CN109116219B - Distributed fault diagnosis method for circuit system - Google Patents

Abstract

The invention discloses a distributed fault diagnosis method of a circuit system, which comprises the following steps: modeling the circuit system, and configuring the causal relationship among all parameters to obtain a global diagnosis bonding diagram model of the circuit system; taking each sensor in the circuit system as a basic unit, taking the measured value of the sensor as the local output of the circuit system, extracting the minimum subsystem based on each sensor from the circuit system, and respectively obtaining the local diagnosis bonding diagram model of each minimum subsystem; obtaining a distributed analytic redundancy relation according to a local diagnostic bond map model of a minimum subsystem, and analyzing the distributed analytic redundancy relation to obtain a distributed fault feature matrix; and carrying out fault diagnosis according to the distributed fault feature matrix. The invention effectively improves the fault isolation performance under the conditions of single fault and multiple faults, reduces the complexity of subsequent fault element identification, and improves the diagnosis speed and the diagnosis precision of the faults.

Description

Distributed fault diagnosis method for circuit system

Technical Field

The invention relates to the field of fault diagnosis of a circuit system, in particular to a distributed fault diagnosis method of the circuit system.

Background

In modern industrial production, many complex electrical and mechanical circuit systems are generated, such as complex integrated circuit systems, electromechanical circuit systems, hydraulic circuit systems, and the importance of fault diagnosis and real-time monitoring of the circuit systems is increasingly highlighted in order to improve the safety and reliability of the circuit systems.

With the increase of the complexity of the circuit system, the difficulty of fault diagnosis of the circuit system is increased, and a common fault diagnosis method is to model the circuit system based on a bonding diagram to obtain a global diagnosis bonding diagram model of the circuit system, directly obtain a global fault feature matrix of the circuit system according to the global diagnosis bonding diagram model of the circuit system, and then perform fault diagnosis on the circuit system according to the global fault feature matrix. However, as the complexity of the circuit system increases, the following main problems occur in this method of global fault diagnosis: firstly, under the condition of single fault, because some elements in a circuit system have the same fault characteristics, fault isolation cannot be realized; second, under multiple fault conditions, fault isolation cannot be achieved because the fault characteristics of some components in the circuitry cancel or mask each other. Meanwhile, the complexity of subsequent fault element identification is affected by the performance of fault isolation, and the more elements are included in a possible fault set obtained by fault isolation, the higher the complexity of fault element identification is, and the lower the identification accuracy is.

Disclosure of Invention

In order to overcome the defects in the prior art, the invention provides a distributed fault diagnosis method for a circuit system, which effectively improves the fault isolation performance under the conditions of single fault and multiple faults, reduces the complexity of subsequent fault element identification, and improves the fault diagnosis speed and the fault diagnosis precision.

In order to achieve the purpose, the invention adopts the following technical scheme that:

a distributed fault diagnosis method of a circuit system, comprising the steps of:

s1, modeling the circuit system, and configuring the causal relationship among all parameters to obtain a global diagnosis bonding diagram model of the circuit system;

s2, taking each sensor in the circuit system as a basic unit, taking the measured value of the sensor as the local output of the circuit system, extracting the minimum subsystem based on each sensor from the circuit system, and respectively obtaining the local diagnosis bond map model of each minimum subsystem;

s3, obtaining a distributed analytic redundancy relationship according to the local diagnostic bond map model of the minimum subsystem, and analyzing the distributed analytic redundancy relationship to obtain a distributed fault feature matrix;

and S4, performing fault diagnosis according to the distributed fault feature matrix.

In step S2, the manner of obtaining the local diagnostic bond map model of the smallest subsystem is specifically as follows: taking a certain sensor of the circuit system as an output sensor, taking the measured value of the output sensor as the local output of the circuit system, namely taking the measured value of the output sensor as the output signal of the minimum subsystem based on the output sensor, starting from the output sensor to the end of another sensor, taking the measured value of the sensor which is the end sensor as the input signal of the minimum subsystem, and obtaining a local diagnostic bonding map model of the minimum subsystem based on the output sensor; in a path from the output sensor to the termination sensor, a parameter associated with the measured value of the output sensor is derived from causal relationships between various parameters on a global diagnostic bond map model of the circuitry, and the parameter associated with the measured value of the output sensor is used as a parameter of a local diagnostic bond map model based on a smallest subsystem of the output sensor.

In step S1, the parameters of the global diagnostic bond map model of the circuit system include: potential source U_M＝V_inResistive element { R₁，R₂，R₃}, capacitive element { C₁，C₂，C₃}, voltage sensor { De₁，De₂}, current sensor Df, switch k₁(ii) a Wherein, the potential source V_inIs the input voltage of the system; resistance R₁And a capacitor C₁Series connected, resistance R₂And a capacitor C₂Series connected, resistance R₃And a capacitor C₃Are connected in series; will resistance R₁And a capacitor C₁Formed series circuit, resistor R₂And a capacitor C₂Formed series circuit, resistor R₃And a capacitor C₃The formed series circuits are connected in parallel; voltage sensor De₁Connected in parallel to a capacitor C₂Both ends of (a); voltage sensor De₂Connected in parallel to a capacitor C₃Both ends of (a); current sensor Df and resistor R₁Potential source V_inAre connected in series; switch k₁And a resistor R₃Are connected in series.

Using current sensor Df as basic unit from current sensor Df to voltage sensor De₁Terminate, extract based on current sensor DfMinimum subsystem

The parameters of the local diagnostic bond map model of the minimal subsystem include: potential source

Resistive element { R₁，R₂}, capacitive element C₁A flow sensor Df;

with a voltage sensor De₁As a basic unit, a slave voltage sensor De₁To start, to the voltage sensor De₂And the current sensor Df is terminated, and the voltage sensor De is extracted₁Minimum subsystem of

The minimum subsystem

The parameters of the local diagnostic bond map model of (a) include: potential source

Resistive element { R₂，R₃}, capacitive element { C₁，C₂}, a voltage sensor De₁Switch k₁；

With a voltage sensor De₂As a basic unit, a slave voltage sensor De₂To start, to the voltage sensor De₁Terminate, extract based on the voltage sensor De₂Minimum subsystem of

The minimum subsystem

The parameters of the local diagnostic bond map model of (a) include: potential source

Resistive element R₃Capacitive element C₃D, a voltage sensor De₂Switch k₁。

In step S3, the method includes the following steps:

s31, obtaining a distributed analytical redundancy relationship according to the local diagnostic bond map models of the three minimum subsystems and the causal relationship among all parameters in the local diagnostic bond map models;

in the above manner, the minimum subsystem is obtained

Distributed analytic redundancy relationship DARR₁Comprises the following steps:

minimum subsystem

Distributed analytic redundancy relationship DARR₂Comprises the following steps:

minimum subsystem

Distributed analytic redundancy relationship DARR₃Comprises the following steps:

wherein, in DARR₁In (e)_1-1＝V_in，e_1-8＝De₁，e_1-1And e_1-8Respectively representing minimum subsystems

Of two inputs, Df^*Representing a minimum subsystem

The measured value of the electric current sensor Df;

in DARR₂In (e)_2-1＝V_in-R₁Df，e_2-10＝De₂，e_2-1And e_2-10Respectively representing minimum subsystems

The value of the two input voltages of (a),

representing a minimum subsystem

Voltage sensor De₁A measured value of (a);

in DARR₃In (e)_3-1＝De₁，e_3-1Representing a minimum subsystem

The value of the input voltage of (a),

representing a minimum subsystem

Voltage sensor De₂A measured value of (a);

s32, according to the distributed analytic redundancy relation DARR of the three minimum subsystems₁、DARR₂、DARR₃Obtaining a distributed fault characteristic matrix, wherein the distributed fault characteristic matrix is as follows:

θ/r	dr1	dr2	dr3
				R1	1	0	0
R 2	1	1	0
				R3	0	k1	k1
C1	1	0	0
				C 2	0	1	0
C 3	0	0	k1

wherein dr is₁、dr₂、dr₃Respectively corresponding to the minimum subsystem

Residual error of (1), R₁、R₂、R₃Resistive elements, C, being circuitry₁、C₂、C₃Capacitive elements of the circuit system, wherein 1 in the distributed fault signature matrix indicates that the residual error of the column is sensitive to the elements of the row, 0 in the distributed fault signature matrix indicates that the residual error of the column is insensitive to the elements of the row, and k in the distributed fault signature matrix₁The switch k represents the sensitivity of the residual of the column to the elements of the row₁Determining;

due to the minimal subsystem

Distributed analytic redundancy relationship DARR₁The element contained in (A) has R₁、R₂、 C₁D is therefore dr₁To the element R₁、R₂、C₁Sensitive, so in the distributed fault signature matrix, the drth₁R of₁Line, R₂Line, C₁The value of the line is 1 at the dr₁R of₃Line, C₂Line, C₃The value of the row is 0;

due to the minimal subsystem

Distributed analytic redundancy relationship DARR₂The element contained in (A) has R₂、R₃、 C₂D is therefore dr₂To the element R₂、R₃、C₂Is sensitive and R is₃And also with a switch k₁Correlation, therefore dr₂To the element R₃Is sensitive to by the switch k₁Determining, therefore, in the distributed fault signature matrix, the drth₂R of₂Line, C₂The value of the line is 1, dr₂R of₃The value of a row is k₁Dr th₂R of₁Line, C₁Line, C₃The value of the row is 0;

due to the minimal subsystem

Distributed analytic redundancy relationship DARR₃The element contained in (A) has R₃、C₃D is therefore dr₂To the element R₃、C₃Is sensitive and R is₃、C₃Are all connected with switch k₁Correlation, therefore dr₃To the element R₃、C₃Is sensitive to by the switch k₁Determining, therefore, in the distributed fault signature matrix, the drth₃R of₃Line, C₃The value of a row is k₁Dr th₂R of₁Line, R₂Line, C₁Line, C₂The row has a value of 0. .

The invention has the advantages that:

(1) the modeling method of the bonding diagram clearly and quantitatively describes the physical relationship among the elements of the circuit system, and is convenient for quickly and accurately positioning the specific fault element during fault diagnosis.

(2) The invention decomposes a complex circuit system into a plurality of simple and independent minimum subsystems, thereby reducing the complexity of fault isolation.

(3) Under the condition of single fault, the fault isolation of the distributed fault diagnosis method is more accurate.

(4) Under the condition of multiple faults, the distributed fault diagnosis method improves the isolation capability of the multiple faults and simultaneously reduces the complexity of fault identification after fault isolation.

(5) Under the conditions of single fault and multiple faults, the distributed fault diagnosis method effectively improves the fault isolation performance, reduces the complexity of subsequent fault element identification, and improves the diagnosis speed and the diagnosis precision of the faults.

Drawings

FIG. 1 is a flow chart of the method of the present invention.

FIG. 2 is a global diagnostic bond map model of the circuit system of the present invention.

FIG. 3(1) is a diagram of the minimal subsystem of the present invention

Local diagnostic bond map model.

FIG. 3(2) is a diagram of the minimal subsystem of the present invention

Local diagnostic bond map model.

FIG. 3(3) is a diagram of the minimal subsystem of the present invention

Local diagnostic bond map model.

FIG. 4(a) shows an injection resistor R according to the present invention₁Global analytical redundancy relationship GARR at fault₁The waveform of (2).

FIG. 4(b) shows an injection resistor R according to the present invention₁Global analytical redundancy relationship GARR at fault₂The waveform of (2).

FIG. 4(c) shows an injection resistor R according to the present invention₁Global analytical redundancy relationship GARR at fault₃The waveform of (2).

FIG. 4(d) shows an injection resistor R according to the present invention₁Distributed analytic redundancy relationship DARR upon failure₁The waveform of (2).

FIG. 4(e) shows an injection resistor R according to the present invention₁Distributed analytic redundancy relationship DARR upon failure₂The waveform of (2).

FIG. 4(f) shows an injection resistor R according to the present invention₁Distributed analytic redundancy relationship DARR upon failure₃The waveform of (2).

FIG. 5(a) shows an injection resistor R according to the present invention₁And a capacitor C₃Global analytical redundancy relationship GARR during simultaneous failure₁The waveform of (2).

FIG. 5(b) shows an injection resistor R according to the present invention₁And a capacitor C₃Global analytical redundancy relationship GARR during simultaneous failure₂The waveform of (2).

FIG. 5(c) shows an injection resistor R according to the present invention₁And a capacitor C₃Global analytical redundancy relationship GARR during simultaneous failure₃The waveform of (2).

FIG. 5(d) shows an injection resistor R according to the present invention₁And a capacitor C₃Distributed analytic redundancy relationship DARR during simultaneous failure₁The waveform of (2).

FIG. 5(e) shows an injection resistor R according to the present invention₁And a capacitor C₃Distributed analytic redundancy relationship DARR during simultaneous failure₂The waveform of (2).

FIG. 5(f) shows an injection resistor R according to the present invention₁And a capacitor C₃Distributed analytic redundancy relationship DARR during simultaneous failure₃The waveform of (2).

Detailed Description

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

In this embodiment, the advantages of the distributed fault diagnosis method of the present invention are highlighted by comparing the fault diagnosis effects based on the global fault feature matrix and the distributed fault feature matrix.

As shown in fig. 1, a distributed fault diagnosis method for a circuit system includes the following steps:

and S1, taking a complex circuit system as an example, modeling the circuit system, configuring the causal relationship among all parameters, and obtaining a global diagnosis bonding diagram model of the circuit system.

And S2, taking each sensor in the circuit system as a basic unit, taking the measured value of the sensor as the local output of the circuit system, extracting the minimum subsystem based on each sensor from the circuit system, and respectively obtaining the local diagnosis bonding diagram model of each minimum subsystem.

And S3, obtaining a distributed analytic redundancy relationship according to the local diagnostic bond map model of the minimum subsystem, and analyzing the distributed analytic redundancy relationship to obtain a distributed fault feature matrix.

And S4, performing fault diagnosis according to the distributed fault feature matrix.

As shown in fig. 2, in step S1, the parameters in the global diagnostic bond map model of the circuit system include: potential source U_M＝V_inResistive element { R₁，R₂，R₃}, capacitive element { C₁，C₂，C₃}, voltage sensor { De₁，De₂}, current sensor Df, switch k₁. Wherein, the potential source V_inIs the input voltage of the system; resistance R₁And a capacitor C₁Series connected, resistance R₂And a capacitor C₂Series connected, resistance R₃And a capacitor C₃Are connected in series; will resistance R₁And a capacitor C₁Formed series circuit, resistor R₂And a capacitor C₂Formed series circuit, resistor R₃And a capacitor C₃The formed series circuits are connected in parallel; voltage sensor De₁Connected in parallel to a capacitor C₂Both ends of (a); voltage sensor De₂Connected in parallel to a capacitor C₃Both ends of (a); current sensor Df and resistor R₁Potential source V_inAre connected in series; switch k₁And a resistor R₃Are connected in series.

In the global diagnosis bond-graph model, directional arrows represent keys, 0 and 1 represent two nodes, namely a common potential node and a common current node, wherein the potentials on the keys connected with the common potential node are the same, and the currents on the keys connected with the common current node are the same. In thatDescribing constraints for nodes and parameters in the diagnostic bond graph, e_iIndicating the potential, i.e. voltage, f, at the bond numbered i_iIndicating a flow, i.e., current, over the key numbered i. In this example, V_inAnd R₁Df are connected in series, the currents flowing through the three elements are the same, and a common current junction 1 is used in a bonding diagram₁Connection, V_inAnd C₁Parallel connection, using a common potential junction 0 in a bonding diagram₂Is connected to R₂In series with the front-end circuit, the same current flows as the front-end circuit, and a common current junction 1 is used in the bonding diagram₃Connected to the front end, C₂、De₁Connected in parallel with the front-end circuit and at the same voltage, and the common potential junction 0 is used in the bonding diagram₄To the front end, R₃、k₁The same current flows in series with the front end, and a common current junction 1 is used in the bonding diagram₅Connected to a front end, De₂、C₃Parallel connection with front end, same voltage, common potential junction 0 in bonding diagram₆Connected to the front end, thereby forming a bond map model of the circuitry used in the present invention.

In the global diagnostic bond map model of the circuit system, the constraint relationship between each node and key is as follows:

in step S2, the manner of obtaining the local diagnostic bond map model of the smallest subsystem is specifically as follows: taking a certain sensor of the circuit system as an output sensor, taking the measured value of the output sensor as the local output of the circuit system, namely taking the measured value of the output sensor as the output signal of the minimum subsystem based on the output sensor, starting from the output sensor to the end of another sensor, taking the measured value of the sensor which is the end sensor as the input signal of the minimum subsystem, and obtaining a local diagnostic bonding map model of the minimum subsystem based on the output sensor; in a path from the output sensor to the termination sensor, a parameter associated with the measured value of the output sensor is derived from causal relationships between various parameters on a global diagnostic bond map model of the circuitry, and the parameter associated with the measured value of the output sensor is used as a parameter of a local diagnostic bond map model based on a smallest subsystem of the output sensor.

As shown in fig. 3(1), the current sensor Df is used as a basic unit from the current sensor Df to the voltage sensor De₁Terminate, extract the smallest subsystem based on the current sensor Df

The parameters of the local diagnostic bond map model of the minimal subsystem include: potential source

Resistive element { R₁,R₂}, capacitive element C₁Flow sensor Df.

Minimum subsystem

The local diagnostic bond map model of (2) wherein the constraint relationship between each node and bond is as follows:

shown in FIG. 3(2), with a voltage sensor De₁As a basic unit, a slave voltage sensor De₁To start, to the voltage sensor De₂And the current sensor Df is terminated, and the voltage sensor De is extracted₁Minimum subsystem of

The minimum subsystem

The parameters of the local diagnostic bond map model of (a) include: potential source

Resistive element { R₂,R₃}, capacitive element { C₁,C₂}, a voltage sensor De₁Switch k₁。

Minimum subsystem

The local diagnostic bond map model of (2) wherein the constraint relationship between each node and bond is as follows:

as shown in the graph of figure 3(3),

with a voltage sensor De₂As a basic unit, a slave voltage sensor De₂To start, to the voltage sensor De₁Terminate, extract based on the voltage sensor De₂Minimum subsystem of

The minimum subsystem

The parameters of the local diagnostic bond map model of (a) include: potential source

Resistive element R₃Capacitive element C₃D, a voltage sensor De₂Switch k₁。

Minimum subsystem

The local diagnostic bond map model of (2) wherein the constraint relationship between each node and bond is as follows:

in step S3, the method includes the following steps:

and S31, obtaining a distributed analytic redundancy relation through a causal path overlay method according to the local diagnostic bond map model of the minimum subsystem and the causal relation among all parameters in the local diagnostic bond map model.

The specific manner of the causal path coverage is as follows: selecting a node in the bonding graph model, listing related analytical redundancy relations through constraint relations between the selected node and the key, replacing unknown variables in the formula with known variables or measurable variables connected with the node through a causal path, removing all the unknown variables, and obtaining an analytical redundancy relation formula after removing all the unknown variables. And after obtaining an analytic redundancy relation, selecting another node, listing the analytic redundancy relation related to the node according to the mode, checking whether the listed analytic redundancy relation is independent from other analytic redundancy relations, if so, keeping the analytic redundancy relation, if not, selecting the next node, and so on until all nodes are selected.

In this embodiment, the minimum subsystem is obtained

Distributed resolving of redundant relationshipsDARR₁Comprises the following steps:

minimum subsystem

Distributed analytic redundancy relationship DARR₂Comprises the following steps:

minimum subsystem

Distributed analytic redundancy relationship DARR₃Comprises the following steps:

wherein, in DARR₁In (e)_1-1＝V_in，e_1-8＝De₁，e_1-1And e_1-8Respectively representing minimum subsystems

Of two inputs, Df^*Representing a minimum subsystem

The measured value of the electric current sensor Df of (a),

represents a differential operation over time;

in DARR₂In (e)_2-1＝V_in-R₁Df，e_2-10＝De₂，e_2-1And e_2-10Respectively representing minimum subsystems

The value of the two input voltages of (a),

representing a minimum subsystem

Voltage sensor De₁Measured value of (a), k₁Represents the open and close state of the switch in the circuit system, the open state is 0, the close state is 1,

represents a differential operation over time;

in DARR₃In (e)_3-1＝De₁，e_3-1Representing a minimum subsystem

The value of the input voltage of (a),

representing a minimum subsystem

Voltage sensor De₂Measured value of (a), k₁Represents the open and close state of the switch in the circuit system, the open state is 0, the close state is 1,

representing a differential operation over time.

S32, according to the distributed analytic redundancy relation DARR of the three minimum subsystems₁、DARR₂、DARR₃Obtaining a distributed fault characteristic matrix, wherein the distributed fault characteristic matrix is as follows:

wherein dr is₁、dr₂、dr₃Respectively corresponding to the minimum subsystem

Residual error of (1), R₁、R₂、R₃Resistive elements, C, being circuitry₁、C₂、C₃Capacitive elements of the circuit system, wherein 1 in the distributed fault signature matrix indicates that the residual error of the column is sensitive to the elements of the row, 0 in the distributed fault signature matrix indicates that the residual error of the column is insensitive to the elements of the row, and k in the distributed fault signature matrix₁The switch k represents the sensitivity of the residual of the column to the elements of the row₁And (6) determining.

Due to the minimal subsystem

Distributed analytic redundancy relationship DARR₁The element contained in (A) has R₁、R₂、 C₁D is therefore dr₁To the element R₁、R₂、C₁Sensitive, so in the distributed fault signature matrix, the drth₁R of₁Line, R₂Line, C₁The value of the line is 1 at the dr₁R of₃Line, C₂Line, C₃The row has a value of 0.

Due to the minimal subsystem

Distributed analytic redundancy relationship DARR₂The element contained in (A) has R₂、R₃、 C₂D is therefore dr₂To the element R₂、R₃、C₂Is sensitive and R is₃And also with a switch k₁Correlation, therefore dr₂To the element R₃Is sensitive to by the switch k₁Determining, therefore, in the distributed fault signature matrix, the drth₂R of₂Line, C₂The value of the line is 1, dr₂R of₃The value of a row is k₁Dr th₂R of₁Line, C₁Line, C₃The row has a value of 0.

Due to the minimal subsystem

Distributed analytic redundancy relationship DARR₃The element contained in (A) has R₃、C₃D is therefore dr₂To the element R₃、C₃Is sensitive and R is₃、C₃Are all connected with switch k₁Correlation, therefore dr₃To the element R₃、C₃Is sensitive to by the switch k₁Determining, therefore, in the distributed fault signature matrix, the drth₃R of₃Line, C₃The value of a row is k₁Dr th₂R of₁Line, R₂Line, C₁Line, C₂The row has a value of 0.

In step S4, the element parameters and sensor data of the minimum subsystem are substituted into the corresponding distributed analytical redundancy relationship to obtain the residual error of the minimum subsystem, and the residual errors dr of the three minimum subsystems are substituted₁、dr₂、 dr₃A set of residues is constructed.

The residual error is a numerical estimation for analyzing the redundancy relation and is a basis for judging whether the circuit system has a fault, if the circuit system has no fault, all residual values in the circuit system are 0, and if the circuit system has a fault, at least one residual value in the circuit system is 1. The residual error fluctuates in a certain range due to the problems of external noise, disturbance, element parameter tolerance and measurement accuracy of the sensor, and in order to prevent misdiagnosis caused by the fluctuation, the residual errors dr are respectively detected₁、dr₂、dr₃Setting a residual threshold, and if the fluctuation of the residual exceeds the residual threshold, determining that the residual is not in the normal stateThe residual error value is 1, namely at least one element in the minimum subsystem corresponding to the residual error has a fault; and if the fluctuation of the residual is within the residual threshold value, setting the residual value to be 0, namely that all the element parameters in the minimum subsystem corresponding to the residual are normal values. The size of the residual threshold is determined according to the difference of the diagnosis object and the difference of the application environment, and specifically, the residual threshold is determined according to the experiment of the diagnosis object and the application environment.

FIG. 4(d) shows a DARR for distributed resolution of redundancy relationships₁The waveform of (1), the fluctuation of which exceeds a residual threshold, the corresponding residual dr₁Has a value of 1; FIG. 4(f) shows a DARR for distributed resolution of redundancy relationships₃The fluctuation of the waveform does not exceed the residual error threshold, corresponding residual error dr₃The value of (d) is 0.

A global fault diagnosis method for a circuit system, comprising the steps of:

and S21, taking a complex circuit system as an example, modeling the circuit system, configuring the causal relationship among all parameters, and obtaining a global diagnosis bonding diagram model of the circuit system.

And S22, obtaining a global analytic redundancy relationship according to the global diagnostic bond diagram model of the circuit system, and analyzing the global analytic redundancy relationship to obtain a global fault feature matrix.

And S23, performing fault diagnosis according to the global fault feature matrix.

Step S21 in the global diagnosis method of a circuit system coincides with step S1 in the distributed diagnosis method of a circuit system.

As shown in fig. 2, in step S21, the parameters in the global diagnostic bond map model of the circuit system include: potential source U_M＝V_inResistive element { R₁，R₂，R₃}, capacitive element { C₁，C₂，C₃}, voltage sensor { De₁，De₂}, current sensor Df, switch k₁. Wherein, the potential source V_inIs the input voltage of the system; resistance R₁And a capacitor C₁Series connected, resistance R₂And a capacitor C₂Series connected, resistance R₃And a capacitor C₃Are connected in series; will resistance R₁And a capacitor C₁Formed series circuit, resistor R₂And a capacitor C₂Formed series circuit, resistor R₃And a capacitor C₃The formed series circuits are connected in parallel; voltage sensor De₁Connected in parallel to a capacitor C₂Both ends of (a); voltage sensor De₂Connected in parallel to a capacitor C₃Both ends of (a); current sensor Df and resistor R₁Potential source V_inAre connected in series; switch k₁And a resistor R₃Are connected in series.

In the global diagnosis bond-graph model, directional arrows represent keys, 0 and 1 represent two nodes, namely a common potential node and a common current node, wherein the potentials on the keys connected with the common potential node are the same, and the currents on the keys connected with the common current node are the same. In describing the constraints for each node and each parameter in the diagnostic bond graph, e_iIndicating the potential, i.e. voltage, f, at the bond numbered i_iIndicating a flow, i.e., current, over the key numbered i. In this example, V_inAnd R₁Df are connected in series, the currents flowing through the three elements are the same, and a common current junction 1 is used in a bonding diagram₁Connection, V_inAnd C₁Parallel connection, using a common potential junction 0 in a bonding diagram₂Is connected to R₂In series with the front-end circuit, the same current flows as the front-end circuit, and a common current junction 1 is used in the bonding diagram₃Connected to the front end, C₂、De₁Connected in parallel with the front-end circuit and at the same voltage, and the common potential junction 0 is used in the bonding diagram₄To the front end, R₃、k₁The same current flows in series with the front end, and a common current junction 1 is used in the bonding diagram₅Connected to a front end, De₂、C₃Parallel connection with front end, same voltage, common potential junction 0 in bonding diagram₆Connected to the front end, thereby forming a bond map model of the circuitry used in the present invention.

In the global diagnostic bond map model of the circuit system, the constraint relationship between each node and key is as follows:

in step S22, the method includes the following steps:

and S221, obtaining a global analysis redundancy relation through a causal path overlay method according to the global diagnosis bonding diagram model of the circuit system and the causal relation among all parameters in the global diagnosis bonding diagram model.

The specific manner of the causal path coverage is as follows: selecting a node in the bonding graph model, listing related analytical redundancy relations through constraint relations between the selected node and the key, replacing unknown variables in the formula with known variables or measurable variables connected with the node through a causal path, removing all the unknown variables, and obtaining an analytical redundancy relation formula after removing all the unknown variables. And after obtaining an analytic redundancy relation, selecting another node, listing the analytic redundancy relation related to the node according to the mode, checking whether the listed analytic redundancy relation is independent from other analytic redundancy relations, if so, keeping the analytic redundancy relation, if not, selecting the next node, and so on until all nodes are selected.

In this embodiment, the obtained global analytic redundancy relationships of the circuit system are respectively:

wherein, V_inIs the input voltage value, R, of the circuit system₁，R₂，R₃Respectively representing the resistance values, C, in the circuit system₁，C₂，C₃Respectively representing the capacitance values, De, in the circuitry₁，De₂Respectively representing the measured value of a voltage sensor in the circuit system, Df representing the measured value of a current sensor in the circuit system, k₁Represents the open and close state of the switch in the circuit system, the open state is 0, the close state is 1,

representing a differential operation over time.

S222, according to the global analysis redundancy relation GARR of the circuit system₁，GARR₂，GARR₃Obtaining a global fault characteristic matrix, wherein the global fault characteristic matrix is as follows:

θ/r	gr1	gr2	gr3
				R1	1	1	0
R 2	1	1	0
				R 3	0	0	k1
C1	1	0	0
				C 2	0	1	0
C3	0	k1	k1

wherein, gr₁、gr₂、gr₃For three residuals of the circuitry, R₁、R₂、R₃Resistive elements, C, being circuitry₁、C₂、C₃The capacitive elements are circuit systems, 1 in the global fault feature matrix indicates that the residual error of the column is sensitive to the elements of the row, 0 in the global fault feature matrix indicates that the residual error of the column is insensitive to the elements of the row, and k in the global fault feature matrix₁The switch k represents the sensitivity of the residual of the column to the elements of the row₁And (6) determining.

In step S23, substituting the element parameters and sensor data of the circuit system into the corresponding global analytic redundancy relationship to obtain the residual error of the circuit system, and obtaining the three residual errors gr of the circuit system₁、gr₂、gr₃A set of residues is constructed.

Based on a global fault signature matrix, R, when a single fault occurs in a circuit system₁And R₂All the same fault signatures in the global fault signature matrix are [ 110 ]]As shown in FIGS. 4(a), (b), and (c), three residuals gr from the circuit system are observed₁、gr₂、gr₃The set of constructed residuals is [ 110 ]]Then the possible faulty element obtained by global fault diagnosis is R₁And R₂Result in R₁And R₂Fault isolation cannot be achieved in the circuit system; based on a distributed fault signature matrix, R₁And R₂Respectively is [100 ]]And [ 110 ]]That is, if the corresponding failure characteristics are different, as shown in fig. 4(d), (e), and (f), the residual dr of the three smallest subsystems is observed₁、dr₂、dr₃The set of constructed residuals is [100 ]]Then the possible failure element obtained by the distributed failure feature matrix is R₁. Therefore, the fault isolation based on the distributed fault signature matrix of the invention is more accurate than the fault isolation based on the global fault signature matrix.

When multiple faults occur in the circuit system, e.g. switch k₁Closed, i.e. k ₁1, element R in the circuit system₁、 C₃The simultaneous failures are based on the global failure signature matrix, as shown in FIGS. 5(a), (b), and (c), from the three residuals gr of the circuitry₁、gr₂、gr₃The set of constructed residuals is [ 111 ]]At this time, it cannot be determined whether other elements in the circuit system are in failure at the same time, so that in the subsequent failure element identification process, all elements in the circuit system need to be identified, thereby increasing the complexity of failure diagnosis; based on the distributed fault feature matrix, shown in the graphs (d), (e) and (f), the residual error dr of the three minimum subsystems₁、dr₂、dr₃The set of constructed residuals is [ 101 ]]The possible failure element obtained by the distributed failure feature matrix is R₁、C₁、C₃. Therefore, the present inventionThe clear distributed fault diagnosis method improves the isolation capability of multiple faults and simultaneously reduces the complexity of fault identification after fault isolation.

The invention is not to be considered as limited to the specific embodiments shown and described, but is to be understood to cover all modifications, equivalents, and alternatives falling within the spirit and scope of the invention as defined by the appended claims.

Claims (4)

1. A distributed fault diagnosis method for a circuit system, comprising the steps of:

s1, modeling the circuit system, and configuring the causal relationship among all parameters to obtain a global diagnosis bonding diagram model of the circuit system;

s2, taking each sensor in the circuit system as a basic unit, taking the measured value of the sensor as the local output of the circuit system, extracting the minimum subsystem based on each sensor from the circuit system, and respectively obtaining the local diagnosis bond map model of each minimum subsystem;

s3, obtaining a distributed analytic redundancy relationship according to the local diagnostic bond map model of the minimum subsystem, and analyzing the distributed analytic redundancy relationship to obtain a distributed fault feature matrix;

s4, performing fault diagnosis according to the distributed fault feature matrix;

in step S2, the manner of obtaining the local diagnostic bond map model of the smallest subsystem is specifically as follows: taking a certain sensor of the circuit system as an output sensor, taking the measured value of the output sensor as the local output of the circuit system, namely taking the measured value of the output sensor as the output signal of the minimum subsystem based on the output sensor, starting from the output sensor to the end of another sensor, taking the measured value of the sensor which is the end sensor as the input signal of the minimum subsystem, and obtaining a local diagnostic bonding map model of the minimum subsystem based on the output sensor; in a path from the output sensor to the termination sensor, a parameter associated with the measured value of the output sensor is derived from causal relationships between various parameters on a global diagnostic bond map model of the circuitry, and the parameter associated with the measured value of the output sensor is used as a parameter of a local diagnostic bond map model based on a smallest subsystem of the output sensor.

2. The distributed fault diagnosis method for circuit systems according to claim 1, wherein in step S1, the parameters of the global diagnostic bond map model for circuit systems include: potential source U_M＝V_inResistive element { R₁，R₂，R₃}, capacitive element { C₁，C₂，C₃}, voltage sensor { De₁，De₂}, current sensor Df, switch k₁(ii) a Wherein, the potential source V_inIs the input voltage of the system; resistance R₁And a capacitor C₁Series connected, resistance R₂And a capacitor C₂Series connected, resistance R₃And a capacitor C₃Are connected in series; will resistance R₁And a capacitor C₁Formed series circuit, resistor R₂And a capacitor C₂Formed series circuit, resistor R₃And a capacitor C₃The formed series circuits are connected in parallel; voltage sensor De₁Connected in parallel to a capacitor C₂Both ends of (a); voltage sensor De₂Connected in parallel to a capacitor C₃Both ends of (a); current sensor Df and resistor R₁Potential source V_inAre connected in series; switch k₁And a resistor R₃Are connected in series.

3. The distributed fault diagnosis method of circuit system according to claim 2, wherein the current sensor Df is used as a basic unit, starting from the current sensor Df to the voltage sensor De₁Terminate, extract the smallest subsystem based on the current sensor Df

The parameters of the local diagnostic bond map model of the minimal subsystem include: potential source

Resistive element { R₁，R₂}, capacitive element C₁A flow sensor Df;

with a voltage sensor De₁As a basic unit, a slave voltage sensor De₁To start, to the voltage sensor De₂And the current sensor Df is terminated, and the voltage sensor De is extracted₁Minimum subsystem of

The minimum subsystem

The parameters of the local diagnostic bond map model of (a) include: potential source

Resistive element { R₂，R₃}, capacitive element { C₁，C₂}, a voltage sensor De₁Switch k₁；

With a voltage sensor De₂As a basic unit, a slave voltage sensor De₂To start, to the voltage sensor De₁Terminate, extract based on the voltage sensor De₂Minimum subsystem of

The minimum subsystem

The parameters of the local diagnostic bond map model of (a) include: potential source

Resistive element R₃Capacitive element C₃D, a voltage sensor De₂Switch k₁。

4. The distributed fault diagnosis method for circuit systems according to claim 3, wherein step S3 includes the following specific steps:

s31, obtaining a distributed analytical redundancy relationship according to the local diagnostic bond map models of the three minimum subsystems and the causal relationship among all parameters in the local diagnostic bond map models;

in the above manner, the minimum subsystem is obtained

Distributed analytic redundancy relationship DARR₁Comprises the following steps:

minimum subsystem

Distributed analytic redundancy relationship DARR₂Comprises the following steps:

minimum subsystem

Distributed analytic redundancy relationship DARR₃Comprises the following steps:

wherein, in DARR₁In (e)_1-1＝V_in，e_1-8＝De₁，e_1-1And e_1-8Respectively representing minimum subsystems

Of two inputs, Df^*Representing a minimum subsystem

The measured value of the electric current sensor Df;

in DARR₂In (e)_2-1＝V_in-R₁Df，e_2-10＝De₂，e_2-1And e_2-10Respectively representing minimum subsystems

The value of the two input voltages of (a),

representing a minimum subsystem

Voltage sensor De₁A measured value of (a);

in DARR₃In (e)_3-1＝De₁，e_3-1Representing a minimum subsystem

The value of the input voltage of (a),

representing a minimum subsystem

Voltage sensor De₂A measured value of (a);

s32, according to the distributed analytic redundancy relation DARR of the three minimum subsystems₁、DARR₂、DARR₃Obtaining a distributed fault characteristic matrix, wherein the distributed fault characteristic matrix is as follows:

θ/r dr₁ dr₂ dr₃ R₁ 1 0 0 R₂ 1 1 0 R₃ 0 k₁ k₁ C₁ 1 0 0 C₂ 0 1 0 C₃ 0 0 k₁

wherein dr is₁、dr₂、dr₃Respectively corresponding to the minimum subsystem

Residual error of (1), R₁、R₂、R₃Resistive elements, C, being circuitry₁、C₂、C₃Capacitive elements of the circuit system, wherein 1 in the distributed fault signature matrix indicates that the residual error of the column is sensitive to the elements of the row, 0 in the distributed fault signature matrix indicates that the residual error of the column is insensitive to the elements of the row, and k in the distributed fault signature matrix₁The switch k represents the sensitivity of the residual of the column to the elements of the row₁Determining;

due to the minimal subsystem

Distributed analytic redundancy relationship DARR₁The element contained in (A) has R₁、R₂、C₁D is therefore dr₁To the element R₁、R₂、C₁Sensitive, so in the distributed fault signature matrix, the drth₁R of₁Line, R₂Line, C₁The value of the line is 1 at the dr₁R of₃Line, C₂Line, C₃The value of the row is 0;

due to the minimal subsystem

Distributed analytic redundancy relationship DARR₂The element contained in (A) has R₂、R₃、C₂D is therefore dr₂To the element R₂、R₃、C₂Is sensitive and R is₃And also with a switch k₁Correlation, therefore dr₂To the element R₃Is sensitive to by the switch k₁Determining, therefore, in the distributed fault signature matrix, the drth₂R of₂Line, C₂The value of the line is 1, dr₂R of₃The value of a row is k₁Dr th₂R of₁Line, C₁Line, C₃The value of the row is 0;

due to the minimal subsystem

Distributed analytic redundancy relationship DARR₃The element contained in (A) has R₃、C₃D is therefore dr₂To the element R₃、C₃Is sensitive and R is₃、C₃Are all connected with switch k₁Correlation, therefore dr₃To the element R₃、C₃Is sensitive to by the switch k₁Determining, therefore, in the distributed fault signature matrix, the drth₃R of₃Line, C₃The value of a row is k₁Dr th₂R of₁Line, R₂Line, C₁Line, C₂The row has a value of 0.

Images