Quick fault detection and isolation method for actuator of dual-rotor unmanned aerial vehicle
Through the design of the state observer based on the LPV model, the rapid detection and isolation of the actuator failure of the dual-rotor UAV is achieved, which solves the signal coupling problem, improves flight safety, and reduces false alarms.
Patent Information
- Application Number
- CN202411058402.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-02
- Publication Date
- 2024-11-08
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
During flight, the double-rotor UAV is difficult to detect faults due to changes in aerodynamic characteristics, electromagnetic interference and component aging. The existing methods are susceptible to noise and interference, resulting in false alarms and reducing flight reliability. and security.
The multihedral LPV model of a dual-rotor drone is established based on linear parameter change method, and the state observer is designed, the residual signal is calculated through the measured values of the fuselage sensor and the estimated values generated by the observer, the dynamic detection threshold is set, and the fault detection and Isolation rules to achieve rapid detection and isolation of actuator failures.
It effectively solves the signal coupling problem, improves the flight safety of the twin-rotor drone, reduces false alarms, and improves the robustness and sensitivity of fault detection.
Smart Images

Figure CN118915811A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of rotorcraft fault diagnosis, and in particular to a method for rapid fault detection and isolation of a dual-rotor unmanned aerial vehicle actuator. Background Art
[0002] The dual-rotor UAV is a typical nonlinear strongly coupled system with obvious underactuated characteristics. The aerodynamic characteristics and mathematical models of the dual-rotor UAV vary greatly in different flight modes. Therefore, compared with traditional coaxial rotor helicopters or quad-rotor UAVs, the control of dual-rotor UAVs is more difficult. At the same time, in the actual flight environment, changes in atmospheric parameters and aerodynamic parameters of the aircraft, electromagnetic interference, and aging of components can lead to helicopter failures, which in turn reduce the flight reliability and safety of the dual-rotor UAV, and even cause economic losses and threaten the safety of people around.
[0003] At present, the methods for fault detection and isolation applied to dual-rotor UAVs are divided into: model-based methods, data-based methods and knowledge-based methods. Among them, the data-based method is suitable for systems with massive data. It uses the data generated during the operation of the equipment for analysis to distinguish the normal operation state of the system from the fault state. At present, there are mainly machine learning methods, signal processing methods and statistical methods. The knowledge-based method is an empirical diagnostic method, which is mainly divided into expert system method, fuzzy reasoning method, fault tree method, neural network method, etc. The model-based method is based on the understanding of the motion mechanism of the physical system, designs a multi-channel fault detection observer or filter, and judges the health state of the system execution components through the generated state residual. It is the mainstream method for multi-rotor UAV fault diagnosis. In addition, the existence of factors such as measurement noise and external interference will affect the performance of the observer-based fault detection method to a certain extent, causing false alarms. To this end, we propose a method for rapid detection and isolation of dual-rotor UAV actuator faults to solve the above problems. Summary of the invention
[0004] The purpose of this section is to summarize some aspects of embodiments of the present invention and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the specification abstract and the invention title of this application to avoid blurring the purpose of this section, the specification abstract and the invention title, and such simplifications or omissions cannot be used to limit the scope of the present invention.
[0005] In view of the above problems in the prior art, the present invention is proposed.
[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: a method for rapid fault detection and isolation of a dual-rotor UAV actuator, comprising the following steps:
[0007] Step 1. Analyze the motion characteristics of the dual-rotor UAV system and establish a dual-rotor UAV polyhedron LPV model based on the linear parameter variation method;
[0008] Step 2. Based on the above model, design a set of state observers for fault detection and isolation;
[0009] Step 3. Calculate the residual signal based on the measured value of the fuselage sensor and the estimated value generated by the observer, and design and calculate the state evaluation function; at the same time, set the dynamic detection threshold according to the known interference boundary conditions;
[0010] Step 4. Based on the state evaluation function and the detection threshold, define the fault detection and isolation rules. If the evaluation function does not exceed the detection threshold, the actuator state of the dual-rotor UAV is normal; otherwise, it means that a fault has occurred, and the system is judged to be faulty;
[0011] Furthermore, the specific faulty actuator is located by analyzing the residuals generated by the two observers.
[0012] As a preferred solution of the method for rapid fault detection and isolation of a dual-rotor UAV actuator described in the present invention, in which: during the step 1, the specific establishment process of the dual-rotor UAV polyhedron LPV model is:
[0013] First, through mechanism analysis, the dynamic equation 1 of the dual-rotor UAV is established as follows:
[0014]
[0015] Among them, ε, θ, ψ represent the pitch angle, roll angle and yaw angle respectively, J ε ,J θ ,J ψ They represent the moment of inertia of the pitch axis, roll axis and yaw axis of the system, respectively, K f represents the thrust mapping coefficient, L a Indicates the distance from the aircraft to the pitch angle, L w Indicates the distance from the balance block to the pitch angle, V f ,V b Respectively represent the voltage values of the front and rear actuators, M h ,M w Respectively represent the mass of the body and the mass of the counterweight, w ε , w θ , w ψ Respectively represent the disturbances in the three channels of pitch, roll and yaw;
[0016] Then transform equation 1 into state space form and define the state vector as Define the input vector as Assuming that additive faults will occur in the actuator, simplifying the yaw angle dynamics at the equilibrium point, the state space model of the dual-rotor UAV is obtained as follows:
[0017]
[0018] y(t)=Cx(t)
[0019] in,
[0020]
[0021] d ψ =w ψ +b ψ (sinε-1)(V f +V b );b ε =K f L a / J ε ; b θ =K f L h / J θ ; b ψ =K f L h / J ψ ;
[0022] g(x,t)=g(M h L a -M w L w )cosε;
[0023] f a =[f a1 ,f a2 ] T Indicates a fault in the front or rear actuator;
[0024] Assume that the function g(x,t) satisfies the Lipschitz nonlinear condition:
[0025]
[0026] Where κ is the Lipschitz constant. Then, the nonlinear state space equation is transformed into LPV form to establish the LPV model of the dual-rotor UAV.
[0027] Due to the limitation of mechanical structure, the variation range of pitch angle is bounded, satisfying -π / 4≤θ≤π / 4. Then a new variable parameter α(t) is defined to describe the variation of variable θ, α(t)=cos(θ). According to the variation range of angle θ, α can be obtained. min ≤α≤α max ,αmin =0.707,α max =1;
[0028] Therefore, the polyhedron LPV model of the dual-rotor UAV obtained by linearization at different working points can be expressed as:
[0029]
[0030] y(t)=Cx(t).
[0031] As a preferred solution of the method for rapid fault detection and isolation of a dual-rotor UAV actuator of the present invention, the state observer in the step 2 includes: state observers corresponding to the front and rear actuators of the UAV respectively;
[0032] By parameter design, the actuator fault and residual are decoupled, so that the residual generated by the first state observer is only sensitive to the rear actuator fault, and the residual generated by the second state observer is only sensitive to the front actuator fault, thereby realizing the detection and isolation of the front and rear actuator faults. When deriving the solution conditions for calculating the gain of the fault detection observer, in addition to satisfying the decoupling conditions, H is incorporated into it. ∞ The robustness index of fault detection is integrated into H - conditions to enhance the sensitivity of fault detection indicators.
[0033] As a preferred solution of the method for rapid fault detection and isolation of a dual-rotor UAV actuator described in the present invention, in order to facilitate the design of the observer in step 2, the dual-rotor UAV LPV model is rearranged as follows:
[0034]
[0035] y(t)=Cx(t)
[0036] Among them, B ij is the jth column of the input matrix B of the i-th LPV system of the helicopter, is the matrix consisting of the remaining columns after removing the jth column from B, f aj is the fault vector of the jth actuator, is the fault vector of all actuators except the jth actuator;
[0037] The form of the observer corresponding to the designed j-th actuator is as follows:
[0038]
[0039] in, is the estimate of the state vector x(t) generated by the jth observer, M ij ,Gij , N ij , L ij and H ij is the gain of the observer to be designed.
[0040] As a preferred solution of the method for rapid fault detection and isolation of a dual-rotor UAV actuator described in the present invention, the specific condition for solving the observer gain is:
[0041]
[0042] In the above formula, H is obtained based on Lyapunov stability theory. ∞ Conditions and H - Conditions are used to implement fault detection, while the fault isolation conditions are:
[0043]
[0044] As a preferred solution of the method for rapid fault detection and isolation of a dual-rotor UAV actuator described in the present invention, wherein: in step 3, fault detection and isolation are implemented based on the logic rule of threshold comparison. First, the state estimation error e is defined. j (t) and the residual signal r j (t):
[0045]
[0046] Secondly, define the state evaluation function as the 2-norm of the state estimation error vector:
[0047] J e (t)=||r(t)||2
[0048] Then, define the monitoring threshold as follows:
[0049]
[0050] in, is the upper bound of the initial value error of the state estimate e0, λ max is the matrix P determined in step 2 ij is the maximum eigenvalue of , γ is the robust performance coefficient defined in step 2, is a custom coefficient.
[0051] As a preferred solution of the method for rapid fault detection and isolation of a dual-rotor UAV actuator described in the present invention, in which: during step 4, the specific fault detection logic rule is:
[0052]
[0053] Beneficial effects of the invention: The invention provides a fault detection and isolation method for a dual-rotor UAV actuator to solve the signal coupling problem in fault detection and positioning of dual-rotor UAV actuators under under-actuated conditions and improve the flight safety of the dual-rotor UAV. BRIEF DESCRIPTION OF THE DRAWINGS
[0054] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. Among them:
[0055] Figure 1 This is an overall block diagram of a method for rapid fault detection and isolation of a dual-rotor UAV actuator proposed by the present invention;
[0056] Figure 2 A schematic diagram of a dual-rotor UAV actuator fault detection and isolation structure of a dual-rotor UAV actuator rapid fault detection and isolation method proposed by the present invention;
[0057] Figure 3 This is a graph of the observed values of the pitch angular velocity of a method for rapid fault detection and isolation of a dual-rotor UAV actuator proposed by the present invention;
[0058] Figure 4 This is a graph of the roll angular velocity observation values of a method for rapid fault detection and isolation of a dual-rotor UAV actuator proposed in the present invention;
[0059] Figure 5 This is a diagram of the yaw angular velocity observation value of a dual-rotor UAV actuator rapid fault detection and isolation method proposed by the present invention;
[0060] Figure 6 The residual error and threshold value diagram of the rear motor of the fast fault detection and isolation method of the dual-rotor UAV actuator proposed in this invention
[0061] Figure 7 This is the front motor residual and threshold diagram of the dual-rotor UAV actuator rapid fault detection and isolation method proposed in the present invention. DETAILED DESCRIPTION
[0062] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.
[0063] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.
[0064] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.
[0065] Example 1, reference Figure 1-Figure 2 , as an embodiment of the present invention, provides a method for rapid fault detection and isolation of a dual-rotor UAV actuator, the method comprising the following steps:
[0066] Step 1. Analyze the motion characteristics of the dual-rotor UAV system and establish a dual-rotor UAV polyhedron LPV model based on the linear parameter variation method;
[0067] Furthermore, in the step 1, the specific process of establishing the dual-rotor UAV polyhedron LPV model is as follows:
[0068] First, through mechanism analysis, the dynamic equation 1 of the dual-rotor UAV is established as follows:
[0069]
[0070] Among them, ε, θ, ψ represent the pitch angle, roll angle and yaw angle respectively, J ε ,J θ ,J ψ They represent the moment of inertia of the pitch axis, roll axis and yaw axis of the system, respectively, K f represents the thrust mapping coefficient, L a Indicates the distance from the aircraft to the pitch angle, L w Indicates the distance from the balance block to the pitch angle, V f ,V b Respectively represent the voltage values of the front and rear actuators, M h ,M w Respectively represent the mass of the body and the mass of the counterweight, w ε , w ψ Respectively represent the disturbances in the three channels of pitch, roll and yaw;
[0071] Then transform equation 1 into state space form and define the state vector as Define the input vector as Assuming that additive faults will occur in the actuator, simplifying the yaw angle dynamics at the equilibrium point, the state space model of the dual-rotor UAV is obtained as follows:
[0072]
[0073] y(t)=Cx(t)
[0074] in,
[0075]
[0076] d ψ =w ψ +b ψ (sinε-1)(V f +V b );b ε =K f L a / J ε ; b θ =K f L h / J θ ; b ψ =K f L h / J ψ ;
[0077] g(x,t)=g(M h L a -M w L w )cosε;
[0078] f a =[f a1 ,f a2 ] T Indicates a fault in the front or rear actuator;
[0079] Assume that the function g(x,t) satisfies the Lipschitz nonlinear condition:
[0080]
[0081] Where κ is the Lipschitz constant. Then, the nonlinear state space equation is transformed into LPV form to establish the LPV model of the dual-rotor UAV.
[0082] Due to the limitation of mechanical structure, the variation range of pitch angle is bounded, satisfying -π / 4≤θ≤π / 4. Then a new variable parameter α(t) is defined to describe the variation of variable θ, α(t)=cos(θ). According to the variation range of angle θ, α can be obtained. min ≤α≤αmax ,α min =0.707,α max =1;
[0083] Therefore, the polyhedron LPV model of the dual-rotor UAV obtained by linearization at different working points can be expressed as:
[0084]
[0085] y(t)=Cx(t).
[0086] Among them B i is a time-invariant matrix, ρ i is the weighting function of each subsystem, ρ2(α)=1-ρ1(α).
[0087] Step 2. Based on the above model, design a set of state observers for fault detection and isolation;
[0088] Includes: state observers for the front and rear actuators of the drone;
[0089] By parameter design, the actuator fault and residual are decoupled, so that the residual generated by the first state observer is only sensitive to the rear actuator fault, and the residual generated by the second state observer is only sensitive to the front actuator fault, thereby realizing the detection and isolation of the front and rear actuator faults. When deriving the solution conditions for calculating the gain of the fault detection observer, in addition to satisfying the decoupling conditions, H is incorporated into it. ∞ The robustness index of fault detection is integrated into H - conditions to enhance the sensitivity of fault detection indicators.
[0090] Specifically, in order to facilitate the design of the observer in step 2, the dual-rotor UAV LPV model is rearranged as follows:
[0091]
[0092] y(t)=Cx(t)
[0093] Among them, B ij is the jth column of the input matrix B of the i-th LPV system of the helicopter, is the matrix consisting of the remaining columns after removing the jth column from B, f aj is the fault vector of the jth actuator, is the fault vector of all actuators except the jth actuator;
[0094] The form of the observer corresponding to the designed j-th actuator is as follows:
[0095]
[0096] in, is the estimate of the state vector x(t) generated by the jth observer, M ij ,G ij , N ij , L ij and H ij is the gain of the observer to be designed.
[0097] The specific conditions for solving the observer gain are:
[0098]
[0099] In the above formula, H is obtained based on Lyapunov stability theory. ∞ Conditions and H - Conditions are used to implement fault detection, while the fault isolation conditions are:
[0100]
[0101] In the above specific implementation, the observer gain M that satisfies the conditions ij ,G ij , N ij , L ij and H ij The calculation can be realized with the help of LMI toolbox.
[0102] Step 3. Calculate the residual signal based on the measured value of the fuselage sensor and the estimated value generated by the observer, and design and calculate the state evaluation function; at the same time, set the dynamic detection threshold according to the known interference boundary conditions;
[0103] Specifically, fault detection and isolation are implemented based on the logic rules of threshold comparison. First, the state estimation error e is defined as j (t) and the residual signal r j (t):
[0104]
[0105] Secondly, define the state evaluation function as the 2-norm of the state estimation error vector:
[0106] J e (t)=||r(t)||2
[0107] Then, define the monitoring threshold as follows:
[0108]
[0109] in, is the upper bound of the initial value error of the state estimate e0, λ max is the matrix P determined in step 2 ij is the maximum eigenvalue of , γ is the robust performance coefficient defined in step 2, is a custom coefficient.
[0110] Step 4. Based on the state evaluation function and the detection threshold, define the fault detection and isolation rules. If the evaluation function does not exceed the detection threshold, the actuator state of the dual-rotor UAV is normal; otherwise, it means that a fault has occurred, and the system is judged to be faulty;
[0111] Furthermore, by analyzing the residuals generated by the two observers, the specific faulty actuator can be located.
[0112] The specific fault detection logic rules are:
[0113]
[0114] Example 2, reference Figures 3 to 7 In order to verify the actual effect of the present invention, step 4 is reproduced and verified through simulation and experimental verification. The results are as follows:
[0115] It can be seen that Figures 3 to 5 The observed angular velocity of the drone is shown. The effect of the proposed fault detection and isolation scheme can be more clearly seen from the angular velocity observations. The fault detection and isolation scheme of the proposed method produces observations that deviate significantly from the actual values during the actuator failure and achieves decoupling of the front and rear actuator fault signals.
[0116] Figure 6 The fault isolation results of the rear actuator are shown. Figure 6 It can be seen that at 60 seconds, the proposed FDI scheme detects a fault and the evaluation function value increases rapidly and exceeds the threshold. At 70 seconds, it reaches the maximum value and ends at 80 seconds.
[0117] Figure 7 The results of the fault detection and isolation of the front actuator are shown. Figure 7 As can be seen in Figure 2, a fault is detected at 20 seconds, the evaluation function value increases rapidly and exceeds the threshold, and ends within 36 seconds.
[0118] In summary, the present invention provides a fault detection and isolation method for a dual-rotor UAV actuator to solve the signal coupling problem in fault detection and positioning of dual-rotor UAV actuators under under-actuated conditions and improve the flight safety of the dual-rotor UAV.
[0119] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.
Claims
1. A method for rapid fault detection and isolation of a dual-rotor UAV actuator, characterized in that: The following steps are involved: Step 1. Analyze the motion characteristics of the dual-rotor UAV system and establish a dual-rotor UAV polyhedron LPV model based on the linear parameter variation method; Step 2. Based on the above model, design a set of state observers for fault detection and isolation; Step 3. Calculate the residual signal based on the measured value of the fuselage sensor and the estimated value generated by the observer, and design and calculate the state evaluation function; at the same time, set the dynamic detection threshold according to the known interference boundary conditions; Step 4. Based on the state evaluation function and the detection threshold, define the fault detection and isolation rules. If the evaluation function does not exceed the detection threshold, the actuator state of the dual-rotor UAV is normal; On the contrary, it means that a fault has occurred, and the system is judged to be faulty; Furthermore, the specific faulty actuator is located by analyzing the residuals generated by the two observers.
2. The method for rapid fault detection and isolation of a dual-rotor UAV actuator according to claim 1, characterized in that: During step 1, the specific process of establishing the dual-rotor UAV polyhedron LPV model is as follows: First, through mechanism analysis, the dynamic equation 1 of the dual-rotor UAV is established as follows: Among them, ε, θ, ψ represent the pitch angle, roll angle and yaw angle respectively, J ε ,J θ ,J ψ They represent the moment of inertia of the pitch axis, roll axis and yaw axis of the system, respectively, K f represents the thrust mapping coefficient, L a Indicates the distance from the aircraft to the pitch angle, L w Indicates the distance from the balance block to the pitch angle, V f ,V b Represent the voltage values of the front and rear actuators respectively, M h ,M w Respectively represent the mass of the body and the mass of the counterweight, w ε , w θ , w ψ Respectively represent the disturbances in the three channels of pitch, roll and yaw; Then transform equation 1 into state space form and define the state vector as Define the input vector as Assuming that additive faults will occur in the actuator, simplifying the yaw angle dynamics at the equilibrium point, the state space model of the dual-rotor UAV is obtained as follows: y(t)=Cx(t) in, d ψ =w ψ +b ψ (sinε-1)(V f +V b ); b ε =K f L a / J ε ;b θ =K f L h / J θ ;b ψ =K f L h / J ψ ;g(x,t)=g(M h L a -M w L w )cosε; f a =[f a1 ,f a2 ] T Indicates a fault in the front or rear actuator; Assume that the function g(x,t) satisfies the Lipschitz nonlinear condition: Where κ is the Lipschitz constant. Then, the nonlinear state space equation is transformed into LPV form to establish the LPV model of the dual-rotor UAV. Due to the limitation of mechanical structure, the variation range of pitch angle is bounded, satisfying -π / 4≤θ≤π / 4. Then a new variable parameter α(t) is defined to describe the variation of variable θ, α(t)=cos(θ). According to the variation range of angle θ, α can be obtained. min ≤α≤α max ,α min =0.707,α max =1; Therefore, the polyhedron LPV model of the dual-rotor UAV obtained by linearization at different working points can be expressed as: y(t)=Cx(t).
3. A method for rapid fault detection and isolation of a dual-rotor UAV actuator according to claim 2, characterized in that: The state observer in step 2 includes: state observers of the front and rear actuators of the drone; Through parameter design, the actuator fault and residual are decoupled, so that the residual generated by the first state observer is only sensitive to the rear actuator fault, and the residual generated by the second state observer is only sensitive to the front actuator fault, thereby realizing the detection and isolation of the front and rear actuator faults; when deriving the solution conditions for calculating the gain of the fault detection observer, in addition to satisfying the decoupling conditions, H ∞ The robustness index of fault detection is integrated into H - conditions to enhance the sensitivity of fault detection indicators.
4. A method for rapid fault detection and isolation of a dual-rotor UAV actuator according to claim 3, characterized in that: In order to facilitate the design of the observer in step 2, the dual-rotor UAV LPV model is rearranged as follows: y(t)=Cx(t Among them, B ij is the jth column of the input matrix B of the i-th LPV system of the helicopter, is the matrix consisting of the remaining columns after removing the jth column from B, f aj is the fault vector of the jth actuator, is the fault vector of all actuators except the jth actuator; The form of the observer corresponding to the designed j-th actuator is as follows: in, is the estimate of the state vector x(t) generated by the jth observer, M ij ,G ij , N ij , L ij and H ij is the gain of the observer to be designed.
5. A method for rapid fault detection and isolation of a dual-rotor UAV actuator according to claim 4, characterized in that: The specific conditions for solving the observer gain are: In the above formula, H is obtained based on Lyapunov stability theory. ∞ Conditions and H - Conditions are used to implement fault detection, while the fault isolation conditions are:
6. A method for rapid fault detection and isolation of a dual-rotor UAV actuator according to claim 5, characterized in that: In step 3, fault detection and isolation are implemented based on the logic rule of threshold comparison. First, the state estimation error e is defined. j (t) and the residual signal r j (t): Secondly, define the state evaluation function as the 2-norm of the state estimation error vector: J e (t)=||r(t)||2 Then, define the monitoring thresholds as follows: in, is the upper bound of the initial value error of the state estimate e0, λ max is the matrix P determined in step 2 ij is the maximum eigenvalue of , γ is the robust performance coefficient defined in step 2, is a custom coefficient.
7. A method for rapid fault detection and isolation of a dual-rotor UAV actuator according to claim 6, characterized in that: In step 4, the specific fault detection logic rules are as follows:
Citation Information
Cited By
Fiber bragg grating sensor measuring method for helicopter blade waving and shimmy decoupling position
CN119223168A
