CN110597230A - Active fault diagnosis method - Google Patents
Active fault diagnosis method Download PDFInfo
- Publication number
- CN110597230A CN110597230A CN201910905908.XA CN201910905908A CN110597230A CN 110597230 A CN110597230 A CN 110597230A CN 201910905908 A CN201910905908 A CN 201910905908A CN 110597230 A CN110597230 A CN 110597230A
- Authority
- CN
- China
- Prior art keywords
- fault diagnosis
- afd
- output
- active fault
- input sequence
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 53
- 238000003745 diagnosis Methods 0.000 title claims abstract description 25
- 238000000926 separation method Methods 0.000 claims abstract description 20
- 230000008859 change Effects 0.000 claims abstract description 5
- 239000011159 matrix material Substances 0.000 claims description 13
- 238000005457 optimization Methods 0.000 claims description 11
- 239000013598 vector Substances 0.000 claims description 6
- 239000000126 substance Substances 0.000 claims description 3
- 230000009471 action Effects 0.000 claims description 2
- 238000004590 computer program Methods 0.000 claims 1
- 238000013459 approach Methods 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 230000008901 benefit Effects 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000014509 gene expression Effects 0.000 description 2
- 238000009499 grossing Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000006467 substitution reaction Methods 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- OAICVXFJPJFONN-UHFFFAOYSA-N Phosphorus Chemical compound [P] OAICVXFJPJFONN-UHFFFAOYSA-N 0.000 description 1
- 238000013476 bayesian approach Methods 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 150000001875 compounds Chemical class 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000002955 isolation Methods 0.000 description 1
- 238000012804 iterative process Methods 0.000 description 1
- 239000011541 reaction mixture Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B23/00—Testing or monitoring of control systems or parts thereof
- G05B23/02—Electric testing or monitoring
- G05B23/0205—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults
- G05B23/0218—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults characterised by the fault detection method dealing with either existing or incipient faults
- G05B23/0221—Preprocessing measurements, e.g. data collection rate adjustment; Standardization of measurements; Time series or signal analysis, e.g. frequency analysis or wavelets; Trustworthiness of measurements; Indexes therefor; Measurements using easily measured parameters to estimate parameters difficult to measure; Virtual sensor creation; De-noising; Sensor fusion; Unconventional preprocessing inherently present in specific fault detection methods like PCA-based methods
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B2219/00—Program-control systems
- G05B2219/20—Pc systems
- G05B2219/24—Pc safety
- G05B2219/24065—Real time diagnostics
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Automation & Control Theory (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
An active fault diagnosis method comprising the steps of: s1, obtaining an invariant set representation form of an output estimation set aiming at a linear discrete time invariant system (LDIT) with multiple faults; s2, converting the output estimation set obtained in the step S1 into an equivalent polyhedral set form through set form change; s3, set separation is achieved by maximizing the distance between the sets of output estimates represented by the invariant set, thereby resulting in a sequence of inputs that can be used to achieve Active Fault Diagnosis (AFD) and achieve control objectives for the system. The input sequence obtained by the method can realize the AFD target at a limited moment, and has the potential of being applied to a control target, so the method has good application prospect.
Description
Technical Field
The invention relates to Fault Diagnosis (FD), in particular to an Active Fault Diagnosis (AFD).
Background
Fault Diagnosis (FD) is an important guarantee of system safety and reliability in a real complex system. In general, FD comprises three tasks: fault detection, fault isolation and fault identification. Model-based FD methods are mainly classified into two categories: passive methods and active methods. Passive Fault Diagnosis (PFD) determines the operational state of a system by detecting input signals, state variables, and output information of the system. This means that even if an optimal PFD is designed, the FD task cannot be completed when the relevant information of the system is insufficient. Compared with the former, Active Fault Diagnosis (AFD) has little influence on actual system operation. Generally, active methods involve injecting signals into the system to improve fault detectability.
AFD is studied for different types of systems, such as linear systems, non-linear systems, description systems, etc. Model features are selected to distinguish AFD methods, which can be divided into two categories: a stochastic method and a deterministic method. The former assumes that the uncertainty of the system can be modeled with a known probability density function. For example, the residual signal is represented by a random process; or a bayesian approach is used for stochastic AFD. Deterministic methods assume that the interference, noise and initial conditions of the system under study can be modeled as bounded signals, and are generally divided into three cases. The first case is energy limited based on the uncertainty of the system. The uncertainty of the second assumption system is limited to the set. In the present invention, the uncertainty of the multi-model is modeled as a convex set. Furthermore, AFD is typically combined with fault tolerant control. In addition to the stochastic and deterministic methods, a third mixed approach was also investigated.
The invariant set based AFD method is a deterministic AFD method. In the existing literature, similar studies exist. For example, the afr is performed by using a zonotope set or a constrained zonotope set to represent the initial conditions of the system. The existing multistep AFD method based on the zonotope set needs to predict multistep information of a system according to the current information, so that an input sequence capable of realizing the AFD purpose is obtained. But the sequence cannot be altered so that the input can only be used to achieve FD, but not effectively to achieve the control goals of the system. In the existing steady-state AFD method based on the invariant set, the influence of the uncertainty of the system on the state set is described by the invariant set, the separation of the steady-state set is ensured at infinite time, and the AFD target is realized. The method can only realize set separation at infinite time; meanwhile, in the actual process, the actual state of the system is difficult to obtain, and the method adopts the state set to carry out AFD, so that the method is not beneficial to the actual application of the method.
Disclosure of Invention
The invention mainly aims to overcome the defects of the prior art and provide an active fault diagnosis method.
In order to achieve the purpose, the invention adopts the following technical scheme:
an active fault diagnosis method, comprising the steps of:
s1, obtaining an invariant set representation form of an output estimation set aiming at a linear discrete time invariant system (LDIT) with multiple faults;
s2, converting the output estimation set obtained in the step S1 into an equivalent polyhedral set form through set form change;
s3, set separation is achieved by maximizing the distance between the sets of output estimates represented by the invariant set, thereby resulting in a sequence of inputs that can be used to achieve Active Fault Diagnosis (AFD) and achieve control objectives for the system.
The invention has the following beneficial effects:
the invention provides an active fault diagnosis method based on set distance increase, and the active fault diagnosis method combines invariant set technology, provides a multistep AFD thought, and finally realizes an AFD target by a two-step distance increase algorithm. The input sequence obtained by the invention can be used for a fault diagnosis target and also has the potential of being applied to a control target. Specifically, the present invention can realize:
1. a two-step distance increase of the output set is achieved;
2. separation of limited time of output set is realized;
3. the purpose of sufficiently smoothing the designed input sequence is achieved.
In embodiments of the invention, smoothness is introduced in the objective function of the optimization problem to maximize the ability of the input sequence to separate the output sets.
Compared with the traditional AFD method, the invariant set-based AFD method has the advantages of different methods, and can obtain an input sequence which can be combined with a control target. The input sequence obtained by the method can realize the AFD target at a limited moment, and has the potential of being applied to a control target, so the method has good application prospect. The advantages of the present invention include:
by the AFD method, the separation of the output set in limited time can be realized;
by the AFD method, the obtained input sequence can be balanced between the minimum energy and the smooth input;
the idea of increasing collective distance as a separation idea is realized by the AFD method of the present invention.
Detailed Description
The embodiments of the present invention will be described in detail below. It should be emphasized that the following description is merely exemplary in nature and is not intended to limit the scope of the invention or its application.
The invention adopts different sets and simultaneously adopts different ideas with the traditional method to realize AFD. Traditional set-based approaches take the separation of an output set or a state set as the basis for implementing AFD. Whereas in stochastic AFD, the input signal is designed based on the lowest probability of making a false decision. The present invention proposes a new approach, the objective of which is to maximize the distance between the output prediction sets represented by the invariant set, whereas the sign of completion of AFD is the separation of the output sets. Thus, the present invention provides a new idea to implement AFD while introducing smoothness in the objective function of the optimization problem to maximize the ability of the input sequence to separate the output sets. The innovative idea of the embodiment of the invention is embodied in that: 1. an AFD strategy taking the increase of the set distance as a separation idea; 2. in the separation process, the strategy of AFD is realized through two-step optimization.
Specific embodiments of the present invention are further described below.
Output invariant set creation
This approach is considered on the basis of a linear discrete time invariant system (LDIT). Equation (1) represents LDIT with multiple failures:
wherein the content of the first and second substances,andrespectively the state vector, the output signal and the input signal of the system.Andis the disturbance and noise of the system, Ai,BiE, C, F are system parameters. i denotes different system models. i ≠ 1 denotes normal system, and i ≠ 1 denotes fault system model.
The initial state of the system and the perturbation and noise vectors are both considered to be energy bounded and may be represented using a zonotope set. The formula (2) gives their specific expressions:
given the initial conditions, an output set can be established to complete the AFD task.
In the present invention, to represent an output set, the system (1) is equivalently split into two parts: a center equation and a perturbation equation. The formula (3) represents the specific contents:
the first equation of the formula (3) represents a central equation and reflects the influence of the input on the system state; equation two represents the disturbance equation, reflecting the effect of the disturbance on the state of the system. The influence of the disturbance on the system is expressed by a set, and an expression form of a system state set can be obtained, as shown in (4):
in the formula (4), the reaction mixture is,to representThe set of state estimates for model i at time k + 1. Further, the output set may be obtained by a state estimation set, which has been formally transformed before.
In order to eliminate the influence of the disturbance on the state estimation set, which exists continuously, in the iterative process, the concept of the invariant set is used to describe the disturbance part. Therefore, the set of the final characterization output set is influenced by the input signal, only the position of the set is changed, the form is not influenced, and the realization of the AFD target is facilitated. Introducing the concept of invariant set, equation (4) is transformed as shown in equation (5):
further, according to the system (1), an invariant set representation of the output estimate set is obtained as shown in equation (6):
in the formula (I), the compound is shown in the specification,representing the set of output estimates for model i at time instant k + 1. The center of the set is affected by the input signal and the morphology does not change.
Set form transformation
The output estimation set obtained in the above process is a zonotope set, and it is difficult to directly use the output estimation set in the input sequence solving process for realizing the AFD purpose, so that it needs to be subjected to formal transformation. In the existing method, the zonotope set may be converted into a form of an equivalent polyhedral set (polytope) by a form change. Briefly, a zonotope set Z may be equivalently transformed as follows:
wherein g and H respectively represent a central vector and a segment matrix, the characteristic of zonotope is reflected, and the parameter matrixes A and b in the new form are derived from g and H according to a certain method.
On the basis, the method for realizing AFD by increasing the distance can be realized by solving an optimization problem.
Design input implementation of AFD
The traditional AFD method realizes the separation of an output set or a state set directly, namely realizes the AFD target, and finally realizes the set separation by means of realizing the maximization of set distance. Therefore, set separation needs to be described. In equation (7), the condition for output set separation is given:
where phi denotes the null set at design input ukUnder the action of the AFD, when any two different output sets are intersected and are empty sets, the AFD task is realized.
When u iskIs an unconstrained input, one can always find the appropriate input such that (7) holds, but based on the characteristics of the input signal of the AFD (with as little interference as possible with the normal operation of the system), plus the input of the actual system is always energy-bounded, so the designed input u is always energy-boundedkThe constraint formula (8) needs to be satisfied:
wherein U represents a positive real number, | · | | non-woven phosphor∞Representing an infinite norm of the vector.
Under the constraint (8), the equation (7) is usually unsolved, but the idea of increasing the collective distance means that there must be an input to achieve the increase in the distance. Under the constraint of (8), realizing the output setThe distance increase can be achieved by solving an unconstrained problem (7). Specifically, for the set distance increase problem including the constraint, the constraint is removed, and the original problem may be represented as (9). The input obtained by the new form (7) cannot be obtainedBut the direction of the set movement embodied by the input is worth using. Therefore, to implement the original constraint problem, it can be implemented by the problem (9):
whereinIntermediate input sequence representing an optimized solution, D1Representing a given symmetric positive definite matrix, usually taken as a unit matrix, U represents the upper bound of the input sequence.
Input sequence u obtained in question (9)kMaximum distance shift of the output set can be achieved untilThe AFD task is implemented. In order to better exploit the input, a two-step improvement is made to the problem (9):
first, considering input smoothing, and taking full advantage of the effect of each dimension of input on the increase of the set distance, smoothness can be generally characterized by the following smoothness function:
wherein xiThe ith value of the sequence of length n is represented, and D represents the formula form and the generated parameter matrix is rewritten.
Therefore, the second formula in formula (9) can be further modified as follows:
wherein D2And in accordance with the meaning of D, both alpha and beta are non-negative real numbers, and the balance between the minimum energy and the smoothness as possible of the input sequence is realized by adjusting the values of the alpha and the beta.
The second point, when the current time can not realize the separation, is the next time set separationAs a target, a two-step separation strategy is referred to. If in the case of (9) above,it is shown that the AFD target cannot be realized at the current time, and the fastest next time is realized, so that the following time set distance increase can be taken as the target, and is prepared in advance, and equation (10) represents the influence of the input sequence on the backward-stepping time output set:
wherein the parameter matrix definition is not changed,representing the center value of the resulting set of system state estimates at time i. The AFD, guided by the distance increase, is finally expressed by equation (10):
so far, the way AFD is implemented has been expressed, which in turn needs to be translated into an optimization problem that can be used for optimization solution.
Problem of optimization
Specifically, the problem (11) is a problem that is used for solving the input and realizing the maximization of the set distance, but the problem cannot be solved and needs to be explicit. Problem(s)Can be equivalently converted into:
further, (12) can be equivalently converted into (13):
whereinRepresenting a parameter matrix, p, reflecting the positional relationship of the prediction sets i and jijIs the row dimension of both, Mr,εrPositive real numbers, Δ, set large enough and small enough, respectivelyijRepresenting the variables introduced in solving the optimization problem.
And (3) combining (11) and (13) to obtain a mixed integer quadratic programming problem which can be used for optimization solution. By solving this problem, an input sequence can be obtained that achieves the purpose of AFD.
The foregoing is a more detailed description of the invention in connection with specific/preferred embodiments and is not intended to limit the practice of the invention to those descriptions. It will be apparent to those skilled in the art that various substitutions and modifications can be made to the described embodiments without departing from the spirit of the invention, and these substitutions and modifications should be considered to fall within the scope of the invention.
Claims (9)
1. An active fault diagnosis method, comprising the steps of:
s1, obtaining an invariant set representation form of an output estimation set aiming at a linear discrete time invariant system (LDIT) with multiple faults;
s2, converting the output estimation set obtained in the step S1 into an equivalent polyhedral set form through set form change;
s3, set separation is achieved by maximizing the distance between the sets of output estimates represented by the invariant set, thereby resulting in a sequence of inputs that can be used to achieve Active Fault Diagnosis (AFD) and achieve control objectives for the system.
2. The active fault diagnosis method according to claim 1, wherein in step S1, the LDIT with multiple faults is defined by equation (1):
wherein the content of the first and second substances,andthe state vector of the system, the output signal and the input signal,andis the disturbance and noise of the system, Ai,BiE, C, F are system parameters, i denotes a different system model, i ≠ 1 denotes a normal system, and i ≠ 1 denotes a faulty system model.
3. The active fault diagnosis method of claim 2, wherein in step S1, the invariant set representation of the set of output estimates is obtained as shown in equation (6):
wherein the content of the first and second substances,representing the set of output estimates of model i at time k +1, the center of the set being affected by the input signal and the morphology not changing.
4. The active fault diagnosis method according to claim 3, wherein in step S2, a zonotope set Z is equivalently transformed as follows:
wherein, A and b are parameter matrixes, g and H respectively represent a central vector and a segment matrix, and reflect the characteristics of zonotope.
5. The active fault diagnosis method according to any one of claims 1 to 4, characterized in that in step S3, the following formula (9) is used:
the resulting input sequence ukMaximum distance shift of the output set can be achieved untilAn active fault diagnosis task is implemented in which,represents the set of output estimates of the model i at time k +1, phi represents the null set, in the input sequence ukUnder the action of (2), when any two different output sets are intersected and are empty sets, the active fault diagnosis task is realized, whereinIntermediate input sequence representing an optimized solution, D1Representing a given symmetric positive definite matrix, usually taken as a unit matrix, U represents the upper bound of the input sequence.
6. The active fault diagnosis method according to claim 5, wherein step S3 includes:
the input smoothness is characterized using the following smoothness function:
wherein xiRepresenting the ith value of a sequence of length n, D representing a parameter generated by a formula form overwritingA matrix of numbers is formed by a matrix of numbers,
the second formula in the formula (9) is further changed as follows:
wherein D2The meaning of the D is consistent, both alpha and beta are non-negative real numbers, and the balance between the minimum energy and the smoothness as much as possible of the input sequence is realized by adjusting the values of the alpha and the beta;
when the separation cannot be realized at the current moment, the next moment is set to separate as a target, which is called a two-step separation strategy; if in the case of the formula (9),explaining that the AFD target cannot be realized at the current time, the following time set distance is increased as a target, and is prepared in advance, and the influence of the input sequence on the output set at the two-step backward time is represented by equation (10):
wherein the parameter matrix definition is not changed,represents the center value of the resulting set of system state estimates at time i,
AFD is determined by equation (11):
thus, AFD is achieved.
7. The active fault diagnosis method according to claim 6, characterized in that in step S3, formula (11) and formula (13) below are combined:
whereinRepresenting a parameter matrix reflecting the positional relationship between the ith and jth prediction sets, pijIs the row dimension of both, Mr,εrPositive real numbers, Δ, set large enough and small enough, respectivelyijRepresenting the variables introduced in solving the optimization problem,
and obtaining a hybrid integer quadratic programming problem which can be used for optimization solution, and obtaining an input sequence for realizing the AFD purpose by solving the problem.
8. A computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out the steps of the method according to any one of claims 1 to 7.
9. A control method, characterized in that the control objectives of the system are achieved with an input sequence obtained by the method according to any of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910905908.XA CN110597230B (en) | 2019-09-24 | 2019-09-24 | Active fault diagnosis method, computer readable storage medium and control method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910905908.XA CN110597230B (en) | 2019-09-24 | 2019-09-24 | Active fault diagnosis method, computer readable storage medium and control method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110597230A true CN110597230A (en) | 2019-12-20 |
| CN110597230B CN110597230B (en) | 2020-08-18 |
Family
ID=68862796
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910905908.XA Active CN110597230B (en) | 2019-09-24 | 2019-09-24 | Active fault diagnosis method, computer readable storage medium and control method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110597230B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112099351A (en) * | 2020-09-03 | 2020-12-18 | 西北工业大学 | Distributed fault diagnosis method based on centrosymmetric polyhedron |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2015181041A (en) * | 2009-09-03 | 2015-10-15 | イー． ラリモア，ウォレス | Method and system for empirical modeling of time-varying, parameter-varying, and nonlinear systems via iterative linear subspace computation |
| CN105700354A (en) * | 2016-01-31 | 2016-06-22 | 南通大学 | Intelligent sampling and detecting system with fault accommodation function |
| CN108594660A (en) * | 2018-05-07 | 2018-09-28 | 华侨大学 | Not the operational modal parameter recognition methods of structure changes and system when a kind of |
| CN109710983A (en) * | 2018-12-04 | 2019-05-03 | 北京大学 | A kind of cylinder of diesel engine hierarchical fault diagnosis method based on Key Performance Indicator |
| CN109799803A (en) * | 2018-12-11 | 2019-05-24 | 大连理工大学 | A kind of aero-engine sensor and actuator failure diagnostic method based on LFT |
-
2019
- 2019-09-24 CN CN201910905908.XA patent/CN110597230B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2015181041A (en) * | 2009-09-03 | 2015-10-15 | イー． ラリモア，ウォレス | Method and system for empirical modeling of time-varying, parameter-varying, and nonlinear systems via iterative linear subspace computation |
| CN105700354A (en) * | 2016-01-31 | 2016-06-22 | 南通大学 | Intelligent sampling and detecting system with fault accommodation function |
| CN108594660A (en) * | 2018-05-07 | 2018-09-28 | 华侨大学 | Not the operational modal parameter recognition methods of structure changes and system when a kind of |
| CN109710983A (en) * | 2018-12-04 | 2019-05-03 | 北京大学 | A kind of cylinder of diesel engine hierarchical fault diagnosis method based on Key Performance Indicator |
| CN109799803A (en) * | 2018-12-11 | 2019-05-24 | 大连理工大学 | A kind of aero-engine sensor and actuator failure diagnostic method based on LFT |
Non-Patent Citations (1)
| Title |
|---|
| FENG XU: "Conservatism comparison of set-based robust fault detection methods: Set-theoretic UIO and interval observer cases", 《AUTOMATICA》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112099351A (en) * | 2020-09-03 | 2020-12-18 | 西北工业大学 | Distributed fault diagnosis method based on centrosymmetric polyhedron |
| CN112099351B (en) * | 2020-09-03 | 2022-09-30 | 西北工业大学 | Distributed fault diagnosis method based on centrosymmetric polyhedron |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110597230B (en) | 2020-08-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| James et al. | Intelligent time-adaptive transient stability assessment system | |
| Brandão de Oliveira et al. | A hybrid search method for the vehicle routing problem with time windows | |
| Attar et al. | A simulation-based optimization approach for free distributed repairable multi-state availability-redundancy allocation problems | |
| Bhusal et al. | Deep ensemble learning-based approach to real-time power system state estimation | |
| Bombara et al. | Offline and online learning of signal temporal logic formulae using decision trees | |
| CN110597230B (en) | Active fault diagnosis method, computer readable storage medium and control method | |
| Zhang et al. | Online power system dynamic security assessment with incomplete PMU measurements: A robust white‐box model | |
| CN110032799A (en) | A kind of the angle similarity divided stages and monitoring method of microbiological pharmacy process | |
| Peng et al. | A new likelihood ratio method for training artificial neural networks | |
| Pesteh et al. | A new interior point solver with generalized correntropy for multiple gross error suppression in state estimation | |
| CN115618195A (en) | Sensor circuit fault diagnosis method, system, medium, and apparatus | |
| Yuan et al. | Improving fault tolerance in diagnosing power system failures with optimal hierarchical extreme learning machine | |
| CN112668697B (en) | Fuzzy test method and system for flight control parameters of unmanned aerial vehicle | |
| Könighofer et al. | Correct-by-Construction Runtime Enforcement in AI–A Survey | |
| Ji et al. | Cc: Causality-aware coverage criterion for deep neural networks | |
| Chumnaul et al. | Generalized confidence interval for the scale parameter of the power-law process with incomplete failure data | |
| Chen et al. | Application of Data‐Driven Iterative Learning Algorithm in Transmission Line Defect Detection | |
| Baraldi et al. | A method for estimating the confidence in the identification of nuclear transients by a bagged ensemble of FCM classifiers | |
| CN116310642A (en) | Variable dynamic discriminator differential privacy data generator based on PATE framework | |
| Zhang et al. | Dynamics-adaptive continual reinforcement learning via progressive contextualization | |
| Huang et al. | Two-stage cost-sensitive local models for heterogeneous cross-project defect prediction | |
| CN115936503A (en) | Power system security evaluation hidden attack method, system and storage medium | |
| CN109815108A (en) | A kind of combined test set of uses case priorization sort method and system based on weight | |
| CN114742644A (en) | Method and device for training multi-scene wind control system and predicting business object risk | |
| Jin et al. | Towards hierarchical fuzzy rule interpolation |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |