right]=\left[\begin{array}{ccc}\cos \theta \cos \psi & \cos \theta \sin \psi & \sin \theta \\ \sin \varphi \sin \theta \cos \psi -\cos \varphi \sin \psi & \sin \varphi \sin \theta \sin \psi +\cos \varphi \cos \psi & -\sin \varphi \cos \theta \\ \cos \varphi \sin \theta \cos \psi +\sin \varphi \sin \psi & \cos \varphi \sin \theta \cos \psi +\sin \varphi \sin \psi & -\cos \varphi \cos \theta \end{array}\right]·\left[\begin{array}{c}{V}_{g_{x_0}}-W_{x_0}\\ V_{g_{y_0}}-W_{y_0}\\ V_{g_{z_0}}-W_{z_0}\end{array}\right],$

• • in which:

$V_{g_{x_0}},V_{g_{y_0}},V_{g_{z_0}}$

are the components of the ground speed in the Earth's reference frame,

• • • ψ corresponds to a heading measurement,
    • φ corresponds to a list angle measurement, and
    • θ corresponds to an inclination measurement.

The side slip has for expression:

${\beta }_m=\mathrm{atan}\left(\frac{v}{\sqrt{u^2+m^2}}\right).$

The static pressure has for expression:

P _{S m} =ζ(z_Pm),

• • in which is a function linking the measurement of the pressure altitude z_Pm to the static pressure P_Sm according to the formula:

$\{\begin{array}{c}\zeta \left(z_{P_m}\ge z_{P_{11}}\right)={10}^{{\log }_{10}\left(P_{11}\right)+\frac{g\left(z_{P_{11}}-z_{P_m}\right)}{{\mathrm{RT}}_{11}\log \left(10\right)}}\\ \zeta \left(z_{P_m}<z_{P_{11}}\right)={P_0\left(1-\frac{G_{T_{z_0}}}{T_{15}}z_{P_m}\right)}^{\frac{g}{{\mathrm{rG}}_{T_{z_0}}}}\end{array},$

• • in which z_P11=11 km corresponds to the standard altitude of the tropopause, P₁₁=226.321 mbar and T₁₁=216.65 k correspond to the standard static pressure and static temperature at the tropopause, G_Tz0=0.0065 k/m corresponds to the standard temperature gradient for z_Pm<z_P11, g corresponds to the acceleration of gravity and r corresponds to the specific constant of the air.

The pressure altitude has for expression:

$z_{P_m}=\frac{z_G}{1+\frac{\Delta \mathrm{ISA}}{T_{15}}}-C_{b_x}$

• • in which T₁₅=288.15 K.

The vertical load factor

$n_{Z_{1_m}}$

has for expression:

$n_{z_{1_m}}=\frac{\gamma {\mathrm{SP}}_SM^2\left(C_L+b_{C_L}\right)}{2\mathrm{mg}\cos \alpha }+n_{X_{1_m}}\tan \alpha ,$

• • in which γ is the adiabatic incidence of the air (equal to 1.4), S is the reference surface of the aircraft, M is the Mach number, C_L is the lift coefficient,

$n_{X_{1_m}}$

is the horizontal load factor, m is the weight of the aircraft, g is the acceleration of gravity. The Mach number M has for expression

$M=\frac{V}{\sqrt{\gamma {\mathrm{rT}}_S}},$

• •  in which T_S is the static temperature and r is the specific constant of the air.

The static temperature T_S has for expression T_s=T₀+G_Tz0z_G+ΔISA, in which T₀=273.15 K, G_Tz0=0.0065 K/m and z_G is the geographic altitude.

The total pressure has for expression:

$P_T={P_S\left(1+\frac{\gamma -1}{2}M^2\right)}^{\frac{\gamma }{\gamma -1}}.$

The total temperature has for expression:

$T_T=T_S\left(1+\frac{\gamma -1}{2}M^2\right).$

The sensors considered are therefore three anemometric sensors (static pressure, total pressure and total temperature sensors), angle of incidence probes C1, positioning sensors such as satellite geolocation systems (GPS for “Global Positioning System”) and inertial units such as IRS (“Inertial Reference System”) systems. A first anemometric sensor measures the total pressure. A second anemometric sensor measures the static pressure. A third anemometric sensor measures the total temperature.

Among the parameters that can be incorrect, there is the parameter P1 corresponding to the weight of the aircraft AC. The parameter P1 is equal to the sum of the weight of the aircraft AC without fuel and of the weight of the fuel. This parameter P1 can be incorrect on take-off of the aircraft AC and until the end of the flight.

The setting of the extended Kalman filter is performed via state and measurement noise covariance matrixes and on the initialization implemented by the initialization module 3 on a state vector X₀ and a covariance matrix of the error P₀.

From an algorithmic point of view, an analytical formulation is preferentially used for the computation of the Jacobian matrixes associated with the Kalman filter.

Preferably, in a Kalman filter correction step, an algorithm of sequential processing of the measurements is used in order to be able to select, on a per-case basis, the valid measurements, without modifying the settings of the measurement noise covariance matrix R. This makes it possible in particular to avoid a more computation time-consuming matrix inversion operation. For the determination of the valid measurements, use will be made of the validity vector f=[FI_sensors] obtained from the combinations of the different fault indicators defined in the detection submodule 51 whose value refers to the validity of the sensors C1, C2, . . . , CN used as input and of the parameter P1 corresponding to the weight. Said vector refers to the value of the variances and validity Booleans. The change of the value of a validity Boolean from 1 to 0 or a drastic increase in the associated variance makes it possible to no longer take account of the invalidated sensor in the updating of the states whereas the change of a validity Boolean from 0 to 1 or the reselection of a standard variance makes it possible to once again select the valid parameter and/or sensor. Adaptive extended Kalman filter is then referred to for all of the estimation module. The validity vector f is kept up to date using the detection module 5. All of the values of the parameters of the observation vector Y(t) and of the auxiliary measurement vector Z(t) define the inputs of the estimation submodule implementing the extended Kalman filter. The extended Kalman filter is formed to deliver in real time, in a flight of the aircraft AC, estimations of the selected flight parameters and of the atmospheric and bias parameters and makes it possible in particular to recompute an estimated calibrated speed CAS according to the formula:

$\mathrm{CAS}=\sqrt{\gamma {\mathrm{RT}}_0}\sqrt{\frac{2}{\gamma -1}\left({\left(\frac{P_T-P_S}{P_0}+1\right)}^{\frac{\gamma -1}{\gamma }}-1\right)}$

With P₀=101325 Pa, T₀=273.15 K and γ=1.4, the adiabatic incidence of the air and r=287.058 J·kg⁻¹·K⁻¹, the specific constant of the air.

The estimation submodule 42 generates, on an iteration k, the following residues.

$r_1\left(t_k\right)={\varepsilon }_{\alpha }\left(t_k\right)+\frac{{\hat{b}}_{C_L}\left(t_k\right)}{C_{L_{\alpha }}\left(\mathrm{conf}\left(t_k\right)\right)}$

A first residue r₁ has for expression in which ε_α(t_k) corresponds to the innovation term associated with the angle of incidence α at a time t_k, {circumflex over (b)}_CL corresponds to the estimated modelling bias of the lift coefficient, and C_Lα(conf(t_k)) corresponds to a tabulated value of the lift coefficient C_Lα depending only on the value of the parameter corresponding to the configuration of the aircraft AC (position of the leading edge and of the foil flaps of the aircraft AC) at a time t_k.

A second residue r₂ has for expression at the time

$t_kr_2\left(t_k\right)=\sqrt{\frac{2}{\gamma -1}\left(\frac{T_{T_m}}{{\hat{T}}_s\left(z_{P_m},t_k\right)}-1\right)}-\sqrt{\frac{2}{\gamma -1}\left({\left(\frac{P_{T_m}}{\zeta \left(z_{P_m}\right)}\right)}^{\frac{\gamma -1}{\gamma }}-1\right)},$

in which:

• • γ corresponds to the adiabatic coefficient of the air,
  • T_Tm corresponds to the measured total temperature flight parameter,
  • P_Tm corresponds to the measured total pressure,
  • z_Pm corresponds to the measured pressure altitude,
  • ζ corresponds to the function linking the measurement of the pressure altitude to the static pressure, and
  • {circumflex over (T)}_s(z_Pm, t_k) corresponds to an estimation of the static temperature.

The estimation of the static temperature {circumflex over (T)}_s(z_Pm, t_k) is recomputed from the parameters estimated at the current time instant t_k and preceding time instant t_k-1 as a function of the value of the residue r₁, according to the formula:

${\hat{T}}_s\left(z_{P_m},t_k\right)=T_{15}+G_{T_{z0}}\left(z_{P_m}\left(t_k\right)+ℒ\left(s\right)\left[{\hat{C}}_{b_x}\left(t_k\right){\Pi }_{r\frac{+}{1}}\left(r_1\left(t_k\right)\right)+{\hat{C}}_{b_x}\left(t_{k-1}\right)\left(1-{\Pi }_{r\frac{+}{1}}\left(r_1\left(t_k\right)\right)\right)\right]\right)+ℒ\left(s\right)\left[\left(t_k\right){\Pi }_{r\frac{+}{1}}\left(r_1\left(t_k\right)\right)+\left(t_{k-1}\right)\left(1-{\Pi }_{r\frac{+}{1}}\left(r_1\left(t_k\right)\right)\right)\right],$

in which is the transfer function of a low-pass filter and

${\Pi }_{r\frac{+}{1}}\left(r\right)=H\left(r+\frac{r\frac{+}{1}}{2}\right)-H\left(r-\frac{r\frac{+}{1}}{2}\right)$

where H is the Heaviside function, and

$r\frac{+}{1}$

corresponds to the limits associated with the residue r₁ defined hereinafter in the description.

A third residue has for expression at the instant t_k,

$r_3\left(t_k\right)=\sqrt{\frac{2}{\gamma -1}\left(\frac{T_{T_m}}{{\hat{T}}_s\left(z_{P_m},t_k\right)}-1\right)}-\sqrt{\frac{2{\mathrm{mgn}}_{z_{1_m}}\cos {\alpha }_m}{\gamma S\zeta \left(z_{P_m}\right){\hat{C}}_z}-n_{X_{1_m}}\tan {\alpha }_m}$

in which:

• • m corresponds to the current weight of the aircraft AC,
  • S corresponds to the reference surface of the aircraft AC,
  • g corresponds to the acceleration of gravity,
  • γ corresponds to the adiabatic coefficient of the air,

$n_{Z_{1_m}}$

• • corresponds to the measured vertical load factor,

$n_{X_{1_m}}$

• • corresponds to the measured longitudinal load factor,
  • α_m corresponds to the measured angle of incidence,
  • z_Pm corresponds to the measured pressure altitude,
  • T_Tm corresponds to the measured total temperature,
  • ζ corresponds to the function linking the measurement of the pressure altitude to the static pressure,
  • {circumflex over (T)}_s(z_Pm, t_k) corresponds to the estimation of the static temperature defined in the residue r₂, and
  • Ĉ_z corresponds to the lift coefficient computed from the estimated parameters relating to the flight.

In a nonlimiting manner, the first residue r₁ is filtered over a time τ_rc of a few seconds, whereas the residues r₂ and r₃ are forced to a nil value as long as the geometrical altitude z_G of the aircraft AC is less than a limit altitude chosen in a nonlimiting manner to be 7000 feet (approximately 2133.6 m) because of the inaccuracies of the expression expressing the pressure altitude

$z_{P_m}=\frac{z_G}{1+\frac{\Delta \mathrm{ISA}}{T_{15}}}-C_{b_x}$

below 7000 feet.

For example, the time τ_rc is equal to 10 s.

For the determination of the status of the sensors C1, C2, . . . CN and of the parameter P1 corresponding to the current weight of the aircraft AC, the detection submodule 51 is configured to:

• • assign to each residue r_i a maximum limit r_i⁺ and a minimum limit r_i⁻ with i lying between 1 and 3;
  • construct two validity indicators in Boolean form V_PS and V_PT associated with the static pressure and total pressure sensors from the estimated parameters and the measurements derived from the sensors of the aircraft;
  • deduce therefrom residue indicators R_i associated with each residue r_i and validity indicators V_i;
  • identify the occurrence of a failure when the sum of the residue indicators is strictly greater than zero (ΣR_i>0) and identify (or isolate), in case of failure, the faulty sensor by comparing the current values of the residue indicators to those itemized in a table identified offline or online itemizing the failure cases as a function of the different combinations of values of the residue indicators R_i.

The validity indicators V_PS and V_PT are constructed as follows.

The validity indicator V_PS has the value 1 if the relationship |_n(s)(z_Pm z_Gm)|≤l_PS is borne out. V_PS has the value 0 if said relationship is not borne out.

The term l_PS corresponds to a limit determined experimentally as a function of the dynamics of the aeroplane. The term _h corresponds to a transfer function of a high-pass filter with a time constant chosen in a nonlimiting manner to be a few seconds, for example 30 s.

The validity indicator V_PT has the value 1 if the residue r₁ crosses its respective limits, at an instant t_k, and that nϵ such that t₀ϵ[t_k−τ_PT, t_k] bearing out |(1−_τ)(M_PT(t_n)−M_PT(t_n-j))|≤l_PT, with:

• • τ_PT corresponding to a constant chosen in a nonlimiting manner to be 120 seconds,
  • _τ corresponding to a transfer function of a low-pass filter having for time constant τ chosen in a nonlimiting manner to be a few hundreds of seconds, for example 700 s,
  • jϵ corresponding to a constant chosen in a nonlimiting manner to be equal to 8, and
  • l_PT corresponding to a limit determined experimentally as a function of the dynamics of the aeroplane.

To deduce a residue indicator R_i associated with a residue r_i, reference is made to the value of the residue r_i. If the value of the residue r_i is respectively greater or smaller than the maximum limit r_i⁺ or the minimum limit r_i⁻, the residue indicator R₁ is equal to 1, otherwise the residue indicator R₁ is equal to zero. In the case of a residue indicator R_i associated with a validity indicator V_i, the residue indicator R_i is equal to the validity indicator V_i.

Each column of the table, itemizing the failure cases as a function of the different combinations of values of the residue indicators R_i, corresponds to a fault index FI_i (fault indicator) and is the combination of residue indicators R₁, R₂ . . . R_m that can take 0 or 1 for value and each row corresponds to the values of a residue indicator R_i for each of the fault indexes FI₁, FI₂, . . . FI_m constructed. A fault index FI_i refers to the status of the sensor Ci, except in a first instance of the sensor measuring the angle of incidence C1 for which the fault index FI_α/m refers to a common status of the measurement of the angle of incidence measured by an angle-of-incidence measurement sensor C1 and of the parameter P1 of current weight of the aircraft.

The maximum limit r_i⁺ and the minimum limit r_i⁻ are determined from thresholds

$r\frac{+}{i},$

tabulated as a function of estimated flight parameters, centred on a central residue value r_im which corresponds to the residue r_i filtered for a time τ_i which depends on each residue. Thus, the maximum limit r_i⁺ has for expression

$r_i^{+}=r_{i_m}+r\frac{+}{i}$

and the minimum limit r_i⁻ has for expression

$r_i^{-}=r_{i_m}-r\frac{+}{i}.$

The definition of the thresholds

$r\frac{+}{i}$

can be done in many ways. Preferentially, the threshold is chosen to be symmetrical about a filtered value, denoted r_im, of the raw residue over a time τ_r that is sufficiently great. The distance

$r\frac{+}{i}-r_{i_m}$

can depend on the accuracy of a modelling and on the accuracy of sensors C1, C2, . . . CN. In the chosen application case, it can be a function of the Mach, of the configuration of the aircraft AC, of the control surface deflections but also of the diagonal elements of the covariance matrix of the error computed via the adaptive extended Kalman filter previously described:

$r\frac{+}{i}-r_{i_m}={\sigma }_i\left(M,\mathrm{conf}\right){\left(1+u_i{\mathrm{Pu}}_i^T\right)}^{1/2},$

• • in which u_i is a dependency vector making it possible to link the influence of each state to the measurement i.

In order to avoid including thresholds in case of failure, it is recommended to introduce saturators to force r_i⁺ and r_i⁻ to remain within a realistic predefined interval:

$\forall i,\{\begin{array}{c}{r}_i^{+}\in \left[r_{i_{\min }}^{+}r_{i_{\max }}^{+}\right]\\ r_i^{-}\in \left[r_{i_{\min }}^{-}r_{i_{\max }}^{-}\right]\end{array}.$

The threshold r₁^± for the first residue r₁ is defined as a function of the confidence that is placed on the model of the lift coefficient C_L. It can for example depend on the configuration of the aircraft AC, on the Mach number, on the position of the landing gears of the aircraft AC (lowered or retracted), on the position of the control surfaces whose effects have not been taken into account in the modelling and on the accuracy of the incidence probes. In a nonlimiting manner, the filtering time τ_r¹ is chosen to be 500 seconds.

The thresholds

$r_2^{\pm }\mathrm{and}r_3^{\pm }$

for the second residue r₂ and the third residue r₃ are defined according to a different logic given that they are associated with Mach measurement deviations. Their value can depend on the accuracy of the anemometric measurement sensors and on the estimated or measured flight parameters such as the geometrical altitude z_G.

For the isolation of a failure, the algorithm of the detection submodule 51 refers to the combinations of the failure table below. The acronym FI refers to the fault indicator whereas the index f=0 refers to the nominal case (that is to say in the absence of failure before the detection of the fault) and f>0 in the case of degraded operation (the measurement of at least one parameter has already been detected as faulty and is therefore no longer used by the system). R_v1 and R_v2 are associated respectively with the validity indicators V_PS and V_PT.

	FIα/m,f=0	FIPT,f=0	FIPs,f=0	FITT,f=0
	R1	1	1	0/1	(0/1)
	R2	0/1	0/1	0/1	1
	R3	0	0/1	0/1	1
	Rv1	0	0	1	0
	Rv2	0	1	0	0

These combinations depend on the setting of the estimator and, more particularly, on the accuracy of the sensors C1, C2, . . . , CN used. They are determined offline or online via a dedicated algorithm. The notation ‘0/1’ refers to the values of the residue indicators that do not have an impact for the isolation of the failure, the isolation being satisfied via the other dependencies.

Depending on the combination of 1 and of 0 obtained on the different failure indicators FI, a deduction is made therefrom as to whether there is a failure and its origin.

For example, the status of the total pressure sensor P_T corresponds to a failure when R₁ and R_v2 are equal to 1. The common status of the sensor of angle of incidence α and of the weight m corresponds to a failure state if R₁ is equal to 1 and R₃, R_v1 and R_v2 are equal to zero.

The submodule 51 for detecting a failing sensor C1, C2, . . . , CN makes it possible both to detect and isolate the fault, with an indeterminacy to be lifted between a fault on the measurement of the angle of incidence α and on the current weight m, reflected by the notation α/m in FI_α/m,f=0 which refers to both FI_α,f=0 and FI_m,f=0 given that they have the same signature and are therefore merged. The validation submodule 52 makes it possible to lift the ambiguity.

In the case where a first failure has already been detected (f>0), the fault indicators for the valid sensors (with which the fault indexes R_j are associated) can be reduced to a simple indicator:

FI_f>0=1 if ΣR_j>0

The detection of the failure will then be possible while the identification of the source will not be able to be made directly without the addition of information external to the system described. Abrupt failures do nevertheless remain detectable and isolatable via the use of the preceding dependency matrix by considering only the sensors C1, C2, . . . , CN, the residues r₁, r₂, r₃ and the validity indicators V_PS, V_PT that are still valid.

These indicators are then communicated to the validation submodule 52 with the current value of the weight m and the estimation of the lift coefficient modelling bias {circumflex over (b)}_CL, of the lift coefficient C_L and of the first residue r₁.

The validation submodule 52 makes it possible to distinguish a fault on the incidence sensors C1 from a weight error. This validation submodule 52 nevertheless requires a minimum of variation of the angle of incidence α to work. An indicator that will be called “dynamic indicator” here is therefore established in order to validate the results obtained.

As soon as the detection submodule returns a fault on the incidence α or the weight m, the maximum deviation of variation of weight m is computed from the instant t₁ of detection.

From the estimation of the first residue r₁, the following weight error m is computed at each instant t:

$\Delta m=m\frac{C_{L_{\alpha }}\left(\mathrm{conf}\right)r_1}{{\hat{C}}_L},$

in which Ĉ_L corresponds to the estimation of the lift coefficient, C_Lα(conf) to the coefficient C_Lα derived from a simplified model that is only a function of the configuration of the aircraft AC and m the current weight of the aircraft AC.

The dynamic indicator is defined on the value of the incidence α. Given that the measurement thereof can be incorrect, a virtual incidence α_v is constructed from the raw measurement of the estimated modelling bias α_m and the residue associated with the incidence r_cα:

α_v=α_m−r_cα

In this way, the virtual incidence α_v will compensate any failure on the measured incidence. The latter will nevertheless still be affected by a weight error m which will introduce a bias but its variation Δα_v will be identical to that of the real incidence.

A first variant for the definition of the virtual incidence corresponds to the integration of the following formula:

{dot over (α)}_v≈q+g(1−n_z1)/{circumflex over (V)},

in which:

• • g is the acceleration of gravity,
  • q corresponds to the angular pitch speed,
  • n_z1 corresponds to the measurement of the vertical load factor expressed in the reference frame associated with the aircraft AC,
  • {circumflex over (V)} corresponds to the estimation of the air speed.

The auxiliary measurement vector Z is then increased by the measurement q corresponding to the angular pitch speed measurement in the reference frame of the aircraft.

The variation of the virtual incidence Δα_v has for expression:

${\mathrm{\Delta \alpha }}_V\left(t\right)=\underset{\left(t_i,t_j\right)\in \left[t_1,t\right]}{\max }\left|{\alpha }_V\left(t_i\right)-{\alpha }_V\left(t_j\right)\right|$

There is then defined a minimum variation threshold that has to be reached by the virtual incidence Δα_v from the detection of the failure in order to allow the isolation of the failure between an angle of incidence α measurement error and an incorrect current weight m, and in the second case, the evaluation of the associated weight error via a second computation. t₂ is then defined such that:

$t_2=\underset{{\mathrm{\Delta \alpha }}_V\left(t_2\right)>{\mathrm{\Delta \alpha }}_{\lim }}{\min }t$

A second threshold on the estimated weight variation makes it possible to deduce therefrom whether the latter is sufficiently great to be associated with a fault of the angle of incidence sensor C1. This threshold depends for example on the quality of the modelling used for the lift coefficient C_L.

The maximum deviation of weight variation Δ²m is deduced therefrom from the instant t₁ of detection to the instant t₂:

${\Delta }^2m=\underset{\left(t_i,t_j\right)\in \left[t_1,t_2\right]}{\max }\left|\Delta m\left(t_i\right)-\Delta m\left(t_j\right)\right|$

Δ²m_lim is defined as the maximum variation limit that makes it possible to deduce therefrom whether the failure can be identified as a fault of the angle of incidence sensors C1 or as a current weight m error:

if Δ²m<Δ²m_lim then

FI_α,f=0=0

FI_m,f=0=1

else

FI_α,f=0=1

FI_m,f=0=0

The validation submodule 52 finally makes it possible to establish the fault indicators associated with the incidence FI_α,f=0 and with the weight FI_m,f=0.

There is nevertheless a type of failure on the incidence α that can lead to a bad interpretation:

$\{\begin{array}{c}{\alpha }_{\mathrm{err}}\left(t\right)=\alpha \left(t\right)+\mathrm{\Delta \alpha }\left(t\right)\\ \mathrm{\Delta \alpha }\left(t\right)=K\left(\alpha \left(t\right)-{\alpha }_0\right)\end{array}\hspace{1em}$

This is a fault of scale factor type on the incidence α. All the other types of failures are correctly identified. In order to minimize this type of error, two additional checks are possible. First of all, check that the estimated weight is situated within the interval between the minimum and maximum weight on take-off defined for the aircraft AC. Then, a request to verify the weight input on take-off by the pilot is possible as is the case applied in certain situations.

In a first variant embodiment, it is possible to integrate the engine measurements and considerably increase the observability of the system and therefore the detection capabilities of the algorithm via the addition of a modelling called engine-nacelle modelling described hereinabove. For that, three new observations are integrated:

• • the virtual total pressure derived from the engine-nacelle modelling P_Teng,
  • the virtual static pressure derived from the engine-nacelle modelling P_Seng,
  • the total engine temperature T_Teng.

These two virtual measurements are derived from the engine-nacelle modelling using only the static engine pressure P_mot and the static pressure measured in the nacelle P_nac and the total engine temperature T_Teng. They exhibit two common failure modes given that a failure of one of the two static pressures leads to a fault of the virtual static pressure and of the virtual total pressure and that a fault of the total engine temperature leads to a fault thereof and of the two virtual measurements.

Thus, the observation vector has for expression:

$Y=\left(\begin{array}{c}{\alpha }_m=\mathrm{atan}\left(\frac{w}{u}\right)\\ {\beta }_m=\mathrm{atan}\left(\frac{v}{\sqrt{u^2+w^2}}\right)\\ P_{S_m}=\zeta \left(z_p\right)\\ n_{Z_m}=\frac{{\mathrm{SP}}_d\left(C_L+b_{C_L}\right)}{\mathrm{mg}\cos \alpha }+n_{X_{1_m}}\tan \alpha \\ P_{T_m}={P_S\left(1+\frac{\gamma -1}{2}M^2\right)}^{\frac{\gamma }{\gamma -1}}\\ T_{T_m}=T_S\left(1+\frac{\gamma -1}{2}M^2\right)\\ P_{T_{\mathrm{eng}}}={P_S\left(1+\frac{\gamma -1}{2}M^2\right)}^{\frac{\gamma }{\gamma -1}}\\ P_{S_{\mathrm{eng}}}=\zeta \left(z_p\right)\\ T_{T_{{\mathrm{eng}}_m}}=T_S\left(1+\frac{\gamma -1}{2}M^2\right)\end{array}\right)$

in which

• • P_Teng corresponds to a total pressure value derived from the nacelle-engine modelling,
  • P_Seng corresponds to a static pressure value derived from the modelling, and

$T_{T_{{\mathrm{eng}}_m}}$

corresponds to a measured total temperature value derived from the total engine temperature measurement.

When setting the Kalman filter, a choice will be made not to use the engine measurements to update the estimated states. They will be used only for the detection and will then replace the measurement missing after certain failure cases. The variance specified on each other sensor is otherwise used for the setting of the filter.

The sensors C1, C2, . . . , CN considered are those of the first variant to which are added the total engine temperature sensor T_Teng and the virtual sensors of static pressure P_Seng, and of total engine pressure P_Teng.

The residues generated are as follows:

• • a fourth residue r₄ having for expression

$r_4\left(t_k\right)={\varepsilon }_{P_{T_{\mathrm{eng}}}}\left(t_k\right)$

in which

${\varepsilon }_{P_{T_{\mathrm{eng}}}}\left(t_k\right)$

corresponds to the difference at a time (t_k) between the measured value of the total pressure and said virtual measurement derived from the engine-nacelle modelling P_Teng,

• • a fifth residue r₅ having for expression

$r_5\left(t_k\right)={\varepsilon }_{P_{s_{\mathrm{eng}}}}\left(t_k\right)\mathrm{in}\mathrm{which}{\varepsilon }_{P_{s_{\mathrm{eng}}}}\left(t_k\right)$

corresponds to the difference at a time (t_k) between the measured static pressure value and said virtual measurement derived from the modelling P_Seng, and

• • a sixth residue r₆ having for expression

$r_6\left(t_k\right)={\varepsilon }_{T_{T_{\mathrm{eng}}}}\left(t_k\right)\mathrm{in}\mathrm{which}{\varepsilon }_{T_{T_{\mathrm{eng}}}}\left(t_k\right)$

corresponds to the difference at a time (t_k) between the measured total temperature value and said engine measurement T_Teng.

In a nonlimiting manner, the residues r_i generated in the estimation substep can also be filtered over a time τ_rc.

The thresholds

$r_4^{\pm },r_5^{\pm }\mathrm{and}r_6^{\pm }$

associated respectively with the residues r₄, r₅ and r₆ to determine the maximum limit and the minimum limit depend on the accuracy of the model associated with the virtual engine sensors and on the accuracy of the engine P_eng, nacelle P_nac static pressures and of the total temperature sensor which feed it.

For the detection of a failure, the algorithm of the detection submodule 51 then refers to the fault indexes defined according to the combinations of the failure table below.

	FIα/m,f=0	FIPT,f=0	FIPs,f=0	FITT,f=0	FIPeng,f=0
R1	1	1	0/1	(0/1)	0	0
R2	0/1	0/1	0/1	1	0	0
R3	0	0/1	0/1	1	0	0
R4	0	0	1	0	1	0/1
R5	0	1	0	0	1	0/1
R6	0	0	0	1	0	1
Rv1	0	0	0/1	0	0	0
Rv2	0	0/1	0	0	0	0

This first variant has the advantage of a great number of observations. Consequently, it is possible to isolate the source of a failure much more rapidly than previously by means of the indicators marked by the notation ‘1’. Conversely, although the notation ‘0/1’ reflects a possible overshoot of the associated threshold, the latter does not need to be identified to rapidly isolate the failure, the isolation being satisfied via the other dependencies. These combinations depend on the setting of the estimator and more particularly on the accuracy of the sensors C1, C2, . . . , CN used, as evidenced by the notation ‘(0/1)’ which reports a possible overshoot as a function of the chosen setting. They are determined offline by the use or online via a dedicated algorithm.

Once a first failure is detected and isolated, it is possible to continue the monitoring by using the residues that are still relevant. For example, in case of failure of the measurements derived from the engine model, it will be reduced to the formulation of the standard embodiment. It is therefore possible to detect and isolate at least two successive sensor failures and to reconstitute the missing parameters. Beyond, the level of observability may no longer be sufficient and some states may be frozen to continue to ensure the stability of the filter. The failure detection will nevertheless still be able to be ensured in some cases.

In a second variant, it is possible to integrate in particular a kinematic model allowing a monitoring and an estimation of the ground parameters including the biases of the accelerometers

$b_{n_{x_1}},b_{n_{y_1}},b_{n_{z_1}}$

on the three axes of the aircraft. This second variant is of interest because air space is these days undergoing a few difficulties with the GPS signals, particularly when flying over territory using GPS scramblers as is the case in some countries at war. In this case, the GPS measurements may no longer be available. In order to avoid an unpredicted removal of the detector estimator filter, it is necessary to ensure the detection of a removal of the GPS measurements in order to leave control to the inertial sensors for the time of their unavailability. The estimation of the accelerometric biases makes it possible to refine the estimation of the ground parameters in the absence of GPS measurements.

In this second variant, the state vector X has for expression:

$X=\left(\begin{array}{c}{W}_{x_0}\\ W_{y_0}\\ W_{z_0}\\ \Delta \mathrm{ISA}\\ b_{C_L}\\ C_{b_x}\\ v_{g_{x_0}}\\ v_{g_{y_0}}\\ v_{g_{z_0}}\\ b_{n_{x_1}}\\ b_{n_{y_1}}\\ b_{n_{z_1}}\\ z_G\end{array}\right),$

in which:

$V_{g_{x_{0_m}}},V_{g_{y_{0_m}}},V_{g_{z_{0_m}}}$

• • correspond to the three components of the ground speed in the Earth's reference frame,

$b_{n_{x_1}},b_{n_{y_1}},b_{n_{z_1}}$

• • correspond to the three bias components of accelerometers in the reference frame associated with the aircraft AC,
  • z_G corresponds to a geometrical altitude.

The derivative of the state vector has for expression:

$\stackrel{.}{X}=\left(\begin{array}{c}{\stackrel{.}{W}}_{x_0}=0\\ {\stackrel{.}{W}}_{y_0}=0\\ {\stackrel{.}{W}}_{z_0}=0\\ \Delta \stackrel{.}{\mathrm{ISA}}=0\\ {\stackrel{.}{b}}_{C_L}=-\frac{b_{C_L}}{{\tau }_b}\\ {\stackrel{.}{C}}_{b_x}=0\\ \begin{array}{c}\begin{array}{c}{\stackrel{.}{v}}_{g_{x_0}}\\ {\stackrel{.}{v}}_{g_{y_0}}\\ {\stackrel{.}{v}}_{g_{z_0}}\end{array}=-g*M_{\mathrm{rot}}\left(\begin{array}{c}\left(n_{x_{1_m}}+b_{\mathrm{nx}}\right)\sin \theta \\ \left(n_{y_{1_m}}+b_{\mathrm{ny}}\right)\cos \theta \sin \phi \\ \left(n_{z_{1_m}}+b_{\mathrm{nz}}\right)\cos \theta \cos \phi \end{array}\right)\\ {\stackrel{.}{b}}_{n_{x_1}}=0\end{array}\\ {\stackrel{.}{b}}_{n_{y_1}}=0\\ {\stackrel{.}{b}}_{n_{z_1}}=0\\ {\stackrel{.}{b}}_{n_{z_1}}=0\\ {\stackrel{.}{z}}_G=V_{g_{z_0}}\end{array}\right),$

in which

$M_{\mathrm{rot}}=\left(\begin{array}{ccc}\cos \theta \cos \doteq & \sin \phi \sin \theta \cos \psi -\cos \phi \sin \psi & \cos \phi \sin \theta \cos \psi +\sin \phi \sin \psi \\ \cos \theta \sin \psi & \sin \phi \sin \theta \sin \psi +\cos \phi \cos \psi & \cos \phi \sin \theta \sin \psi -\sin \phi \cos \psi \\ \sin \theta & -\sin \phi \cos \theta & -\cos \phi \cos \theta \end{array}\right).$

The auxiliary measurement vector Z is increased by the measurement

$n_{y_{1_m}}$

corresponding to the measurement of the lateral load factor in the reference frame of the aircraft.

The observation vector Y therefore has for expression:

$Y=\left(\begin{array}{c}{\alpha }_m=\mathrm{atan}\left(\frac{w}{u}\right)\\ {\beta }_m=\mathrm{atan}\left(\frac{v}{\sqrt{u^2+w^2}}\right)\\ P_{S_m}=\zeta \left(z_P\right)\\ n_{Z_{1_m}}=\frac{\gamma {\mathrm{SP}}_sM^2\left(C_L+b_{C_L}\right)}{2\mathrm{mg}\cos \alpha }+n_{X_{1_m}}\tan \alpha \\ P_{T_m}={P_S\left(1+\frac{\gamma -1}{2}M^2\right)}^{\frac{\gamma }{\gamma -1}}\\ T_{T_m}=T_S\left(1+\frac{\gamma -1}{2}M^2\right)\\ V_{g_{x_{0_m}}}=V_{g_{x_0}}\\ V_{g_{y_{0_m}}}=V_{g_{y_0}}\\ V_{g_{z_{0_m}}}=V_{g_{z_0}}\\ z_{G_m}=z_G\end{array}\right),$

in which:

$V_{g_{x_{o_m}}},V_{g_{y_{o_m}}},V_{g_{z_{o_m}}}$

• • correspond to measurements of the three components of the ground speed, and
  • z_Gm corresponds to a measurement of the geometrical altitude.

This second variant can be coupled or not with the first variant for a more comprehensive formulation and therefore greater performance. This nevertheless has the disadvantage of being more complex and therefore of having a much higher computational impact as well as a more difficult setting of the estimator filter.

The knowledge of the kinematic model makes it possible to perform the monitoring of the GPS measurements via the study of the residues associated with the new states introduced and of the trend of the accelerometric biases

$b_{n_{x_1}},b_{n_{y_1}},b_{n_{z_1}}.$

In nominal operation, this is done at low frequency and with a low amplitude. Conversely, any failure of the GPS measurements leads to abrupt modifications of great amplitude.

According to a third variant, the lateral component of the air speed v is assumed nil, which amounts to considering the side slip p as nil. This is true most of the time. Some situations such as engine failure can invalidate this assumption. Nevertheless the side slip values in all cases remain relatively low and even in case of strong variations of the side slip angle, that affects the estimation and detection process and the performance levels thereof only very little. This assumption therefore only introduces very little in the way of errors whereas it makes it possible to dispense with any monitoring on the side slip probes (which would otherwise have required establishing a lateral model of the aircraft AC). A coupling with the previous variants described is possible.

According to a second embodiment (FIG. 1), the statuses associated with some of said sensors C1, C2, . . . CN are also determined from auxiliary statuses associated with these sensors that can be sent to the detection module by an external monitoring module COMP49. The detection submodule 51 then takes care of translating the statuses sent by the external monitoring module 9 before transmitting the translated statues to the estimation module 4. This second embodiment can be combined with the other embodiments.

According to a third embodiment represented in FIG. 4, in the case where the aircraft AC uses hardware redundancy, the inputs of the estimation module 4 correspond initially to the consolidated measurements as they are derived from the output of a vote module 11 based on a majority-based vote. It is however possible to imagine cases of figures where the consolidated measurement is invalidated by the vote module 11 and is therefore faulty, but that one of the sensors of the corresponding type is still valid, which may be the case with some common failure modes. In this case, it is advantageous to have the capacity to recover the measurement that is still valid. FIG. 4 represents several sets of redundant sensors C1, C2, C3, C4, C5, C6 comprising one or more sensors. For example, the set of sensors C1 comprises sensors C1a, C1b, C1c of angle of incidence. The set of sensors C2 comprises sensors C2a, C2b, C2c of static pressure. The set of sensors C3 comprises sensors C3a, C3b, C3c of total pressure. The set of sensors C4 comprises sensors C4a, C4b, C4c of total temperature. The set of sensors C5 comprises sensors C5a and C5b correspondingly respectively to a satellite geolocation system assembly and an inertial unit assembly. The set of sensors C6 corresponds to virtual sensors of virtual total pressure P_Teng, of the virtual static pressure P_Seng derived from the engine-nacelle modelling and from the total engine temperature T_Teng.

The monitoring and estimation device 1 comprises a unitary verification module COMP510. When a failure is detected for a measurement used as input for the estimation module, the unitary verification module is configured to:

• • reconfigure the estimation submodule 42 to not take account of the faulty measurement,
  • after a convergence time τ_cv, compute the differences between the estimations of said parameters relating to the flight and each of the measurements derived from the plurality of sensors. For each of the measurements derived from the plurality of sensors, if the absolute value of the difference is less than a predetermined value, the corresponding measurement is retained for the estimation submodule 42.

A possible variant relates to the use of methods similar to the extended Kalman filter, using the same set of equations as the integration of an unscented Kalman filter, of particle filters, or any similar variant by using their respective procedures and the equations given hereinabove.

The device 1 for monitoring and estimating parameters relating to the flight of an aircraft AC as described hereinabove implements a method (FIG. 2) comprising:

• • an initialization step E1, implemented by the initialization module 3, consisting in initializing statuses of sensors C1, C2, . . . , CN configured to determine the flight parameters of the aircraft AC and a parameter P1 corresponding to the weight of said aircraft AC and in initializing parameters used in the implementation of the monitoring and estimation method 1.

The method further comprises the following steps, implemented iteratively:

• • an estimation step E2, implemented by the estimation module 4, consisting in determining an estimation of the values of the parameters relating to the flight of the aircraft AC and an estimation of the error of said weight, from:
    • measurements of the parameters relating to the flight supplied by the sensors C1, C2, . . . CN,
    • parameters relating to the flight initialized in the initialization step or estimated on the preceding iteration of the estimation step E2 and
    • statuses associated with each of said sensors C1, C2, . . . , CN,
  •  the estimation step E2 consisting also in generating residues r_i which are functions of the measured and estimated values of the parameters relating to the flight and of innovation terms which correspond to the difference between a measured flight parameter value and said estimated value;
  • a first transmission step E3, implemented by the first transmission module 7, consisting:
    • in transmitting to a user device 6 and to the detection module 5 a signal representative of the estimation of the values of the parameters relating to the flight of the aircraft AC and of the weight error, determined in the estimation step E2,
    • in sending a signal representative of the residues generated in the estimation step E2 to said detection module 5;
  • a detection step E4, implemented by the detection module 5, consisting in determining the different statuses associated with each of said sensors C1, C2, . . . , CN and with a parameter P1 corresponding to a current weight of the aircraft AC, from:
    • said residues determined in the estimation step,
    • the estimation of the values of the parameters relating to the flight and the estimated weight error, determined in the estimation step E2,
    • said statuses determined on the preceding iteration or initialized in the initialization step;
  • a second transmission step E5, implemented by the second transmission module 8, consisting in transmitting to the user device 6 and, on the next iteration, to the estimation module 4 the different statuses associated with each of said sensors C1, C2, . . . , CN and the status associated with said parameter P1 corresponding to the current weight.

Furthermore, the estimation step E2 comprises the following substeps:

• • an adaptation substep E21, implemented by the adaptation submodule 41, consisting in determining a variance and/or a validity Boolean associated with each of the measurements of the parameters relating to the flight supplied by the sensors C1, C2, . . . , CN and of the setting parameters associated with the estimation algorithm used in the estimation substep E22, from:
    • said measurements of the parameters relating to the flight and
    • from the statuses associated with each of said sensors C1, C2, . . . , CN; the adaptation substep E21 consisting also in correcting the current weight from a weight error estimated on the preceding iteration or initialized in the initialization step E1, and from a status associated with the parameter corresponding to the weight,
  • an estimation substep E22, implemented by the estimation submodule 42, consisting in determining the estimation of the values of the parameters relating to the flight and an estimation of the error of said weight from:
    • the measurements of the parameters relating to the flight supplied by said sensors C1, C2, . . . , CN,
    • the parameters relating to the flight estimated on the preceding iteration or initialized in the initialization step E1 and
    • from the variance and/or the validity Boolean of each of the measurements of the parameters relating to the flight and of the setting parameters determined in the adaptation substep E21, the estimation substep E22 consisting also in generating the residues from the estimated and measured parameters relating to the flight and the innovation terms.

Furthermore, the detection step E4 comprises the following substeps:

• • a substep E41 of detection of a failing sensor and of incorrect parameters relating to the flight, implemented by a detection submodule 51, consisting in determining:
    • the common status associated with the sensor C1 configured to measure the angle of incidence α of the aircraft AC and with the parameter P1 corresponding to the current weight and
    • the status associated with the other sensors (C2, C3, . . . , CN),
  •  from:
    • the measurements of the parameters relating to the flight supplied by said sensors C1, C2, . . . , CN,
    • the estimation of the values of the parameters relating to the flight,
    • the statuses associated with each of said sensors C1, C2, . . . , CN and with a parameter P1 corresponding to a current weight of the aircraft (AC) determined on the preceding iteration or initialized in the initialization step E1, and
    • the residues;
  • a substep E42 of validation of the angle of incidence and of the current weight, implemented by a validation submodule 52, consisting in determining the status associated with the parameter P1 corresponding to the current weight and the status associated with the sensor C1 configured to measure the angle of incidence α, from:
    • the common status associated with the sensor C1 configured to measure the angle of incidence α of the aircraft AC and with the parameter P1 corresponding to the current weight,
    • the statuses associated with the other sensors C2, C3, . . . , CN,
    • the estimated parameters relating to the flight,
    • the estimated weight error,
    • the measured parameters relating to the flight,
    • the residues generated in the estimation substep E22 and
    • a lift coefficient C_L supplied from an embedded modelling fed by the parameters relating to the flight estimated and measured by the sensors C1, C2, . . . , CN.

The determination of the status of the sensors C2, . . . , CN and of the common status associated with the status of the sensor C1 and with the parameter P1 corresponding to the weight, in the failing sensor detection substep E41, comprises the following substeps:

• • a substep E411 of assignment to each residue of a maximum limit r_i⁺ and a minimum limit r_i⁻, of construction of validity indicators V_i then of residue indicators R_i associated with each residue r_i and validity indicator V_i;
  • a substep E412 of identification of failures (E R_i>0) and of isolation of the source thereof by identifying the combination of the current values of the residue indicators R_i with those itemized in a table pre-identified offline in a preferred embodiment, itemizing the failure case as a function of the different combinations of values of the residue indicators R_i. Each column of the table corresponds to a fault index FI_i and is the combination of residue indicators R₁, R₂ . . . R_m that can take 0 or 1 for value and each row corresponds to the values of a residue indicator R_i for each of the fault indexes FI₁, FI₂, . . . FI_m constructed. A fault index FI_i refers to the status of the sensor Ci, except in a first instance of the sensor measuring the incidence C1 for which the fault index FI_α/m refers to a common status of the measurement of the angle of incidence measured by an angle-of-incidence measurement sensor C1 and of the parameter P1 corresponding to the weight of the aircraft.

For a flight parameter that is consolidated and therefore measured from a plurality of sensors (FIG. 3), the method comprises the following substeps implemented by a unitary verification module 10, when a failure is detected for the measurement used as input for the estimation submodule 42 for said flight parameter:

• • a substep E61 of reconfiguration of the estimation submodule 42 to not take account of the incorrect measurement previously used as input for the estimation submodule 42,
  • a substep E62 of computation of the difference, after a convergence time τ_cvf, between the estimation of said flight parameter and of the measurement derived from the plurality of sensors, for each of the measurements derived from the plurality of sensors,if the absolute value of the difference is less than a predetermined value, the measurement being retained for the estimation module 4.

While at least one exemplary embodiment of the present invention(s) is disclosed herein, it should be understood that modifications, substitutions and alternatives may be apparent to one of ordinary skill in the art and can be made without departing from the scope of this disclosure. This disclosure is intended to cover any adaptations or variations of the exemplary embodiment(s). In addition, in this disclosure, the terms “comprise” or “comprising” do not exclude other elements or steps, the terms “a” or “one” do not exclude a plural number, and the term “or” means either or both. Furthermore, characteristics or steps which have been described may also be used in combination with other characteristics or steps and in any order unless the disclosure or context suggests otherwise. This disclosure hereby incorporates by reference the complete disclosure of any patent or application from which it claims benefit or priority.

Claims

1. A method for monitoring and estimating: parameters relating to the flight of an aircraft (AC); statuses of sensors (C1, C2, . . . , CN), the statuses being representative of an operation of said sensors (C1, C2, . . . , CN); and a status of a parameter (P1) corresponding to the current weight of the aircraft (AC), the status being representative of the validity of said parameter, the method comprising: an initialization step (E1), implemented by an initialization module (3), including initializing the statuses of sensors (C1, C2, . . . , CN) configured to determine flight parameters of the aircraft (AC) and the status of the parameter (P1) corresponding to the current weight of the aircraft (AC) and in initializing parameters used in the implementation of the monitoring and estimation device (1);

2. The method according to claim 1, wherein the estimation step (E2) comprises the following substeps: an adaptation substep (E21), implemented by an adaptation submodule (41), including determining a variance and/or a validity Boolean associated with each of the measurements of the parameters relating to the flight supplied by the sensors (C1, C2, . . . , CN) and of the setting parameters associated with the estimation algorithm used in an estimation substep (E22), from: said measurements of the parameters relating to the flight, andthe statuses associated with each of said sensors (C1, C2, . . . , CN);

3. The method according to claim 1, wherein the detection step (E4) comprises the following substeps: a substep (E41) of detection of a failing sensor and of incorrect parameters relating to the flight, implemented by a detection submodule (51), including determining: the common status associated with the sensor (C1) configured to measure the angle of incidence (a) of the aircraft (AC) and with the parameter (P1) corresponding to the current weight; andthe status associated with the other sensors (C2, C3, . . . , CN), from: the measurements of the parameters relating to the flight supplied by said sensors (C1, . . . , CN);the estimation of the values of the parameters relating to the flight and the weight error;the statuses associated with each of said sensors (C1, . . . , CN) and with the parameter (P1) corresponding to a current weight of the aircraft (AC) determined on the preceding iteration or initialized in the initialization step (E1;) andthe residues (ri);a substep (E42) of validation of the angle of incidence and of the current weight, implemented by a validation submodule (52), including determining the status associated with the parameter (P1) corresponding to the current weight and the status associated with the sensor (C1) configured to measure the angle of incidence (α), from: the common status associated with the sensor (C1) configured to measure the angle of incidence (a) of the aircraft (AC) and with the parameter (P1) corresponding to the current weight;the statuses associated with the other sensors (C2, C3, . . . , CN);the estimated parameters relating to the flight;the estimated weight error;the measured parameters relating to the flight;the residues generated in the estimation substep (E22); anda lift coefficient (CL) supplied from an embedded modelling fed by the parameters relating to the flight estimated and measured by the sensors (C1, C2, . . . CN).

4. The method according to claim 2, wherein the estimation substep (E22) corresponds to an extended Kalman filter associated with a state vector (X), an observation vector (Y) and an auxiliary measurement vector (Z), the auxiliary measurement vector (Z) having for expression:

5. The method according to claim 4, wherein the observation vector (Y) has for expression:

6. The method according to claim 4, wherein the estimation substep (E22) corresponds to an extended Kalman filter associated with a state vector (X) and an observation vector (Y) and an auxiliary measurement vector (Z), the auxiliary measurement vector (Z) having for expression:

7. The method according to claim 6, wherein the lateral component v of the air speed V is assumed nil.

8. The method according to claim 3, wherein the determination of the common status and of the status associated with the other sensors of the failing sensor detection substep (E41) comprises the following substeps: a substep (E411) of assignment to each residue of a maximum limit (ri+) and a minimum limit (ri−) from the estimated flight parameters;the substep (E411) being also a substep of construction of a first validity indicator (VPS) associated with the static pressure sensor and of a second validity indicator (VPT) associated with the total pressure sensor from the estimated parameters and the measurements derived from the sensors of the aircraft: the first validity indicator (VPS) having the value 1 if the relationship |h(s)(zPm−zGm)|≤lPS is borne out,the first validity indicator (VPS) having the value 0 if the relationship |h(s)(zPm−zGm)|≤lPS is not borne out,in which lPS corresponds to a limit determined experimentally as a function of the dynamics of the aircraft (AC), h corresponding to the transfer function of a high-pass filter,the second validity indicator (VPT) having the value 0 by default and 1 if the residue r1 crosses its respective limits, at an instant tk, determined later and that nϵ such that tnϵ[tk−τPT, tk]**bearing out |(1−τ)(MPT(tn)−MPT(tn-j))|≤lPT, with τPT a time constant, τ a transfer function of a low-pass filter having for time constant τ, jϵ a constant, and lPT a limit determined experimentally as a function of the dynamics of the aeroplane;residue indicators (Ri) being computed and associated with each residue (rj) and validity indicators (Vk);a substep (E412) of identification of the occurrence of failures when the sum of the residue indicators (ΣRi) is strictly greater than zero and identification, in case of failure, of the faulty sensor by comparing the current values of the residue indicators to those itemized in a table identified offline and online itemizing the failure cases as a function of the different combinations of values of the residue indicators (Ri).

9. The method according to claim 1, wherein the statuses associated with each of said sensors (C1, C2, . . . , CN) are also determined from auxiliary statuses associated with each of said sensors that can be sent to the detection module (5) by a monitoring module (9).

10. The method according to claim 2, wherein, for a measured consolidated flight parameter from a plurality of sensors, the method comprises the following substeps implemented by a unitary verification module (10), when a failure is detected for a sensor measuring said flight parameter for which said measurement is used as input for the estimation module (4): a substep (E61) of reconfiguration of the estimation submodule to not take account of the incorrect measurement of the flight parameter used hitherto in the estimation substep (E22),a substep (E62) of computation of the difference between the estimation of said flight parameter and of the measurement of one of said sensors derived from the plurality of sensors, for each of the measurements derived from the plurality of sensors,if the absolute value of the difference is less than a predetermined value, the measurement being retained in the estimation substep (E22).

11. A device for monitoring and estimating: parameters relating to the flight of an aircraft (AC); statuses of sensors, the statutes being representative of an operation of said sensors (C1, C2, . . . , CN); and a status of a parameter (P1) corresponding to the current weight of the aircraft (AC), the status being representative of the validity of said parameter (P1), wherein the device comprises:an initialization module (3), configured to initialize the statuses of sensors (C1, C2, . . . , CN) configured to determine parameters relating to the flight of the aircraft (AC) and the status of the parameter (P1) corresponding to the current weight of the aircraft and to initialize parameters used in the implementation of the monitoring and estimation device (1);

12. The device according to claim 11, wherein the estimation module (4) comprises: an adaptation submodule (41) configured to determine a variance and/or a validity Boolean associated with each of the measurements of the parameters relating to the flight supplied by the sensors (C1, C2, . . . , CN) and of the setting parameters associated with the estimation algorithm used in an estimation submodule (42), from: said measurements of the parameters relating to the flight andthe statuses associated with each of said sensors (C1, C2, . . . , CN);wherein the adaptation submodule (41) is also configured to correct the current weight of the aircraft (AC) from a weight error estimated on the preceding iteration or initialized by the initialization module (3), and from a status associated with the parameter (P1) corresponding to the weight,wherein the estimation submodule (42) is configured to determine the estimation of the values of the parameters relating to the flight and an estimation of the error of said weight, from: the measurements of the parameters relating to the flight supplied by said sensors (C1, C2, . . . , CN),the parameters relating to the flight estimated on the preceding iteration or initialized in the initialization step (E1), andthe variance and/or the validity Boolean of each of the measurements of the parameters relating to the flight and of the setting parameters determined in the adaptation substep (E21),wherein the estimation submodule (42) is also configured to generate residues from the estimated and measured parameters relating to the flight and from the innovation terms.

13. The device according to claim 11, wherein the detection module (5) comprises: a detection submodule (51) configured to determine: the common status associated with the sensor (C1) configured to measure the angle of incidence (α) of the aircraft (AC) and with the parameter (P1) corresponding to the current weight, andthe status associated with the other sensors (C2, C3, . . . , CN), from: the measurements of the parameters relating to the flight supplied by said sensors (C2, C3, . . . , CN),the estimation of the values of the parameters relating to the flight and the weight error,the statuses associated with each of said sensors (C2, C3, . . . , CN) and with a parameter (P1) corresponding to a current weight of the aircraft (AC) determined on the preceding iteration or initialized by the initialization module (3),the residues; anda validation submodule (52) configured to determine the status associated with the parameter (P1) corresponding to the current weight and the status associated with the sensor (C1) configured to measure the angle of incidence (α), from: the common status associated with the sensor (C1) configured to measure the angle of incidence (α) of the aircraft (AC) and with the parameter (P1) corresponding to the current weight,the statuses associated with the other sensors (C2, C3, . . . , CN),the estimated parameters relating to the flight,the estimated weight error,the measured parameters relating to the flight,the residues generated in the estimation substep (E22), anda lift coefficient (CL) supplied from an embedded modelling fed by the parameters relating to the flight estimated and measured by said sensors (C1, C2, . . . , CN).

14. An aircraft comprising a device (1) for monitoring and estimating flight parameters of an aircraft (AC), according to claim 11.