CN110244697B - Complex fault diagnosis and identification method for vertical take-off and landing aircraft based on composite observer - Google Patents

Complex fault diagnosis and identification method for vertical take-off and landing aircraft based on composite observer Download PDF

Info

Publication number
CN110244697B
CN110244697B CN201910561695.3A CN201910561695A CN110244697B CN 110244697 B CN110244697 B CN 110244697B CN 201910561695 A CN201910561695 A CN 201910561695A CN 110244697 B CN110244697 B CN 110244697B
Authority
CN
China
Prior art keywords
fault
observer
engine
diagnosis
channel
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201910561695.3A
Other languages
Chinese (zh)
Other versions
CN110244697A (en
Inventor
韦常柱
崔乃刚
李源
陈嘉凯
关英姿
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Harbin Institute of Technology
Original Assignee
Harbin Institute of Technology
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Harbin Institute of Technology filed Critical Harbin Institute of Technology
Priority to CN201910561695.3A priority Critical patent/CN110244697B/en
Publication of CN110244697A publication Critical patent/CN110244697A/en
Application granted granted Critical
Publication of CN110244697B publication Critical patent/CN110244697B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B23/00—Testing or monitoring of control systems or parts thereof
    • G05B23/02—Electric testing or monitoring
    • G05B23/0205—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults
    • G05B23/0259—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults characterized by the response to fault detection
    • G05B23/0262—Confirmation of fault detection, e.g. extra checks to confirm that a failure has indeed occurred
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B2219/00—Program-control systems
    • G05B2219/20—Pc systems
    • G05B2219/24—Pc safety
    • G05B2219/24065—Real time diagnostics

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • Automation & Control Theory (AREA)
  • Feedback Control In General (AREA)

Abstract

The invention discloses a complex fault diagnosis and identification method of a vertical take-off and landing aircraft based on a composite observer, which comprises the following steps: step one, establishing a system fault state equation; establishing a detection observer to quickly realize fault judgment, namely judging whether the system has faults or not; establishing a group of single-channel diagnosis observers and multi-channel coupling separation observers, preliminarily extracting fault information by using the single-channel diagnosis observers, and realizing fault positioning and accurate diagnosis by using the multi-channel coupling separation observers; and step four, rapidly diagnosing a fault mode and fault information based on the observation results of the step two and the step three. According to the method, fault identification can be rapidly and accurately realized only by using the attitude angular velocity information of the vertical take-off and landing aircraft, and a foundation is laid for control reconstruction and task reconstruction design.

Description

Complex fault diagnosis and identification method for vertical take-off and landing aircraft based on composite observer
Technical Field
The invention belongs to the technical field of aircraft control, and relates to a method for diagnosing and identifying a composite fault of a power system and a servo system of a vertical take-off and landing aircraft.
Background
The vertical take-off and landing aircraft is a high-precision complex aircraft with a fast ballistic change, and demands for high precision and high stability are provided for a power system and a servo system. However, the flying environment of the vertical take-off and landing aircraft is complex, the probability of faults occurring in the task load of a power system and a servo system is high, and the control performance, the flying stability performance, the flying capacity and the like of the vertical take-off and landing aircraft are seriously influenced. The servo mechanism power of the vertical take-off and landing aircraft is often provided by a power system, so that the servo mechanism fault is often accompanied with the synchronous occurrence of the power system fault, namely, when the vertical take-off and landing aircraft causes the power system fault due to the turbine pump fault, the thrust chamber fault or the pipeline and valve fault, the power-generating thrust is often caused to descend, and the corresponding servo mechanism is stuck or loosened and floated. According to the control mechanism, the fault modes between the power system and the servo system are seriously coupled, the synchronous realization of the accurate fault diagnosis and identification under the complex fault mode is difficult to realize, the failure of the accurate diagnosis and identification under the complex fault mode can seriously affect the flight reliability of the vertical take-off and landing aircraft, and the development of the vertical take-off and landing aircraft in China is greatly affected.
Disclosure of Invention
Aiming at the problems, the invention provides a complex fault diagnosis and identification method for a vertical take-off and landing aircraft based on a composite observer. According to the method, fault identification can be rapidly and accurately realized only by using the attitude angular velocity information of the vertical take-off and landing aircraft, and a foundation is laid for control reconstruction and task reconstruction design.
The purpose of the invention is realized by the following technical scheme:
a complex fault diagnosis and identification method for a vertical take-off and landing aircraft based on a composite observer comprises the following steps:
step one, establishing a system fault state equation:
Figure GDA0002431552180000021
in the formula:
x is the state vector:
x＝[ωxωyωz]T；
ωx,ωy,ωzthe actual attitude angular velocity of the vertical take-off and landing aircraft;
y represents the output of the system:
y＝[ωxωyωz]T；
can be directly output by an inertial navigation system;
c is the output matrix of the system, including the measurement error of the system;
a is a state transition matrix of the system:
Figure GDA0002431552180000022
q is dynamic pressure, V is aircraft flight speed, SmIs a reference area, l is a reference length,
Figure GDA0002431552180000023
is the three-axis moment of inertia of the aircraft,
Figure GDA0002431552180000024
three channel damping coefficients for the aircraft;
d (t) is the perturbation vector of the system:
Figure GDA0002431552180000031
mRmass of a single engine, JRFor moment of inertia about the hinge axis of the engine,/RIs the distance from the center of mass of the engine to the hinge axis,
Figure GDA0002431552180000032
δψ,δγthree-channel engine equivalent swing angle, xRAs engine hinge position, xTAs the engine centroid position, MBX,MBY,MBZFor structural disturbance of moment, MKY,MKZAn engine disturbance torque;
Biindicating to control the input matrix BδThe matrix with zero set in the ith column;
u＝[δ1δ2δ3δ4]Trepresenting the swing angles of four engines;
birepresenting a control input matrix BδThe ith column;
pkui＝kiδki；
δkirepresenting the fault angle of the servo mechanism i;
kirepresenting the thrust loss coefficient of the ith single-engine of the power system;
Figure GDA0002431552180000033
representing the disturbance moment caused by the asymmetry of the thrust layout caused by the thrust loss of the engine I,
Figure GDA0002431552180000034
r is the engine thrust to axis distance;
Tithrust of the ith single-engine;
establishing a detection observer to quickly realize fault judgment, namely judging whether the system has faults or not; the method comprises the following specific steps:
designing a detection observer, and determining whether the system is in fault or not by using a residual signal generated by the detection observer, wherein the observation form is as follows:
Figure GDA0002431552180000041
in the formula: z (t) is the state vector of the Luenberger detection observer, r (t) is the fault detection residual vector, uδ(t) represents a servo commanded swing angle;
introducing an error vector e (t):
e(t)＝z(t)-Gx(t)；
observer matrices F, G, L, M and H satisfy the following condition:
LC+FG-GA＝0；
MG+HC＝0；
Re(λ(F))＜0；
MGEd＝0；
in the formula: λ (F) is recorded as the eigenvalue of the matrix F, Re (-) represents the real part of the variable;
when the system is fault-free, the theoretical input of the system is equal to the actual input, and at the moment
Figure GDA0002431552180000042
The residual vector satisfies:
Figure GDA0002431552180000043
when the system has a fault, the theoretical input and the actual input of the system are not equal, and at the moment, the error equation and the residual error equation are expressed as follows:
Figure GDA0002431552180000051
the system is not able to converge, i.e.
Figure GDA0002431552180000052
The system diverges;
primarily completing fault diagnosis according to a residual signal output by a detection observer, and judging that a system has a fault when the residual signal is greater than a set threshold value;
establishing a group of single-channel diagnosis observers and multi-channel coupling separation observers, preliminarily extracting fault information by using the single-channel diagnosis observers, and realizing fault positioning and accurate diagnosis by using the multi-channel coupling separation observers; the method comprises the following specific steps:
(1) 4 single-channel diagnosis observers are established by utilizing the rolling channel and correspond to four different engines, and the form of the single-channel diagnosis observer is as follows:
Figure GDA0002431552180000053
Bxmindicating to control the input matrix BxThe m-th column of (a) is set to zero; bmAn mth column representing a control data matrix;
Figure GDA0002431552180000054
F. g, L, M and H are both parameters to be designed;
when the system has faults, the parameter p is realized through the single-channel diagnosis observerkui(ii) an estimate of (d);
(2) establishing 4 groups of multi-channel coupling separation observers in the following form, respectively corresponding to complex fault modes of four engines, and outputting corresponding residual signals:
Figure GDA0002431552180000061
Bma control matrix representing the zeroth column of the control input matrix B; bmAn mth column representing a control data matrix; p is a radical ofkum＝kmδkmThe single-channel diagnosis observer can be used for accurate estimation;
Figure GDA0002431552180000062
to represent
Figure GDA0002431552180000063
An estimated value of, and
Figure GDA0002431552180000064
is measured by
Figure GDA0002431552180000065
Determining that F, G, L, M and H are both matrices to be designed;
when the system has a fault in the form of step one, only the m (i) th observer in the 4 separation filters has the same control input structure with the actual fault, and when the fault occurs in the form of step one, the control input structure is the same as that of the actual fault
Figure GDA0002431552180000066
Converge to pkuiAnd is
Figure GDA0002431552180000067
Converge to kiIts residual signal rm(t) | | will approach zero, fault information kiBy online adjustment using the adaptive law
Figure GDA0002431552180000068
Estimated that the other observer output residuals diverge, i.e.
Figure GDA0002431552180000069
Step four, rapidly diagnosing a fault mode and fault information based on the observation results of the step two and the step three; the method comprises the following specific steps:
judging the failure mode of the servo mechanism:
setting the desired swing angle output by the control system to δkiThe estimated value of the swing angle is
Figure GDA00024315521800000610
The deviation of the swing angle can be expressed as
Figure GDA00024315521800000611
Setting a given threshold value delta0；
(1) Without failure
When the swing angle difference delta is smaller than a given threshold value delta0Judging that the servo mechanism has no fault;
(2) fail to work
When the swing angle difference delta is larger than the loosening threshold delta0And the swing angle of the fault servo mechanism changes along with the change of time, and then the fault mode at the moment is judged to be the failure fault of the engine servo mechanism;
(3) blocking is dead
When the swing angle difference delta is larger than the loosening threshold delta0If the swing angle of the failed engine does not change along with the change of time, judging that the failure mode is failure of the engine servo mechanism;
and (3) judging whether the power system has a fault:
(1) without failure
Coefficient of thrust loss
Figure GDA0002431552180000071
When the estimated value is close to 1, judging that the power system has no fault;
(2) partial failure fault
Coefficient of thrust loss
Figure GDA0002431552180000072
When the estimated value is less than 1, the power system is judged to be in fault, and the thrust loss coefficient is
Figure GDA0002431552180000073
Compared with the prior art, the invention has the following advantages:
according to the invention, fault diagnosis and identification under a complex fault mode can be realized only by utilizing the angular velocity information observation quantity of the vertical take-off and landing aircraft, the use information is less, a sensor is not required to be additionally designed, and the cost can be reduced; the faults are diagnosed and separated through a plurality of observers, and the diagnosis result is accurate; the method realizes fault diagnosis under the condition of considering the thrust loss and the servo mechanism coupling, and has innovativeness.
Drawings
FIG. 1 shows an engine pivot angle relationship.
Detailed Description
The technical solution of the present invention is further described below with reference to the accompanying drawings, but not limited thereto, and any modification or equivalent replacement of the technical solution of the present invention without departing from the spirit and scope of the technical solution of the present invention shall be covered by the protection scope of the present invention.
The invention provides a complex fault diagnosis and identification method of a vertical take-off and landing aircraft based on a composite observer, aiming at the design requirement of rapid diagnosis of complex faults of the vertical take-off and landing aircraft. Firstly, establishing a detection observer to quickly realize fault judgment, namely judging whether a system has a fault; then a group of single-channel diagnosis observers and multi-channel coupling separation observers are established, fault information is preliminarily extracted by the single-channel diagnosis observers, fault positioning and accurate diagnosis are realized by the multi-channel coupling separation observers, and finally, fault modes and fault information are quickly diagnosed based on observation results. The specific implementation steps are as follows:
1. system equation of state establishment
The vertical take-off and landing aircraft rotation dynamics equation around the center of mass can be expressed as:
Figure GDA0002431552180000081
Figure GDA0002431552180000082
the layout of the power system of the vertical take-off and landing aircraft is mostly in the form shown in figure 1.
The above dynamic model around the centroid can be abbreviated as:
Figure GDA0002431552180000083
in the formula (I), the compound is shown in the specification,
x is the state vector:
x＝[ωxωyωz]T；
y represents the output of the system:
y＝[ωxωyωz]T；
c is the output matrix of the system, including the measurement error of the system;
a is a state transition matrix of the system:
Figure GDA0002431552180000091
d (t) is the perturbation vector of the system:
Figure GDA0002431552180000092
bu (t) is the control input of the system, with the servo swing angle as the control input vector, Bu (t) can be expressed as Bδuδ(t)，uδ＝[δ1δ2δ3δ4]TThe control matrix B here has the following form:
Bδ＝[M1M2M3M4]；
Figure GDA0002431552180000093
Figure GDA0002431552180000094
when the system has complex faults, namely the faults of the servo mechanism and the thrust loss of the power system occur on the same engine at the same time, the ith servo mechanism is set to have faults, and the fault angle is deltakiCoefficient of thrust loss of kiThen the system state equation can be expressed as:
Figure GDA0002431552180000101
in the formula:
Biindicating to control the input matrix BδThe matrix with zero set in the ith column;
birepresenting a control input matrix BδThe ith column;
δkirepresenting the fault angle of the servo mechanism i;
kirepresenting a thrust loss coefficient of the power system;
Figure GDA0002431552180000102
representing the disturbance moment caused by the asymmetry of the thrust layout caused by the thrust loss of the engine I,
Figure GDA0002431552180000103
when the No. 1 engine has thrust loss, k2＝k3＝k41, then
Figure GDA0002431552180000104
Can be expressed as:
Figure GDA0002431552180000105
when the No. 2 engine has thrust loss, k1＝k3＝k41, then
Figure GDA0002431552180000106
Can be expressed as:
Figure GDA0002431552180000107
when the No. 3 engine has thrust loss, k1＝k2＝k41, then
Figure GDA0002431552180000108
Can be expressed as:
Figure GDA0002431552180000111
when the No. 4 engine has thrust loss, k1＝k2＝k31, then
Figure GDA0002431552180000112
Can be expressed as:
Figure GDA0002431552180000113
if the No. 1 engine has thrust loss fault and servo mechanism jamming fault at the same time, the interference moment M caused by the faultz1Can be expressed as:
Figure GDA0002431552180000114
by Mz1The form of the interference torque is known, the component of the interference torque on the x axis is caused by the fault angle of the servo mechanism, the components of the interference torque on the y axis and the z axis are caused by the clamping fault of the servo mechanism and the thrust loss fault together, meanwhile, the form of the interference torque caused by the clamping fault of the servo mechanism is known, the interference torque is influenced by the fault angle of the servo mechanism and the thrust loss coefficient together, and the parameter k of the thrust loss coefficient is visible1Angle delta to servo failurek1Is a parameter pku1Convergence estimation is carried out, the influence of the product coupling of the dead angle of the servo mechanism and the thrust loss coefficient on the parameter convergence speed and precision is reduced, and at the moment, the parameter to be estimated is changed into pkuiAnd kiThe state equation of the system can also be expressed as:
Figure GDA0002431552180000115
reduce thrust loss and simultaneous output of servo mechanismNow, the parameter coupling influence on the same engine only needs to be optimized by the parameter pkuAnd k1Namely, simultaneously using the formula
pkui＝kiδki(6)
The failure angle of the servo mechanism can be calculated.
From the above formula, the component of the disturbance torque in the x-axis is only equal to pkuiRelated to the interference moment, and the components of the interference moment in the y-axis and the z-axis are represented by pkuiAnd k1The interaction results, and therefore the quantity of state x (t) that can be taken is the roll angular velocity of the liquid rocket, at which point the equation of state in the roll channel can be expressed as:
Figure GDA0002431552180000121
in the formula: the system state quantity x (t) is the roll angular velocity omegax；
Figure GDA0002431552180000122
d (t) represents the effect of the perturbation; b isx＝[-T1r -T2r -T3r -T4r]。
The system state equation in the roll channel in the complex fault mode can be expressed as:
Figure GDA0002431552180000123
in the formula: b isxiRepresents the matrix BxThe ith column of (1) is set to zero, bxiRepresentation matrix BxiThe ith element of (1).
2. Detection observer design
Designing a detection observer, and determining whether the system is in fault or not by using a residual signal generated by the detection observer, wherein the observation form is as follows:
Figure GDA0002431552180000124
in the formula: z (t) is the state vector of the Luenberger detection observer, and r (t) isFault detection residual vector, uδ(t) represents the servomechanism command pivot angle, and both matrices F, G, L, M and H are matrices to be designed. The matrix G is designed to eliminate the influence of the disturbance vector on the identification precision of the observer, and the specific form is given in equation (14).
The observer tracks the state quantity Gx (t) by using z (t), when the system has a fault, the actual input of the carrier rocket system deviates from the input of the observer, the observer cannot track the state quantity Gx (t), and the output residual quantity r (t) cannot be converged, namely
Figure GDA0002431552180000131
At this time, the system is diagnosed as malfunctioning.
Introducing an error vector e (t):
e(t)＝z(t)-Gx(t) (10)。
the observer matrices F, G, L, M and H are designed to satisfy the following condition:
LC+FG-GA＝0 (11)；
MG+HC＝0 (12)；
Re(λ(F))＜0 (13)；
MGEd＝0 (14)。
in the formula: λ (F) is noted as the eigenvalue of the matrix F, and Re (-) represents the real part of the variable.
When the system fails, the error equation and the residual equation can be expressed as follows:
Figure GDA0002431552180000132
in the formula: b isδ0Representing a control input matrix after a fault; u. ofδ0Indicating the control input after the fault.
Applying equations (11) to (14), the error equation and the residual equation can be expressed as:
Figure GDA0002431552180000133
when the system is fault-free, the theoretical input of the system is equal to the actual input, and at the moment
Figure GDA0002431552180000141
The residual vector satisfies:
Figure GDA0002431552180000142
when the system fails, the theoretical input and the actual input of the system are not equal, and the system can not be converged, i.e. the system fails to work
Figure GDA0002431552180000143
The system diverges.
The fault diagnosis can be preliminarily finished according to the residual signal output by the detection observer, and when the residual signal is greater than a set threshold value, the system is judged to have faults. And after the system fault is judged, activating the preset design of a single-channel diagnosis observer and a multi-channel coupling separation observer to realize the diagnosis and identification of the complex fault.
3. Single channel diagnostic observer design
The system state equation in the roll channel can be expressed as:
Figure GDA0002431552180000144
in the formula: b isx＝[-T1r -T2r -T3r -T4r]，BxiRepresents the matrix BxThe ith column of (1) is set to zero, bxiRepresentation matrix BxiThe ith element of (1).
To obtain the coefficient pkuiThe estimated value of (1) is to establish 4 single-channel diagnosis observers by utilizing a rolling channel, and the single-channel diagnosis observers correspond to four different engines, and have the following forms:
Figure GDA0002431552180000145
the observer establishes a state input matrix of BxmDenotes that the control is input to the matrix BxThe m-th column of (a) is set to zero; bmAn mth column representing a control data matrix;
Figure GDA0002431552180000151
F. g, L, M and H are both parameters to be designed (equation (11) -equation (14)).
Design of self-adaptation law in single-channel diagnosis observer
Figure GDA0002431552180000152
To pair
Figure GDA0002431552180000153
Adaptive adjustment is performed, ηm＝Mbm，ρmFor adaptive parameter, determine
Figure GDA0002431552180000154
The convergence speed of (2). When in use
Figure GDA0002431552180000155
Estimated value and actual
Figure GDA0002431552180000156
When different, the system outputs residual rm(t) ≠ 0, in which case the adaptation law (equation (22)) uses the system output residual
Figure GDA0002431552180000157
To pair
Figure GDA0002431552180000158
Carrying out self-adaptive adjustment; up to
Figure GDA0002431552180000159
Converge to pkuiAt this time, the system outputs a residual rm(t) will also converge to zero.
From the system input matrix Bx＝[-T1r -T2r -T3r -T4r]The 1 st, 2 nd, 3 rd and 4 th columns of the matrix have the same form, that is, four observers in the roll channel have the same form, and after the system has a fault, the four observers can all realize the parameter p through the observer (formula (21))kuiIs estimated.
4. Design of multichannel coupling separation observer
In order to realize accurate estimation of the thrust loss coefficient and fault location, the following 4 groups of multi-channel coupling separation observers are established, which respectively correspond to complex fault modes of four engines and are used for outputting corresponding residual signals:
Figure GDA00024315521800001510
compared to the fault detection filter (equation (9)), the state input matrix established by the split observer is BmIndicating a control matrix that zeroes the mth column of the control input matrix B; bmAn mth column representing a control data matrix; p is a radical ofkum＝kmδkmAccurate estimation can be performed by an observer (equation (21));
Figure GDA0002431552180000161
to represent
Figure GDA0002431552180000162
An estimated value of, and
Figure GDA0002431552180000163
is measured by
Figure GDA0002431552180000164
And (6) determining. F. G, L, M and H are both matrices to be designed (equation (11) -equation (14)).
The separation observer is designed with self-adaptive law
Figure GDA0002431552180000165
Coefficient of thrust loss to power system
Figure GDA0002431552180000166
Performing adaptive adjustment, wherein ηn＝M(MTM)-1PG, P are
Figure GDA0002431552180000167
A symmetric positive definite matrix. RhonDetermining the estimated value of the thrust loss coefficient of the power system for the adaptive parameter
Figure GDA0002431552180000168
The convergence speed of (2). When in use
Figure GDA0002431552180000169
Does not converge on pkuiOr
Figure GDA00024315521800001610
Does not converge on kiTime, system output residual rm(t) ≠ 0, in which case the adaptation law (equation (24)) uses the system output residual
Figure GDA00024315521800001611
To pair
Figure GDA00024315521800001612
Carrying out self-adaptive adjustment; up to
Figure GDA00024315521800001613
Converge on kiAt this time, the system outputs a residual rm(t) will also converge to zero.
The systematic error equation and the residual equation can be expressed as follows:
Figure GDA00024315521800001614
match with the actual fault modelIn the matched observer Bi、biAnd Bm、bmIn the same way, the first and second,
Figure GDA00024315521800001615
and
Figure GDA00024315521800001616
having the same structure, definition
Figure GDA00024315521800001617
The error equation and residual equation can be expressed as:
Figure GDA00024315521800001618
when the system has a fault in the form of formula (5), only the m (i) th observer in the 4 separation filters has the same control input structure as the actual fault, namely Bi、biAnd Bm、bmIn the same way, the first and second,
Figure GDA0002431552180000171
and
Figure GDA0002431552180000172
has the same structure as that of
Figure GDA0002431552180000173
Converge to pkuiAnd is
Figure GDA0002431552180000174
Converge to kiIts residual signal rm(t) | | will approach zero, fault information kiIt can also be adjusted on-line by using the adaptive law (equation (24))
Figure GDA0002431552180000175
And (4) estimating. While the other observer output residuals diverge, i.e.
Figure GDA0002431552180000176
The above schemes are combined to benefitFault separation is realized by system residual signal and on-line adjustment is realized by self-adaptive law
Figure GDA0002431552180000177
And
Figure GDA0002431552180000178
and realizing the estimation of the fault.
5. Failure mode determination
The method can obtain the estimated value of the actual swing angle of the servo mechanism, and in order to realize the fault mode judgment, the swing angle and the estimated value of the swing angle of the fault servo mechanism need to be output by a control system, and finally the fault mode is determined through threshold logic judgment. Setting the desired swing angle output by the control system to δcThe estimated value of the swing angle is deltafThen the yaw angle deviation can be expressed as delta-deltac-δf。
(1) Without failure
And when the swing angle difference is smaller than a given threshold value, judging that the servo mechanism is not in fault.
(2) Fail to work
And when the swing angle difference delta is larger than the loosening and floating threshold value and the swing angle of the fault servo mechanism changes along with the change of time, judging that the fault mode is the failure fault of the engine servo mechanism at the moment.
(3) Blocking is dead
And when the swing angle difference delta is larger than the loosening and floating threshold value and the swing angle of the fault engine does not change along with the change of time, judging that the fault mode is the failure fault of the engine servo mechanism at the moment.
The failure mode of the servo mechanism of the system can be judged through the logic.
Whether the power system is in failure can be judged through the estimated value of the thrust loss coefficient:
(1) without failure
And when the estimated value of the thrust loss coefficient is close to 1, judging that the power system has no fault.
(2) Partial failure fault
When the estimated value of the thrust loss coefficient is less than 1, judging that the power system has a fault, and determining that the thrust loss coefficient is
Figure GDA0002431552180000181

Claims (2)

1. A complex fault diagnosis and identification method for a vertical take-off and landing aircraft based on a composite observer is characterized by comprising the following steps:
step one, establishing a system fault state equation, wherein the system fault state equation is as follows:
Figure FDA0002431552170000011
in the formula:
x is the state vector:
x＝[ωxωyωz]T；
ωx,ωy,ωzthe actual attitude angular velocity of the vertical take-off and landing aircraft;
y represents the output of the system:
y＝[ωxωyωz]T；
c is the output matrix of the system, including the measurement error of the system;
a is a state transition matrix of the system:
Figure FDA0002431552170000012
q is dynamic pressure, V is aircraft flight speed, SmIs a reference area, l is a reference length,
Figure FDA0002431552170000013
is the three-axis moment of inertia of the aircraft,
Figure FDA0002431552170000014
three channel damping coefficients for the aircraft;
d (t) is the perturbation vector of the system:
Figure FDA0002431552170000021
mRmass of a single engine, JRFor moment of inertia about the hinge axis of the engine,/RIs the distance from the center of mass of the engine to the hinge axis,
Figure FDA0002431552170000022
δψ,δγthree-channel engine equivalent swing angle, xRAs engine hinge position, xTAs the engine centroid position, MBX,MBY,MBZFor structural disturbance of moment, MKY,MKZAn engine disturbance torque;
Biindicating to control the input matrix BδThe matrix with zero set in the ith column;
u＝[δ1δ2δ3δ4]Trepresenting the swing angles of four engines;
birepresenting a control input matrix BδThe ith column;
pkui＝kiδki；
δkirepresenting the fault angle of the servo mechanism i;
kirepresenting the thrust loss coefficient of the ith single-engine of the power system;
Figure FDA0002431552170000023
representing the disturbance moment caused by the asymmetry of the thrust layout caused by the thrust loss of the engine I,
Figure FDA0002431552170000024
r is the engine thrust to axis distance;
Tithrust of the ith single-engine;
step two, establishing a detection observer to quickly realize fault judgment, namely judging whether the system has faults or not, and specifically comprising the following steps:
designing a detection observer, and determining whether the system is in fault or not by using a residual signal generated by the detection observer, wherein the observation form is as follows:
Figure FDA0002431552170000031
in the formula: z (t) is the state vector of the Luenberger detection observer, r (t) is the fault detection residual vector, uδ(t) represents a servo commanded swing angle;
introducing an error vector e (t):
e(t)＝z(t)-Gx(t)；
observer matrices F, G, L, M and H satisfy the following condition:
LC+FG-GA＝0；
MG+HC＝0；
Re(λ(F))＜0；
MGEd＝0；
in the formula: λ (F) is recorded as the eigenvalue of the matrix F, Re (-) represents the real part of the variable;
when the system is fault-free, the theoretical input of the system is equal to the actual input, and at the moment
Figure FDA0002431552170000032
The residual vector satisfies:
Figure FDA0002431552170000033
when the system has a fault, the theoretical input and the actual input of the system are not equal, and at the moment, the error equation and the residual error equation are expressed as follows:
Figure FDA0002431552170000034
the system is not able to converge, i.e.
Figure FDA0002431552170000041
The system diverges;
primarily completing fault diagnosis according to a residual signal output by a detection observer, and judging that a system has a fault when the residual signal is greater than a set threshold value;
establishing a group of single-channel diagnosis observers and multi-channel coupling separation observers, preliminarily extracting fault information by using the single-channel diagnosis observers, and realizing fault positioning and accurate diagnosis by using the multi-channel coupling separation observers, wherein the specific steps are as follows:
(1) 4 single-channel diagnosis observers are established by utilizing the rolling channel and correspond to four different engines, and the form of the single-channel diagnosis observer is as follows:
Figure FDA0002431552170000042
Bxmindicating to control the input matrix BxThe m-th column of (a) is set to zero; bmAn mth column representing a control data matrix;
Figure FDA0002431552170000043
F. g, L, M and H are both parameters to be designed; r ism(t) is the system output residual ηm＝Mbm，ρmIs an adaptive parameter;
when the system has faults, the parameter p is realized through the single-channel diagnosis observerkui(ii) an estimate of (d);
(2) establishing 4 groups of multi-channel coupling separation observers in the following form, respectively corresponding to complex fault modes of four engines, and outputting corresponding residual signals:
Figure FDA0002431552170000044
Bma control matrix representing the zeroth column of the control input matrix B; bmAn mth column representing a control data matrix; p is a radical ofkum＝kmδkmThe single-channel diagnosis observer can be used for accurate estimation;
Figure FDA0002431552170000051
to represent
Figure FDA0002431552170000052
An estimated value of, and
Figure FDA0002431552170000053
is measured by
Figure FDA0002431552170000054
Determining that F, G, L, M and H are both matrices to be designed;
Figure FDA0002431552170000055
the estimated value of the thrust loss coefficient of the power system is obtained; rhonFor adaptive parameters, ηn＝M(MTM)-1PG, P are
Figure FDA0002431552170000056
A symmetric positive definite matrix of (a);
when the system has a fault in the form of step one, only the m (i) th observer in the 4 separation filters has the same control input structure with the actual fault, and when the fault occurs in the form of step one, the control input structure is the same as that of the actual fault
Figure FDA0002431552170000057
Converge to pkuiAnd is
Figure FDA0002431552170000058
Converge to kiIts residual signal rm(t) | | will approach zero, fault information kiBy online adjustment using the adaptive law
Figure FDA0002431552170000059
Is estimated to beIt observes the observer output residual divergence, i.e.
Figure FDA00024315521700000510
And step four, rapidly diagnosing a fault mode and fault information based on the observation results of the step two and the step three.
2. The complex fault diagnosis and identification method for the VTOL aerial vehicle based on the composite observer of claim 1, wherein the detailed steps of the fourth step are as follows:
judging the failure mode of the servo mechanism:
setting the desired swing angle output by the control system to δkiThe estimated value of the swing angle is
Figure FDA00024315521700000511
The deviation of the swing angle can be expressed as
Figure FDA00024315521700000512
Setting a given threshold value delta0；
(1) Without failure
When deviation of swing angle
Figure FDA00024315521700000513
Less than a given threshold delta0Judging that the servo mechanism has no fault;
(2) fail to work
When deviation of swing angle
Figure FDA00024315521700000514
Greater than a given threshold delta0And the swing angle of the fault servo mechanism changes along with the change of time, and then the fault mode at the moment is judged to be the failure fault of the engine servo mechanism;
(3) blocking is dead
When deviation of swing angle
Figure FDA0002431552170000061
Greater than a given threshold value Δδ0If the swing angle of the failed engine does not change along with the change of time, judging that the failure mode is failure of the engine servo mechanism;
and (3) judging whether the power system has a fault:
(1) without failure
Coefficient of thrust loss
Figure FDA0002431552170000062
When the estimated value is close to 1, judging that the power system has no fault;
(2) partial failure fault
Coefficient of thrust loss
Figure FDA0002431552170000063
When the estimated value is less than 1, the power system is judged to be in fault, and the thrust loss coefficient is
Figure FDA0002431552170000064
CN201910561695.3A 2019-06-26 2019-06-26 Complex fault diagnosis and identification method for vertical take-off and landing aircraft based on composite observer Active CN110244697B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910561695.3A CN110244697B (en) 2019-06-26 2019-06-26 Complex fault diagnosis and identification method for vertical take-off and landing aircraft based on composite observer

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910561695.3A CN110244697B (en) 2019-06-26 2019-06-26 Complex fault diagnosis and identification method for vertical take-off and landing aircraft based on composite observer

Publications (2)

Publication Number Publication Date
CN110244697A CN110244697A (en) 2019-09-17
CN110244697B true CN110244697B (en) 2020-05-26

Family

ID=67889648

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910561695.3A Active CN110244697B (en) 2019-06-26 2019-06-26 Complex fault diagnosis and identification method for vertical take-off and landing aircraft based on composite observer

Country Status (1)

Country Link
CN (1) CN110244697B (en)

Families Citing this family (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111221345B (en) * 2020-01-23 2023-04-14 北京航天自动控制研究所 Decision tree-based aircraft power system fault online identification method
CN112526884A (en) * 2020-12-04 2021-03-19 北京航空航天大学 Fault system self-adaptive fault tolerance method and system
CN112528492B (en) * 2020-12-09 2022-09-20 中国航空工业集团公司沈阳飞机设计研究所 Fault detection method and device under wing damage condition
CN113221719B (en) * 2021-05-06 2023-04-07 北京航空航天大学 Aircraft fault diagnosis method and device and electronic equipment
CN113778115A (en) * 2021-08-30 2021-12-10 北京三快在线科技有限公司 Multi-rotor unmanned aerial vehicle control method, device, medium, equipment and unmanned aerial vehicle
CN114019936B (en) * 2021-09-30 2023-12-29 北京控制工程研究所 Translation thruster fault diagnosis and recombination method suitable for deep space landing

Family Cites Families (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101578584A (en) * 2005-09-19 2009-11-11 克利夫兰州立大学 Controllers, observers, and applications thereof
DE102012104783B4 (en) * 2012-06-01 2019-12-24 Quantum-Systems Gmbh Aircraft, preferably UAV, drone and / or UAS
CN103838145B (en) * 2014-01-17 2016-08-17 南京航空航天大学 VTOL aircraft Robust Fault-Tolerant Control Systems based on cascade observer and method
CN103970997B (en) * 2014-05-06 2016-05-25 南昌华梦达航空科技发展有限公司 A kind of depopulated helicopter sensor fault fast diagnosis method
CN105204499B (en) * 2015-10-09 2018-01-02 南京航空航天大学 Helicopter collaboration formation method for diagnosing faults based on Unknown Input Observer
CN108647442B (en) * 2018-05-11 2020-06-19 北京航空航天大学 Auxiliary output-based six-rotor unmanned aerial vehicle fault estimation method
CN109343369B (en) * 2018-11-19 2021-08-10 南京邮电大学 Design method of four-rotor fault-tolerant controller based on non-linear observer

Also Published As

Publication number Publication date
CN110244697A (en) 2019-09-17

Similar Documents

Publication Publication Date Title
CN110244697B (en) Complex fault diagnosis and identification method for vertical take-off and landing aircraft based on composite observer
Samy et al. Survey and application of sensor fault detection and isolation schemes
CN107608367B (en) Multivariable interference compensation quadrotor unmanned aerial vehicle trajectory and attitude cooperative control method
CN106742067B (en) Attitude control engine on-orbit fault Fault Diagnostic Expert System and method based on reference model
Madani et al. Sliding mode observer and backstepping control for a quadrotor unmanned aerial vehicles
CN111309042B (en) Aircraft overload tracking method taking overload and angular speed as outer loop
Hamel et al. Advances in rotorcraft system identification
CN103676941B (en) Satellite control system method for diagnosing faults based on kinematics and dynamics modeling
CN113485304B (en) Aircraft hierarchical fault-tolerant control method based on deep learning fault diagnosis
CN111024143A (en) Hypersonic aircraft sensor cascading failure diagnosis and fault-tolerant control method
CN103389088A (en) Determination method of optimal configuration scheme of four-redundancy strapdown inertial navigation system (RFINS)
CN101713654A (en) Fusing method of atmospheric attack angle and inertia attack angle in transonic flight stage
CN111024124A (en) Multi-sensor information fusion combined navigation fault diagnosis method
CN111208844A (en) Aircraft longitudinal overload nonlinear control method based on overload angular velocity
CN110553642A (en) Method for improving inertial guidance precision
Freeman Reliability assessment for low-cost unmanned aerial vehicles
Wu et al. Simultaneous state and parameter estimation based actuator fault detection and diagnosis for an unmanned helicopter
Saied et al. Differential flatness-based approach for sensors and actuators fault diagnosis of a multirotor UAV
Bao et al. Aerodynamic model/INS/GPS failure-tolerant navigation method for multirotor UAVs based on federated Kalman Filter
Nejati et al. Robust three stage central difference kalman filter for helicopter unmanned aerial vehicle actuators fault estimation
Baldi et al. Combined geometric and neural network approach to generic fault diagnosis in satellite reaction wheels
Litt et al. A demonstration of a retrofit architecture for intelligent control and diagnostics of a turbofan engine
Abraham et al. In-flight estimation of helicopter gross weight and mass center location
Freeman et al. Analytical fault detection for a small UAV
CN114578837A (en) Active fault-tolerant anti-interference trajectory tracking control method of unmanned helicopter considering actuator faults

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant