CN113723546A - Bearing fault detection method and system based on discrete hidden Markov model - Google Patents
Bearing fault detection method and system based on discrete hidden Markov model Download PDFInfo
- Publication number
- CN113723546A CN113723546A CN202111031722.XA CN202111031722A CN113723546A CN 113723546 A CN113723546 A CN 113723546A CN 202111031722 A CN202111031722 A CN 202111031722A CN 113723546 A CN113723546 A CN 113723546A
- Authority
- CN
- China
- Prior art keywords
- hidden markov
- markov model
- bearing
- discrete hidden
- detected
- 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
- 238000001514 detection method Methods 0.000 title claims abstract description 59
- 238000005457 optimization Methods 0.000 claims abstract description 39
- 239000013598 vector Substances 0.000 claims abstract description 36
- 238000000034 method Methods 0.000 claims abstract description 25
- 238000000605 extraction Methods 0.000 claims abstract description 22
- 238000012549 training Methods 0.000 claims abstract description 16
- 238000012545 processing Methods 0.000 claims description 25
- 238000000354 decomposition reaction Methods 0.000 claims description 10
- 238000007476 Maximum Likelihood Methods 0.000 claims description 8
- 230000002860 competitive effect Effects 0.000 claims description 6
- 238000003064 k means clustering Methods 0.000 claims description 6
- 238000010276 construction Methods 0.000 claims description 3
- 230000006870 function Effects 0.000 description 29
- 239000000126 substance Substances 0.000 description 6
- 238000003745 diagnosis Methods 0.000 description 5
- 238000013139 quantization Methods 0.000 description 5
- -1 ORF _2 Proteins 0.000 description 4
- 101100038645 Streptomyces griseus rppA gene Proteins 0.000 description 4
- 239000013256 coordination polymer Substances 0.000 description 4
- 238000012360 testing method Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 230000003993 interaction Effects 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 230000035772 mutation Effects 0.000 description 2
- 230000008707 rearrangement Effects 0.000 description 2
- 238000009827 uniform distribution Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/29—Graphical models, e.g. Bayesian networks
- G06F18/295—Markov models or related models, e.g. semi-Markov models; Markov random fields; Networks embedding Markov models
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
- G01M13/00—Testing of machine parts
- G01M13/04—Bearings
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Probability & Statistics with Applications (AREA)
- Complex Calculations (AREA)
Abstract
The invention provides a bearing fault detection method and a system based on a discrete hidden Markov model, wherein the method comprises the following steps: establishing an optimization algorithm of a discrete hidden Markov model; acquiring data of a bearing to be detected; carrying out feature extraction on the data of the bearing to be detected to obtain feature vectors of different fault types of the bearing to be detected; determining a codebook of a discrete hidden Markov model to obtain observation sequences of different fault types of a bearing to be detected; training a discrete hidden Markov model according to an optimization algorithm and observation sequences of different fault types of a bearing to be detected; and carrying out fault detection on the bearing to be detected according to the feature vectors of different fault types of the bearing to be detected and the trained discrete hidden Markov model. The invention can reduce the dependency on the initial value and improve the searching capability, thereby improving the learning precision of the discrete hidden Markov model parameters and ensuring the fault detection precision.
Description
Technical Field
The invention relates to the technical field of fault detection, in particular to a bearing fault detection method based on a discrete hidden Markov model and a bearing fault detection system based on the discrete hidden Markov model.
Background
The reliability and safety of mechanical equipment and complex electromechanical systems not only play a crucial role in the efficient operation of the whole industrial system, but also are particularly critical to personal safety and environmental protection. The fault detection and diagnosis technology of the dynamic system is an effective method for improving the reliability of the system and reducing the accident risk. With the successful application of HMM (Hidden Markov Model) in various fields and the higher requirements of the industrial community on reliability and safety, HMM is receiving more and more attention as a data-driven fault diagnosis method.
At present, the HMM according to the expression method of the observation variable can be classified into a discrete HMM (discrete hidden markov model) and a continuous HMM (continuous hidden markov model). However, there are three basic problems to be solved in the practical application of DHMM, namely, the evaluation problem, the decoding problem, and the learning problem, which is the most difficult problem of DHMM to solve. The BW algorithm (Baum-Welch algorithm) and the gradient algorithm are most widely used at present for the learning problem, and although the two algorithms can provide effective estimation of DHMM parameters, the BW algorithm and the gradient algorithm are local optimization methods and have strong dependence on initial values; in addition, in the current fault diagnosis application based on the DHMM, no practical situation is provided for the fault data insufficiency.
Disclosure of Invention
The present invention is directed to solving, at least to some extent, one of the technical problems in the art described above. Therefore, an object of the present invention is to provide a method for detecting a bearing fault based on a discrete hidden markov model, which can reduce the dependency on an initial value and improve the search capability, thereby improving the accuracy of parameter learning of the discrete hidden markov model and ensuring the accuracy of fault detection.
The second purpose of the invention is to provide a bearing fault detection system based on a discrete hidden Markov model.
In order to achieve the above object, a first aspect of the present invention provides a method for detecting a bearing fault based on a discrete hidden markov model, including the following steps: establishing an optimization algorithm of the discrete hidden Markov model; acquiring data of a bearing to be detected; extracting characteristics of the bearing data to be detected to obtain characteristic vectors of different fault types of the bearing to be detected; determining a codebook of the discrete hidden Markov model to obtain observation sequences of different fault types of the bearing to be detected; training the discrete hidden Markov model according to the optimization algorithm and observation sequences of different fault types of the bearing to be detected; and carrying out fault detection on the bearing to be detected according to the feature vectors of different fault types of the bearing to be detected and the trained discrete hidden Markov model.
According to the bearing fault detection method based on the discrete hidden Markov model, which is provided by the embodiment of the invention, the optimization algorithm of the discrete hidden Markov model is established, the characteristic extraction is carried out on the bearing data to be detected, the codebook of the discrete hidden Markov model is determined, the discrete hidden Markov model is trained according to the observation sequences of different fault types of the bearing to be detected of the optimization algorithm, and the fault detection is carried out on the bearing to be detected according to the characteristic vectors of the different fault types of the bearing to be detected and the trained discrete hidden Markov model, so that the dependency on the initial value can be reduced, the searching capability can be improved, the parameter learning precision of the discrete hidden Markov model can be improved, and the fault detection precision can be ensured.
In addition, the discrete hidden markov model-based bearing fault detection method proposed according to the above embodiment of the present invention may further have the following additional technical features:
further, the optimization algorithm for establishing the discrete hidden markov model comprises the following steps: constructing a topological mesh by adopting a plurality of discrete hidden Markov models; initializing the topological mesh by adopting an orthogonal metrology method; optimizing each discrete hidden Markov model in each topological grid by adopting double-strategy competitive learning to obtain a first sequence topological grid; judging whether the random real number of each discrete hidden Markov model uniformly distributed in the interval 0 to 1 in the first sequence topological grid is smaller than the parameter in the defined interval 0 to 1; if so, optimizing each discrete hidden Markov model in the first sequence topological grid by adopting cooperative learning to obtain a second sequence topological grid; judging whether the random real number of each discrete hidden Markov model in the second sequence topological grid, which is uniformly distributed in the interval 0 to 1, is smaller than the parameter in the defined interval 0 to 1; if so, optimizing each discrete hidden Markov model in the second sequence topological grid by adopting Gaussian variation to obtain a topological grid of the next iteration; and determining the discrete hidden Markov model with the maximum likelihood value in the topological mesh of the next iteration and executing a self-learning algorithm on the discrete hidden Markov model until the obtained discrete hidden Markov model meets a termination standard.
Further, the method for extracting the characteristics of the data of the bearing to be detected to obtain the characteristic vectors of different fault types of the bearing to be detected comprises the following steps: and performing feature extraction on the bearing data to be detected by adopting Daubechies wavelet and wavelet packet decomposition to obtain feature vectors of different fault types of the bearing to be detected.
Further, determining a codebook of the discrete hidden markov model to obtain observation sequences of different fault types of the bearing to be detected, comprising the following steps: and determining the codebook of the discrete hidden Markov model by adopting a k-means clustering algorithm to obtain observation sequences of different fault types of the bearing to be detected.
Further, training the discrete hidden Markov model according to the optimization algorithm and the observation sequence of the bearing to be detected with different fault types comprises the following steps: establishing an initial discrete hidden Markov model for the observation sequences of different fault types of the bearing to be detected; and training the initial discrete hidden Markov model by adopting the optimization algorithm until the probability of generating the observation sequence reaches the maximum value.
In order to achieve the above object, a second aspect of the present invention provides a bearing fault detection system based on a discrete hidden markov model.
A bearing fault detection system based on a discrete hidden Markov model is characterized by comprising: a building module for building an optimization algorithm of the discrete hidden Markov model; the acquisition module is used for acquiring the data of the bearing to be detected; the extraction module is used for extracting the characteristics of the data of the bearing to be detected so as to obtain the characteristic vectors of different fault types of the bearing to be detected; the first processing module is used for determining a codebook of the discrete hidden Markov model to obtain observation sequences of different fault types of the bearing to be detected; the second processing module is used for training the discrete hidden Markov model according to the optimization algorithm and observation sequences of different fault types of the bearing to be detected; and the third processing module is used for carrying out fault detection on the bearing to be detected according to the feature vectors of different fault types of the bearing to be detected and the trained discrete hidden Markov model.
According to the discrete hidden Markov model-based bearing fault detection system provided by the embodiment of the invention, an optimization algorithm of a discrete hidden Markov model is established through a construction module, and the extraction module is used for extracting the characteristics of the bearing data to be detected, and determining a codebook of the discrete hidden Markov model through a first processing module, training the discrete hidden Markov model through a second processing module according to the observation sequence of different fault types of the bearing to be detected by the optimization algorithm, the bearing to be detected is subjected to fault detection through the third processing module according to the feature vectors of different fault types of the bearing to be detected and the trained discrete hidden Markov model, so that the dependency on the initial value can be reduced, the searching capability can be improved, therefore, the accuracy of discrete hidden Markov model parameter learning can be improved, and the accuracy of fault detection can be ensured.
In addition, the discrete hidden markov model-based bearing fault detection system proposed according to the above embodiment of the present invention may further have the following additional technical features:
further, the building module is specifically configured to: constructing a topological mesh by adopting a plurality of discrete hidden Markov models; initializing the topological mesh by adopting an orthogonal metrology method; optimizing each discrete hidden Markov model in each topological grid by adopting double-strategy competitive learning to obtain a first sequence topological grid; judging whether the random real number of each discrete hidden Markov model uniformly distributed in the interval 0 to 1 in the first sequence topological grid is smaller than the parameter in the defined interval 0 to 1; if so, optimizing each discrete hidden Markov model in the first sequence topological grid by adopting cooperative learning to obtain a second sequence topological grid; judging whether the random real number of each discrete hidden Markov model in the second sequence topological grid, which is uniformly distributed in the interval 0 to 1, is smaller than the parameter in the defined interval 0 to 1; if so, optimizing each discrete hidden Markov model in the second sequence topological grid by adopting Gaussian variation to obtain a topological grid of the next iteration; and determining the discrete hidden Markov model with the maximum likelihood value in the topological mesh of the next iteration and executing a self-learning algorithm on the discrete hidden Markov model until the obtained discrete hidden Markov model meets a termination standard.
Further, the extraction module is specifically configured to: and performing feature extraction on the bearing data to be detected by adopting Daubechies wavelet and wavelet packet decomposition to obtain feature vectors of different fault types of the bearing to be detected.
Further, the first processing module is specifically configured to: and determining the codebook of the discrete hidden Markov model by adopting a k-means clustering algorithm to obtain observation sequences of different fault types of the bearing to be detected.
Further, the second processing module is specifically configured to: establishing an initial discrete hidden Markov model for the observation sequences of different fault types of the bearing to be detected; and training the initial discrete hidden Markov model by adopting the optimization algorithm until the probability of generating the observation sequence reaches the maximum value.
Drawings
FIG. 1 is a flow chart of a discrete hidden Markov model based bearing fault detection method according to an embodiment of the present invention;
FIG. 2 is a flow diagram of an optimization algorithm for building a discrete hidden Markov model, in accordance with one embodiment of the present invention;
FIG. 3 is a diagram of a discrete hidden Markov model optimization algorithm in accordance with one embodiment of the present invention in comparison with a prior art BW algorithm;
FIG. 4 is a block diagram of a discrete hidden Markov model based bearing fault detection system according to an embodiment of the present invention.
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.
Fig. 1 is a flowchart of a discrete hidden markov model-based bearing fault detection method according to an embodiment of the present invention.
As shown in fig. 1, a discrete hidden markov model-based bearing fault detection method according to an embodiment of the present invention includes the following steps:
and S1, establishing an optimization algorithm of the discrete hidden Markov model.
Specifically, as shown in fig. 2, the method comprises the following steps:
s101, constructing a topological mesh by adopting a plurality of discrete hidden Markov models.
More specifically, the topological mesh G may be constructed using a plurality of discrete hidden markov models, and the size of the topological mesh G may be S × S. Each discrete hidden Markov model can be fixed at a set position of the topological mesh G and can perform information interaction with the discrete hidden Markov models in the neighborhood. By constructing the topological mesh G by using a plurality of discrete hidden Markov models, extremely limited fault data can be fully mined under a certain topological structure, so that the performance of the fault diagnosis system is improved.
For the sake of illustration, the parameter set of each discrete hidden markov model is converted into a K-dimensional vector as a whole, and the discrete hidden markov model that can be set at grid coordinates (m, n) is Gmn＝[z1,z2,…,zK](m, n is 1,2, …, S), and the objective function value of the discrete hidden markov model may be set to Obj (G)mn) In addition, a topological mesh G formed by the discrete hidden Markov model, namely the discrete hidden Markov model G arranged on the mesh coordinate (m, n) can be arrangedmnThe neighborhood of (c) is:
wherein the content of the first and second substances,
further, G may be providedtIs the topology mesh obtained after the t-th iteration and can be set at GtAnd Gt+1The obtained topological mesh is G in sequencet+1/3And Gt+2/3。BtIs GtOf all discrete hidden markov models of (1) the model with the smallest value of the objective function, BTtIs from G0Up to GtOf all discrete hidden Markov models, PS、PCAnd PGIs a predefined intra-interval (0,1) parameter.
And S102, initializing the topological mesh by adopting an orthogonal metrology method.
More specifically, z can be definediIs the ith factor, and has a quantization domain of [ ci,di](wherein, 0. ltoreq. ci≤diLess than or equal to 1), and optionally [ c)i,di]The quantization was at the following H (required odd) levels:
wherein j is a variable.
Further, the orthogonal group L can be constructed by using an arrangement methodJ(HQ) Wherein J ═ Hl，Q＝(Hl-1)/(H-1), l being such that
Is the smallest positive integer.
Further, from orthogonal set LJ(HQ) In the random selection of K columns to obtain LJ(HK) And the obtained L can be usedJ(HK) Application to deltai,jThe following J vectors are obtained:
further, S can be randomly selected from the J vectors2The vectors are normalized and restored to be in the form of (pi, A, B), so as to obtain the initial topological mesh G0At the same time, can update B0And t ← 0.
S103, optimizing each discrete hidden Markov model in each topological mesh by adopting double-strategy competitive learning to obtain a first sequence topological mesh.
More specifically, the topological mesh G may be pairedtEach discrete hidden Markov model in (a) performs a two-strategy competition learning to obtain a first sequence topological mesh Gt+1/3。
For example, the topological mesh G may be determined firsttEach discrete hidden Markov model in (1), Gij＝[z1,z2,…,zK](i, j ═ 1,2, …, S) of the model M in the neighborhood with the smallest value of the objective functionij＝[α1,α2,…,αK]And then compare Gi,jAnd Mi,jThe objective function value of (1). Wherein, if Gi,jHas an objective function value of less than Mi,jIf the objective function value is less than the target function value, the discrete hidden Markov model is kept; if G isi,jHas an objective function value of M or morei,jThe objective function value of (1) is Mi,jSub-model Sub generatedij＝[β1,β2,…,βK]Replacing the discrete hidden markov model. A
Further, Mi,jCan adopt the followingTwo strategies are used to generate the submodel:
strategy one: subij＝[β1,β2,…,βK]Is generated by the following formula:
wherein χ ═ αi+U(-1,1)×(αi-zi) U (-1,1) denotes a uniform distribution in [ -1, 1]]Random real numbers of (2).
And (2) strategy two: subij＝[β1,β2,…,βK]Is generated by the following formula
Wherein U (0,1) represents a random real number uniformly distributed over [0,1 ].
It should be noted that the two strategies mentioned above are represented by the probability PSIt is determined if U (0,1) < PSIf not, executing strategy two.
S104, judging whether the random real number of each discrete hidden Markov model uniformly distributed in the interval 0 to 1 in the first sequence topological grid is smaller than the parameter in the defined interval 0 to 1.
And S105, if so, optimizing each discrete hidden Markov model in the first sequence topological grid by adopting cooperative learning to obtain a second sequence topological grid.
More specifically, for the first sequence the topology grid Gt+1/3Can judge that U (0,1) < P for each discrete hidden Markov model in (1)cIf yes, then adopting cooperative learning to optimize the first sequence topological grid Gt+1/3To obtain a second sequence of topological meshes Gt+2/3. It should be noted that the process of collaborative learning is as follows:
first, can be set from GijAnd MijThe generated search space is
Wherein the content of the first and second substances,ε＝[min(α1,z1),…,min(αK,zK)]，
further, can be
Is quantized to the following F levels:
further, the variable z may be set1,z2,…,zKDivided into groups V, each group can be regarded as a factor, and V-1 integers e can be randomly generated firstlyi(i-1, 2, …, V-1) so as to satisfy 1 < e1＜e2＜…＜eV-1< K, the following V factors can then be generated:
wherein the ith factor gammaiF levels of (a) are defined as:
wherein, [ eta ](i,1),η(i,2),…,η(i,F)]Is to [ eta ]i,1,ηi,2,…,ηi,F]Random rearrangement of the components and e0Is 0.
Further, L may be constructed with reference to step S102U(FV)＝[wij]U×VAnd can use LU(FV) The following U models were generated:
further, among these models, the model having the largest likelihood value may be selected instead of Gi, j.
S106, judging whether the random real number of each discrete hidden Markov model uniformly distributed in the interval 0 to 1 in the second sequence topological mesh is smaller than the parameter in the defined interval 0 to 1.
And S107, if so, optimizing each discrete hidden Markov model in the second sequence topological mesh by adopting Gaussian variation to obtain the topological mesh of the next iteration.
More specifically, for the second sequence topology mesh Gt+2/3Can judge that U (0,1) < P for each discrete hidden Markov model in (1)cIf true, the second sequence topology mesh G can be optimized by Gaussian variationt+2/3To obtain the topological mesh G of the next iterationt+1. The process of gaussian mutation is:
first, t can be set as the number of iterations and G (0,1/t) can be set as a Gaussian random number generator, thus, for Gij＝[z1,z2,…,zK]Each component of (i, j ═ 1,2, …, S), that is, each discrete hidden markov model, can determine UmWhether (0,1) ≥ 1/K is true, if true, zm←zm+ G (0,1/t), (m ═ 1,2, …, K), where U ism(0,1) is a random number generated independently for each component, i.e., each discrete hidden Markov model. Thus, a new discrete hidden Markov model can be used instead of the original model.
And S108, determining the discrete hidden Markov model with the maximum likelihood value in the topological mesh of the next iteration and executing a self-learning algorithm on the discrete hidden Markov model until the obtained discrete hidden Markov model meets the termination standard.
More specifically, the topological mesh G that can be used in the next iterationt+1Discrete hidden Markov models with maximum likelihood value determined in, Bt+1And a self-learning algorithm is performed thereon.It should be noted that the process of the self-learning algorithm is as follows:
first, B can bet+1Is marked as Gij＝[z1,z2,…,zK]Thus, a small-scale initial topological mesh sG with the size of S 'multiplied by S' can be generated0And the initial topological mesh sG0In (1) discrete hidden Markov models, i.e. sgkl(k, l ═ 1,2, …, S') is:
wherein the content of the first and second substances,
the elements of (A) are:
where σ is the search radius.
Further, an initial topological mesh sG is generated0Then, sG can be settFor the initial topological mesh sG0The grid obtained after t iterations and can be set to sBTtIs (sG)0,sG1,…,sGt) The model with the minimum objective function value in all the discrete hidden Markov models can also be used for converting the topological mesh sGtThe discrete hidden Markov model with the minimum value of the target function is marked as sBtFurthermore, sT can be set as iteration termination criterion, sPGIs a predefined intra-interval (0,1) parameter.
Further, the topological mesh sG can be paired firsttEach discrete hidden Markov model in the topology network carries out a strategy two in the double-strategy competition learning to obtain a topological mesh sGt+1/2(ii) a Further, it can be judged that U (0,1) < sPGIf yes, then the topological mesh sG is checkedt+1/2Each model of (1) performs Gaussian variation to obtain a topological mesh sGt+1(ii) a Finally, the topological grid sG can be judgedt+1Discrete hidden horse with minimum medium objective function valueErkov model sBt+1Whether the objective function value of (a) is greater than sBTtIf yes, sBTt+1←sBt +1If not, sBTt+1←sBTt、sBt+1←sBTt。
Further, it can be determined whether the iteration number t < sT is true, if yes, t ← t +1, and continue to execute the above steps; if not, sBT is usedtSubstituted GijAnd terminates the self-learning process.
The utility of the discrete hidden markov model optimization algorithm of the present invention will be further explained in conjunction with fig. 3.
As shown in fig. 3, when the objective function values of the sphere fault, the outer ring fault and the inner ring fault at 1 horsepower and 2 horsepower are obtained, that is, BF _1, BF _2, ORF _1, ORF _2, IRF _1 and IRF _2, the values of the objective function values are significantly smaller than the objective function values obtained by the BW algorithm at 1 horsepower and 2 horsepower in the prior art, that is, BF _1, BF _2, ORF _1, ORF _2, IRF _1 and IRF _2, so that it can be seen that the discrete hidden markov model optimization algorithm of the present invention can overcome the disadvantage that the BW algorithm in the prior art only has local search capability and has strong dependency on the initial value, and thus can improve the accuracy of parameter learning of the discrete hidden markov model.
And S2, acquiring the data of the bearing to be detected.
Specifically, the data of the bearing to be detected can be obtained through the bearing data center.
And S3, performing feature extraction on the bearing data to be detected to obtain feature vectors of different fault types of the bearing to be detected.
Specifically, Daubechies wavelet and wavelet packet decomposition can be adopted to perform feature extraction on the data of the bearing to be detected so as to obtain feature vectors of different fault types of the bearing to be detected, for example db4 and packet decomposition of 3 decomposition levels can be adopted to perform feature extraction on the data of normal conditions and each fault condition of the bearing to be detected.
And S4, determining a codebook of the discrete hidden Markov model to obtain observation sequences of different fault types of the bearing to be detected.
In particular, a codebook of discrete hidden markov models may be determined using a k-means clustering algorithm to obtain observation sequences of different fault types of the bearing to be detected, for example, the observation sequence O ═ O1o2…oT。
And S5, training the discrete hidden Markov model according to the optimization algorithm and the observation sequence of the bearing to be detected with different fault types.
Specifically, an initial discrete hidden markov model may be established for observation sequences of different fault types of the bearing to be detected, and the initial discrete hidden markov model may be trained using an optimization algorithm until a probability of generating the observation sequences reaches a maximum value, for example, the observation sequence O ═ O1o2…oTThe probability P (O | λ) of (a, B, pi) is the trained discrete hidden markov model.
And S6, carrying out fault detection on the bearing to be detected according to the feature vectors of different fault types of the bearing to be detected and the trained discrete hidden Markov model.
Specifically, the feature vectors of different fault types of the bearing to be detected can be input into the trained discrete hidden markov model, and then the quantization indexes can be adopted: CP ═ logP (o)t|o1o2…ot-1λ) so that the corresponding fault type can be selected as the final diagnostic decision based on the highest CP value.
The utility of the discrete hidden markov model based bearing fault detection method of the present invention will be further explained with reference to tables 1 and 2.
As shown in table 1, when the discrete hidden markov model-based bearing fault detection method of the present invention is used to detect a ball fault, an outer ring fault and an inner ring fault of a bearing, the training operating conditions and the testing operating conditions are 1772rpm, 1hp and 1750rpm, and 2hp, respectively, the control limit is 2.0577, and the fault detection rate is 100%; as further shown in table 2, when the ball fault, the outer ring fault, and the inner ring fault of the bearing are tested, that is, when the ball fault, the outer ring fault, and the inner ring fault of the bearing are respectively tested through 475, 477, and 475 test samples, the fault location accuracy rates are 100%, 98.95%, and 99.79%, respectively, and thus it can be seen that the discrete hidden markov model-based bearing fault detection method of the present invention has high fault detection accuracy rate and location accuracy.
TABLE 1
TABLE 2
According to the bearing fault detection method based on the discrete hidden Markov model, which is provided by the embodiment of the invention, the optimization algorithm of the discrete hidden Markov model is established, the characteristic extraction is carried out on the bearing data to be detected, the codebook of the discrete hidden Markov model is determined, the discrete hidden Markov model is trained according to the observation sequences of different fault types of the bearing to be detected of the optimization algorithm, and the fault detection is carried out on the bearing to be detected according to the characteristic vectors of the different fault types of the bearing to be detected and the trained discrete hidden Markov model, so that the dependency on the initial value can be reduced, the searching capability can be improved, the parameter learning precision of the discrete hidden Markov model can be improved, and the fault detection precision can be ensured.
Corresponding to the embodiment, the invention further provides a bearing fault detection system based on the discrete hidden Markov model.
As shown in fig. 4, the discrete hidden markov model based bearing fault detection system of the present invention includes a building module 10, an obtaining module 20, an extracting module 30, a first processing module 40, a second processing module 50, and a third processing module 60. Wherein the building module 10 is configured to build an optimization algorithm of the discrete hidden markov model; the acquiring module 20 is used for acquiring data of the bearing to be detected; the extraction module 30 is configured to perform feature extraction on the data of the bearing to be detected to obtain feature vectors of different fault types of the bearing to be detected; the first processing module 40 is configured to determine a codebook of the discrete hidden markov model to obtain observation sequences of different fault types of the bearing to be detected; the second processing module 50 is configured to train the discrete hidden markov model according to the optimization algorithm and the observation sequences of different fault types of the bearing to be detected; the third processing module 60 is configured to perform fault detection on the bearing to be detected according to the feature vectors of different fault types of the bearing to be detected and the trained discrete hidden markov model.
In an embodiment of the present invention, the building module 10 is specifically configured to: constructing a topological mesh by adopting a plurality of discrete hidden Markov models; initializing the topological mesh by adopting an orthogonal metrology method; optimizing each discrete hidden Markov model in each topological grid by adopting double-strategy competitive learning to obtain a first sequence topological grid; judging whether the random real number of each discrete hidden Markov model uniformly distributed in the interval 0 to 1 in the first sequence topological grid is smaller than the parameter in the defined interval 0 to 1; if so, optimizing each discrete hidden Markov model in the first sequence topological grid by adopting cooperative learning to obtain a second sequence topological grid; judging whether the random real number of each discrete hidden Markov model in the second sequence topological grid, which is uniformly distributed in the interval 0 to 1, is smaller than the parameter in the defined interval 0 to 1; if so, optimizing each discrete hidden Markov model in the second sequence topological grid by adopting Gaussian variation to obtain a topological grid of the next iteration; and determining the discrete hidden Markov model with the maximum likelihood value in the topological mesh of the next iteration and executing a self-learning algorithm on the discrete hidden Markov model until the obtained discrete hidden Markov model meets a termination standard.
More specifically, the building module 10 may build the topological mesh G using a plurality of discrete hidden markov models, and the size of the topological mesh G may be S × S. Each discrete hidden Markov model can be fixed at a set position of the topological mesh G and can perform information interaction with the discrete hidden Markov models in the neighborhood. By constructing the topological mesh G by using a plurality of discrete hidden Markov models, extremely limited fault data can be fully mined under a certain topological structure, so that the performance of the fault diagnosis system is improved.
For the sake of illustration, the parameter set of each discrete hidden markov model is converted into a K-dimensional vector as a whole, and the discrete hidden markov model that can be set at grid coordinates (m, n) is Gmn＝[z1,z2,…,zK](m, n is 1,2, …, S), and the objective function value of the discrete hidden markov model may be set to Obj (G)mn) In addition, a topological mesh G formed by the discrete hidden Markov model, namely the discrete hidden Markov model G arranged on the mesh coordinate (m, n) can be arrangedmnThe neighborhood of (c) is:
wherein the content of the first and second substances,
further, G may be providedtIs the topology mesh obtained after the t-th iteration and can be set at GtAnd Gt+1The obtained topological mesh is G in sequencet+1/3And Gt+2/3。BtIs GtOf all discrete hidden markov models of (1) the model with the smallest value of the objective function, BTtIs from G0Up to GtOf all discrete hidden Markov models, PS、PCAnd PGIs a predefined intra-interval (0,1) parameter.
More specifically, build module 10 may define ziIs the ith factor, and has a quantization domain of [ ci,di](wherein, 0. ltoreq. ci≤diLess than or equal to 1), and optionally [ c)i,di]Quantified as H (to)Odd) levels:
wherein j is a variable.
Further, the orthogonal group L can be constructed by using an arrangement methodJ(HQ) Wherein J ═ Hl，Q＝(Hl-1)/(H-1), l being such that
Is the smallest positive integer.
Further, from orthogonal set LJ(HQ) In the random selection of K columns to obtain LJ(HK) And the obtained L can be usedJ(HK) Application to deltai,jThe following J vectors are obtained:
further, S can be randomly selected from the J vectors2The vectors are normalized and restored to be in the form of (pi, A, B), so as to obtain the initial topological mesh G0At the same time, can update B0And t ← 0.
More specifically, the building module 10 may be directed to the topological mesh GtEach discrete hidden Markov model in (a) performs a two-strategy competition learning to obtain a first sequence topological mesh Gt+1/3。
For example, the topological mesh G may be determined firsttEach discrete hidden Markov model in (1), Gij＝[z1,z2,…,zK](i, j ═ 1,2, …, S) of the model M in the neighborhood with the smallest value of the objective functionij＝[α1,α2,…,αK]And then compare Gi,jAnd Mi,jThe objective function value of (1). Wherein, if Gi,jHas an objective function value of less than Mi,jThe objective function value of (2) thenLeaving the discrete hidden Markov model; if G isi,jHas an objective function value of M or morei,jThe objective function value of (1) is Mi,jSub-model Sub generatedij＝[β1,β2,…,βK]Replacing the discrete hidden markov model. A
Further, Mi,jThe sub-models can be generated using the following two strategies:
strategy one: subij＝[β1,β2,…,βK]Is generated by the following formula:
wherein χ ═ αi+U(-1,1)×(αi-zi) U (-1,1) denotes a uniform distribution in [ -1, 1]]Random real numbers of (2).
And (2) strategy two: subij＝[β1,β2,…,βK]Is generated by the following formula
Wherein U (0,1) represents a random real number uniformly distributed over [0,1 ].
It should be noted that the two strategies mentioned above are represented by the probability PSIt is determined if U (0,1) < PSIf not, executing strategy two.
More specifically, for the first sequence the topology grid Gt+1/3Can judge that U (0,1) < P for each discrete hidden Markov model in (1)cIf yes, then adopting cooperative learning to optimize the first sequence topological grid Gt+1/3To obtain a second sequence of topological meshes Gt+2/3. It should be noted that the process of collaborative learning is as follows:
first, can be set from GijAnd MijThe generated search space is
Wherein the content of the first and second substances,ε＝[min(α1,z1),…,min(αK,zK)]，
further, can be
Is quantized to the following F levels:
further, the variable z may be set1,z2,…,zKDivided into groups V, each group can be regarded as a factor, and V-1 integers e can be randomly generated firstlyi(i-1, 2, …, V-1) so as to satisfy 1 < e1＜e2＜…＜eV-1< K, the following V factors can then be generated:
wherein the ith factor gammaiF levels of (a) are defined as:
wherein, [ eta ](i,1),η(i,2),…,η(i,F)]Is to [ eta ]i,1,ηi,2,…,ηi,F]Random rearrangement of the components and e0Is 0.
Further, L may be constructed with reference to step S102U(FV)＝[wij]U×VAnd can use LU(FV) The following U models were generated:
further, among these models, the model having the largest likelihood value may be selected instead of Gi, j.
More specifically, for the second sequence topology mesh Gt+2/3Can judge that U (0,1) < P for each discrete hidden Markov model in (1)cIf true, the second sequence topology mesh G can be optimized by Gaussian variationt+2/3To obtain the topological mesh G of the next iterationt+1. The process of gaussian mutation is:
first, t can be set as the number of iterations and G (0,1/t) can be set as a Gaussian random number generator, thus, for Gij＝[z1,z2,…,zK]Each component of (i, j ═ 1,2, …, S), that is, each discrete hidden markov model, can determine UmWhether (0,1) ≥ 1/K is true, if true, zm←zm+ G (0,1/t), (m ═ 1,2, …, K), where U ism(0,1) is a random number generated independently for each component, i.e., each discrete hidden Markov model. Thus, a new discrete hidden Markov model can be used instead of the original model.
More specifically, build module 10 may build topology mesh G at the next iterationt+1Discrete hidden Markov models with maximum likelihood value determined in, Bt+1And a self-learning algorithm is performed thereon. It should be noted that the process of the self-learning algorithm is as follows:
first, B can bet+1Is marked as Gij＝[z1,z2,…,zK]Thus, a small-scale initial topological mesh sG with the size of S 'multiplied by S' can be generated0And the initial topological mesh sG0In (1) discrete hidden Markov models, i.e. sgkl( k ,1,2, …, S') is:
wherein the content of the first and second substances,
the elements of (A) are:
where σ is the search radius.
Further, an initial topological mesh sG is generated0Then, sG can be settFor the initial topological mesh sG0The grid obtained after t iterations and can be set to sBTtIs (sG)0,sG1,…,sGt) The model with the minimum objective function value in all the discrete hidden Markov models can also be used for converting the topological mesh sGtThe discrete hidden Markov model with the minimum value of the target function is marked as sBtFurthermore, sT can be set as iteration termination criterion, sPGIs a predefined intra-interval (0,1) parameter.
Further, the topological mesh sG can be paired firsttEach discrete hidden Markov model in the topology network carries out a strategy two in the double-strategy competition learning to obtain a topological mesh sGt+1/2(ii) a Further, it can be judged that U (0,1) < sPGIf yes, then the topological mesh sG is checkedt+1/2Each model of (1) performs Gaussian variation to obtain a topological mesh sGt+1(ii) a Finally, the topological grid sG can be judgedt+1Discrete hidden Markov model sB with minimum intermediate objective function valuet+1Whether the objective function value of (a) is greater than sBTtIf yes, sBTt+1←sBt +1If not, sBTt+1←sBTt、sBt+1←sBTt。
Further, it can be determined whether the iteration number t < sT is true, if yes, t ← t +1, and continue to execute the above steps; if not, sBT is usedtSubstituted GijAnd terminates the self-learning process.
The utility of the discrete hidden markov model optimization algorithm of the present invention will be further explained in conjunction with fig. 3.
As shown in fig. 3, when the objective function values of the sphere fault, the outer ring fault and the inner ring fault at 1 horsepower and 2 horsepower are obtained, that is, BF _1, BF _2, ORF _1, ORF _2, IRF _1 and IRF _2, the values of the objective function values are significantly smaller than the objective function values obtained by the BW algorithm at 1 horsepower and 2 horsepower in the prior art, that is, BF _1, BF _2, ORF _1, ORF _2, IRF _1 and IRF _2, so that it can be seen that the discrete hidden markov model optimization algorithm of the present invention can overcome the disadvantage that the BW algorithm in the prior art only has local search capability and has strong dependency on the initial value, and thus can improve the accuracy of parameter learning of the discrete hidden markov model.
In one embodiment of the present invention, the obtaining module 20 may obtain the data of the bearing to be detected through a bearing data center.
In an embodiment of the present invention, the extraction module 30 may perform feature extraction on the data of the bearing to be detected by using Daubechies wavelet and wavelet packet decomposition to obtain feature vectors of different fault types of the bearing to be detected, for example, db4 and 3 decomposition levels of packet decomposition may be used to perform feature extraction on the data of the normal condition of the bearing to be detected and each fault condition.
In an embodiment of the present invention, the first processing module 40 may determine the codebook of the discrete hidden markov model by using a k-means clustering algorithm to obtain an observation sequence of different fault types of the bearing to be detected, for example, the observation sequence O ═ O1o2…oT。
In an embodiment of the present invention, the second processing module 50 may establish an initial discrete hidden markov model for observation sequences of different fault types of the bearing to be detected, and may train the initial discrete hidden markov model using an optimization algorithm until a probability of generating the observation sequences reaches a maximum value, for example, the observation sequence O ═ O1o2…oTThe probability P (O | λ) of (a, B, pi) is the trained discrete hidden markov model.
In one embodiment of the invention, the third processing moduleThe feature vectors of different fault types of the bearing to be detected can be input into the trained discrete hidden markov model, and then the quantization indexes can be adopted: CP ═ logP (o)t|o1o2…ot-1λ) so that the corresponding fault type can be selected as the final diagnostic decision based on the highest CP value.
The utility of the discrete hidden markov model based bearing fault detection system of the present invention will be further explained in conjunction with tables 1 and 2.
As shown in table 1, when the discrete hidden markov model-based bearing fault detection system of the present invention detects a ball fault, an outer ring fault, and an inner ring fault of a bearing, the training operating conditions and the testing operating conditions are 1772rpm, 1hp, 1750rpm, and 2hp, respectively, the control limit is 2.0577, and the fault detection rate is 100%; as further shown in table 2, when the ball fault, the outer ring fault, and the inner ring fault of the bearing are tested, that is, when the ball fault, the outer ring fault, and the inner ring fault of the bearing are respectively tested through 475, 477, and 475 test samples, the fault location accuracy rates are 100%, 98.95%, and 99.79%, respectively, and thus it can be seen that the discrete hidden markov model-based bearing fault detection system of the present invention has higher fault detection accuracy rate and location accuracy rate.
TABLE 1
TABLE 2
According to the discrete hidden Markov model-based bearing fault detection system provided by the embodiment of the invention, an optimization algorithm of a discrete hidden Markov model is established through a construction module, and the extraction module is used for extracting the characteristics of the bearing data to be detected, and determining a codebook of the discrete hidden Markov model through a first processing module, training the discrete hidden Markov model through a second processing module according to the observation sequence of different fault types of the bearing to be detected by the optimization algorithm, the bearing to be detected is subjected to fault detection through the third processing module according to the feature vectors of different fault types of the bearing to be detected and the trained discrete hidden Markov model, so that the dependency on the initial value can be reduced, the searching capability can be improved, therefore, the accuracy of discrete hidden Markov model parameter learning can be improved, and the accuracy of fault detection can be ensured.
In the description of the present invention, the terms "first" and "second" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implying any number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. The meaning of "plurality" is two or more unless specifically limited otherwise.
In the present invention, unless otherwise expressly stated or limited, the terms "mounted," "connected," "secured," and the like are to be construed broadly and can, for example, be fixedly connected, detachably connected, or integrally formed; can be mechanically or electrically connected; either directly or indirectly through intervening media, either internally or in any other relationship. The specific meanings of the above terms in the present invention can be understood by those skilled in the art according to specific situations.
In the present invention, unless otherwise expressly stated or limited, the first feature "on" or "under" the second feature may be directly contacting the first and second features or indirectly contacting the first and second features through an intermediate. Also, a first feature "on," "over," and "above" a second feature may be directly or diagonally above the second feature, or may simply indicate that the first feature is at a higher level than the second feature. A first feature being "under," "below," and "beneath" a second feature may be directly under or obliquely under the first feature, or may simply mean that the first feature is at a lesser elevation than the second feature.
In the description herein, references to the description of the term "one embodiment," "some embodiments," "an example," "a specific example," or "some examples," etc., mean that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the invention. In this specification, the schematic representations of the terms used above do not necessarily refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples. Furthermore, various embodiments or examples and features of different embodiments or examples described in this specification can be combined and combined by one skilled in the art without contradiction.
Claims (10)
1. A bearing fault detection method based on a discrete hidden Markov model is characterized by comprising the following steps:
establishing an optimization algorithm of the discrete hidden Markov model;
acquiring data of a bearing to be detected;
extracting characteristics of the bearing data to be detected to obtain characteristic vectors of different fault types of the bearing to be detected;
determining a codebook of the discrete hidden Markov model to obtain observation sequences of different fault types of the bearing to be detected;
training the discrete hidden Markov model according to the optimization algorithm and observation sequences of different fault types of the bearing to be detected;
and carrying out fault detection on the bearing to be detected according to the feature vectors of different fault types of the bearing to be detected and the trained discrete hidden Markov model.
2. The discrete hidden markov model based bearing fault detection method according to claim 1, wherein the optimization algorithm for establishing the discrete hidden markov model comprises the steps of:
constructing a topological mesh by adopting a plurality of discrete hidden Markov models;
initializing the topological mesh by adopting an orthogonal metrology method;
optimizing each discrete hidden Markov model in each topological grid by adopting double-strategy competitive learning to obtain a first sequence topological grid;
judging whether the random real number of each discrete hidden Markov model uniformly distributed in the interval 0 to 1 in the first sequence topological grid is smaller than the parameter in the defined interval 0 to 1;
if so, optimizing each discrete hidden Markov model in the first sequence topological grid by adopting cooperative learning to obtain a second sequence topological grid;
judging whether the random real number of each discrete hidden Markov model in the second sequence topological grid, which is uniformly distributed in the interval 0 to 1, is smaller than the parameter in the defined interval 0 to 1;
if so, optimizing each discrete hidden Markov model in the second sequence topological grid by adopting Gaussian variation to obtain a topological grid of the next iteration;
and determining the discrete hidden Markov model with the maximum likelihood value in the topological mesh of the next iteration and executing a self-learning algorithm on the discrete hidden Markov model until the obtained discrete hidden Markov model meets a termination standard.
3. The discrete hidden Markov model-based bearing fault detection method according to claim 2, wherein the step of performing feature extraction on the bearing data to be detected to obtain feature vectors of different fault types of the bearing to be detected comprises the following steps:
and performing feature extraction on the bearing data to be detected by adopting Daubechies wavelet and wavelet packet decomposition to obtain feature vectors of different fault types of the bearing to be detected.
4. The discrete hidden Markov model-based bearing fault detection method according to claim 2, wherein determining the codebook of the discrete hidden Markov model to obtain the observation sequences of different fault types of the bearing to be detected comprises the following steps:
and determining the codebook of the discrete hidden Markov model by adopting a k-means clustering algorithm to obtain observation sequences of different fault types of the bearing to be detected.
5. The discrete hidden Markov model-based bearing fault detection method according to claim 3, wherein the training of the discrete hidden Markov model according to the optimization algorithm and the observation sequence of different fault types of the bearing to be detected comprises the following steps:
establishing an initial discrete hidden Markov model for the observation sequences of different fault types of the bearing to be detected;
and training the initial discrete hidden Markov model by adopting the optimization algorithm until the probability of generating the observation sequence reaches the maximum value.
6. A discrete hidden markov model based bearing fault detection system, comprising:
a building module for building an optimization algorithm of the discrete hidden Markov model;
the acquisition module is used for acquiring the data of the bearing to be detected;
the extraction module is used for extracting the characteristics of the data of the bearing to be detected so as to obtain the characteristic vectors of different fault types of the bearing to be detected;
the first processing module is used for determining a codebook of the discrete hidden Markov model to obtain observation sequences of different fault types of the bearing to be detected;
the second processing module is used for training the discrete hidden Markov model according to the optimization algorithm and observation sequences of different fault types of the bearing to be detected;
and the third processing module is used for carrying out fault detection on the bearing to be detected according to the feature vectors of different fault types of the bearing to be detected and the trained discrete hidden Markov model.
7. The discrete hidden markov model based bearing fault detection system of claim 6, wherein the construction module is specifically configured to:
constructing a topological mesh by adopting a plurality of discrete hidden Markov models;
initializing the topological mesh by adopting an orthogonal metrology method;
optimizing each discrete hidden Markov model in each topological grid by adopting double-strategy competitive learning to obtain a first sequence topological grid;
judging whether the random real number of each discrete hidden Markov model uniformly distributed in the interval 0 to 1 in the first sequence topological grid is smaller than the parameter in the defined interval 0 to 1;
if so, optimizing each discrete hidden Markov model in the first sequence topological grid by adopting cooperative learning to obtain a second sequence topological grid;
judging whether the random real number of each discrete hidden Markov model in the second sequence topological grid, which is uniformly distributed in the interval 0 to 1, is smaller than the parameter in the defined interval 0 to 1;
if so, optimizing each discrete hidden Markov model in the second sequence topological grid by adopting Gaussian variation to obtain a topological grid of the next iteration;
and determining the discrete hidden Markov model with the maximum likelihood value in the topological mesh of the next iteration and executing a self-learning algorithm on the discrete hidden Markov model until the obtained discrete hidden Markov model meets a termination standard.
8. The discrete hidden markov model based bearing fault detection system of claim 7, wherein the extraction module is specifically configured to: and performing feature extraction on the bearing data to be detected by adopting Daubechies wavelet and wavelet packet decomposition to obtain feature vectors of different fault types of the bearing to be detected.
9. The discrete hidden markov model based bearing fault detection system of claim 8, wherein the first processing module is specifically configured to: and determining the codebook of the discrete hidden Markov model by adopting a k-means clustering algorithm to obtain observation sequences of different fault types of the bearing to be detected.
10. The discrete hidden markov model based bearing fault detection system of claim 9, wherein the second processing module is specifically configured to:
establishing an initial discrete hidden Markov model for the observation sequences of different fault types of the bearing to be detected;
and training the initial discrete hidden Markov model by adopting the optimization algorithm until the probability of generating the observation sequence reaches the maximum value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111031722.XA CN113723546B (en) | 2021-09-03 | 2021-09-03 | Bearing fault detection method and system based on discrete hidden Markov model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111031722.XA CN113723546B (en) | 2021-09-03 | 2021-09-03 | Bearing fault detection method and system based on discrete hidden Markov model |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113723546A true CN113723546A (en) | 2021-11-30 |
| CN113723546B CN113723546B (en) | 2023-12-22 |
Family
ID=78681430
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111031722.XA Active CN113723546B (en) | 2021-09-03 | 2021-09-03 | Bearing fault detection method and system based on discrete hidden Markov model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113723546B (en) |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100310157A1 (en) * | 2009-06-05 | 2010-12-09 | Samsung Electronics Co., Ltd. | Apparatus and method for video sensor-based human activity and facial expression modeling and recognition |
| CN103778919A (en) * | 2014-01-21 | 2014-05-07 | 南京邮电大学 | Speech coding method based on compressed sensing and sparse representation |
| CN107102223A (en) * | 2017-03-29 | 2017-08-29 | 江苏大学 | NPC photovoltaic DC-to-AC converter method for diagnosing faults based on improved hidden Markov model GHMM |
| CN107274011A (en) * | 2017-06-05 | 2017-10-20 | 上海电力学院 | The equipment state recognition methods of comprehensive Markov model and probability net |
| CN107378641A (en) * | 2017-08-23 | 2017-11-24 | 东北电力大学 | A kind of Monitoring Tool Wear States in Turning based on characteristics of image and LLTSA algorithms |
| CN109142969A (en) * | 2018-07-20 | 2019-01-04 | 西南交通大学 | A kind of power transmission line fault phase selection based on Continuous Hidden Markov Model |
| WO2019019252A1 (en) * | 2017-07-28 | 2019-01-31 | 平安科技（深圳）有限公司 | Acoustic model training method, speech recognition method and apparatus, device and medium |
| US20200019153A1 (en) * | 2018-07-13 | 2020-01-16 | United States Of America, As Represented By The Secretary Of The Navy | Real-Time Automation for Monitor and Control of Electro-Mechanical System |
| CN110705581A (en) * | 2019-07-18 | 2020-01-17 | 浙江运达风电股份有限公司 | Pitch bearing fault identification method based on improved hidden Markov model |
| CN111740582A (en) * | 2020-03-13 | 2020-10-02 | 东南大学 | Wireless charging system PFC fault detection method based on improved HMM |
| CN112345249A (en) * | 2020-07-28 | 2021-02-09 | 江苏理工学院 | Rolling bearing fault diagnosis method based on optimized variational modal decomposition |
| CN113139251A (en) * | 2021-04-23 | 2021-07-20 | 桂林电子科技大学 | Variable working condition rolling bearing fault diagnosis method for optimizing theme correlation analysis |
-
2021
- 2021-09-03 CN CN202111031722.XA patent/CN113723546B/en active Active
Patent Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100310157A1 (en) * | 2009-06-05 | 2010-12-09 | Samsung Electronics Co., Ltd. | Apparatus and method for video sensor-based human activity and facial expression modeling and recognition |
| CN103778919A (en) * | 2014-01-21 | 2014-05-07 | 南京邮电大学 | Speech coding method based on compressed sensing and sparse representation |
| CN107102223A (en) * | 2017-03-29 | 2017-08-29 | 江苏大学 | NPC photovoltaic DC-to-AC converter method for diagnosing faults based on improved hidden Markov model GHMM |
| CN107274011A (en) * | 2017-06-05 | 2017-10-20 | 上海电力学院 | The equipment state recognition methods of comprehensive Markov model and probability net |
| WO2019019252A1 (en) * | 2017-07-28 | 2019-01-31 | 平安科技（深圳）有限公司 | Acoustic model training method, speech recognition method and apparatus, device and medium |
| CN107378641A (en) * | 2017-08-23 | 2017-11-24 | 东北电力大学 | A kind of Monitoring Tool Wear States in Turning based on characteristics of image and LLTSA algorithms |
| US20200019153A1 (en) * | 2018-07-13 | 2020-01-16 | United States Of America, As Represented By The Secretary Of The Navy | Real-Time Automation for Monitor and Control of Electro-Mechanical System |
| CN109142969A (en) * | 2018-07-20 | 2019-01-04 | 西南交通大学 | A kind of power transmission line fault phase selection based on Continuous Hidden Markov Model |
| CN110705581A (en) * | 2019-07-18 | 2020-01-17 | 浙江运达风电股份有限公司 | Pitch bearing fault identification method based on improved hidden Markov model |
| CN111740582A (en) * | 2020-03-13 | 2020-10-02 | 东南大学 | Wireless charging system PFC fault detection method based on improved HMM |
| CN112345249A (en) * | 2020-07-28 | 2021-02-09 | 江苏理工学院 | Rolling bearing fault diagnosis method based on optimized variational modal decomposition |
| CN113139251A (en) * | 2021-04-23 | 2021-07-20 | 桂林电子科技大学 | Variable working condition rolling bearing fault diagnosis method for optimizing theme correlation analysis |
Non-Patent Citations (4)
| Title |
|---|
| JIAWEI LIU等: "A discrete hidden Markov model fault diagnosis strategy based on K-means clustering dedicated to PEM fuel cell systems of tramways", 《INTERNATIONAL JOURNAL OF HYDROGEN ENERGY》, vol. 43, no. 27, pages 12428 - 12441 * |
| 毛明明 等: "基于L-M神经网络的齿轮故障诊断", 《计算机技术与发展》, vol. 21, no. 1, pages 210 - 213 * |
| 赵钰: "基于隐马尔科夫模型的风电机组轴承状态诊断方法", 《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》, no. 4, pages 042 - 251 * |
| 饶雷 等: "基于CNN-SVM和特征融合的齿轮箱故障诊断", 《组合机床与自动化加工技术》, no. 8, pages 130 - 133 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113723546B (en) | 2023-12-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Wang et al. | Pruning from scratch | |
| CN110175386B (en) | Method for predicting temperature of electrical equipment of transformer substation | |
| JP2018022473A (en) | Learning device, identification device, learning identification system, and program | |
| CN107832789B (en) | Feature weighting K nearest neighbor fault diagnosis method based on average influence value data transformation | |
| CN110889111A (en) | Power grid virtual data injection attack detection method based on deep belief network | |
| CN111260620B (en) | Image anomaly detection method and device and electronic equipment | |
| CN115235612B (en) | Intelligent fault diagnosis system and method for servo motor | |
| Fei | The Hybrid Method of VMD‐PSR‐SVD and Improved Binary PSO‐KNN for Fault Diagnosis of Bearing | |
| CN112414715B (en) | Bearing fault diagnosis method based on mixed feature and improved gray level symbiosis algorithm | |
| CN111260024A (en) | Fault detection method and system based on combination of long-term and short-term memory and typical correlation | |
| CN111126560A (en) | Method for optimizing BP neural network based on cloud genetic algorithm | |
| CN114398992A (en) | Intelligent fault diagnosis method based on unsupervised domain adaptation | |
| CN116827685B (en) | Dynamic defense strategy method of micro-service system based on deep reinforcement learning | |
| CN113723546A (en) | Bearing fault detection method and system based on discrete hidden Markov model | |
| WO2023231374A1 (en) | Semi-supervised fault detection and analysis method and apparatus for mechanical device, terminal, and medium | |
| CN115936926A (en) | SMOTE-GBDT-based unbalanced electricity stealing data classification method and device, computer equipment and storage medium | |
| CN116215292A (en) | Charging pile and charging voltage stabilization control method thereof | |
| CN116400168A (en) | Power grid fault diagnosis method and system based on depth feature clustering | |
| Nag et al. | Reconstruction guided meta-learning for few shot open set recognition | |
| CN115422831A (en) | Power grid stability discrimination method and system considering cost sensitivity | |
| CN110472272B (en) | Structural damage identification method based on multi-parameter and convolutional neural network | |
| Goel et al. | Locally optimal detection of stochastic targeted universal adversarial perturbations | |
| Abbas et al. | An adaptive evolutionary algorithm for Volterra system identification | |
| CN111863134B (en) | Method and device for obtaining silencing efficiency prediction result of siRNA sequence | |
| CN115345202B (en) | Method and system for detecting interaction data abnormality of third party load aggregation platform |
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 |