CN114329971A - Controller reference power supply fault diagnosis and correction method based on compressed sensing - Google Patents
Controller reference power supply fault diagnosis and correction method based on compressed sensing Download PDFInfo
- Publication number
- CN114329971A CN114329971A CN202111634691.7A CN202111634691A CN114329971A CN 114329971 A CN114329971 A CN 114329971A CN 202111634691 A CN202111634691 A CN 202111634691A CN 114329971 A CN114329971 A CN 114329971A
- Authority
- CN
- China
- Prior art keywords
- fault
- reference power
- power supply
- model
- component
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Landscapes
- Testing And Monitoring For Control Systems (AREA)
Abstract
The invention relates to a controller reference power supply fault diagnosis and correction method based on compressed sensing, which comprises the following steps: carrying out system analysis on a reference power supply circuit inside an aircraft engine controller, establishing a correlation model, and obtaining a correlation graphic model and a mathematical model matrix; establishing a linearized fault equation according to the correlation model; based on an orthogonal matching pursuit algorithm, reconstructing the measured value and the fault characteristic matrix to obtain a fault mode, and realizing fault diagnosis of the reference power supply; and evaluating the deviation proportion of all the components by calculating the projection length of the measured value in different fault subspaces, and further realizing the correction of the measured value by correcting the reference power supply. The invention improves the fault detection rate of the system, can reduce the requirements of test points, can realize the positioning of single-point faults and multi-point combined faults simultaneously, provides an effective means for the health management of the system, provides a reference power supply deviation correction method and improves the measurement precision of the controller.
Description
Technical Field
The invention belongs to the field of overall design of an aircraft engine control system, and particularly relates to a compressed sensing-based controller reference power supply fault diagnosis and correction method.
Background
The control system of the aero-engine is a control center of the aero-engine, the electronic controller is used as a main device of the control system, and the reliability of the control system has important significance for the safe operation of the aero-engine. The voltage reference source is a high-precision and high-stability power supply in an internal circuit of the controller and generally provides a reference for other working acquisition voltages. The accuracy of the reference power supply is related to the temperature, humidity, long-term working stability and the like of a working environment, and the influence of external or internal factors may cause the attenuation, drift and even failure and other abnormalities of the reference power supply, so that the deviation of a large number of related measurement parameters is caused, and the control accuracy and performance of the whole system are seriously influenced.
The monitoring, positioning and isolation of partial power supply faults can be realized through resources such as BIT and power supply warning circuits in the current controller, but the following two defects still exist: on one hand, the system only judges and processes the normal and failure modes of the power supply in the controller, but analyzes the influence of the system hierarchy in abnormal and decline states except the normal and failure modes, detects the abnormality and corrects the countermeasure; on the other hand, part of the non-testable reference power supply exists in the controller, forward testability design and verification are required to be carried out according to the relevance of the internal power supply, test resources in the controller are fully utilized, the coverage rate of the key power supply fault detection is 100%, and the testability and reliability of the system are improved.
Disclosure of Invention
The invention aims to overcome the defects in the prior art and provide a reference power supply fault diagnosis and correction method of a controller based on compressed sensing, which can improve the testability and reliability of a control system.
According to the technical scheme provided by the invention, the controller reference power supply fault diagnosis and correction method based on compressed sensing comprises the following steps:
step 1: performing system analysis on a reference power circuit inside an aircraft engine controller, establishing a correlation model, and obtaining a correlation graphic model and a mathematical model D matrix;
step 2: and establishing a linearized fault equation according to the correlation model:
y＝Ax+w
wherein y is an M-dimensional measurement value, A belongs to CM multiplied by N and is a fault characteristic matrix, and w is measurement noise;
and step 3: based on an orthogonal matching tracking algorithm, reconstructing the measured value y and the fault feature matrix A to obtain a fault mode x, and realizing fault diagnosis of the reference power supply;
and 4, step 4: and evaluating the offset proportion k of all the components by calculating the projection length l of the measured value in different fault subspaces, and further realizing the correction of the measured value by correcting the reference power supply.
Preferably, the step 1 specifically comprises:
1-1) determining all the components F which may be faulty in the reference power supply circuit by means of a system analysisnAnd available test points TmWherein N is 1,2, …, N, M is 1,2, …, M, N represents the total number of components, M represents the total number of test points;
1-2) analyzing the correlation between the component and the test point, and establishing a correlation graphic model; in the correlation graph model, a block is used to represent a component FnThe circle box represents the test point TmArrows indicate functional information flow;
1-3) building component F according to the flow direction of function information in the correlation graph modelnAnd a test point TmD matrix model D betweenM×NThe D matrix model is a correlation mathematical model and adopts DmnRepresentation matrix DM×NM-th row and n columns of elements, dmnThe values of (A) are as follows:
preferably, the step 2 specifically comprises:
for a complex system, generalThere are often multiple failure types; the failure mode is represented by a vector x of dimension N, xnDenotes the nth fault, N is 1,2, …, N, and x is assumed to be binarynThe values of (a) can be expressed as,
for failure mode xiE {0,1}, the linearized fault equation is,
y＝Ax+w
wherein y is a measurement value in M dimension, and A belongs to CM×NAnd w is the fault characteristic matrix and the measurement noise.
The compressed sensing principle is a method for obtaining an original signal through reconstruction of a small number of measurement values when the original signal has sparsity; the sparsity refers to that most elements in a vector are 0;
the above fault equation satisfies the following condition:
because the probability of each fault is very low, only a few position elements of the vector x are not zero, and the vector x is a sparse vector;
due to the limitations of sensor setting and the like, the parameters of objects which can be measured are very limited, so that the dimension of a measured value is far smaller than that of a fault mode, namely M & lt N is satisfied;
the fault equation meets the compressed sensing condition, a compressed sensing algorithm can be adopted, and a fault mode is obtained through reconstruction of the measured value y and the fault characteristic matrix A, so that abnormal detection and fault diagnosis are realized;
the specific process of establishing the fault equation by the correlation model is as follows:
2-1) the measured value y is formed by all test points T in the step 1)mM1, 2, …, M is the relative deviation component of the measured data, i.e. M is the relative deviation of the measured data
Wherein, tmRepresenting a test point TmTest data of (d), t0mRepresenting a test point TmA reference value of (d);
2-2) failure mode x represents component F in step 1)nN is 1,2, …, N is failed, i.e.
2-3) the fault characteristic matrix A is formed by a D matrix model D in the step 1)M×NEach column vector is respectively multiplied by the influence ratio of the corresponding fault to the measured value y, and then the obtained result is obtained after normalization,
wherein N is 1,2, …, N, AnAnd DnRespectively represent A and DM×NN-th column vector, KnFor M-dimensional column vectors, representing component FnThe ratio is influenced for each element of the measured value y.
Preferably, the step 3 specifically comprises:
3-1) setting initial values: residual r0Y, failure mode x 0, index set
The iteration counter q is 0;
3-2) update counter: q is q + 1;
3-3) finding the index value λq: solve the optimization problem lambdaq＝arg maxn＝1,…,N|<An,rq>If any, one solution is selected;
3-4) updating parameters: omegaq＝Ωq-1∪{λq}；
3-5) updating the failure mode: solve the optimization problem
3-6) updating residual error: r isq＝y-Axq；
3-7) when Q ═ Q, finish the algorithm, xQI.e. the desired failure mode, otherwise repeating steps 3-2) to 3-7).
Preferably, the step 4 specifically comprises:
4-1) calculating projection length l and recording fault mode xQThe index vector corresponding to the next faulty component is f, AQ＝AfThe projection length l of the measured value y in the subspace corresponding to the column vector of A is expressed as,
wherein l is an n-dimensional column vector, and the projection length of the subspace corresponding to the faultless component is 0;
4-2) evaluation of component offset ratio, | knThe larger |, the component FnThe greater the deviation, knBy a 1nThe calculation results in that,
4-3) the step is mainly to correct the reference power supply and calculate the test point T according to the offset proportionmMeasured data t ofmThe correction value of (a) is,
where c denotes a component index value corresponding to the reference power supply.
The invention has the following advantages:
1. according to the invention, the correlation analysis is carried out on the controller reference power supply, a correlation model between the components and the test points is established, so that the components which are not directly measurable in the system are dominant, the fault detection of the components is realized by establishing a fault strategy, and the fault detection rate of the system is improved.
2. The invention provides a controller reference power supply fault diagnosis based on a compressed sensing algorithm, which finishes the mapping from all related components of a system to a test point by establishing a fault characteristic matrix, realizes the positioning of the fault positions of a plurality of components by adopting a small amount of test point data by utilizing the sparsity of fault mode vectors, and compared with the traditional threshold method, fault tree method and other methods, on one hand, the requirement of the test point can be reduced, on the other hand, the single-point fault can be positioned, and the multi-point combined fault can also be positioned.
3. The invention evaluates the deviation degree of the components by calculating the deviation proportion of each component, further realizes the monitoring of the attenuation, drift and other abnormalities of all the components and provides an effective means for the health management of the system.
4. The invention provides a reference power supply correction method based on anomaly detection and fault diagnosis, which realizes the correction of measurement data by calculating the offset proportion of a reference power supply, can avoid the measurement deviation of a system caused by the attenuation of the reference power supply, and improves the measurement precision of a controller.
Drawings
FIG. 1 is a general flow diagram of the process of the present invention.
FIG. 2 is a controller reference power model according to an embodiment of the present invention.
FIG. 3 is a correlation graph model of one embodiment of the present invention.
FIG. 4 is a comparison graph of the calculated value and the actual value of the relative deviation of the reference power of the test case 8 according to the present invention.
FIG. 5 is a comparison graph of the calculated value and the actual value of the relative deviation of the reference power of the test case 9 according to the present invention.
FIG. 6-1 is a graph of raw voltage measurements for U3, U4, and U5.
FIG. 6-2 is a graph comparing the U3 raw voltage measurement to the U3 calibrated measurement.
6-3 are graphs comparing the U4 raw voltage measurement to the U4 calibrated measurement.
6-4 are graphs comparing the U5 raw voltage measurement to the U5 calibrated measurement.
Detailed Description
The invention is further described with reference to the following figures and specific examples.
The controller power supply model illustrated in fig. 2 is used as an object, and the model is characterized in that:
the power model comprises regulated power supplies U1-U5, the voltage values of which are marked in the figure and are dimensionless processed values. The measurement noise w is white noise. The principle of the A/D acquisition module is that analog quantity is converted into digital quantity, the digital quantity converted by the reference power supply U1 is used as a reference, and the digital quantity converted by other analog quantities is in proportional relation with the digital quantity to convert to obtain corresponding measured value, so that the reference power supply U1 cannot measure the measured value, but the measured value of the A/D acquisition module has measurement deviation due to fault or attenuation, and the A/D acquisition module is a key component of the system. The deviation of the measured values of the regulated power supplies U3, U4 and U5 can be used for carrying out fault monitoring on U1, but the measured values of U3, U4 and U5 are related to the actual values of the power supplies U2-U5, so that the deviation between fault location and a reference power supply U1 needs to be corrected by comprehensively considering the correlation of models.
The specific embodiment comprises the following steps:
1) performing system analysis on a reference power circuit inside an aircraft engine controller, establishing a correlation model, and obtaining a correlation graphic model and a mathematical model D matrix; the method comprises the following specific steps:
1-1) determining all the components F which may be faulty in the reference power supply circuit by means of a system analysisnAnd available test points TmAs follows below, the following description will be given,
F1: reference power supply U1
F2: regulated power supply U2
F3: regulated power supply U3
F4: regulated power supply U4
F5: regulated power supply U5
T1: u3 measured value
T2: u4 measured value
T3: u5 measured value
Wherein the total number of components is 5, and the total number of test points is 3.
1-2) establishing a correlation graphic model according to the circuit connection relation and the principle of an A/D acquisition module, and representing a component F by a boxnThe circle box represents the test point TmThe arrows indicate the flow of functional information, as shown in fig. 3.
1-3) building component F according to the flow direction of function information in the correlation graph modelnAnd a test point TmD matrix model D betweenM×N。
2) And establishing a linearized fault equation based on a compressed sensing principle according to the correlation model.
2-1) the measured value y is formed by all test points T in the step 1)mThe relative deviation of the measured data of 1,2,3, i.e. m
Wherein, tmRepresenting a test point TmTest data of (d), t 0m3, 1 and 2 in sequence.
2-2) representing the component F in step 1) by means of a failure mode xnIf n is 1,2, …,5, i.e. fails
2-3) the fault characteristic matrix A is formed by a D matrix model D in the step 1)M×NAnd multiplying each column vector by the influence ratio of the corresponding fault on the measured value y, and normalizing to obtain the fault-free measured value. That is to say that the first and second electrodes,
A1＝[1 1 1]T/|[1 1 1]T|＝[0.5774 0.5774 0.5774]T
A2＝[1 1 0]T/|[1 1 0]T|＝[0.7071 0.7071 0]T
A3＝[1 0 0]T
A4＝[0 1 0]T
A5＝[0 0 1]T
thereby obtaining a measurement matrix
3) And reconstructing the measured value y and the fault characteristic matrix A by adopting an OMP (Orthogonal Matching Pursuit) algorithm to obtain a fault mode x. Where Q is 2, i.e. at most 2 components are considered to fail simultaneously. The fault diagnosis results of the embodiment of the invention are shown in table 1, and the reconstructed fault mode x is consistent with the actual fault injection position for the test cases of fault injection of single-point faults, two-point combined faults, attenuation, disconnection and the like of different components, which shows that the fault diagnosis of the control reference power supply is realized and the fault detection rate of the system is improved based on the compressed sensing principle.
TABLE 1 Fault diagnosis results
4) Calculating the projection length l of the measured value in different fault subspaces according to the test cases in the table 1, calculating the offset proportion k of all the components, wherein the result is shown in the last column of the table 1, comparing and injecting the deviation, the calculated offset proportion is basically consistent with the actual deviation value, in the test cases 8 and 9, the reference power supply injects the deviation decaying along with the time, and k is1The calculated values over time are shown in fig. 4 and 5, respectively, and the actual deviation values substantially coincide with the trend of the calculated deviation values. The above results show that the component offset proportion k calculated by the method can accurately reflect the actual deviation condition of the component.
The reference power supply and the measured values were corrected for the case of test case 9 in table 1, i.e., the combined failure of the attenuation of the reference power supply U1 and the deviation of regulated power supply U4, and the test results of the embodiment of the present invention are shown in fig. 6-1 to 6-4. Due to the attenuation of the reference power supply U1, the acquired value of the A/D conversion voltage is deviated, and by applying the correction method, the corrected acquired value is consistent with the actual value, so that the influence of the attenuation of the reference power supply U1 on the acquired value is eliminated, and the acquisition precision of the controller is improved.
Claims (5)
1. A controller reference power supply fault diagnosis and correction method based on compressed sensing is characterized by comprising the following steps:
step 1: performing system analysis on a reference power circuit inside an aircraft engine controller, establishing a correlation model, and obtaining a correlation graphic model and a mathematical model D matrix;
step 2: and establishing a linearized fault equation according to the correlation model:
y＝Ax+w
wherein y is a measurement value in M dimension, and A belongs to CM×NA fault characteristic matrix is adopted, x represents a fault mode, and w represents measurement noise;
and step 3: based on an orthogonal matching tracking algorithm, reconstructing the measured value y and the fault feature matrix A to obtain a fault mode x, and realizing fault diagnosis of the reference power supply;
and 4, step 4: and evaluating the offset proportion k of all the components by calculating the projection length l of the measured value in different fault subspaces, and further realizing the correction of the measured value by correcting the reference power supply.
2. The compressed sensing-based controller reference power failure diagnosis and correction method of claim 1, wherein: the step 1 specifically comprises the following steps:
1-1) determining all the components F which may be faulty in the reference power supply circuit by means of a system analysisnAnd available test points TmWherein N is 1,2, …, N, M is 1,2, …, M, N represents the total number of components, M represents the total number of test points;
1-2) analyzing the correlation between the component and the test point, and establishing a correlation graphic model; in the correlation graph model, a block is used to represent a component FnThe circle box represents the test point TmArrows indicate functional information flow;
1-3) according to correlationFunctional information flow in graphic model, building component FnAnd a test point TmD matrix model D betweenM×NThe D matrix model is a correlation mathematical model and adopts DmnRepresentation matrix DM×NM-th row and n columns of elements, dmnThe values of (A) are as follows:
3. the compressed sensing-based controller reference power failure diagnosis and correction method of claim 1, wherein: the step 2 specifically comprises the following steps:
2-1) the measured value y is formed by all test points T in the step 1)mM1, 2, …, M is the relative deviation component of the measured data, i.e. M is the relative deviation of the measured data
Wherein, tmRepresenting a test point TmTest data of (d), t0mRepresenting a test point TmA reference value of (d);
2-2) failure mode x represents component F in step 1)nN is 1,2, …, N is failed, i.e.
2-3) the fault characteristic matrix A is formed by a D matrix model D in the step 1)M×NEach column vector is respectively multiplied by the influence ratio of the corresponding fault to the measured value y, and then the obtained result is obtained after normalization,
wherein N is 1,2, …, N, AnAnd DnAre respectively provided withDenotes A and DM×NN-th column vector, KnFor M-dimensional column vectors, representing component FnThe ratio is influenced for each element of the measured value y.
4. The compressed sensing-based controller reference power failure diagnosis and correction method of claim 1, wherein: the step 3 specifically comprises the following steps:
3-1) setting initial values: residual r0Y, failure mode x 0, index set
The iteration counter q is 0;
3-2) update counter: q is q + 1;
3-3) finding the index value λq: solve the optimization problem lambdaq＝argmaxn＝1,…,N|<An,rq>If any, one solution is selected;
3-4) updating parameters: omegaq＝Ωq-1∪{λq}；
3-5) updating the failure mode: solve the optimization problem
3-6) updating residual error: r isq＝y-Axq；
3-7) when Q ═ Q, finish the algorithm, xQI.e. the desired failure mode, otherwise repeating steps 3-2) to 3-7).
5. The compressed sensing-based controller reference power failure diagnosis and correction method of claim 1, wherein: the step 4 specifically comprises the following steps:
4-1) calculating projection length l and recording fault mode xQThe index vector corresponding to the next faulty component is f, AQ＝AfThe projection length l of the measured value y in the subspace corresponding to the column vector of A is expressed as,
wherein l is an n-dimensional column vector, and the projection length of the subspace corresponding to the faultless component is 0;
4-2) evaluation of component offset ratio, | knThe larger |, the component FnThe greater the deviation, knBy a 1nThe calculation results in that,
4-3) the step is mainly to correct the reference power supply and calculate the test point T according to the offset proportionmMeasured data t ofmThe correction value of (a) is,
where c denotes a component index value corresponding to the reference power supply.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111634691.7A CN114329971A (en) | 2021-12-29 | 2021-12-29 | Controller reference power supply fault diagnosis and correction method based on compressed sensing |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111634691.7A CN114329971A (en) | 2021-12-29 | 2021-12-29 | Controller reference power supply fault diagnosis and correction method based on compressed sensing |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114329971A true CN114329971A (en) | 2022-04-12 |
Family
ID=81016672
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111634691.7A Pending CN114329971A (en) | 2021-12-29 | 2021-12-29 | Controller reference power supply fault diagnosis and correction method based on compressed sensing |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114329971A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115712840A (en) * | 2022-11-15 | 2023-02-24 | 中国人民解放军陆军工程大学 | Multi-fault diagnosis method and system for electronic information system |
-
2021
- 2021-12-29 CN CN202111634691.7A patent/CN114329971A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115712840A (en) * | 2022-11-15 | 2023-02-24 | 中国人民解放军陆军工程大学 | Multi-fault diagnosis method and system for electronic information system |
| CN115712840B (en) * | 2022-11-15 | 2023-06-13 | 中国人民解放军陆军工程大学 | Multi-fault diagnosis method and system for electronic information system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US9239760B2 (en) | Systems and methods for detecting, correcting, and validating bad data in data streams | |
| CN109990922B (en) | Fault diagnosis method and system for temperature sensor with reduced temperature sensor redundancy | |
| CN106250709A (en) | Gas turbine abnormality detection based on sensors association network and fault diagnosis algorithm | |
| CN102081350B (en) | Method for determining equivalent aeroservoelasticity (ASE) robust stability of statically unstable aircraft through flight test | |
| CN109060899A (en) | Compensation method, device and the equipment of electrochemical sensor measurement result | |
| CN110275129B (en) | Method and system for determining synthetic error of high-voltage electric energy metering device | |
| US11550983B2 (en) | Method for determining an electrical model of a string of photovoltaic modules, diagnostic method and device associated therewith | |
| US11846680B2 (en) | Battery resistance diagnosis device and method | |
| Li et al. | Signal frequency domain analysis and sensor fault diagnosis based on artificial intelligence | |
| CN114329971A (en) | Controller reference power supply fault diagnosis and correction method based on compressed sensing | |
| US20050171680A1 (en) | Method of diagnosing the operating state of a motor vehicle diesel engine | |
| CN111046331A (en) | Health monitoring method based on signal abnormity | |
| CN110161181A (en) | The concentration of component recognition methods of mixed gas and system | |
| CN113610266B (en) | Method and device for predicting failure of automobile part, computer equipment and storage medium | |
| CN113283113B (en) | Solar cell array power generation current prediction model training method, abnormality detection method, device and medium | |
| CN114511025A (en) | Fan fault diagnosis method and device based on weighted multi-sensor fusion filtering | |
| CN113189513A (en) | Ripple-based redundant power supply current sharing state identification method | |
| Zhao et al. | Error theory study on EKF-based SOC and effective error estimation strategy for Li-ion batteries | |
| JP2003271231A (en) | Estimation device of detector drift and monitor system of detector | |
| CN112052496A (en) | Operation method of high arch dam valley amplitude deformation influence factor analysis system based on VAR model | |
| CN117031294A (en) | Battery multi-fault detection method, device and storage medium | |
| TWI810098B (en) | battery management device, battery management program | |
| CN115935519A (en) | Grouping weighting online least square identification method for longitudinal aerodynamic parameters of aircraft | |
| CN115659271A (en) | Sensor abnormality detection method, model training method, system, device, and medium | |
| CN114818881A (en) | Fault detection and positioning method for voltage sensor of vehicle-mounted power battery pack |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |