Fault propagation analysis method and system based on a cause-effect hierarchical topology network model

Abstract

The application relates to the technical field of electromechanical equipment fault analysis, and discloses a fault propagation analysis method and system based on a cause-effect hierarchical topology network model, which can more accurately determine the actual fault hierarchical propagation path of a source fault node under a specific fault mode. The method comprises the following steps: according to the propagation relationship of signals between longitudinal nodes during actual operation of electromechanical equipment, connecting the longitudinal nodes through directed edges to form a topology network, wherein the directed edges comprise horizontal cause-effect edges and longitudinal cause-effect edges, and establishing a fault propagation model based on the cause-effect hierarchical topology network; then, for a specific fault mode, starting from the fault disturbance suffered by a bottom element, calculating whether each potential directed edge can propagate the fault mode of the cause-effect fault node to the result fault node, and obtaining the actual hierarchical propagation path of the fault based on the cause-effect hierarchical topology network fault propagation model.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of electromechanical equipment failure analysis, and particularly relates to a failure propagation analysis method and system based on a cause-and-effect hierarchical topology network model. BACKGROUND

[0002] The functional, structural and hierarchical correlation relationship in modern complex electromechanical equipment is complex. Once a fault occurs somewhere, the fault will spread horizontally along the functional structural network and vertically evolve in a hierarchical manner, eventually causing a large area and serious fault to occur.

[0003] The users of the equipment mainly focus on the influence of the fault on the function and performance of the equipment, while the maintenance personnel of the equipment mainly focus on the bottom-level cause of the fault of the equipment. Therefore, it is necessary to clearly understand the correlation relationship between the bottom-level physical fault and the function and performance of the equipment (i.e. the hierarchical propagation rule of the electromechanical equipment fault), which has important value for the use and maintenance support of the electromechanical equipment. Through the research on the fault propagation of the electromechanical equipment, not only can the hierarchical and step-by-step propagation rule and characteristics of the fault be determined, but also a theoretical basis can be laid for fault prediction and fault diagnosis.

[0004] At present, the model-based method is usually used to study the fault propagation problem. The fault propagation model can be divided into three categories: based on cellular automata, based on graph theory and based on topology network. The cellular automata has the ability to model large-scale systems, and is a kind of net format dynamic model with spatial and temporal discretization and finite state. However, the model requires all bottom-level objects contained therein to comply with the same rule. For electromechanical equipment, it usually contains a large number of bottom-level elements of different types, and it is difficult to ensure that all elements comply with the same rule. Therefore, the fault propagation model based on cellular automata is not suitable for studying the fault propagation rule of electromechanical equipment. In addition, the fault propagation model based on cellular automata does not consider the influence of the cell structure on the state change of the unit, and is not suitable for studying the fault propagation problem from the overall perspective of the equipment. The fault propagation models based on graph theory such as multi-signal flow graph, fault tree, Petri net and Bayesian network are simple to use, and can intuitively describe the fault propagation path. However, they are limited in describing and analyzing research objects with high complexity, and are difficult to apply to complex equipment. At present, the design of the equipment is mainly based on the hierarchical thinking. The hierarchical modeling can clearly indicate the propagation relationship of the fault in each level of the electromechanical equipment. Although these models can intuitively describe the fault "horizontal propagation" relationship between nodes in each level through the directed graph on the basis of the hierarchical research object, they cannot intuitively describe the fault "vertical propagation" relationship between nodes in each level through the directed graph.

[0005] The basic idea of the topological network is to abstract each part of the electromechanical equipment as a node, and connect each node by using a line that can represent the structure and functional relationship of each part of the electromechanical equipment, and form a network through the nodes and the lines between the nodes. The topological network model is usually represented in the form of a directed graph network when it is constructed. The direction of the directed graph represents the signal transmission direction between each part of the electromechanical equipment, and also represents the fault propagation direction between each part. The most fundamental cause of the fault of the electromechanical equipment is usually the fault of the bottommost basic unit, i.e., the element, and the propagation caused by the fault. Therefore, it is necessary to study the propagation rule and influence of the fault by taking the element of the electromechanical equipment as the most basic unit. However, due to the large number of bottom elements of the equipment, the complexity of the mapping from the actual equipment to the topological network is extremely high, the network scale is large, and the model operation is cumbersome, which brings great difficulties to modeling and model verification.

[0006] The fault propagation in the complex electromechanical equipment has the following characteristics: 1. The fault can be propagated through the medium of function, performance or signal, and the hierarchical propagation rule of the fault in the electromechanical equipment can be better reflected from the signal level; 2. According to the fault propagation result, the fault propagation can be divided into fault information propagation and fault state propagation, and according to the characteristics of the fault, the hierarchical propagation form of the fault can be divided into direct propagation and evolution propagation; 3. The hierarchical fault propagation has four path modes, i.e., single-line propagation path, multi-line propagation path, jump propagation path and collapse propagation path; 4. The hierarchical fault propagation has strong uncertainty. Considering the advantages and disadvantages of various methods and the characteristics of the fault propagation in the complex electromechanical equipment, a method combining the topological network and hierarchical modeling is proposed. The electromechanical equipment is abstracted into a causal hierarchical topological network according to the functional structure relationship, and the fault modes and propagation parameters of each node of the network are associated and fused to form a fault propagation model based on the causal hierarchical topological network. A hierarchical fault propagation reasoning method of the electromechanical equipment on the global hierarchical topological network is proposed, and the quantitative reasoning of the hierarchical fault propagation path in the electromechanical equipment is realized, so as to provide an effective technical method for analyzing the hierarchical fault propagation path of the electromechanical equipment. SUMMARY

[0007] The purpose of the present application is to disclose a fault propagation analysis method and system based on a causal hierarchical topological network model, so as to more accurately determine the actual hierarchical fault propagation path of the source fault node under a specific fault mode.

[0008] In order to achieve the above purpose, the present application discloses a fault propagation analysis method based on a causal hierarchical topological network model, comprising:

[0009] Step S1, on the basis of analyzing the functional structure of the electromechanical equipment, the electromechanical equipment is vertically divided into levels, and the equipment entities corresponding in each level are abstracted into nodes to form a vertical node set;

[0010] Step S2, analyze the fault mode and influence corresponding to each longitudinal node to form a longitudinal node fault mode set;

[0011] Step S3, according to the propagation relationship of signals between each longitudinal node when the electromechanical equipment is actually working, connect each longitudinal node through a directed edge to form a topological network, the directed edge includes a horizontal causal edge and a longitudinal causal edge, and establish a fault propagation model based on the causal hierarchical topological network; the fault propagation model based on the causal hierarchical topological network is represented by a four-tuple composed of a hierarchy set, a node set, a fault mode set and a causal edge set, and each causal edge in the causal edge set is a directed edge between two nodes with a potential fault propagation path;

[0012] Step S4, for a specific fault mode, starting from the fault disturbance suffered by the bottom element, based on the fault strength and fault propagation ability of the cause fault node, the actual propagation coefficient corresponding to the fault strength and the tolerance ability of the effect fault node, calculate whether each potential directed edge can propagate the fault mode of the cause fault node to the effect fault node one by one, until all nodes based on the causal hierarchical topological network are traversed, thereby obtaining the actual hierarchical propagation path selected from each potential propagation path for any target fault mode in the fault propagation model based on the causal hierarchical topological network (in other words, the actual hierarchical propagation path of the fault is obtained based on the fault propagation model based on the causal hierarchical topological network).

[0013] Preferably, it is assumed that for a specific fault mode, the fault strength of the cause fault node A is ε A , the actual fault propagation coefficient between node A and the effect fault node B under the fault strength is RS AB , and the fault propagation ability of node A is PC A ; then the fault strength of the effect fault node B ε B can be represented by the following formula:

[0014] ε B = ε A RS AB PC A

[0015] If the fault strength of the effect fault node B exceeds its own fault tolerance ability, it is judged that the effect fault node B produces actual fault propagation; otherwise, it is judged that the fault does not propagate to the effect fault node B.

[0016] Preferably, the calculation of the fault strength of the node specifically includes:

[0017] Selecting the related signal parameters of the device entity output end corresponding to the node;

[0018] Obtaining the index XT = [XT-, XT+ XT is the lower limit value of the node state characteristic parameter when the node is normal, XT + XT is the upper limit value of the node state characteristic parameter when the node is normal;

[0019] The device is working in actual working conditions, the state characteristic parameter value XP of the node is acquired, if The fault intensity of the node is the deviation degree between the actual node state characteristic parameter value XP of the node and the node normal index XT when the device is working in actual working conditions, which can be calculated by the following formula:

[0020]

[0021] To achieve the above purpose, the application further discloses a fault propagation analysis system based on a causal hierarchical topology network model, which comprises a memory, a processor, and a computer program stored in the memory and capable of running on the processor, and the processor implements the steps corresponding to the above method when executing the computer program.

[0022] The application has the following beneficial effects:

[0023] 1. The fault propagation analysis method based on the causal hierarchical topology network model establishes a causal hierarchical topology network model in a longitudinal hierarchical manner according to the functional structural relationship when the mechanical and electrical equipment is actually working.

[0024] 2. The method can determine whether the fault is propagated between nodes from the aspects of the fault disturbance intensity of the bottom element and the node tolerance, can solve the problems of strong data dependency and insufficient engineering application of the current model from the probability angle, and can more accurately determine the actual fault hierarchical propagation path of the source fault node under a specific fault mode.

[0025] The application will be further described in detail below with reference to the accompanying drawings. BRIEF DESCRIPTION OF DRAWINGS

[0026] The accompanying drawings, which form a part of the present application, are included to provide a further understanding of the application, and are incorporated herein for purposes of illustrating the illustrative embodiments of the present application and the explanations provided herein, and are not intended as a limitation. In the drawings:

[0027] Figure 1 is a schematic diagram of a single-line propagation path mode of two adjacent nodes according to the embodiment of the present application.

[0028] Figure 2A flow chart of a fault hierarchical propagation algorithm in a single-line propagation path mode of two adjacent nodes disclosed in an embodiment of the present application.

[0029] Figure 3 A flow chart of a fault hierarchical propagation quantitative reasoning disclosed in an embodiment of the present application.

[0030] Figure 4 A functional and structural block diagram of a certain millimeter wave radar disclosed in an embodiment of the present application.

[0031] Figure 5 A fault propagation model diagram of a certain millimeter wave radar based on a causal hierarchical topology network disclosed in an embodiment of the present application. DETAILED DESCRIPTION

[0032] Embodiments of the present application are described in detail below with reference to the accompanying drawings, but the present application can be implemented in various different ways limited and covered by the claims.

[0033] Embodiment 1

[0034] This embodiment discloses a fault propagation analysis method based on a causal hierarchical topology network model.

[0035] The fault propagation model based on a causal hierarchical topology network mainly consists of a causal hierarchical topology network and model parameters closely related to model operation. Specifically, the fault propagation model based on a causal hierarchical topology network is described as follows:

[0036] (1) The hierarchical set L of the research object is L = {L1, L2,..., L i ,...,L n}. Wherein L1 is the bottom layer, L n is the highest layer. On the basis of analyzing the working principle of the research object and obtaining the structure and function block diagram / signal flow diagram, the research object is hierarchically divided from top to bottom, and then the fault hierarchical propagation behavior is studied from bottom to top. Referring to the specific examples below, generally, the bottom layer is the component layer, the highest layer is the device layer, and the middle layer is divided into functional circuit layer, sub-functional module layer and functional module layer according to the coverage of the function.

[0037] (2) The node set A = {A1, A2,..., A i ,...,A n}, wherein A1 to A n represent the node set in L1 level to L n level, A i is the node set in L i level, to represent the node set in L iThe 1st and the mth nodes in the layer. A node represents a device entity with complete function in the electronic device, and also represents the location where the fault occurs. Since the range attribute of the high-level node is greater than that of the low-level node, the number of the bottom layer node to the top layer node is gradually converging. The high-level node is composed of the relevant nodes in its next level, for example, the nodes in the 3rd level consisting of 5 nodes in the 2nd level The relationship between them can be described as: In order to quantitatively analyze the law and influence of fault hierarchical propagation, the following node attributes need to be defined:

[0038] ① The level L where the node is located i : Identify any node The level to which it belongs is L i Level.

[0039] ② The physical meaning of the node: represents the device entity part corresponding to the node.

[0040] ③ The reachable node set of the node The set of all nodes to which the fault may propagate after the node fails. Since there is no possibility of fault propagation between nodes connected by causal edges, the reachable node set of the node can be directly obtained from the causal hierarchical topology network.

[0041] ④ The fault propagation capacity PC of the node: The fault propagation capacity of the node represents the ability of the fault to propagate to the related nodes after the node fails. The fault propagation capacity of the node is usually represented by the importance of the node, and the fault propagation capacity of the node is related to the structural parameters of the node in the topology network.

[0042] ⑤ The state representation parameter x of the node and its value in normal state: The state representation parameter x of the node is a parameter based on the signal layer, and the state of the node can be judged by monitoring the relevant parameters in the signal of the node. The common state representation parameters in analog signals are usually amplitude, frequency, phase, etc. The specific node state representation parameter is determined according to the actual situation. The state representation parameter value y of the node in normal state x represents the standard value of the node state representation parameter under ideal working conditions without any fault disturbance. It can usually be obtained by testing or analyzing the signal after testing.

[0043] ⑥ Node fault tolerance The node fault tolerance represents the fault intensity that the node can tolerate. According to the idea of "load-capacity model" in cascading fault propagation, when the fault intensity of the node exceeds the fault tolerance of the node Right now When a node fails, its fault tolerance is divided into upper tolerance and lower tolerance. and lower tolerance

[0044] ⑦ Node state F or NF: The state of a node is either a fault state (represented by F) or a normal state (represented by NF). By identifying the state of a node, the path of fault propagation can be described more clearly.

[0045] (3) Fault mode set F = {F1, F2, ..., F i ,...,F n}, Where F is the set of failure modes of the entire research object, F i For the Lth i The set of failure modes in the hierarchy, For the Lth i The set of failure modes for the j-th node in the hierarchy. Failure modes are specific to a particular node; since nodes are hierarchical, failure modes are also hierarchical, and failures at different levels exhibit different failure modes. The propagation patterns and impacts of failures under the same node and the same failure mode differ depending on their attributes. Therefore, to quantitatively analyze the propagation patterns and impacts of failures, the following failure mode attributes are defined:

[0046] Fault strength Fault intensity represents the degree of deviation of a node from the normal signal under a fault disturbance in a given fault mode. The fault intensity of the source fault node is obtained by testing the signal state characterization parameters of the source fault node and calculating the degree of deviation from the normal signal state characterization parameters. The fault intensity of other nodes besides the source fault node is calculated by the fault hierarchical propagation algorithm.

[0047] Fault severity level Fault intensity level represents the severity of a fault under a given fault mode. Related research indicates that the fault propagation coefficient between nodes A and B is affected by the fault intensity of node A. Since it is difficult to accurately calculate the quantitative relationship between the fault propagation coefficient between nodes A and B and the fault intensity of node A, fuzzy mathematics is used to fuzzify the fault intensity into a fault intensity level, dividing the fault intensity into 5 levels: fs = {low, lower, moderate, higher, high}.

[0048] reachable fault mode set This indicates the possible failure modes that can occur in other nodes after a certain failure mode occurs in a certain node. The set of reachable failure modes for a certain failure mode of a node can be obtained through FMEA (Failure Mode and Effect Analysis).

[0049] Actual Fault Propagation Node Set and actual fault propagation mode set This indicates the actual fault modes and corresponding nodes that exist in other nodes after a certain fault mode occurs in a node, under the same fault mode attribute. By propagating the set of faulty nodes and the set of propagated fault modes, the hierarchical propagation path information of the fault can be obtained.

[0050] (4) Causal boundary set E={e i,j A causal edge is a directed edge connecting two nodes. When the two nodes connected by a causal edge are at the same level, the causal edge is called a "horizontal causal edge"; when the two nodes connected by a causal edge are at different levels, the causal edge is called a "vertical causal edge". A causal edge represents both the direction of fault propagation (i.e., from the "cause fault node" to the "effect fault node") and the signal correlation between the two nodes. The properties of a causal edge are defined as follows:

[0051] ① Inherent fault propagation coefficient S. This is related to the inherent signal correlation between two nodes and is an inherent attribute value between the nodes. The fault propagation coefficient between nodes can usually be obtained in three ways: derived through mathematical formulas based on the working principles of the corresponding equipment entities; given by relevant experience or expert knowledge; or obtained through a large amount of input-output experimental data.

[0052] ② Actual Fault Propagation Coefficient RS. The actual fault propagation coefficient represents the degree of influence between two nodes under actual fault disturbance conditions. It is related to both the inherent fault propagation coefficient and the fault intensity. If the influence of the fault intensity under a certain fault mode on the fault propagation coefficient between two nodes is represented by parameter I, then the actual fault propagation coefficient RS between the two nodes is RS = S × I.

[0053] (5) Based on the above description, the fault propagation model based on causal hierarchical topology network is represented by a quadruple, CHTNM={L,A,F,E}.

[0054] Furthermore, the relevant parameters of CHTNM are defined and calculated as follows:

[0055] like Figure 1 As shown, the source fault node A (which can also be called the cause fault node based on the causal relationship, and will not be elaborated on later) and the effect fault node B follow a single-line propagation path pattern. The only neighbor of node A is node B, and the only neighbor of node B is node A.

[0056] In research such as Figure 1 When quantifying the fault propagation relationship between two adjacent nodes in the single-line propagation path mode shown, it can be done using, for example... Figure 2The algorithm flow is shown to analyze whether the fault state at node A is propagated to node B. Whether the fault state at node A can be propagated to node B is related to three factors, which are: the fault intensity ε A of node A, the actual fault propagation coefficient RS AB between the two nodes under the fault intensity, and the fault propagation ability PC A of node A. The fault intensity ε B of node B can be expressed by the following formula:

[0057] ε B = ε A RS AB PC A

[0058] Node state representation parameter X: the relevant signal parameter of the output end of the device entity corresponding to the node is selected, and if the state of the node can be judged through the signal parameter, the signal parameter is called the node state representation parameter.

[0059] Node normal index XT: the node normal index is an index of the node in the normal state range under the ideal working condition of the device, which can be judged by the node state representation parameter. The node normal index is expressed as XT = [XT - , XT + ], wherein XT - is the lower limit value of the node state representation parameter when the node is normal, and XT + is the upper limit value of the node state representation parameter when the node is normal.

[0060] Actual node state representation parameter value XP: the value of the node state representation parameter of the node when the device works under the actual working condition, if then the node is faulty.

[0061] Node fault intensity ε: the fault intensity of a node refers to the deviation degree between the actual node state representation parameter value XP of the node and the node normal index XT when the device works under the actual working condition. It can be calculated by the following formula:

[0062]

[0063] Fault intensity level fs: the fault intensity level is a quantity for describing the size of the fault intensity, and the greater the fault intensity, the higher the fault intensity level.

[0064] Since the inherent failure propagation coefficient between nodes is an inherent attribute value between nodes, which is a quantitative value, the actual failure propagation coefficient RS between two nodes is mainly affected by the variable of node failure intensity ε. The influence parameter I is used to represent the influence of failure intensity ε on the actual failure propagation coefficient RS between nodes. Since the influence of failure intensity ε on the actual failure propagation coefficient RS between nodes has a fuzzy relationship, which is difficult to accurately describe by mathematical formula, the method of fuzzy mathematics is used to describe the influence of failure intensity ε on the actual failure propagation coefficient RS between nodes. Generally, the failure intensity of burst-type failure is divided into two levels, fs={none, high}, and the failure intensity of evolution-type failure is divided into five levels, fs={low, lower, moderate, higher, high}. By dividing the levels of failure intensity, the quantitative relationship between the actual failure propagation coefficient RS between nodes and the failure intensity ε can be converted into the quantitative relationship between the actual failure propagation coefficient RS between nodes and the failure intensity level fs. The actual failure propagation coefficient RS between two nodes can be calculated by the following formula:

[0065] RS=S×I (3)

[0066] The influence of failure intensity level fs on the actual failure propagation coefficient RS between nodes can be given by a large number of historical data or failure experts. It is assumed that under different failure intensity levels, the influence of node failure intensity ε on the actual failure propagation coefficient RS between nodes is I={I1, I2, I3, I4, I5}. Only the membership degree of actual failure intensity to each failure intensity level is required to obtain the influence parameter I, and the membership degree of actual failure intensity to each failure intensity level can be obtained by membership function, which is not described here.

[0067] Further, after obtaining the failure propagation rule between two adjacent nodes, the failure propagation rule in the global hierarchical topology network of devices can be obtained by further analyzing the failure hierarchical propagation reasoning, and the flow of failure hierarchical propagation quantitative reasoning is shown in Figure 3 .

[0068] The fault hierarchical propagation path information contains two aspects: one is the actual fault propagation node set after the fault propagation; the other is the actual fault propagation mode set after the fault propagation. Since the set nodes have only a "one-to-one" relationship between the fault modes, the elements in the actual fault propagation node set and the actual fault propagation mode set are consistent in number and are mapped to each other. Through the foregoing analysis, it can be seen that the nodes connected by the cause-effect edge have the possibility of fault propagation, but the fault propagation behavior does not necessarily occur. Whether the fault propagation behavior occurs is related to the fault propagation ability of the node, the actual fault propagation coefficient between nodes, the fault strength and the node tolerance. Among them, the fault propagation ability and tolerance of the node are inherent attribute values, and the actual fault propagation coefficient between nodes is also only affected by the fault strength variable. Therefore, the essence of fault hierarchical propagation path reasoning is to study the actual fault propagation node set and the actual fault propagation mode set under the same fault mode and different fault strengths. The basic flow of fault hierarchical propagation path reasoning is given as follows:

[0069] 1. Select the fault source node i and the fault mode f i , judge the fault type (sudden fault or evolving fault), and calculate the fault propagation ability PC(i) of the fault source node i.

[0070] 2. Calculate the fault strength ε i of the fault source node under the fault mode f i . When f i is a sudden fault, the fault strength is a constant value; when f i is an evolving fault, the fault strength of the node i will change with the strength of the fault disturbance suffered by the node i.

[0071] 3. Obtain the inherent fault propagation coefficient between nodes in the hierarchical topology network and the fault strength level boundary value, and calculate the actual fault propagation coefficient RS(ε i ) between nodes under the fault strength ε i .

[0072] 4. Calculate the fault strength ε j of the node j.

[0073] 5. Obtain the fault tolerance T i of the node j, and judge whether the node j has a state fault.

[0074] 6. Traverse all nodes in the hierarchical topology network according to the above steps, form the actual fault propagation node set rsa(i) and the actual fault propagation mode set rsf(i) of the fault source under the fault mode f i and the fault strength ε i . When the fault source is a sudden fault, only rsa(i) and rsf(i) are obtained; when the fault source is an evolving fault, rsa(i) and rsf(i) will change with the change of εi The changes will result in different rsa(i) and rsf(i).

[0075] To avoid ambiguity for those skilled in the art, the method of this embodiment is further illustrated below with a specific example:

[0076] Instance object is Figure 4 The image shows an onboard millimeter-wave radar used for distance measurement and collision avoidance. The vehicle's DC power supply is connected to the millimeter-wave radar. The external DC power supply is first regulated to 4V DC by the radar's DC power supply voltage regulator module. This 4V DC voltage is then further converted to 3.3V, 1.8V, and 3.3V. The converted 3.3V DC voltage is input to a crystal oscillator circuit to generate the fundamental frequency REF. REF is then synthesized by a frequency synthesizer to generate a charge pump voltage V. CO V CO The voltage-controlled oscillator (VCO) input to the RF transceiver generates radio frequency (RF). This RF is amplified and transmitted through the transmitting antenna, mixed with the received RF signal to obtain an intermediate frequency (IF), and then divided by a frequency divider before being input to a frequency synthesizer. Feedback adjusts the division value N of the frequency synthesizer, and the voltage V is adjusted by the value of N. CO The intermediate frequency (IF) is then divided into two channels for input amplification circuits, which are then input to a digital signal processor. The processed signal is further processed by the back-end and connected to the host computer software to display the working results of the millimeter-wave radar.

[0077] Hierarchical Division and Node Selection of a Millimeter-Wave Radar

[0078] Based on the analysis of the working principle of a millimeter-wave radar and the generation of its functional and structural block diagrams, the radar was further divided into vertical hierarchical levels and node selections. Using the functional and structural block diagrams, the millimeter-wave radar was divided into five levels from top to bottom: equipment layer, functional module layer, sub-functional module layer, functional circuit layer, and bottom-level component layer. Based on these levels, nodes were selected for each level according to the functional and structural block diagrams. As the node hierarchy attributes indicate, higher levels contain fewer nodes. The nodes in each level and their corresponding equipment entities are shown in Table 1.

[0079] Table 1. Vertical Hierarchical Division and Node Selection of a Millimeter-Wave Radar

[0080]

[0081]

[0082] After the hierarchical division, node selection and FMEA of a certain millimeter wave radar, further according to the function, signal flow relationship in the structure diagram of the millimeter wave radar, each node with a fault mode is connected through a causal directed edge to form a millimeter wave radar causal hierarchical topology network model diagram as shown in Figure 5 The causal hierarchical topology network fault propagation model diagram can not only clearly observe the causal analysis between nodes, but also can obtain related topology network structure parameters through the causal hierarchical topology network fault propagation model diagram, which lays a foundation for quantitative analysis of fault hierarchical propagation.

[0083] According to the causal hierarchical topology network diagram of the millimeter wave radar, the topology structure parameters of each node in the diagram can be obtained, and thus the fault propagation capability PC of each node is obtained. The fault propagation capability of each node is shown in Table 2.

[0084] Table 2 Fault propagation capability PC of each node

[0085] Node PC Node PC Node PC 2 0.107 11 0.071 20 0.143 3 0.143 12 0.143 21 0.107 4 0.107 13 0.107 22 0.143 5 0.143 14 0.143 23 0.179 6 0.107 15 0.143 24 0.071 7 0.071 16 0.071 25 0.071 8 0.071 17 0.107 26 0.071 9 0.107 18 0.214 27 0.071 10 0.214 19 0.143 28 0.071

[0086] On the basis of constructing the fault propagation model based on the causal hierarchical topology network, the parameters related to the model running are solved, and then the fault propagation path of all nodes in the hierarchical topology network is inferred. It should be pointed out that the fault hierarchical propagation reasoning analysis of this section for the case object mainly analyzes the propagation path rule of the primary power input intermittent circuit fault in the core module of the millimeter wave radar, i.e. the radio frequency transceiver module and the subsequent modules. Therefore, only the data of the corresponding nodes in the radio frequency transceiver module, the digital signal control and processing module, and the backend processing module are obtained.

[0087] The CHTNM related parameter acquisition process of the millimeter wave radar is as follows:

[0088] On the basis of completing the construction of the causal hierarchical topology network of the millimeter wave radar, the output end of the device entity corresponding to the nodes of the causal hierarchical topology network can be tested and analyzed by using an oscilloscope. Through the test, the state representation parameters of each node can be obtained, and then the corresponding normal indicators are given according to the expert experience. The related data are arranged as shown in Table 3.

[0089] Table 3 Test data table of inherent attribute values of each node

[0090]

[0091] The data in the CHTNM of the millimeter wave radar that cannot be obtained by testing need to be given by analyzing a large amount of historical data of the millimeter wave radar or by a fault expert. It is assumed that the fault tolerance of each node given by the related fault expert is shown in Table 4.

[0092] Table 4 Fault tolerance T of each node

[0093]

[0094]

[0095] The typical failure mode propagation path inference analysis of the millimeter wave radar is as follows:

[0096] The millimeter wave radar has sudden faults such as open circuit, short circuit, no output, and evolving faults such as intermittent fault and out-of-tolerance fault. Since the hierarchical propagation rules of these sudden faults in the millimeter wave radar are relatively certain, the hierarchical propagation of the evolving faults has strong uncertainty, and since the intermittent fault belongs to a typical evolving fault, this section selects an actual intermittent open circuit fault of the power supply in the millimeter wave radar as a typical failure mode to perform hierarchical propagation inference analysis.

[0097] In order to better study the hierarchical propagation rules of the intermittent open circuit fault of the power supply, the intermittent fault is divided into short intermittent fault (duration t < 100 ns), medium intermittent fault (101 ns < duration t < 500 μs) and long intermittent fault (501 μs < duration t < 5 ms) according to the duration of the intermittent fault. The intermittent fault is generally described by two characteristic parameters: intermittent fault active duration and intermittent fault inactive duration. The duration in which the device exhibits fault characteristics is called intermittent fault active duration, denoted by T a , and the duration in which the device exhibits normality is called intermittent fault inactive duration, denoted by T r .

[0098] On the basis of establishing the CHTNM of the millimeter wave radar and obtaining the inherent attribute parameters of the related nodes in the network, the propagation path rules of any fault can be studied. In this case, the propagation rules of the intermittent open circuit fault of the power supply in the radio frequency transceiver module and thereafter are mainly studied, so it is necessary to inject intermittent faults of different fault intensities of the same failure mode at the power supply. According to the classification of the duration of the intermittent fault, intermittent faults of different intensity types are injected. Since the equivalent fault injection of short intermittent fault is limited, 2 long intermittent faults and 3 medium intermittent faults are selected. On the basis of selecting the duration T of a single intermittent fault, the proportion Ta of the active duration of the intermittent fault also needs to be determined. According to the principle of equal interval selection, 4 proportions of the active duration of the intermittent fault are selected in this example, which are 25%, 50%, 75% and 90%. According to the duration of the intermittent open circuit fault at the power supply and the corresponding proportion of the active duration of the intermittent fault, 20 intermittent open circuit faults of the power supply shown in Table 5 are injected.

[0099] Table 5 Injection table of intermittent open circuit fault of power supply

[0100]

[0101] The intermittent fault of the input primary power line causes the intermittent fault of node 11. The fault intensity of node 11 under different fault codes is shown in Table 6.

[0102] Table 6 Fault intensity table of node 11 under different fault codes

[0103]

[0104]

[0105] After obtaining the fault intensity of node 11, the influence parameter I of different fault intensities on the actual fault propagation coefficient between node 11 and node 21 can be obtained according to the fuzzy mathematics method. Then, the actual propagation coefficient RS between node 11 and node 21 under different fault intensities can be calculated by combining the inherent propagation coefficient S between node 11 and node 21. The actual propagation coefficient obtained is shown in Table 7.

[0106] Table 7 Actual propagation coefficient table between node 11 and node 21 under fault intensity

[0107]

[0108] According to the formula ε 21 = ε 11 · RS 11-21 · PC 11 , the fault intensity of node 21 under different intermittent fault disturbance codes can be obtained. For example, when the fault code of the input primary power line is ④-①, the fault intensity of node 21 is ε 21 = 0.281 × 8.320 × 0.071 = 0.166. Since ε 21 = 0.166 > T 21 = 0.165, the intermittent fault of node 11 will propagate to node 21, causing intermittent fault of node 21. Similarly, the fault intensity of node 21 under intermittent fault disturbance of any other fault code can be obtained. By comparing the fault intensity of node 21 with the fault tolerance of node 21, it can be determined whether the intermittent fault of node 11 will propagate to node 21. Through the judgment, it is known that under the intermittent fault disturbance of fault codes ④-① ~ ④-④ and ⑤-① ~ ⑤-④, the intermittent fault of node 11 can propagate to node 21. Under the intermittent fault disturbance of other fault codes, the intermittent fault of node 11 cannot propagate to node 21.

[0109] After the fault strength of the node 21 is obtained, the fault strength of other nodes can be obtained according to the fault propagation ability of the node 21 and the actual fault propagation coefficient between the node 21 and other nodes connected with the node 21 by the causal edge under the fault strength (the inherent propagation coefficient is given by an expert, and the actual fault propagation coefficient between nodes can be obtained according to a fuzzy mathematics method), and then compared with the corresponding node fault tolerance, so that the propagation of the 20 intermittent faults in Table 6 in the nodes 22, 23, 24, 25, 26, 27, 9, 10, 6, 5, 7 and 1 can be judged. According to the corresponding calculation rules, under the intermittent fault disturbance of the fault numbers ④-①~④-④ and ⑤-①~⑤-④, the intermittent fault propagates to the nodes 22, 23, 24, 25, 26, 27, 9, 10, 6, 5, 7 and 1, and under the fault disturbance of other fault numbers, the intermittent fault does not propagate.

[0110] Embodiment 2

[0111] Corresponding to the above-mentioned embodiments, the present embodiment discloses a fault propagation analysis system based on a causal hierarchical topology network model, comprising a memory, a processor and a computer program stored in the memory and executable on the processor, and the processor implements the steps corresponding to the above-mentioned method when executing the computer program.

[0112] In summary, the fault propagation analysis method and system based on the causal hierarchical topology network model disclosed in each of the above-mentioned embodiments of the present application have the following beneficial effects:

[0113] 1. The fault propagation analysis method based on the causal hierarchical topology network model disclosed in the present application establishes a causal hierarchical topology network model in a longitudinal layered manner according to the functional structure relationship of the actual work of the electromechanical equipment. The topology network is closer to the physical structure of the electromechanical equipment, and the longitudinal layering can solve the problems of complex modeling and low model precision existing in the topology network model.

[0114] 2. The method of judging whether the fault propagates between nodes from the aspects of the fault disturbance strength of the bottom element and the node tolerance can solve the problems of strong data dependence and insufficient engineering application of the current method of judging whether the fault propagates from the probability angle, and can more accurately determine the actual fault hierarchical propagation path of the source fault node under a specific fault mode.

[0115] The above only describes the preferred embodiments of the present application and is not used to limit the present application. For those skilled in the art, the present application can have various modifications and changes. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A fault propagation analysis method based on a causal hierarchical topology network model, characterized in that, include: Step S1: Based on the analysis of the functional structure of the electromechanical equipment, the electromechanical equipment is divided into vertical levels, and the corresponding equipment entities in each level are abstracted into nodes to form a set of vertical nodes. Step S2: Analyze and obtain the failure modes and effects corresponding to each vertical node to form a set of vertical node failure modes; Step S3: Based on the signal propagation relationship between each longitudinal node during actual operation of the electromechanical equipment, connect each longitudinal node through directed edges to form a topology network. The directed edges include horizontal causal edges and vertical causal edges. Establish a fault propagation model based on the causal hierarchical topology network. The fault propagation model based on the causal hierarchical topology network is represented by a quadruple consisting of a hierarchy set, a node set, a fault mode set, and a causal edge set. Each causal edge in the causal edge set is a directed edge connecting two nodes with potential fault propagation paths. Step S4: For a specific fault mode, starting from the fault disturbance experienced by the underlying components, calculate one by one whether each potential directed edge can propagate the fault mode of the cause fault node to the effect fault node based on the fault strength and fault propagation capability of the cause fault node, the actual propagation coefficient corresponding to the fault strength, and the tolerance capability of the effect fault node, until all nodes based on the causal hierarchical topology network have been traversed, thereby obtaining the actual hierarchical propagation path selected from each potential propagation path for any target fault mode in the fault propagation model based on the causal hierarchical topology network.

2. The fault propagation analysis method based on a causal hierarchical topology network model according to claim 1, characterized in that, The hierarchical set of the quadruples is L = {L1, L2, ..., L...} i ,...,L n }, where L1 is the lowest level, L n The highest level; Node set A = {A1, A2, ..., A} i ,...,A n }, Where A1 to A n These represent levels L1 to L1 respectively. n The set of nodes in the hierarchy, A i For L i The set of nodes in the layer arrive L i The first and mth nodes in the layer; Fault mode set F = {F1, F2, ..., F...} i ,...,F n }, Where F is the failure mode set of the entire research object, F i For the Lth i The set of failure modes in the hierarchy, For the Lth i The set of failure modes for the j-th node in the hierarchy.

3. The fault propagation analysis method based on a causal hierarchical topology network model according to claim 1, characterized in that, Assumption: For a specific failure mode, the failure strength of the faulty node A is ε. A The actual fault propagation coefficient between node A and the faulty node B under this fault intensity is RS. AB And the fault propagation capability PC of node A. A The fault intensity ε of fault node B is then determined. B This can be expressed by the following formula: e B =e A RS AB PC A If the fault intensity of fault node B exceeds its own fault tolerance capability, it is determined that fault node B has actually propagated the fault; otherwise, it is determined that the fault has not propagated to fault node B.

4. The fault propagation analysis method based on a causal hierarchical topology network model according to claim 3, characterized in that, The calculation of node fault intensity specifically includes: Select the relevant signal parameters at the output end of the device entity corresponding to the node; The index XT = [XT-, XT] represents the node's normal operating condition under ideal conditions. + ], where XT- is the lower limit of the node state characterization parameter when the node is normal, XT + This represents the upper limit of the node state characterization parameter when the node is normal. When the equipment is working under actual conditions, it acquires the node's state representation parameter value XP. If the node fails, it is determined that a fault has occurred. Correspondingly, the fault severity of this node refers to the deviation between the actual node state characteristic parameter value XP and the normal node index XT when the equipment is operating under actual conditions. This deviation can be calculated using the following formula:

5. A fault propagation analysis system based on a causal hierarchical topology network model, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 4.

Images