CN112906282B - Method and system for identifying working mode parameters of Sanger neural network parallel principal component extraction - Google Patents
Method and system for identifying working mode parameters of Sanger neural network parallel principal component extraction Download PDFInfo
- Publication number
- CN112906282B CN112906282B CN202110313302.4A CN202110313302A CN112906282B CN 112906282 B CN112906282 B CN 112906282B CN 202110313302 A CN202110313302 A CN 202110313302A CN 112906282 B CN112906282 B CN 112906282B
- Authority
- CN
- China
- Prior art keywords
- time
- invariant
- working mode
- sanger
- neural network
- 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.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/23—Design optimisation, verification or simulation using finite element methods [FEM] or finite difference methods [FDM]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/14—Force analysis or force optimisation, e.g. static or dynamic forces
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/08—Feature extraction
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T90/00—Enabling technologies or technologies with a potential or indirect contribution to GHG emissions mitigation
Abstract
The invention discloses a working mode parameter identification method and a system for Sanger neural network parallel principal component extraction, and relates to a three-dimensional time-invariant structure working mode parameter identification method adopting equidistant feature mapping. Because the equidistant feature mapping algorithm uses the close-proximity relation between the measurement points of the near-ground distance measurement, nonlinear features in the three-dimensional structure can be well expressed, so that the method has a good identification effect on the working modal parameters of the complex three-dimensional structure, and can be used for equipment fault diagnosis, health monitoring and system structure analysis and optimization. The method also comprises a Sanger neural network time-invariant structure working mode parameter identification method, and a Sanger neural network-based parallel principal component extraction working mode parameter identification method is used. The multi-order principal component is obtained through parallel iteration according to the learning rate and the learning rule, is easy to be embedded into multi-core hardware, and has high engineering value.
Description
Technical Field
The invention relates to the field of modal parameter identification, in particular to a method and a system for identifying working modal parameters of a time-invariant structure, comprising a method for identifying the working modal parameters of a three-dimensional time-invariant structure with equidistant feature mapping, a method for identifying the working modal parameters of the time-invariant structure based on Sanger neural network, a method for diagnosing equipment faults and monitoring health states based on equidistant feature mapping, a method for diagnosing equipment faults and monitoring health states based on Sanger neural network and a system for identifying the working modal parameters of the time-invariant structure.
Background
Along with the development and progress of scientific technology, engineering structures in the fields of aerospace, construction, bridges, oceans, machinery and the like gradually develop to large, complex and intelligent directions, loads born by the structures are difficult to measure, if a dynamic model is to be established, a traditional method for obtaining system modal parameters based on measurement input and output is difficult to be applied, and only a working modal parameter identification method requiring measurement output response can be adopted. A time-invariant structure refers to a structure whose kinetic parameters are fixed and do not change over time. According to the mode theory, the vibration characteristics can realize the decoupling of each step in the mode space, and the key of the decoupling is the mode shape vector of each step of time-invariant mode. So far, a one-dimensional time-invariant structure has a relatively complete method system, and is mainly divided into a frequency domain method and a time domain method.
In real engineering application, algorithms are mostly needed to be directly embedded into independent hardware equipment, and with the development of multi-core parallel computing, compared with the common algorithms with serial structures, the parallel algorithms have a larger speed advantage. Neural network algorithms are an important implementation of parallel algorithms due to their unique algorithm structure. The Hebb rule was first proposed in 1949, and Oja et al in 1982 proposed a principal component extraction algorithm based on the Hebb rule, which is called Oja algorithm. However, the Oja algorithm only extracts the first-order principal component, and cannot meet the actual application requirements. In order to extract more order principal components in parallel, t.d. Sanger proposed using a forward neural network that contained only two layers of neurons for input and output, and wang et al proposed new learning rules such that when the input of Sanger network contained multiple eigenvalues, its intermediate weights could also converge to a normalized matrix with orthonormal vectors. The adaptive principal component network (APEX) is proposed by s.y.kung, which greatly reduces the computational effort. Yang et al applied a neural network-based principal component algorithm to the prediction of genetic data, and Jia et al applied it to the extraction of image contours. Principal component extraction based on a neural network is applied to working mode parameter identification, and mode parameters can be extracted in parallel.
Compared with a one-dimensional time-invariant structure, the vibration characteristics of the three-dimensional structure are more complex, and the identification of the mode shape of the three-dimensional structure is more a difficult point and focus of attention in the field of mode analysis. The actual engineering structure is three-dimensional, from one dimension to three dimension, and is a step from scientific and technical research to engineering application based on a modal parameter analysis method. In recent years, a new breakthrough and development are made on the basis of a working mode parameter identification method of a complex three-dimensional continuum structure. In 2014, wang Cheng et al first tried to build principal component analysis models for complex three-dimensional continuum structures, solve the models, and put forward an effective method of assembling three-dimensional mode shapes. However, the principal component analysis model solving method based on eigenvalue decomposition may have a pathological problem, and the three-dimensional modal shape solving method uses a matrix inversion method, so that a solving result is easy to cause a larger error. In 2017, wang Jianying et al propose a method for solving a complex three-dimensional structure by using second-order blind identification, and a matrix inversion method is adopted in an assembly method of a three-dimensional mode shape although a disease state problem solved by a principal component analysis algorithm is avoided. Bai Junqing et al propose a working mode parameter identification method based on local linear embedding, which is the first attempt in the field of working mode parameter identification based on a nonlinear manifold learning algorithm, but the authors do not give detailed explanation on the physical meaning of variables in the solving process. Zhang Jingjing et al optimize the selection method of nearest neighbors in the working mode parameter identification method based on local linear embedding, so that the algorithm is more robust. Dong Longlei et al, on the basis of this, have been based on a composite plate as the object, verify that the working mode parameter identification method based on local advanced embedding has higher robustness to noise. The working mode parameter identification method based on local linear embedding is an algorithm based on nonlinear manifold learning, but mainly aims at the structural nonlinearity caused by the increase of the damping ratio of the structure, and does not deeply excavate the nonlinearity characteristic in the three-dimensional mode response data.
Disclosure of Invention
The invention mainly aims to overcome the defects in the prior art and provides a three-dimensional time-invariant structure working mode parameter identification method based on equidistant feature mapping, a Sanger neural network parallel principal component extraction working mode parameter identification method, a device fault diagnosis and health state monitoring method and a time-invariant structure working mode parameter identification system.
The invention adopts the following technical scheme:
the method for identifying the working modal parameters of the three-dimensional time-invariant structure by equidistant feature mapping is characterized by comprising the following steps of:
s1: acquiring vibration response signal data matrix of stable time domains of a plurality of vibration sensors in three directions under environmental excitation of three-dimensional time-invariant structureAnd->Let the final embedding dimension be d
Wherein (1)>A matrix with dimension D multiplied by T is represented, D represents the number of detection points of the vibration sensor arranged on the time-invariant structure, and T represents the number of time domain sampling points; j is more than or equal to 1 and less than or equal to D; i is more than or equal to 1 and less than or equal to T;
s2: the vibration response signal data matrixes in three directions are assembled to be used as the overall modal response of the three-dimensional time-invariant structure, then the overall modal response is solved, and the new time domain displacement is approximately expressed by modal coordinates:
U three ,V three ,W three in X as a three-dimensional time-invariant structure three ,Y three ,Z three The separation matrix A (t) is a vibration response matrix of the whole complex three-dimensional time-invariant structure, and +.>Is a modal coordinate response;
s3: manifold embedding using equidistant feature map solution modelWhich is the modal coordinate response of a three-dimensional time-invariant structure, the corresponding low-dimensional embedding is +.>
S4: solving a three-dimensional mode shape: according toAnd the dynamics of the three-dimensional time-invariant structure, to obtain the equation +.>Obtaining three-dimensional time-invariant structure three-direction assembly matrix U three ,V three ,W three ,U three ,V three And W is three Respectively representing the mode shapes corresponding to the three directions;
s5: and the modal coordinate response S (t) is calculated to obtain the natural frequency of the three-dimensional time-invariant structure through a single-degree-of-freedom system parameter identification technology.
Is provided withIs a data set comprising T D-dimensional real value vectors sampled on a smooth manifold in a three-dimensional time-invariant structure, the embedded dimension is D, and the S3 specifically comprises the following steps:
s31: constructing a neighborhood graph to obtain each sample pointIs a neighbor of k: sample point->K sample points with nearest Euclidean distance, then the neighborhood graph has edges +.>k is a neighborhood parameter, G is a neighborhood graph formed by k neighbors of all sample points, i is more than or equal to 1 and less than or equal to D, and j is more than or equal to 1 and less than or equal to D;
s32: calculating the shortest path: if neighborhood graph G has edgesThe Euclidean distance between the two points is the shortest path +.>Otherwise the shortest path is set to +.>Let l=1, 2,..
The distance matrix comprises squares of shortest paths among all sample points in the neighborhood graph;
S34: calculating a low-dimensional embedded subspace: calculating d-dimensional embedding by multi-dimensional scaling, i.e. decomposing the eigenvalues of matrix P, the first d eigenvalues lambda 1 ,λ 2 ,…,λ d Corresponding feature vectorIs a characteristic directionAnd (3) embedding the manifold obtained by the eigenvector matrix formed by the quantities into: />
The method for identifying the working mode parameters of the time-invariant structure based on the Sanger neural network is characterized by comprising the following steps of:
vibration response signal data matrix under environment excitation assuming time-invariant structure as inputm represents the number of detection points of the vibration sensor arranged on the time-invariant structure, T represents the number of time domain sampling points, and the first n-order main component +.>n is the number of principal components calculated, m input neurons and n output neurons are set, and { w } ij I=1, 2,. }, m; j=1, 2, n. represents the link between the ith input neuron and the jth output neuron, weight matrixInitializing learning rate lr and loss function as +.>
1) Calculating the current network output Y (J) (t)=W (J) X (t), J is the number of iterations;
2) Obtaining the update amount of the current weight matrix according to the learning rule and the learning rate, wherein
3) Updating the weight matrix W in parallel according to the updating amount of the weight matrix obtained in the step 2) J+1 ,W (J+1) =W (J) +ΔW (J +1) ;
4) Calculating the current error lossWherein lambda is i Is the first n-order principal component->Middle->Is a characteristic value of (2);
5) If e < eta or J < J max Eta is the error threshold, J max If the number of iterations is the maximum, j=j+1 and returns to step 2), otherwise turning to step 7);
6) Obtaining a final weight matrix W, w=w (J+1) The mode shape is the mode shape;
7) Calculating a final principal component Y (t) =WX (t), namely a modal coordinate response;
8) And calculating the natural frequency of the time-invariant structure from the modal coordinate response Y (t) by a single-degree-of-freedom system parameter identification technology.
The equipment fault diagnosis and health state monitoring method based on the equidistant feature mapping is realized based on the three-dimensional time-invariant structure working mode parameter identification method based on the equidistant feature mapping, and comprises the following steps:
step a), collecting a group of vibration response data matrixes in three directions as sample data, and carrying out normalization processing to determine the final embedded dimension as d;
step b), the vibration response data matrixes in three directions are assembled into a vibration response matrix of a three-dimensional time-invariant structure through a direct assembly method;
step c) solving the modal coordinate response of the whole model by using equidistant feature mapping, identifying the modal coordinate response of the three-dimensional time-invariant structure, and inversely substituting the modal coordinate response into the assembled vibration response matrix to obtain the three-dimensional modal shape; the modal coordinate response is calculated by a single-degree-of-freedom system parameter identification technology to obtain the natural frequency of the three-dimensional time-invariant structure.
Step d), analyzing and comparing the mode shape and the natural frequency of the three-dimensional time-invariant structure obtained through calculation with the mode parameters before the tested equipment fails, and determining whether the equipment fails or not and the position of the failure;
step e) when new sample data is introduced, repeating steps b) -d) until the sample is over.
The method for diagnosing equipment faults and monitoring health states based on the Sanger neural network is realized based on the method for identifying the working mode parameters of the time-invariant structure of the Sanger neural network, and comprises the following steps:
step a), collecting a group of unidirectional vibration response data matrixes as sample data, and carrying out normalization processing to determine a final error threshold eta;
step b), using Sanger neural network to solve the mode coordinate response Y (t) and the working mode shape W of the whole model;
step c), calculating the natural frequency of the time-invariant structure from the modal coordinate response Y (t) through a single-degree-of-freedom system parameter identification technology;
step d), analyzing and comparing the mode shape and the natural frequency of the three-dimensional time-invariant structure obtained through calculation with the mode parameters before the tested equipment fails, and determining whether the equipment fails or not and the position of the failure;
step e) when new sample data is introduced, repeating steps b) -d) until the sample is over.
Preferably, the mode parameters include instantaneous mode frequency, instantaneous mode shape.
The working mode parameter identification system of the time-invariant structure is characterized by comprising the time-invariant engineering structure and a time-invariant engineering structure, wherein the time-invariant engineering structure is used for simulating the time-invariant engineering structure of the working mode parameter identification to be identified;
the three-way vibration response signal sensor is arranged on the working structure, and the stable vibration response signal of the measured unchanged engineering structure is obtained by measuring the three-way vibration response signal sensor arranged on the working structure;
the excitation device is used for simulating environmental excitation in a working state;
the vibration data acquisition module is used for vibration data input, signal conditioning and A/D data acquisition and conversion;
the control and data processing module is provided with an OMAP processor, and adopts an equidistant feature mapping three-dimensional time-invariant structure working mode parameter identification method or a Sanger neural network time-invariant structure working mode parameter identification method to identify working mode parameters and obtain diagnosis information;
the communication module is used for uploading the vibration data and the diagnosis information to the upper computer for storage and analysis;
the storage module is used for storing vibration data;
the working mode parameter identification module is used for identifying working mode parameters of the engineering structure;
the liquid crystal display module is used for displaying the diagnosis result and the waveform information;
the power module is used for providing power;
the key control module and the reset module are used for inputting and resetting parameters.
As can be seen from the above description of the present invention, compared with the prior art, the present invention has the following advantages:
the invention discloses a three-dimensional time-invariant structure working mode parameter identification method based on equidistant feature mapping, which is used for identifying time-variant transient working mode parameters (transient working mode shape and transient working mode natural frequency) of a time-invariant structure on line in real time only by actually measuring stable vibration response signals. Firstly, the difference between a one-dimensional time-invariant structure and a three-dimensional time-invariant structure is analyzed, and the difficulty of the three-dimensional time-invariant structure is analyzed. A new method of assembling a three-dimensional mode shape is proposed and compared to previous methods of assembling a three-dimensional mode shape. And solving by using a nonlinear equidistant feature mapping algorithm on the basis of a novel three-dimensional mode shape assembling method. Finally, a finite element calculation model of the three-dimensional thin-wall cylindrical shell is established at the free two ends, vibration response simulation calculation under the load condition with white noise is carried out, the reliability of an algorithm is verified through a simulation result, and compared with the previous method, the method provided by the chapter has higher identification precision.
The working mode parameter identification method based on the Sanger neural network, disclosed by the invention, has the advantages that the principle and physical significance of the proposed algorithm are deduced and analyzed in detail, the method is a parallel principal component analysis algorithm, the working mode parameter identification method based on the algorithm has high identification precision, and the common pathological problems in the traditional principal component analysis algorithm can be avoided. The method can be further embedded into multi-core hardware to improve the algorithm efficiency greatly.
According to the equipment fault diagnosis and health state monitoring method based on equidistant feature mapping and the Sanger neural network, provided by the invention, the overall modal parameters of the three-dimensional structure can be measured and whether faults exist can be determined. Arranging a plurality of three-way vibration response signal sensors on key points of a measurement structure, identifying working mode parameters of the vibration response signals in three directions obtained through measurement through a direct assembly mode and an equidistant feature mapping method, detecting the mode parameters of a three-dimensional system structure, and applying the mode parameters to health state monitoring and fault diagnosis of a large-scale industrial structure.
The invention relates to a working mode parameter identification system with a time-invariant structure, which takes a time-invariant engineering structure, a three-way vibration response signal sensor arranged on the working structure, excitation equipment and an OMAP5912 embedded processor as cores, and integrates units such as data acquisition, liquid crystal display, data storage, control, data processing, working mode parameter identification and the like. The device simulates environmental excitation under the working state by using excitation equipment, and obtains a stable vibration response signal of a measured unchanged engineering structure by using a three-way vibration response signal sensor arranged on the working structure. The design of the device fully utilizes the dual-core structure (ARM core and DSP core) of the OMAP, has the functions of ARM core such as low power consumption, flexible task scheduling, high processing speed and the like, and has the function of powerful digital processing analysis of the DSP core, and the dual-core structure and the DSP core are effectively combined to realize real-time online acquisition, processing, transmission and analysis of vibration signals. Meanwhile, the Ethernet is adopted for data transmission, so that the rapid and efficient transmission of data is realized, the loss of signals in transmission is avoided, the remote diagnosis, the monitoring and the resource sharing are realized, and the defects of offline and delayed data acquisition and the like are overcome. The design of the device effectively combines the signal processing technology, the circuit design, the computer technology, the algorithm design and the working mode parameter identification and fault analysis technology, realizes the digitization, the automation and the intellectualization of a diagnosis system, and has potential application value.
Drawings
FIG. 1 is a flow of direct assembly to solve a three-dimensional structural mode shape;
FIG. 2 is a flow of an algorithm for identifying parameters of a three-dimensional working mode based on equidistant feature mapping;
FIG. 3 is a schematic flow chart of a three-dimensional time-invariant structure working mode parameter identification method according to an embodiment of the present invention;
FIG. 4 is a schematic view of a three-dimensional cylindrical shell structure;
fig. 5 is displacement response data in the X direction of the 1118 th observation point;
fig. 6 is displacement response data in the Y direction of the 1118 th observation point;
fig. 7 is displacement response data in the Z direction of the 1118 th observation point;
FIG. 8 is a true first, second, and third order mode shapes;
FIG. 9 is a true fourth, fifth, and sixth order mode shapes;
FIG. 10 is a result obtained after a modal coordinate response FFT;
FIG. 11 is the result of an identified mode shape;
FIG. 12 is a flow chart for identifying operational mode parameters based on a neural network principal component algorithm;
FIG. 13 is a schematic representation of the Sanger neural network principal component algorithm;
FIG. 14 is a Sanger neural network modal response FFT variation result;
FIG. 15 is a comparison of a Sanger neural network identified mode shape with a true mode shape;
FIG. 16 is a block diagram of an experimental setup;
FIG. 17 is a schematic illustration of a simple beam structure;
FIG. 18 is a diagram of the experimental set-up system;
FIG. 19 is a system for identifying parameters of the working mode of the thin-walled cylindrical shell in a cantilever state;
Detailed Description
The present invention will be described in further detail below with reference to the attached drawings and examples, wherein the examples are described as being only some, but not all, examples of the present invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
The invention aims to realize the identification of the modal parameters of a complex three-dimensional time-invariant structure, in particular to the identification of the three-dimensional modal shape and modal frequency, and in order to make the invention clear and easy to understand, the invention is further described in detail below with reference to the accompanying drawings and the detailed description.
The method for identifying the working mode parameters of the three-dimensional time-invariant structure based on manifold learning combines an equidistant feature mapping method and a direct assembly method, wherein the method can identify the working mode parameters of the time-invariant structure, and nonlinear relations exist in mode response data of a three-dimensional complex continuous structure.
The method comprises the following specific steps:
s1: acquiring stable time domain vibration response signal data matrixes of multiple vibration sensors in three directions under environmental excitation of three-dimensional time-invariant structureAnd->Let the final embedding dimension be d->
Wherein (1)>A matrix with dimension D multiplied by T is represented, D represents the number of detection points of the vibration sensor arranged on the time-invariant structure, and T represents the number of time domain sampling points; j is more than or equal to 1 and less than or equal to D; i is more than or equal to 1 and less than or equal to T;
s2: the vibration response signal data matrixes in three directions are assembled to be used as the overall modal response of the three-dimensional time-invariant structure, then the overall modal response is solved, and the new time domain displacement is approximately expressed by modal coordinates:
U three ,V three ,W three in X as a three-dimensional time-invariant structure three ,Y three ,Z three The separation matrix A (t) is a vibration response matrix of the whole complex three-dimensional time-invariant structure, and +.>Is a modal coordinate response;
s3: manifold embedding using equidistant feature map solution modelThe modal coordinate response of the three-dimensional time-invariant structure is that d is the embedded dimension, k is the neighborhood parameter, and the corresponding low-dimensional embedding isG is a neighborhood graph of k neighbors of all sample points. Equidistant feature mapping (ISOMAP algorithm) is as follows:
s31: is provided withIs a data set comprising T D-dimensional real-valued vectors sampled on a smooth manifold, with embedded dimension D. Constructing a neighborhood graph to obtain each sample point +.>Is a neighbor of k: sample point->K sample points with nearest Euclidean distance, then the neighborhood graph has edges +.>k is a neighborhood parameter, G is a neighborhood graph formed by k neighbors of all sample points, i is more than or equal to 1 and less than or equal to D, and j is more than or equal to 1 and less than or equal to D. Smooth manifold refers to a manifold with a smooth structure in a three-dimensional time-invariant structure, so that the manifold can be subjected to calculus. Also called differential manifold. The latter example uses a cylindrical shell. The sensors are uniformly arranged on the surface of the object, so that vibration responses in three directions are obtained. Here, X is the direction in which the vibration response is largest among the three directions of X, Y, and Z.
S32: calculating the shortest path: if neighborhood graph G has edgesThe Euclidean distance between the two points is the shortest path +.>Otherwise the shortest path is set to +.>Let l=1, 2,..
The distance matrix comprises squares of shortest paths among all sample points in the neighborhood graph;
S34: calculating a low-dimensional embedded subspace: calculating d-dimensional embedding by multi-dimensional scaling, i.e. decomposing the eigenvalues of matrix P, the first d eigenvalues lambda 1 ,λ 2 ,…,λ d Corresponding feature vectorIs a feature vector matrix formed by feature vectors, and the manifold is embedded as follows: />
S4: solving a three-dimensional mode shape: according toAnd the dynamics of the three-dimensional time-invariant structure, to obtain the equation +.>Assembly matrix U capable of obtaining three directions of three-dimensional time-invariant structure three ,V three ,W three ,U three ,V three And W is three Respectively representing the mode shapes corresponding to the three directions;
s5: the modal coordinate response S (t) can be calculated to obtain the natural frequency of the three-dimensional structure through a single-degree-of-freedom system parameter identification technology.
Example 1:
as shown in fig. 4, is a cylindrical shell having a complex three-dimensional structure, wherein the boundary condition of the cylindrical shell is a simple two-end branch. A number of vibration sensors are arranged on the surface of the cylindrical shell, and vibration responses in three directions are recorded. Wherein, the parameters of the cylindrical shell are set as follows: the cylindrical shell had a thickness of 0.005 m, a length of 0.37 m, a radius of 0.1825 m, an elastic modulus of 205Gpa, a poisson's ratio of 0.3, and a density of 7850.
In the simulation, the modal damping ratio η is 0.03 and 0.1 in total. The cylindrical shell was equally divided into 38 circles around its axial direction, with 115 observation points uniformly disposed in each circle, so d=38×115=4370 observation points in total, and the excitation applied to the cylindrical shell was gaussian white noise excitation. The sampling frequency was set to 5120Hz and the sampling time was 1 second, so t=5120. And calculating by using an LMS finite element method, and acquiring structural displacement response data of 3 dimensions of X, Y and Z under different damping ratios from each observation point to form a response data set of 3 directions, wherein the response data set is displacement response data of the 1118 th observation point in three dimensions as shown in fig. 5, 6 and 7.
Setting the sample point close proximity number k=40, embedding dimension d=5, and at a damping ratio of 0.03, fig. 10 is the result obtained after a modal coordinate response FFT, and fig. 9 and 11 are the true and identified modal shapes, respectively.
Table 1 is a comparison of the performance of the two solutions, table 2 is a quantitative comparison of the identification mode frequencies, table 3 is a quantitative comparison of the identification results of the mode shapes, and table 4 is a comparison of the different neighbor values of the MACs.
Table 1 comparison of the performance of the two solutions
Table 2 quantitative comparison of identification modality frequencies
Table 3 quantitative comparison of the modal shape recognition results
Table 4 comparison of MAC different neighbor values
Referring to fig. 12 to fig. 18, the invention further provides a method for identifying working mode parameters of a time-invariant structure based on a Sanger neural network, which specifically comprises the following steps:
vibration response signal data matrix under environment excitation assuming time-invariant structure as inputm represents the number of detection points of the vibration sensor arranged on the time-invariant structure, and T represents the number of time domain sampling points. Solving the first n-order principal component +.>Setting m input neurons and n output neurons, then collecting { w } ij I=1, 2,. }, m; j=1, 2, n. represents the link between the ith input neuron and the jth output neuron, weight matrixInitializing learning rate lr and loss function as +.>
The principal component analysis method based on the neural network mainly comprises three types of an Oja algorithm, a Sanger algorithm and an APEX algorithm, wherein the Oja algorithm only extracts first-order principal components, and the method has great limitation on the practical application of the working mode parameter identification; the APEX algorithm has lower calculation amount, but focuses on the separated principal component, and the middle weight of the APEX algorithm has no actual physical meaning, namely if the APEX algorithm is applied to the identification of working mode parameters, the mode shape with important meaning cannot be directly obtained. The Sanger algorithm not only extracts multi-order principal components in parallel, but also has one-to-one correspondence between the middle weight and the mode shape, so that the Sanger algorithm is selected to be more suitable for the identification of working mode parameters;
1) Calculating the current network output Y (J) (t)=W (J) X (t), J is the number of iterations;
2) Obtaining the update amount of the current weight matrix according to the learning rule and the learning rate, wherein
3) Updating the weight matrix W in parallel according to the updating amount of the weight matrix obtained in the step 2) J+1 ,W (J+1) =W (J) +ΔW (J +1) ;
4) Calculating the current error lossWherein lambda is i Is the first n-order principal component->Middle->Is a characteristic value of (2);
5) If e < eta or J < J max Eta is the error threshold, J max If the number of iterations is the maximum, j=j+1 and returns to step 2), otherwise turning to step 7);
6) Obtaining a final weight matrix W, w=w (J+1) The mode shape is the mode shape;
7) Calculating a final principal component Y (t) =WX (t), namely a modal coordinate response;
8) And calculating the natural frequency of the time-invariant structure from the modal coordinate response Y (t) by a single-degree-of-freedom system parameter identification technology.
Examples:
the method comprises the steps of selecting a time-invariant simply supported beam structure as a research object, dispersing the simply supported beam structure into a limited multi-degree-of-freedom system, solving a dynamic system response by means of a Simulink/Matlab platform and obtaining vibration response data by means of a Newmark-beta integration method. Equally dividing an undamped simply supported beam with the length of 1 meter into 1000 equal parts at equal intervals to generate 1001 response measuring points;
1) Sinusoidal excitation is applied to the undamped simply supported beams;
2) Concentrated application of multi-frequency sinusoidal excitation at 0.2 m and obtaining response data;
3) The sampling frequency is 4096Hz, and the time domain sampling point T=20481;
4) The data are obtained through Matlab 7.0 simulation;
5) And taking the undamped result obtained by finite element solution as the real mode shape and the mode natural frequency.
6) Computer configuration:
operating System Windows 7 flagship edition 64 bit (6.1, version 7601)
Manufacturer Dell Inc. model Inpipron N5110
Processor Intel (R) Core (TM) i5-2430M
CPU@2.40Hz(4CPUs),~2.4Ghz
Memory 4096MB of RAM.
7) A mode shape evaluation criterion (MAC) is used to compare the linear transformation vector identified from the iterative principal component extraction algorithm with the true mode shape as follows:
wherein the method comprises the steps ofIs the identified mode shape, < >>Representing the true mode shape +.>Representing the inner product of the two vectors. The value range of the MAC value is +.>When the MAC value is closer to 1, the identification mode shape is closer to the true mode shape.
Setting parameters of a neural network principal component algorithm: setting a learning rate lr=0.01; setting the number n=5 of output neurons; setting an error threshold η=0.001; maximum number of iterations J max =10000; initial weight matrix w=0.5×i n×m =0.5*I 6 ×1001 。
Table 5 comparison between three neural network principal component algorithms
Table 6 quantitative comparison of Sanger neural network frequency identification results:
table 7 quantitative comparison of MAC values for Sanger identified mode shapes:
table 8Sanger neural network identifies the mode shape from the MAC comparison results
The invention also provides a device fault diagnosis and health state monitoring method based on equidistant feature mapping, which is realized based on the three-dimensional time-invariant structure working mode parameter identification method based on equidistant feature mapping, and comprises the following steps:
step a), collecting a group of vibration response data matrixes in three directions as sample data, and carrying out normalization processing to determine the final embedded dimension as d;
step b), the vibration response data matrixes in three directions are assembled into a vibration response matrix of a three-dimensional time-invariant structure through a direct assembly method;
step c) solving the modal coordinate response of the whole model by using equidistant feature mapping, identifying the modal coordinate response of the three-dimensional time-invariant structure, and inversely substituting the modal coordinate response into the assembled vibration response matrix to obtain the three-dimensional modal shape; the modal coordinate response is calculated by a single-degree-of-freedom system parameter identification technology to obtain the natural frequency of the three-dimensional time-invariant structure.
Step d), analyzing and comparing the mode shape and the natural frequency of the three-dimensional time-invariant structure obtained through calculation with the mode parameters before the tested equipment fails, and determining whether the equipment fails or not and the position of the failure;
step e) when new sample data is introduced, repeating steps b) -d) until the sample is over.
The invention also provides a Sanger neural network-based equipment fault diagnosis and health state monitoring method, which is realized based on the Sanger neural network time-invariant structure working mode parameter identification method, and comprises the following steps:
step a), collecting a group of unidirectional vibration response data matrixes as sample data, and carrying out normalization processing to determine a final error threshold eta;
step b), using Sanger neural network to solve the mode coordinate response Y (t) and the working mode shape W of the whole model;
step c), calculating the natural frequency of the time-invariant structure from the modal coordinate response Y (t) through a single-degree-of-freedom system parameter identification technology;
step d), analyzing and comparing the mode shape and the natural frequency of the three-dimensional time-invariant structure obtained through calculation with the mode parameters before the tested equipment fails, and determining whether the equipment fails or not and the position of the failure;
step e) when new sample data is introduced, repeating steps b) -d) until the sample is over.
As shown in fig. 16 and 19, the invention further provides a working mode parameter identification system with a time-invariant structure, which comprises a time-invariant engineering structure, a three-way vibration response signal sensor 10, vibration excitation equipment 20, a vibration data acquisition module, a working mode parameter identification module, a communication module, a storage module, a liquid crystal display module, a power supply module, a control module, a reset module, an OMAP processor (with a dual-core structure, an ARM core and a DSP core, and has the characteristics of low power consumption and high data processing capability), and the like. The control and data processing module fully plays the role of signal processing and ARM control of the DSP, adopts a three-dimensional time-invariant structure working mode parameter identification method of distance feature mapping or a Sanger neural network time-invariant structure working mode parameter identification method to identify working mode parameters, and obtains diagnosis information.
The three-dimensional vibration response signal sensor is used for measuring and obtaining a stable vibration response signal of a measured unchanged engineering structure, and the excitation equipment is used for simulating environmental excitation in a working state; the vibration data acquisition module comprises functions of signal input, signal conditioning, A/D data acquisition and conversion and the like; the working mode parameter identification module is used for identifying working mode parameters of the engineering structure. A storage module storing a large amount of vibration data; the liquid crystal display module uses an LCD liquid crystal screen to output and display the diagnosis result and waveform information; the power module is responsible for supplying power to the whole system; the reset and control key module is responsible for the functions of resetting the system, inputting parameters and the like; and the communication module is responsible for uploading the acquired data and diagnosis information to the upper computer for storage and analysis.
The device carries out filtering, amplifying, sampling and other treatments on the collected signals in a certain sampling mode and a data organization mode, and transmits the signals to an OMAP processor DSP core for processing, and the three-dimensional time-invariant structure working mode parameter identification method and the Sanger neural network time-invariant structure working mode parameter identification method which are mapped by the equidistant features are adopted to respectively identify the three-dimensional time-invariant structure and the one-dimensional time-invariant structure working mode parameters and obtain diagnosis information.
The data processing is mainly completed by a DSP unit, the DSP performs time domain, frequency domain and algorithm analysis based on improved limited memory principal component analysis on the acquired data, and then the data is sent to an ARM control module through an SPI interface. The control module mainly completes real-time data storage, stores the stored data in a certain format, and transmits the processed data and the original data to the upper computer through the Ethernet for analysis, waveform display and storage.
Examples:
FIG. 19 is a system for identifying parameters of a working mode of a thin-walled cylindrical shell in a cantilever state, comprising a time-invariant engineering structure of the thin-walled cylindrical shell in the cantilever state in a simulated working state, a three-way vibration response signal sensor arranged on the working structure, a small-sized vibration table as excitation equipment for simulating and generating environmental excitation, an OMAP processor for applying an algorithm to obtain diagnosis information, a computer for displaying diagnosis results and the like.
In the present specification, each embodiment is described in a progressive manner, and each embodiment is mainly described in a different point from other embodiments, and identical and similar parts between the embodiments are all enough to refer to each other. For the system disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant points refer to the description of the method section.
The principles and embodiments of the present invention have been described herein with reference to specific examples, the description of which is intended only to assist in understanding the methods of the present invention and the core ideas thereof; also, it is within the scope of the present invention to be modified by those of ordinary skill in the art in light of the present teachings. In view of the foregoing, this description should not be construed as limiting the invention.
Claims (4)
1. The working mode parameter identification method of the time-invariant structure based on the Sanger neural network is characterized by comprising the following steps of:
vibration response signal data matrix under environment excitation assuming time-invariant structure as inputm represents the number of detection points of the vibration sensor arranged on the time-invariant structure, T represents the number of time domain sampling points, and the first n-order main component +.>n is the number of principal components calculated, m input neurons and n output neurons are set, and { w } ij I=1, 2,. }, m; j=1, 2, …, n, representing the link between the ith input neuron and the jth output neuron, weight matrixInitializing learning rate lr and loss function as +.>
1) Calculating the current network output Y (J) (t)=W (J) X (t), J is the iteration numberA number;
2) Obtaining the update amount of the current weight matrix according to the learning rule and the learning rate, wherein
3) Updating the weight matrix W in parallel according to the updating amount of the weight matrix obtained in the step 2) J+1 ,W (J+1) =W (J) +ΔW (J+1) ;
4) Calculating the current error lossWherein lambda is i Is the first n-order principal component->Middle->Is a characteristic value of (2);
5) If e < eta or J < J max Eta is the error threshold, J max If the number of iterations is the maximum, j=j+1 and returns to step 2), otherwise turning to step 7);
6) Obtaining a final weight matrix W, w=w (J+1) The mode shape is the mode shape;
7) Calculating a final principal component Y (t) =WX (t), namely a modal coordinate response;
8) And calculating the natural frequency of the time-invariant structure from the modal coordinate response Y (t) by a single-degree-of-freedom system parameter identification technology.
2. The method for diagnosing equipment faults and monitoring health states based on the Sanger neural network is realized based on the method for identifying working mode parameters of a time-invariant structure of the Sanger neural network, and comprises the following steps of:
step a), collecting a group of unidirectional vibration response data matrixes as sample data, and carrying out normalization processing to determine a final error threshold eta;
step b), using Sanger neural network to solve the mode coordinate response Y (t) and the working mode shape W of the whole model;
step c), calculating the natural frequency of the time-invariant structure from the modal coordinate response Y (t) through a single-degree-of-freedom system parameter identification technology;
step d), analyzing and comparing the mode shape and the natural frequency of the three-dimensional time-invariant structure obtained through calculation with the mode parameters before the tested equipment fails, and determining whether the equipment fails or not and the position of the failure;
step e) when new sample data is introduced, repeating steps b) -d) until the sample is over.
3. A Sanger neural network based device failure diagnosis and health status monitoring method as claimed in claim 2, wherein the modal parameters include instantaneous modal frequency, instantaneous modal shape.
4. The working mode parameter identification system of the time-invariant structure is characterized by comprising the time-invariant engineering structure and a time-invariant engineering structure, wherein the time-invariant engineering structure is used for simulating the time-invariant engineering structure of the working mode parameter identification to be identified;
the three-way vibration response signal sensor is arranged on the working structure, and the stable vibration response signal of the measured unchanged engineering structure is obtained by measuring the three-way vibration response signal sensor arranged on the working structure;
the excitation device is used for simulating environmental excitation in a working state;
the vibration data acquisition module is used for vibration data input, signal conditioning and A/D data acquisition and conversion;
the control and data processing module is provided with an OMAP processor, and the working mode parameter identification method based on the Sanger neural network time-invariant structure is adopted to identify the working mode parameter and obtain diagnosis information;
the communication module is used for uploading the vibration data and the diagnosis information to the upper computer for storage and analysis;
the storage module is used for storing vibration data;
the working mode parameter identification module is used for identifying working mode parameters of the engineering structure;
the liquid crystal display module is used for displaying the diagnosis result and the waveform information;
the power module is used for providing power;
the key control module and the reset module are used for inputting and resetting parameters.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110313302.4A CN112906282B (en) | 2018-05-07 | 2018-05-07 | Method and system for identifying working mode parameters of Sanger neural network parallel principal component extraction |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810428108.9A CN108594660B (en) | 2018-05-07 | 2018-05-07 | Working modal parameter identification method and system of time invariant structure |
CN202110313302.4A CN112906282B (en) | 2018-05-07 | 2018-05-07 | Method and system for identifying working mode parameters of Sanger neural network parallel principal component extraction |
Related Parent Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810428108.9A Division CN108594660B (en) | 2018-05-07 | 2018-05-07 | Working modal parameter identification method and system of time invariant structure |
Publications (2)
Publication Number | Publication Date |
---|---|
CN112906282A CN112906282A (en) | 2021-06-04 |
CN112906282B true CN112906282B (en) | 2023-05-26 |
Family
ID=63635630
Family Applications (2)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810428108.9A Active CN108594660B (en) | 2018-05-07 | 2018-05-07 | Working modal parameter identification method and system of time invariant structure |
CN202110313302.4A Active CN112906282B (en) | 2018-05-07 | 2018-05-07 | Method and system for identifying working mode parameters of Sanger neural network parallel principal component extraction |
Family Applications Before (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810428108.9A Active CN108594660B (en) | 2018-05-07 | 2018-05-07 | Working modal parameter identification method and system of time invariant structure |
Country Status (1)
Country | Link |
---|---|
CN (2) | CN108594660B (en) |
Families Citing this family (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110037681B (en) * | 2019-03-15 | 2021-07-09 | 上海数创医疗科技有限公司 | Method and device for recognizing heart rhythm type based on improved self-encoder network |
CN112257213B (en) * | 2019-07-02 | 2024-02-09 | 大连民族大学 | Method for describing large-deflection vibration of rubber cylindrical shell |
CN110597230B (en) * | 2019-09-24 | 2020-08-18 | 清华大学深圳国际研究生院 | Active fault diagnosis method, computer readable storage medium and control method |
CN112629786A (en) * | 2020-12-03 | 2021-04-09 | 华侨大学 | Working mode parameter identification method and equipment fault diagnosis method |
CN114675547B (en) * | 2022-05-30 | 2022-09-23 | 华中科技大学 | MIMO active vibration reduction control method and system with deep learning automatic diagnosis mechanism |
CN117668597A (en) * | 2023-11-13 | 2024-03-08 | 华南理工大学 | Structural modal parameter identification method and system considering excitation non-stationarity |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO1999044106A1 (en) * | 1998-02-27 | 1999-09-02 | MTU MOTOREN- UND TURBINEN-UNION MüNCHEN GMBH | System and method for diagnosing the conditions of a powerplant |
CN106555788A (en) * | 2016-11-11 | 2017-04-05 | 河北工业大学 | Application of the deep learning based on Fuzzy Processing in hydraulic equipment fault diagnosis |
CN107271127A (en) * | 2017-06-27 | 2017-10-20 | 华侨大学 | Based on the operational modal parameter recognition methods and device extracted from iteration pivot |
Family Cites Families (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP2024862A4 (en) * | 2006-05-02 | 2011-08-10 | Invidi Tech Corp | Fuzzy logic based viewer identification for targeted asset delivery system |
CN104112072B (en) * | 2014-07-15 | 2017-04-05 | 华侨大学 | Operational modal parameter recognition methods based on the principal component analysis of wavelet threshold denoising |
CN104166804B (en) * | 2014-08-20 | 2018-01-30 | 中国科学技术大学 | A kind of operation mode discrimination method based on time-frequency domain list source point sparse component analysis |
CN104698837B (en) * | 2014-12-11 | 2017-10-20 | 华侨大学 | A kind of time-varying linear structure operational modal parameter recognition methods, device and application |
CN106446503B (en) * | 2016-07-21 | 2019-02-22 | 华侨大学 | Forget the time-varying operation mode recognition methods of auto-covariance matrix recursion pivot |
CN106446502B (en) * | 2016-07-21 | 2019-04-26 | 华侨大学 | The time-varying operation mode online recognition method of feature vector recursion with forgetting factor |
-
2018
- 2018-05-07 CN CN201810428108.9A patent/CN108594660B/en active Active
- 2018-05-07 CN CN202110313302.4A patent/CN112906282B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO1999044106A1 (en) * | 1998-02-27 | 1999-09-02 | MTU MOTOREN- UND TURBINEN-UNION MüNCHEN GMBH | System and method for diagnosing the conditions of a powerplant |
CN106555788A (en) * | 2016-11-11 | 2017-04-05 | 河北工业大学 | Application of the deep learning based on Fuzzy Processing in hydraulic equipment fault diagnosis |
CN107271127A (en) * | 2017-06-27 | 2017-10-20 | 华侨大学 | Based on the operational modal parameter recognition methods and device extracted from iteration pivot |
Non-Patent Citations (5)
Title |
---|
DE-ICA优化算法在工作模态参数识别的应用;雷宇翔 等;《华侨大学学报》;第286-292页 * |
Moving window self-iteration PCE based OMA for slow linear time-varying structures;Tianshu Zhang et.al;《JOURNAL OF VIBROENGINEERING》;全文 * |
Three-dimensional operational modal analysis based on self-iteration principal component extraction and direct matrix assembly;Tianshu Zhang et.al;《JOURNAL OF VIBROENGINEERING》;第6262-6276页 * |
基于流形学习的线性结构工作模态参数识别;张天舒;《中国优秀硕士学位论文全文数据库工程科技Ⅱ辑》;全文 * |
基于自迭代主元抽取的工作模态参数识别;张天舒;王成;官威;王建英;刘艳;谢晓东;;数据采集与处理(02);全文 * |
Also Published As
Publication number | Publication date |
---|---|
CN112906282A (en) | 2021-06-04 |
CN108594660A (en) | 2018-09-28 |
CN108594660B (en) | 2021-04-30 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN112906282B (en) | Method and system for identifying working mode parameters of Sanger neural network parallel principal component extraction | |
WO2021004154A1 (en) | Method for predicting remaining life of numerical control machine tool | |
CN107271127B (en) | Based on the operational modal parameter recognition methods extracted from iteration pivot and device | |
CN110377942B (en) | Multi-model space-time modeling method based on finite Gaussian mixture model | |
Zhu et al. | A rapid structural damage detection method using integrated ANFIS and interval modeling technique | |
CN109063818B (en) | Thermal process model online identification method and device | |
CN111461187A (en) | Intelligent building settlement detection system | |
Saravanakumar et al. | Hierarchical symbolic analysis and particle swarm optimization based fault diagnosis model for rotating machineries with deep neural networks | |
CN109033021B (en) | Design method of linear equation solver based on variable parameter convergence neural network | |
Sun et al. | Study on fault diagnosis algorithm in WSN nodes based on RPCA model and SVDD for multi-class classification | |
Nandanoori et al. | Graph neural network and Koopman models for learning networked dynamics: A comparative study on power grid transients prediction | |
CN116227689A (en) | Photovoltaic panel pollution degree state evaluation and cleaning decision system | |
Lai et al. | Digital twin-based structural health monitoring by combining measurement and computational data: An aircraft wing example | |
Wang et al. | A probabilistic detectability-based structural sensor network design methodology for prognostics and health management | |
Hong et al. | Dynamic cooperative identification based on synergetics for pipe structural health monitoring with piezoceramic transducers | |
CN113486573A (en) | Cable core temperature prediction method | |
CN111626000A (en) | Bridge state evaluation method, device, equipment and storage medium | |
CN116364203A (en) | Water quality prediction method, system and device based on deep learning | |
CN115618735A (en) | Facility structure health monitoring method based on digital twins and related device | |
CN112016235B (en) | Impact load identification method and system for flexible antenna structure | |
CN115099268A (en) | Wheeled robot intelligent fault diagnosis method and system based on graph convolution network | |
CN114638048A (en) | Three-dimensional spray pipe flow field rapid prediction and sensitivity parameter analysis method and device | |
Wang | [Retracted] Recognition, Processing, and Detection of Sensor Fault Signal Based on Genetic Algorithm | |
CN109993211B (en) | Damage identification method based on structural acceleration AR model coefficient | |
Yang et al. | Statistical pattern recognition for Structural Health Monitoring using ESN feature extraction method |
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 |