Fan transmission system fault estimation method and device

Abstract

The invention relates to the technical field of fault diagnosis, in particular to a fan transmission system fault estimation method and device, and aims to solve the technical problem of simultaneously estimating unknown input observers of parameter faults and additive faults. The invention discloses a fan transmission system fault estimation method, which comprises the following steps of: 1, constructing a fan transmission system fault mathematical model; 2, constructing an augmentation system of the fan transmission system fault mathematical model; 3, designing an adaptive law, and constructing an adaptive unknown input observer; 4, constructing a state estimation error, a measurement output error, a fault estimation error and an adaptive rate, and obtaining an error dynamic equation; 5, calculating the gain of the adaptive unknown input observer according to a Lyapunov stability theory and a linear matrix inequality method LMI, and obtaining a target observer; and 6, inputting the state data of the fan transmission system into the target observer, and judging whether the fan transmission system fails or not.

Description

Technical Field

[0001] The present invention relates to the technical field of fault diagnosis, and in particular to a method and device for estimating faults of a fan transmission system. Background Art

[0002] With the rapid development of science and technology, wind turbine systems have attracted more and more attention from researchers. Wind turbines integrate multiple subsystems such as pitch, transmission, generator and converter. If a subsystem fails, it will affect the normal operation of other subsystems, thus causing the entire wind turbine to fail to operate normally. Therefore, the fault-tolerant control technology for wind turbines has been rapidly developed. Fault estimation is an important part of fault-tolerant control technology. The quality of fault-tolerant control depends on the accuracy of fault estimation. An excellent fault estimation algorithm is very important.

[0003] The fan drive shaft is prone to multiplicative failures due to wear and other reasons. A type of parameter failure is manifested as a proportional decrease in system performance. At the same time, additive failures may also occur inside the fan, and unknown signals suddenly appear in certain signal channels of the fan system. Summary of the invention

[0004] In view of this, the object of the present invention is to provide a method and device for fault estimation of a wind turbine transmission system to solve the technical problem of an unknown input observer that simultaneously estimates parameter faults and additive faults.

[0005] A method for estimating a fault of a fan transmission system, comprising:

[0006] The first step is to build a mathematical model of wind turbine transmission system failure;

[0007] The second step is to construct an augmented system of the mathematical model of the fan transmission system fault;

[0008] The third step is to design the adaptive law and construct an adaptive unknown input observer;

[0009] Step 4: construct the state estimation error, measurement output error, fault estimation error and adaptive rate to obtain the error dynamic equation;

[0010] The fifth step is to calculate the gain of the adaptive unknown input observer based on Lyapunov stability theory and linear matrix inequality method LMI to obtain the target observer;

[0011] In the sixth step, the status data of the fan transmission system is input into the target observer to determine whether the fan transmission system is faulty.

[0012] As a further improvement of the present application, the establishment of a mathematical model of the fan transmission system fault includes:

[0013]

[0014] in: is the state vector, denote the input vector and the measured output vector respectively, is a bounded vector, representing unknown disturbances and modeling errors, is an additive fault, which represents the additive fault vector of the pitch system; A, B, C, B d ,B f is the constant coefficient matrix known to the system;

[0015] Perturbation B d =[B d1 B d2 ], d1 is decoupled, d2 is not decoupled, in the mathematical model is a nonlinear vector function with a Lipschiz coefficient θ and satisfies the following Lipschiz relation

[0016]

[0017] As a further improvement of the present application, the augmented system for constructing the mathematical model of the fan transmission system fault includes:

[0018] Augmented mathematics for building mathematical models (1)

[0019]

[0020] in:

[0021]

[0022] As a further improvement of the present application, the design of the adaptive law constructs an adaptive unknown input observer.

[0023] Based on the augmented mathematical model (3), an adaptive unknown input observer is established:

[0024]

[0025] Respectively represent the intermediate transition state quantity, The estimator of and the estimator of E; K=K1+K2、 represents the gain matrix to be designed and calculated;

[0026] The following Lipschiz relation is satisfied. According to (2), we can conclude that:

[0027]

[0028] As a further improvement of the present application, the state estimation error, the measurement output error, the fault estimation error and the adaptive rate are constructed to obtain the error dynamic equation, including

[0029] The state estimation error is

[0030] Measurement output error

[0031] Fault estimation error

[0032] Adaptive rate

[0033] in: represents the gain matrix to be designed, δ represents the adaptive learning law, which is a constant;

[0034] The error dynamic equation is constructed as

[0035]

[0036] As a further improvement of the present application, the gain of the adaptive unknown input observer is calculated based on the Lyapunov stability theory and the linear matrix inequality method LMI to obtain the target observer, including

[0037] The gains of the adaptive unknown input observer are

[0038] L1, R, T, K2, K, according to Lyapunov stability theory and linear matrix inequality (LMI), calculate the gain Q, the symmetric positive definite matrix P, so that the following linear matrix inequality is satisfied:

[0039]

[0040] in:

[0041]

[0042] express Dimensional identity matrix

[0043] After calculating the matrix L1, the values ​​of the gain matrices R, T, K2, K can be further calculated by the following formulas, thereby obtaining the fault estimator of the mathematical model (1);

[0044]

[0045]

[0046]

[0047] A fault estimation device for a fan transmission system, characterized by comprising:

[0048] The preprocessing module is used to collect the electrical parameter data of the pumping unit in real time online, perform preprocessing, vectorize the high-dimensional signal data, and construct the original data space of the electrical parameter data;

[0049] A construction module is used to determine the k nearest sample points to each sample point through Euclidean distance to construct a local neighborhood;

[0050] An integration module is used to take any sample point in the same local neighborhood, except the central sample point, as its local mutual representation neighborhood by the remaining sample points in the local neighborhood; the local neighborhood and the local mutual representation neighborhood obtained by each sample point through the construction module are integrated to form a multi-structure neighborhood;

[0051] The reconstruction module is used to calculate the weight coefficients of the local linear relations of the multi-structured neighborhood in high-dimensional space by minimizing the local reconstruction error;

[0052] The feature fusion module is used to fuse the features of multiple local neighborhood structures in a low-dimensional space, build a global reconstruction model, and obtain the final low-dimensional features by determining the optimal solution of the global reconstruction model;

[0053] The fault diagnosis module is used to input the obtained low-dimensional feature results into the classifier and complete the fault diagnosis of the oil pump according to its classification results.

[0054] Based on the above purpose, the present invention provides a fault estimation method for a wind turbine transmission system. The unknown input observer has advantages in decoupling disturbances. In the current research on fault estimation of unknown input observers, most of them use state observation schemes to achieve fault estimation, which has a wide range of applications. However, the unknown input state observer is poor in observation accuracy, and the fault must be a type of slowly varying disturbance, which reduces the application breadth of this type of algorithm to a certain extent. This paper designs a new observer by combining an adaptive observer and an unknown input observer, which can not only decouple disturbances, but also improve the fault estimation accuracy, and can simultaneously estimate parameter faults and additive faults. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only for the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0056] Figure 1 The present invention is a flowchart of a method for estimating a fault in a wind turbine transmission system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0057] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with specific embodiments.

[0058] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the present invention should be understood by people with ordinary skills in the field to which the present invention belongs. The "first", "second" and similar words used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0059] like Figure 1 As shown, a method for estimating a fault of a wind turbine transmission system includes:

[0060] The first step is to build a mathematical model of the fan transmission system fault; including

[0061]

[0062] in: is the state vector, denote the input vector and the measured output vector respectively, is a bounded vector, representing unknown disturbances and modeling errors, is an additive fault, which represents the additive fault vector of the pitch system; A, B, C, B d ,B f is the constant coefficient matrix known to the system;

[0063] Perturbation B d =[B d1 B d2 ], d1 is decoupled, d2 is not decoupled, in the mathematical model is a nonlinear vector function with a Lipschiz coefficient θ and satisfies the following Lipschiz relation

[0064] ||Φ(t,x,u)||≤θ||x||

[0065]

[0066] The second step is to build an augmented system of the mathematical model of the fan transmission system fault; including

[0067] Augmented mathematics for building mathematical models (1)

[0068]

[0069] in:

[0070]

[0071] The third step is to design the adaptive law and construct an adaptive unknown input observer;

[0072] Based on the augmented mathematical model (3), an adaptive unknown input observer is established:

[0073]

[0074] Respectively represent the intermediate transition state quantity, The estimator of and the estimator of E; K=K1+K2、 represents the gain matrix to be designed and calculated;

[0075] The following Lipschiz relation is satisfied. According to (2), we can conclude that:

[0076]

[0077] Step 4: construct the state estimation error, measurement output error, fault estimation error and adaptive rate to obtain the error dynamic equation;

[0078] The state estimation error is

[0079] Measurement output error

[0080] Fault estimation error

[0081] Adaptive rate

[0082] in: represents the gain matrix to be designed, δ represents the adaptive learning law, which is a constant;

[0083] The error dynamic equation is constructed as

[0084]

[0085] The fifth step is to calculate the gain of the adaptive unknown input observer based on Lyapunov stability theory and linear matrix inequality method LMI to obtain the target observer; including

[0086] The gains of the adaptive unknown input observer are

[0087] L1, R, T, K2, K, according to Lyapunov stability theory and linear matrix inequality (LMI), calculate the gain Q, the symmetric positive definite matrix P, so that the following linear matrix inequality is satisfied:

[0088]

[0089] in:

[0090]

[0091] express Dimensional identity matrix

[0092] After calculating the matrix L1, the values ​​of the gain matrices R, T, K2, K can be further calculated by the following formulas, thereby obtaining the fault estimator of the mathematical model (1);

[0093]

[0094]

[0095]

[0096] In the sixth step, the status data of the fan transmission system is input into the target observer to determine whether the fan transmission system is faulty.

[0097] A wind turbine transmission system estimation device, comprising:

[0098] The preprocessing module is used to collect the electrical parameter data of the pumping unit in real time online, perform preprocessing, vectorize the high-dimensional signal data, and construct the original data space of the electrical parameter data;

[0099] A construction module is used to determine the k nearest sample points to each sample point through Euclidean distance to construct a local neighborhood;

[0100] An integration module is used to take any sample point in the same local neighborhood, except the central sample point, as its local mutual representation neighborhood by the remaining sample points in the local neighborhood; the local neighborhood and the local mutual representation neighborhood obtained by each sample point through the construction module are integrated to form a multi-structure neighborhood;

[0101] The reconstruction module is used to calculate the weight coefficients of the local linear relations of the multi-structured neighborhood in high-dimensional space by minimizing the local reconstruction error;

[0102] The feature fusion module is used to fuse the features of multiple local neighborhood structures in a low-dimensional space, build a global reconstruction model, and obtain the final low-dimensional features by determining the optimal solution of the global reconstruction model;

[0103] The fault diagnosis module is used to input the obtained low-dimensional feature results into the classifier and complete the fault diagnosis of the oil pump according to its classification results.

[0104] Those skilled in the art should understand that the discussion of any of the above embodiments is merely illustrative and is not intended to imply that the scope of the present invention (including the claims) is limited to these examples. Under the concept of the present invention, the technical features in the above embodiments or different embodiments may be combined, the steps may be implemented in any order, and there are many other variations of the different aspects of the present invention as described above, which are not provided in detail for the sake of simplicity.

[0105] The present invention is intended to cover all such substitutions, modifications and variations that fall within the broad scope of the appended claims. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for estimating a fault in a fan transmission system, characterized in that: include: The first step is to build a mathematical model of fan transmission system failure; The second step is to construct an augmented system of the mathematical model of the fan transmission system fault; The third step is to design the adaptive law and construct an adaptive unknown input observer; Step 4: construct the state estimation error, measurement output error, fault estimation error and adaptive rate to obtain the error dynamic equation; The fifth step is to calculate the gain of the adaptive unknown input observer based on Lyapunov stability theory and linear matrix inequality method LMI to obtain the target observer; In the sixth step, the status data of the fan transmission system is input into the target observer to determine whether the fan transmission system is faulty.

2. The method for estimating a fault in a wind turbine transmission system according to claim 1, characterized in that: The mathematical model of the fan transmission system fault is established, including in: is the state vector, denote the input vector and the measured output vector respectively, is a bounded vector, representing unknown disturbances and modeling errors, is an additive fault, which represents the additive fault vector of the pitch system; A, B, C, B d ,B f is the constant coefficient matrix known to the system; Perturbation d1 is decoupled, d2 is not decoupled, in the mathematical model is a nonlinear vector function with a Lipschiz coefficient θ and satisfies the following Lipschiz relation 3. The method for estimating a fault in a wind turbine transmission system according to claim 1, characterized in that: The augmented system for constructing the mathematical model of the fan transmission system fault includes: Augmented mathematics for building mathematical models (1) in:

4. The method for estimating a fault in a wind turbine transmission system according to claim 3, characterized in that: The design of the adaptive law constructs an adaptive unknown input observer. Based on the augmented mathematical model (3), an adaptive unknown input observer is established: Respectively represent the intermediate transition state quantity, The estimator of and the estimator of E; K=K1+K2、 represents the gain matrix to be designed and calculated; The following Lipschiz relation is satisfied. According to (2), we can conclude that:

5. The method for estimating a fault in a wind turbine transmission system according to claim 1, characterized in that: The state estimation error, the measurement output error, the fault estimation error and the adaptive rate are constructed to obtain the error dynamic equation, including The state estimation error is Measurement output error Fault estimation error Adaptive rate in: represents the gain matrix to be designed, δ represents the adaptive learning law, which is a constant value; The error dynamic equation is constructed as 6. The method for estimating a fault in a wind turbine transmission system according to claim 1, characterized in that: The gain of the adaptive unknown input observer is calculated based on the Lyapunov stability theory and the linear matrix inequality method LMI to obtain the target observer, including The gains of the adaptive unknown input observer are L1, R, T, K2, K, according to Lyapunov stability theory and linear matrix inequality (LMI), calculate the gain Q, the symmetric positive definite matrix P, so that the following linear matrix inequality is satisfied: in: express Dimensional identity matrix After calculating the matrix L1, the values ​​of the gain matrices R, T, K2, K can be further calculated by the following formulas, thereby obtaining the fault estimator of the mathematical model (1); 7. A fan transmission system fault estimation device, characterized in that: include: The preprocessing module is used to collect the electrical parameter data of the pumping unit in real time online, perform preprocessing, vectorize the high-dimensional signal data, and construct the original data space of the electrical parameter data; A construction module is used to determine the k nearest sample points to each sample point through Euclidean distance to construct a local neighborhood; An integration module is used to take any sample point in the same local neighborhood, except the central sample point, as its local mutual representation neighborhood by the remaining sample points in the local neighborhood; the local neighborhood and the local mutual representation neighborhood obtained by each sample point through the construction module are integrated to form a multi-structure neighborhood; The reconstruction module is used to calculate the weight coefficients of the local linear relations of the multi-structured neighborhood in high-dimensional space by minimizing the local reconstruction error; The feature fusion module is used to fuse the features of multiple local neighborhood structures in a low-dimensional space, build a global reconstruction model, and obtain the final low-dimensional features by determining the optimal solution of the global reconstruction model; The fault diagnosis module is used to input the obtained low-dimensional feature results into the classifier and complete the fault diagnosis of the oil pump according to its classification results.

Images