A Fault Propagation Diagnosis Method for Complex Electromechanical Systems

Abstract

The present invention provides a fault propagation diagnosis method for complex electromechanical systems, mainly including: based on the mechanical, information, and electrical connection relationships existing among components within a complex electromechanical system, constructing a topological network model for the complex electromechanical system; according to historical fault data and the complex electromechanical system fault diagnosis manual, constructing a complex electromechanical system fault propagation network in combination with the topological network model; calculating the topological attributes and functional attributes of the complex electromechanical system, and then constructing a step-by-step fault propagation and diffusion model to obtain a fault propagation path set; finally, constructing a calculation method for the propagation intensity of different fault propagation paths to judge the fault propagation paths. The fault diagnosis method constructed by the present invention can effectively identify which components will be affected in a chain when a certain component fails. Furthermore, performing preventive maintenance on this series of components in advance can effectively prevent the spread of the fault, thereby avoiding the chain loss of economic property caused by this fault.

Description

Technical Field

[0001] The present invention relates to the technical field of complex electromechanical systems, and in particular to a fault propagation diagnosis method for complex electromechanical systems. Background Art

[0002] Complex electromechanical systems involve the mutual integration of multiple disciplines such as control, electrical, mechanical, telecommunications, systems, software, etc. They form a unified whole through mechanical, information, and electrical connections between components and rely on the joint action of each part to complete the functions set by the whole. Complex electromechanical systems with high precision, high performance, and high intelligence have been integrated into all aspects of production and life. For example, high-speed train systems, the aviation industry, and civilian appliances, etc. can all be regarded as the manifestation carriers of complex electromechanical systems. Complex electromechanical systems have extremely high complexity. Therefore, stricter standards are required for their reliability, safety, etc.

[0003] Although complex electromechanical systems have relatively high implementation standards and relatively strict system supervision systems in terms of maintenance and operation, the occurrence of internal system faults is still difficult to completely avoid. At the same time, due to the extremely high complex correlations between the components within the system, some minor faults or abnormal states are very likely to lead to the accumulation, propagation, and spread of system faults, thereby endangering the safe and reliable operation of the entire system. However, due to the high complexity of the internal correlation relationships of complex electromechanical systems and the lack of effective intelligent detection functions for most of the internal sub-components of the system, it will be very difficult to diagnose internal faults of complex electromechanical systems. Therefore, it is of great significance to study the fault propagation mechanism to ensure that the root cause of the fault can be traced in time after the fault occurs and the fault propagation link can be effectively controlled.

[0004] However, in existing research, it is generally based on the assumption that components are independent of each other, and the fault mechanism is analyzed from the functional perspective and expert experience, such as the Fault Tree Analysis (FTA), Hazard and Operability Study (HAZOP), Petri nets, etc. Although the occurrence mechanism of faults can be analyzed to a certain extent, the consideration of the mutual relationships between various parts within the system is lacking, and the multiple spread and propagation of faults are not considered either. Summary of the Invention

[0005] Embodiments of the present invention provide a fault propagation diagnosis method for complex electromechanical systems to solve the current situation of difficult fault diagnosis for complex electromechanical systems mentioned in the above background. In addition, through the fault diagnosis method proposed in this article, the propagation path of faults can be effectively identified, so that corresponding control measures can be taken to block the propagation and diffusion of faults and reduce the chain losses caused by local faults.

[0006] To achieve the above object, the present invention adopts the following technical solutions.

[0007] A fault propagation diagnosis method for complex electromechanical systems, comprising:

[0008] S1 Based on the connection relationships among the components within the complex electromechanical system, construct a topological network model for the entire complex electromechanical system; based on the fault diagnosis manual and fault history data of the complex electromechanical system, construct a fault propagation network model for the complex electromechanical system;

[0009] S2 Based on the topological network model and the fault propagation network model, calculate to obtain a fault propagation matrix for the complex electromechanical system; perform deformation and decomposition operations on the fault propagation matrix, and combine the hierarchical analysis method of triangular fuzzy numbers to calculate and obtain the topological attribute values of the complex electromechanical system;

[0010] S3 Based on the historical fault data, in combination with the design parameters and performance indicators of each component of the complex electromechanical system, fit the fault distribution of each component of the complex electromechanical system through the Weibull distribution model, and then solve each parameter in the Weibull distribution model through maximum likelihood estimation to calculate and obtain the functional attribute values of the complex electromechanical system;

[0011] S4 Based on the topological attribute values and functional attribute values of the complex electromechanical system, calculate to obtain the single - fault propagation coefficient; based on the topological network model and the fault propagation network model, construct a multi - fault propagation matrix through the fault propagation matrix; based on the multi - fault propagation matrix, calculate to obtain the multi - fault propagation coefficient;

[0012] S5 Based on the single - fault propagation coefficient and the multi - fault propagation coefficient, update the fault propagation matrix for the complex electromechanical system; based on the different propagation path sets caused by the failure of a single component in the updated fault propagation matrix for the complex electromechanical system, construct a fault propagation intensity calculation model, and based on the fault propagation intensity calculation model, obtain the propagation intensity of each propagation path in the different propagation path sets.

[0013] Preferably, step S1 includes:

[0014] S11 Based on the mechanical, information, and electrical connection relationships among the components within the complex electromechanical system, construct a topological network model for the entire complex electromechanical system

[0015] G(V,E,A)

[0016]

[0017] In the formula, G refers to a complex network structure diagram based on mechanical, information, and electrical connection relationships, V is the node set of the topological network model of the complex electromechanical system, E is the edge set in the topological network model of the complex electromechanical system, A is the adjacency matrix in the network model. If for any two points v i , v j∈V, if there exists e in the network ij ∈E, then a ij = 1, otherwise a ij = 0, e ij is the edge connecting node i and node j; N represents the total number of nodes in the network;

[0018] S12 Based on the fault diagnosis manual and fault history data of complex electromechanical systems, set the incidence matrix A f as the incidence matrix of the fault propagation network model, and obtain the relational expression

[0019]

[0020] S13 Through the formula

[0021] (I + A f ) p-2 ≠(I + A f ) p-1 =(I + A f ) p , p ≤ n - 1 (3)

[0022] Calculate the fault propagation reachability matrix M f [i, j] of the fault propagation network model, and construct a fault propagation network model for complex electromechanical systems; where, n is the number of nodes in the network model, I is the n - order identity matrix; p is a positive integer; if there exists a p that satisfies the relational expression (2), then the M of the fault propagation reachability matrix f =(I + A f ) p , and satisfies the formula

[0023]

[0024] Preferably, step S2 includes:

[0025] S21 Perform deformation and decomposition operations on the fault propagation reachability matrix M f [i, j] to obtain the fault propagation reachable set R(v i ) and the fault propagation antecedent set A(v i ); The fault propagation reachable set R(v i ) satisfies R(v i ) = {v j |v j ∈V, M f (i, j) = 1}, and the fault propagation antecedent set A(v i ) satisfies A(v i ) = {v j |v j ∈V, M f (j, i) = 1};

[0026] S22 traverses all elements of the fault propagation network model for complex mechatronic systems. If a certain node v i satisfies R(v i ) ∩ A(v i ) = R(v i ), then extract this certain node v i and move it into the first layer of the hierarchical structure. Repeat this process to obtain the elements of the first layer of the hierarchical structure;

[0027] S23 traverses the remaining elements of the fault propagation network model for complex mechatronic systems after performing sub-step S22, and moves the elements that meet the composite criteria into the second layer of the hierarchical structure to obtain the elements of the second layer of the hierarchical structure;

[0028] S24 repeats the process of sub-step S23 until all elements of the fault propagation network model for complex mechatronic systems have been hierarchically operated to obtain the hierarchical structure model of the fault propagation of complex mechatronic systems

[0029] Π(L) = [L 1 , L 2 , …, L s (5);

[0030] In the formula, Π represents the set of the fault propagation hierarchical structure of the complex mechatronic system, and L i represents the i-th layer of the fault propagation layer of the constructed complex mechatronic system;

[0031] S25 Based on the hierarchical structure model of the fault propagation of the complex mechatronic system, combined with historical maintenance data and empirical data, obtain the triangular fuzzy evaluation value T i = (θ i , m i , μ i ) of the fault propagation influence degree of the complex mechatronic system, and there is θ i ≤ m i ≤ μ i ; According to the triangular fuzzy evaluation value of the fault propagation influence degree, obtain the fuzzy judgment matrix Q

[0032]

[0033] In the formula, θ i represents the lower limit of the fault propagation influence degree of the i-th evaluation index in the hierarchical structure network; μ i represents the upper limit of the fault propagation influence degree of the i-th evaluation index in the hierarchical structure network; m i represents the middle value of the fault propagation influence degree of the i-th evaluation index in the hierarchical structure network;

[0034] S26 Transform equation (5) into a matrix P = (p with a diagonal of 1 through column transformation operations of the matrix ij ) n×n ; Based on the matrix P = (p ij ) n×n , and through equations

[0035]

[0036] Obtain the influence degree values of each layer on the fault propagation in the hierarchical structure model of fault propagation for complex mechatronic systems In the formula, p kj represents the value at the k-th row and j-th column of the matrix P = (p ij ) n×n ; p ik represents the value at the i-th row and k-th column of the matrix P = (p ij ) n×n ; p mk represents the value at the m-th row and k-th column of the matrix P = (p ij ) n×n ;

[0037] S27 If Then calculate and obtain the topological attribute value W of the complex mechatronic system through equations

[0038]

[0039] ; Otherwise, re-execute sub-step S22. i ;

[0040] Preferably, step S3 includes:

[0041] S31 Based on historical fault data, calculate and obtain all fault data F(v

[0042] F(v i ,r j )=(j - 0.3) / (N i + 0.4) (9)

[0043] of component v i through the mid-rank method empirical distribution i ,r j ), and draw a scatter plot between the operating mileage and the empirical function value of the fault distribution based on F(v i ,r j );

[0044] S32 Perform a fitting operation on the scatter plot between the operating mileage and the empirical function value of the fault distribution through the Weibull distribution model, and obtain the Weibull distribution function formula of each component of the complex mechatronic system in combination with the maximum likelihood estimation method

[0045] F i F(k)=1 - exp[-(k / α) β , k≥0 (10);

[0046] wherein, F i (k) represents the cumulative failure distribution function of component v when the operating mileage (or usage time) is k; k represents the operating mileage (or usage time); α and β represent the parameters of the Weibull distribution, which are obtained by the maximum likelihood estimation method; i

[0047] S33 calculates the functional attribute value of the complex mechatronic system based on Equation (10).

[0048] Preferably, step S4 includes:

[0049] S41 constructs a direct fault propagation ability coefficient matrix for the complex mechatronic system based on the topological attribute value and functional attribute value of the complex mechatronic system, in combination with the fault propagation network model

[0050]

[0051] S42 calculates the single - fault propagation coefficient matrix R of the complex mechatronic system based on Equation (11) through the formula

[0052] R 1 =DD - I 1 (12)

[0053] ; wherein, I 1 is the diagonal matrix of matrix DD; 1

[0054] S42 calculates the single - fault propagation coefficient R

[0055]

[0056] between components v i and v j of the complex mechatronic system through the formula 1 [i, j];

[0057] S43 calculates the r - step fault propagation coefficient matrix R of the complex mechatronic system through the formula

[0058] R r =(R r-1 +D f )(R 1 +D f ) - I r (14)

[0059] ; wherein, I r is r ​​Denote the matrix (R r-1 + D f )(R 1 + D f ); D f is a diagonal matrix with the functional attribute values of each element on the diagonal, and is obtained by the formula

[0060]

[0061] Calculated.

[0062] Preferably, step S5 includes:

[0063] S51 Obtain the fault propagation coefficient matrix R

[0064]

[0065] of the r-th step of the complex electromechanical system when the fault propagation coefficient ≥ 10 -8 through the formula r ;

[0066] S52 Based on formula (16), the fault direct propagation ability coefficient matrix and the fault single propagation coefficient matrix, through the formula

[0067] CR[i,j] = D[i,j] + R 1 [i,j] + R 2 [i,j] + … + R r [i,j] (17)

[0068] Calculate the fault propagation coefficient CR[i,j] of nodes v i and v j ;

[0069] S53 Based on the fault propagation coefficient CR[i,j] of nodes v i and v j , through the formula

[0070]

[0071] Calculate the propagation intensity of each propagation path in different propagation path sets.

[0072] As can be seen from the technical solutions provided by the embodiments of the present invention described above, the present invention provides a fault propagation diagnosis method for a complex electromechanical system, which mainly includes: based on the mechanical, information, and electrical connection relationships among components in the complex electromechanical system, constructing a topological network model for the complex electromechanical system; according to historical fault data and the fault diagnosis manual of the complex electromechanical system, combining with the topological network model to construct a fault propagation network for the complex electromechanical system; calculating the topological attributes and functional attributes of the complex electromechanical system, and then constructing a step-by-step fault propagation and diffusion model to obtain a set of fault propagation paths; finally, constructing a calculation method for the propagation intensity of different fault propagation paths to judge the fault propagation paths. The fault diagnosis method constructed by the present invention can effectively determine which components will be affected by a chain reaction when a certain component fails. Furthermore, performing preventive maintenance on this series of components in advance can effectively prevent the spread of faults, thereby avoiding the chain loss of economic property caused by this fault.

[0073] Additional aspects and advantages of the present invention will be given in part in the following description, and these will become apparent from the following description, or can be understood through the practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0074] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0075] Figure 1 It is a processing flowchart of a fault propagation diagnosis method for a complex electromechanical system provided by the present invention;

[0076] Figure 2 It is a logical block diagram of a fault propagation diagnosis method for a complex electromechanical system provided by the present invention;

[0077] Figure 3 It is a calculation flowchart of the topological attribute values of a complex electromechanical system for a fault propagation diagnosis method provided by the present invention;

[0078] Figure 4 It is a topological structure diagram of a bogie system for a fault propagation diagnosis method for a complex electromechanical system provided by the present invention;

[0079] Figure 5 It is a fault propagation diagram of a bogie system for a fault propagation diagnosis method for a complex electromechanical system provided by the present invention;

[0080] Figure 6The fault propagation hierarchical structure diagram of the bogie system for a fault propagation diagnosis method of a complex electromechanical system provided by the present invention;

[0081] Figure 7 The calculation results of the topological attributes of each component of the bogie system for a fault propagation diagnosis method of a complex electromechanical system provided by the present invention;

[0082] Figure 8 The schematic diagram of the cumulative fault occurrence probability values of each component of the bogie system for a fault propagation diagnosis method of a complex electromechanical system provided by the present invention at 120 million kilometers. Detailed implementation manners

[0083] The following details the implementation manners of the present invention. The examples of the implementation manners are shown in the accompanying drawings, where the same or similar reference numerals represent the same or similar elements or elements with the same or similar functions throughout. The implementation manners described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention and should not be construed as a limitation of the present invention.

[0084] Those skilled in the art of the present technology can understand that, unless specifically stated otherwise, the singular forms "a", "an", "the", and "said" used herein may also include the plural forms. It should be further understood that the term "including" used in the description of the present invention means the presence of the described features, integers, steps, operations, elements, and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or their groups. It should be understood that when we say an element is "connected" or "coupled" to another element, it can be directly connected or coupled to other elements, or there may also be intermediate elements. In addition, the "connection" or "coupling" used herein may include wireless connection or coupling. The phrase "and / or" used herein includes any unit and all combinations of one or more of the associated listed items.

[0085] Those skilled in the art of the present technology can understand that, unless otherwise defined, all terms (including technical terms and scientific terms) used herein have the same meaning as the general understanding of those of ordinary skill in the art to which the present invention belongs. It should also be understood that terms such as those defined in a general dictionary should be understood to have a meaning consistent with the meaning in the context of the prior art, and will not be interpreted with an idealized or overly formal meaning unless defined as here.

[0086] For the convenience of understanding the embodiments of the present invention, the following will further explain with several specific embodiments as examples in conjunction with the accompanying drawings, and each embodiment does not constitute a limitation on the embodiments of the present invention.

[0087] See Figure 1, the present invention provides a fault propagation diagnosis method for complex electromechanical systems, including the following processes:

[0088] S1 Based on the connection relationships among the components within the complex electromechanical system, construct a topological network model for the entire complex electromechanical system; based on the fault diagnosis manual and fault history data of the complex electromechanical system, construct a fault propagation network model for the complex electromechanical system;

[0089] S2 Based on the topological network model and the fault propagation network model, calculate and obtain a fault propagation matrix for the complex electromechanical system; perform deformation and decomposition operations on the fault propagation matrix, and combine the analytic hierarchy process method of triangular fuzzy numbers to calculate and obtain the topological attribute values of the complex electromechanical system;

[0090] S3 Based on the historical fault data, combined with the design parameters and performance indicators of each component of the complex electromechanical system, fit the fault distribution of each component of the complex electromechanical system through the Weibull distribution model, and then solve the parameters in the Weibull distribution model through maximum likelihood estimation to calculate and obtain the functional attribute values of the complex electromechanical system;

[0091] S4 Based on the topological attribute values and functional attribute values of the complex electromechanical system, calculate and obtain the single - fault propagation coefficient; based on the topological network model and the fault propagation network model, construct a multi - fault propagation matrix through the fault propagation matrix; based on the multi - fault propagation matrix, calculate and obtain the multi - fault propagation coefficient;

[0092] S5 Based on the single - fault propagation coefficient and the multi - fault propagation coefficient, update the fault propagation matrix for the complex electromechanical system; based on the set of different propagation paths caused by the failure of a single component in the updated fault propagation matrix for the complex electromechanical system, construct a fault propagation intensity calculation model, and based on the fault propagation intensity calculation model, obtain the propagation intensity of each propagation path in the set of different propagation paths.

[0093] Based on the research of the fault propagation mechanism, the present invention proposes a fault propagation diagnosis method for complex electromechanical systems, comprehensively considering the system fault propagation mechanism from two aspects of the topological attributes and functional attributes of the complex electromechanical system, so as to realize the tracing of the root cause of the fault and the discrimination of the possible propagation links, and finally provide an important basis for formulating practical control strategies after the fault occurs to reduce the chain losses caused by the fault.

[0094] In the preferred embodiment provided by the present invention, step S1 specifically includes the following processes:

[0095] According to the mechanical, information, and electrical connection relationships among the components within a complex electromechanical system, a topological network model for the complex electromechanical system is constructed. Among them, the component parts are regarded as the nodes in the network. If there are mechanical, information, or electrical connection relationships between the component parts, it is considered that there is an edge between the component parts. Accordingly, a topological network model G(V, E, A) is formed as follows:

[0096] G(V, E, A) (1)

[0097]

[0098] Among them, G refers to the complex network structure diagram based on mechanical, information, and electrical connection relationships; e ij refers to the edge between node i and node j, V is the set of nodes of the topological network model of the complex electromechanical system; E is the edge set in the topological network model of the complex electromechanical system; A is the adjacency matrix in the network model. If for any two points v i , v j ∈V, if there exists e ij ∈E in the network, then a ij = 1, otherwise a ij = 0, and the subscript N of a 1N etc. has the same meaning as N where j ≤ N, representing the total number of nodes in the network.

[0099] Furthermore, based on the actual historical fault data and the descriptions in the relevant fault diagnosis manuals, a fault propagation network model diagram based on the topological network model is constructed. Assuming that the association matrix A f is the association matrix of the fault propagation network model, then there are the following relationships:

[0100]

[0101] Furthermore, solve the fault propagation reachability matrix M f [i, j] in the fault propagation network model. The solution process is as follows:

[0102] (I + A f ) p-2 ≠(I + A f ) p-1 =(I + A f ) p , p ≤ n - 1 (3)

[0103] Among them, n is the number of nodes in the network model, I is the n-order identity matrix; p is a positive integer;

[0104] If there exists a p that satisfies the above relationship, then the M of the fault propagation reachability matrix f =(I + a f ) p . Its meaning is as follows:

[0105]

[0106] Furthermore, the process of calculating the topological properties of a complex mechatronic system is as follows. As Figure 3 shown, based on the fault propagation reachability matrix M f of the complex mechatronic system, according to matrix decomposition, the fault propagation reachable set and the fault propagation antecedent set of each node can be obtained. Among them, taking the node v i as an example, the definitions of its fault propagation reachable set R(v i ) and the fault propagation antecedent set A(v i ) are as follows:

[0107] (1) Fault propagation reachable set of v i : It refers to the set of nodes that can be ultimately affected by the node v i through the mutual correlation relationship between nodes after the node v i fails, that is, R(v i ) = {v j |v j ∈V, M f (i, j) = 1}.

[0108] (2) Fault propagation antecedent set of v i : It refers to the set of nodes that can ultimately affect the node v i through the mutual correlation relationship between nodes after other nodes fail, that is, A(v i ) = {v j |v j ∈V, M f (j, i) = 1}

[0109] Furthermore, based on the fault propagation reachable set R(v i ) and the fault propagation antecedent set A(v i ) of the node, the fault propagation model can be classified, and thus a hierarchical network model for fault propagation in complex mechatronic systems can be constructed to study the influence degree of elements at different layers on fault propagation. The specific method is as follows: First, traverse all elements in the network. If for a certain node v i in the network, there is R(v i ) ∩ A(v i ) = R(v i ), then the element v i is extracted and placed in the first layer of the hierarchical structure; furthermore, after traversing and finding all elements in the first layer, the elements belonging to the first layer are removed from the fault propagation matrix M fRemove it, and then repeat the search process in the previous step. Traverse the remaining elements after removing the elements included in the first layer, find the elements that meet the criteria, and place them in the second layer of the hierarchical network; and so on until all the elements in the fault propagation network are layered; finally, establish a hierarchical structure model for the fault propagation of complex mechatronic systems, denoted as:

[0110] Π(L) = [L 1 , L 2 , …, L s (5)

[0111] where s is the number of hierarchical structures obtained by dividing the fault propagation model of the complex mechatronic system according to the above layering steps, Π represents the set of fault propagation hierarchical structures of the complex mechatronic system; L i represents the i-th layer of the fault propagation layer of the constructed complex mechatronic system.

[0112] Furthermore, based on the fault maintenance history records and expert experience of the complex mechatronic system, use the TFN-AHP method to determine the importance value of the impact of each layer on faults in the hierarchical structure model of the complex mechatronic system. The specific method is as follows: First, determine the triangular fuzzy evaluation value of the fault propagation impact degree of each node in the fault propagation network of the complex mechatronic system. Taking the L i -th layer as an example, according to the historical maintenance records and expert experience, determine its triangular fuzzy evaluation value of the fault propagation impact degree as T i = (θ i , m i , μ i ), and there is θ i ≤ m i ≤ μ i . Furthermore, the fuzzy judgment matrix Q of the fault propagation network can be obtained as follows:

[0113]

[0114] In the formula, θ i represents the lower limit (minimum value) of the fault propagation impact degree of the i-th evaluation index in the hierarchical structure network (referring to the i-th layer structure in this embodiment); μ i represents the upper limit (maximum value) of the fault propagation impact degree of the i-th evaluation index in the hierarchical structure network (referring to the i-th layer structure in this embodiment); m i represents the most likely value (intermediate value) of the fault propagation impact degree of the i-th evaluation index in the hierarchical structure network (referring to the i-th layer structure in this embodiment).

[0115] Furthermore, transform the matrix Q into a matrix P = (p ij ) n×nThen, the influence degree values of each layer on the fault propagation in the hierarchical structure model of the complex electromechanical system fault propagation can be obtained. As follows:

[0116]

[0117] In the formula, p kj represents the value at the k-th row and j-th column of the matrix P = (p ij ); p n×n represents the value at the i-th row and k-th column of the matrix P = (p ik ); p ij represents the value at the m-th row and k-th column of the matrix P = (p n×n ). mk In the formula, p ij represents the value at the k-th row and j-th column of the matrix P = (p n×n ); p

[0118] Furthermore, combined with the triangular fuzzy evaluation values of the influence degree of fault propagation of each layer determined initially, a rationality test is carried out. Taking the L i -th layer as an example, if , then it is considered that the value is reasonable; otherwise, it is necessary to reselect the triangular fuzzy evaluation values of the influence degree of fault propagation of each layer with experts. If the influence degree values of each layer on the fault propagation all meet the rationality test after traversing, then the influence degree value W i of different nodes on the fault propagation in the complex electromechanical system fault propagation network model can be obtained. This value is the topological attribute value of the complex electromechanical system, and the calculation formula is as follows:

[0119]

[0120] Since the fault data records the faults that each component may cause due to its own properties or environmental factors (such as fault time, fault description, operating distance, fault consequences, etc.), the preferred embodiment provided by the present invention believes that the functional attributes of the complex electromechanical system are reflected in the fault data and can be represented by the fault distribution model.

[0121] The preferred embodiment provided by the present invention fits the historical fault data of the complex electromechanical system by the two-parameter Weibull distribution. According to the fault data, the middle-rank method is selected to estimate the empirical function of the fault distribution. Taking the component v i as an example, the empirical distribution formula of the middle-rank method is as follows:

[0122] F(v i , r j ) = (j - 0.3) / (N i + 0.4) (9)

[0123] In the formula, j represents the j-th number. For example, r jDenote the failure occurrence time (or mileage) corresponding to the j-th value after sorting the failure data generation time (or mileage) in ascending order. If j = 5, it is the occurrence time (or mileage) corresponding to the 5th failure data after sorting.

[0124] Among them, F(v i ,r j ) represents the value of the failure distribution empirical function corresponding to the j-th failure data among all the failure data of component v i ; r j represents the operating mileage (or usage time) corresponding to the j-th position after sorting all the failure data of this component in ascending order according to the operating mileage (or usage time); N i represents what v i component has. Based on this, a scatter plot between the operating mileage (or usage time) and the value of the failure distribution empirical function can be drawn according to the failure data.

[0125] Furthermore, use the two-parameter Weibull distribution to fit the image, and use the maximum likelihood estimation to obtain the Weibull distribution function of each component, so as to obtain the failure distribution function corresponding to the operating mileage (or time) of each component of the complex electromechanical system; taking v i as an example, the expression of its two-parameter Weibull distribution function is as follows:

[0126] F i (k) = 1 - exp[-(k / α) β , k ≥ 0 (10)

[0127] Among them, F i (k) represents the cumulative failure distribution function of component v i when the operating mileage (or usage time) is k; k represents the operating mileage (or usage time); α, β represent the parameters of the Weibull distribution, which can be obtained through maximum likelihood estimation.

[0128] Based on this, the cumulative failure occurrence probability of each component of the complex electromechanical system at a certain operating mileage (or usage time) can be obtained. This value is the calculated value of the functional attribute of the complex electromechanical system.

[0129] Comprehensively consider the calculated value of the topological attribute and the calculated value of the functional attribute of the complex electromechanical system, and combine the fault propagation network model to construct the direct fault propagation ability coefficient matrix D for the complex electromechanical system. The calculation formula is as follows:

[0130]

[0131] Based on this, the single-fault propagation system matrix R 1 of the complex electromechanical system can be obtained, as follows:

[0132] R 1 = DD - I 1 (12)

[0133] where I 1 is the diagonal matrix of matrix DD; then the single - fault propagation coefficient R i between components v j and v 1 [i, j] is as follows:

[0134]

[0135] Furthermore, based on the single - fault propagation coefficient, the multi - fault propagation coefficient value of a complex mechatronic system can be calculated, that is, the fault propagates from the fault node to another node through r connecting edges according to the association relationship between nodes in the network. Let R r represent the r - th step fault propagation coefficient matrix of the complex mechatronic system, then there is:

[0136] R r = (R r-1 + D f )(R 1 + D f ) - I r (14)

[0137] Accordingly, the fault propagation coefficient value after r - times of fault propagation can be obtained. Where r = 1, 2,..., representing 1 - time propagation, 2 - time propagation...; I r represents the diagonal matrix of matrix (R r-1 + D f )(R 1 + D f ); D f is a diagonal matrix with the functional attribute values of each element on the diagonal, as follows:

[0138]

[0139] According to the existing literature, when the fault propagation coefficient is less than 10 -8 , it is considered that the fault no longer propagates. Accordingly, for the r - th step propagation coefficient matrix, there is the following formula:

[0140]

[0141] Based on the fault direct - propagation ability coefficient matrix, the single - fault propagation coefficient matrix, and the multi - fault propagation coefficient matrix, the fault propagation matrix CR for a complex mechatronic system can be obtained. Assuming that there can be at most r - step propagation between v i and v j , then v i and vj The fault propagation coefficient CR[i,j] is as follows:

[0142] CR[i,j] = D[i,j] + R 1 [i,j] + R 2 [i,j] + … + R r [i,j] (17)

[0143] According to the above matrix, it can be found that a fault in a certain node will cause cascading failures in multiple nodes. Therefore, it is necessary to calculate the propagation intensity of different propagation paths. Taking v i as an example, the calculation of the propagation intensity of different propagation paths caused by the fault of this node is as follows:

[0144]

[0145] Finally, according to the method proposed by the present invention, the following functions can be realized: ① When a certain node fails, multiple fault propagation paths caused thereby can be determined; ② When a certain node shows a fault state, it can be judged by this method which node faults cause it.

[0146] The present invention also provides an embodiment for exemplarily showing the diagnosis process of the method provided by the present invention.

[0147] Since the high-speed train bogie system is a typical system in complex mechatronic systems, this project selects the high-speed train bogie system as a typical example to explain and illustrate the fault diagnosis method proposed by the present invention. Since the probability of faults in the bogie changes continuously with the operating mileage, in this embodiment, when the operating mileage is 120 million kilometers, the fault propagation mechanism of the system is studied.

[0148] As shown in Table 1, it shows the component composition of the high-speed train bogie system. Based on the mechanical, information, and electrical connection relationships between components, a topological structure diagram such as Figure 4 can be constructed; furthermore, based on historical fault data, fault diagnosis manuals, and expert engineering inspections, a fault propagation diagram such as Figure 5 can be constructed;

[0149] Table 1 Bogie system component list

[0150]

[0151] Accordingly, a connection matrix A f based on the fault propagation network can be constructed, and the reachability matrix M f of the fault propagation network can be obtained. Furthermore, by performing matrix decomposition on the reachability matrix, the fault propagation reachable set and the fault propagation antecedent set of each component can be obtained, thereby constructing a fault propagation hierarchical structure model based on the bogie system, such asFigure 6 As shown. Based on the fault history data and expert experience, the triangular fuzzy evaluation values of the fault propagation influence degree of each layer are set, and the TFN-AHP method is used to conduct the fault propagation influence degree and inspection of each layer. The final results are shown in Table 2. Furthermore, the fault propagation influence degree values of each node can be obtained, as Figure 7 shown, which are the calculated values of the topological attributes of the bogie system.

[0152] Table 2 Triangular fuzzy evaluation values of the fault propagation influence degree of each layer

[0153]

[0154] Based on the fault history data, the Weibull distribution is used to fit the function, and each parameter in the Weibull distribution is solved by maximum likelihood estimation. Finally, the Weibull distribution functions of each component shown in Table 3 below are calculated.

[0155] Table 3 Weibull distribution parameter values of each component of the bogie system

[0156]

[0157] Based on the fault distribution functions of each component, when k = 12 million kilometers, the cumulative fault occurrence probabilities of each component can be obtained, and the results are as Figure 8 shown. Using formulas (11)-(17), the fault propagation coefficients between any two nodes can be obtained, as shown in Table 4 below.

[0158] Table 4 Fault propagation coefficients between components of the bogie system

[0159]

[0160]

[0161]

[0162] In this embodiment, taking v 2 failure as an example, from the above table, it can be seen that there are two fault propagation paths here, namely v 2 →v 3 →v 4 and v 2 →v 5 , which are denoted as path1 and path2 respectively. Based on formula (18), the fault propagation intensities of different paths can be obtained, that is, FPS(path1) = 61.81%, FPS(path2) = 38.19%. It shows that when the train travels 12 million kilometers, when node v 2 fails, there is a 61.81% probability that it will follow v 2 →v3 →v 4 propagates, and there is a 38.19% probability that it will follow v 2 →v 5 .

[0163] In summary, the present invention provides a method for diagnosing fault propagation in a complex electromechanical system, mainly including: constructing a topological network model for the complex electromechanical system based on the mechanical, information, and electrical connection relationships among components in the complex electromechanical system; constructing a fault propagation network for the complex electromechanical system by combining the topological network model with historical fault data and a fault diagnosis manual for the complex electromechanical system; calculating the topological attributes and functional attributes of the complex electromechanical system, and then constructing a step-by-step fault propagation and diffusion model to obtain a set of fault propagation paths; finally, constructing a calculation method for the propagation intensity of different fault propagation paths to judge the fault propagation paths. The fault diagnosis method constructed by the present invention can effectively identify which components will be affected in a chain reaction when a certain component fails. Furthermore, performing preventive maintenance on this series of components in advance can effectively prevent the spread of the fault, thereby avoiding the chain loss of economic property caused by this fault.

[0164] Those of ordinary skill in the art can understand that the drawings are only schematic diagrams of an embodiment, and the modules or processes in the drawings are not necessarily essential for implementing the present invention.

[0165] From the description of the above embodiments, those skilled in the art can clearly understand that the present invention can be implemented by means of software plus a necessary general hardware platform. Based on such an understanding, the technical solution of the present invention, in essence, or the part that makes a contribution to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in various embodiments or some parts of the embodiments of the present invention.

[0166] Each embodiment in this specification is described in a progressive manner. For the same or similar parts among the embodiments, reference can be made to each other. Each embodiment focuses on the differences from other embodiments. In particular, for the device or system embodiments, since they are basically similar to the method embodiments, they are described relatively simply. For the relevant parts, reference can be made to the description of the method embodiments. The device and system embodiments described above are only illustrative. The units described as separate components may or may not be physically separated. The components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. A person of ordinary skill in the art can understand and implement it without creative work.

[0167] As described above, the above are only the preferred specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by those skilled in the art within the technical scope disclosed by the present invention should be covered by the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims

1. A fault propagation diagnosis method for complex electro - mechanical systems, characterized in that, it includes: S1 Based on the connection relationships among the components within the complex electro - mechanical system, construct a topological network model for the entire complex electro - mechanical system; Based on the fault diagnosis manual and fault history data of the complex electro - mechanical system, construct a fault propagation network model for the complex electro - mechanical system; S2 Based on the topological network model and the fault propagation network model, calculate and obtain the fault propagation matrix for the complex electro - mechanical system; Perform deformation and decomposition operations on the fault propagation matrix, and combine with the hierarchical analysis method of triangular fuzzy numbers to calculate and obtain the topological attribute values of the complex electro - mechanical system; S3 Based on the historical fault data, combine the design parameters and performance indicators of each component of the complex electro - mechanical system, fit the fault distribution of each component of the complex electro - mechanical system through the Weibull distribution model, and then solve each parameter in the Weibull distribution model through maximum likelihood estimation to calculate and obtain the functional attribute values of the complex electro - mechanical system; S4 Based on the topological attribute values and functional attribute values of the complex electro - mechanical system, calculate and obtain the single - fault propagation coefficient; based on the topological network model and the fault propagation network model, construct a multi - fault propagation matrix through the fault propagation matrix; based on the multi - fault propagation matrix, calculate and obtain the multi - fault propagation coefficient; S5 Based on the single - fault propagation coefficient and the multi - fault propagation coefficient, update the fault propagation matrix for the complex electro - mechanical system; based on the different propagation path sets caused by the failure of a single component in the updated fault propagation matrix for the complex electro - mechanical system, construct a fault propagation intensity calculation model, and based on the fault propagation intensity calculation model, obtain the propagation intensity of each propagation path in the different propagation path sets.

2. The fault propagation diagnosis method according to claim 1, characterized in that, step S1 includes: S11 Based on the mechanical, information, and electrical connection relationships among the components within the complex electro - mechanical system, construct a topological network model for the entire complex electro - mechanical system wherein, G refers to a complex network structure diagram based on mechanical, information, and electrical connection relationships, V is the node set of the topological network model of the complex mechatronic system, E is the edge set in the topological network model of the complex mechatronic system, A is the adjacency matrix in the network model. If for any two points v i , v j ∈V, if there exists e ij ∈E in the network, then a ij = 1; otherwise, a ij = 0. e ij is the connecting edge between node i and node j, and N represents the total number of nodes in the network; S12 Based on the fault diagnosis manual and fault history data of complex electromechanical systems, set up the incidence matrix A f as the incidence matrix of the fault propagation network model, and obtain the relational expression S13 Through the formula (I + A f ) p-2 ≠(I + A f ) p-1 =(I + A f ) p , p ≤ n - 1 (3) Calculate the fault propagation reachability matrix M of the fault propagation network model f [i, j], construct a fault propagation network model for complex mechatronic systems; where n is the number of nodes in the network model, I is the n-order identity matrix; p is a positive integer; if there exists a p that satisfies relation (2), then the M of the fault propagation reachability matrix f =(I + A f ) p , and satisfies the equation 3. The fault propagation diagnosis method according to claim 2, characterized in that, step S2 includes: S21 performs deformation and decomposition operations on the fault propagation reachability matrix M f [i, j] to obtain the fault propagation reachable set R(v i ) and the fault propagation antecedent set A(v i ); the fault propagation reachable set R(v i ) satisfies R(v i ) = {v j |v j ∈V, M f (i, j) = 1}, and the fault propagation antecedent set A(v i ) satisfies A(v i ) = {v j |v j ∈V, M f (j, i) = 1}; S22 traverses all elements of the fault propagation network model for complex mechatronic systems. If a certain node v i satisfies R(v i ) ∩ A(v i ) = R(v i ), then extract this certain node v i and move it into the first layer of the hierarchical structure. Repeat this process to obtain the elements of the first layer of the hierarchical structure; S23 Traverse the remaining elements of the fault propagation network model for the complex electro - mechanical system after executing sub - step S22, move the elements that meet the criteria into the second layer of the hierarchical structure to obtain the elements of the second layer of the hierarchical structure; S24 Repeat the process of sub - step S23 until all elements of the fault propagation network model for the complex electro - mechanical system have been hierarchically operated to obtain the hierarchical structure model of the fault propagation for the complex electro - mechanical system Π(L)=[L 1 ,L 2 ,…,L s (5); where Π represents the set of the fault propagation hierarchy of the complex mechatronic system, and L i represents the i-th layer of the fault propagation layer of the constructed complex mechatronic system; S25 Based on the hierarchical structure model of fault propagation for complex mechatronic systems, combined with historical maintenance data and empirical data, obtain the triangular fuzzy evaluation value T of the fault propagation influence degree of the complex mechatronic system i =(θ i ,m i ,μ i ), and there is θ i ≤m i ≤μ i ; According to the triangular fuzzy evaluation value of the fault propagation influence degree, obtain the fuzzy judgment matrix Q of the fault propagation network where, θ i represents the lower limit of the failure propagation influence degree of the i-th evaluation index in the hierarchical network; μ i represents the upper limit of the failure propagation influence degree of the i-th evaluation index in the hierarchical network; m i represents the intermediate value of the failure propagation influence degree of the i-th evaluation index in the hierarchical network; S26 transforms equation (5) into a matrix P = (p with a diagonal of 1 through column transformation operations of the matrix ij ) n×n ; Based on the matrix P = (p ij ) n×n , and through equation Obtain the influence degree values of each layer on the fault propagation in the hierarchical structure model of the fault propagation of complex mechatronic systems Wherein, p kj represents the value of the k-th row and j-th column in the matrix P = (p ij ) n×n ; p ik represents the value of the i-th row and k-th column in the matrix P = (p ij ) n×n ; p mk represents the value of the m-th row and k-th column in the matrix P = (p ij ) n×n ; If S27 then through the formula Calculate to obtain the topological attribute value W of the complex electromechanical system i ; Otherwise, re-execute sub-step S22.

4. The fault propagation diagnosis method according to claim 3, characterized in that, step S3 includes: S31 Based on the historical fault data, through the mid - rank method empirical distribution F(v i ,r j ) = (j - 0.3) / (N i + 0.4)(9) Calculate to obtain component v i All the fault data F(v i , r j ), and based on F(v i , r j ), plot a scatter diagram between the operating mileage and the empirical function values of the fault distribution; S32 Fit the scatter plot between the operating mileage and the empirical function values of the fault distribution through the Weibull distribution model, and combine with the maximum likelihood estimation method to obtain the Weibull distribution function formula of each component of the complex electro - mechanical system F i (k) = 1 - exp[-(k / α) β , k ≥ 0 (10); where F i (k) represents the cumulative failure distribution function of the v i component when the operating mileage or usage time is k; k represents the operating mileage or usage time; α and β represent the parameters of the Weibull distribution, which are obtained by the maximum likelihood estimation method; S33 Calculate and obtain the functional attribute values of the complex electro - mechanical system based on formula (10).

5. The fault propagation diagnosis method according to claim 4, characterized in that, Step S4 includes: S41 Based on the topological attribute values and functional attribute values of the complex mechatronic system, combined with the fault propagation network model, construct a fault direct propagation ability coefficient matrix for the complex mechatronic system S42 Based on Equation (11), through Equation R 1 = DD - I 1 (12) Calculate to obtain the single - fault propagation coefficient matrix R of a complex electromechanical system 1 ; where I 1 is the diagonal matrix of matrix D S42 Through Equation Calculate to obtain the failure single propagation coefficient R i between components v j and v 1 [i, j]; S43 Through Equation R r = (R r-1 + D f )(R 1 + D f ) - I r (14) Calculate and obtain the fault propagation coefficient matrix R of the complex electromechanical system at the r-th step r ; where I r represents the diagonal matrix of the matrix (R r-1 + D f )(R 1 + D f ); D f is the diagonal matrix with the functional attribute values of each element on the diagonal, through the formula Calculated and obtained.

6. The fault propagation diagnosis method according to claim 5, characterized in that, Step S5 includes: S51 Through Equation Obtain the fault propagation coefficient matrix \(R\) of the \(r\)th step of the complex mechatronic system when the fault propagation coefficient \(\geq 10\) -8 ; r ; S52 Based on Equation (16), the fault direct propagation ability coefficient matrix and the fault single - propagation coefficient matrix, through Equation CR[i,j] = D[i,j] + R 1 [i,j] + R 2 [i,j] + … + R r [i,j] (17) Calculate to obtain node v i and v j 's fault propagation coefficient CR[i, j]; S53 is based on the node v i and v j of the fault propagation coefficient CR[i,j], through the formula Calculate the propagation intensity of each propagation path in different propagation path sets.

Images