CN114530163A - Method and system for recognizing life cycle of equipment by adopting voice based on density clustering - Google Patents
Method and system for recognizing life cycle of equipment by adopting voice based on density clustering Download PDFInfo
- Publication number
- CN114530163A CN114530163A CN202111679584.6A CN202111679584A CN114530163A CN 114530163 A CN114530163 A CN 114530163A CN 202111679584 A CN202111679584 A CN 202111679584A CN 114530163 A CN114530163 A CN 114530163A
- Authority
- CN
- China
- Prior art keywords
- life cycle
- clustering
- equipment
- tree
- sample
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 50
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 19
- 238000007781 pre-processing Methods 0.000 claims abstract description 14
- 238000012549 training Methods 0.000 claims abstract description 13
- 238000000605 extraction Methods 0.000 claims abstract description 11
- 230000005236 sound signal Effects 0.000 claims abstract description 10
- 238000010276 construction Methods 0.000 claims abstract description 5
- 238000012545 processing Methods 0.000 claims description 17
- 238000004590 computer program Methods 0.000 claims description 10
- 238000005192 partition Methods 0.000 claims description 9
- 238000010845 search algorithm Methods 0.000 claims description 8
- 230000008569 process Effects 0.000 claims description 7
- 238000009432 framing Methods 0.000 claims description 6
- 230000011218 segmentation Effects 0.000 claims description 6
- 238000000926 separation method Methods 0.000 claims description 6
- 238000003860 storage Methods 0.000 claims description 6
- 239000013598 vector Substances 0.000 claims description 6
- 238000005520 cutting process Methods 0.000 claims description 3
- 238000012544 monitoring process Methods 0.000 description 10
- 238000010586 diagram Methods 0.000 description 9
- 238000004519 manufacturing process Methods 0.000 description 6
- 238000012986 modification Methods 0.000 description 5
- 230000004048 modification Effects 0.000 description 5
- 230000006870 function Effects 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 3
- 230000002159 abnormal effect Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000006978 adaptation Effects 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- YHXISWVBGDMDLQ-UHFFFAOYSA-N moclobemide Chemical compound C1=CC(Cl)=CC=C1C(=O)NCCN1CCOCC1 YHXISWVBGDMDLQ-UHFFFAOYSA-N 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
- G10L25/03—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the type of extracted parameters
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
- G10L25/03—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the type of extracted parameters
- G10L25/18—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 characterised by the type of extracted parameters the extracted parameters being spectral information of each sub-band
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
- G10L25/48—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use
- G10L25/51—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use for comparison or discrimination
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Multimedia (AREA)
- Acoustics & Sound (AREA)
- Human Computer Interaction (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Health & Medical Sciences (AREA)
- Signal Processing (AREA)
- Computational Linguistics (AREA)
- Theoretical Computer Science (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a method and a system for recognizing a life cycle of equipment by adopting sound based on density clustering, which comprises a sound signal acquisition step, a data preprocessing step, a feature extraction step, a KD tree construction step, a clustering algorithm model training step and a fuzzy matching recognition life cycle step; and acquiring the life cycle of the equipment to be detected according to the clustering result. The life cycle of the equipment is judged based on the sound information, and the method has the advantages of non-contact, non-stop, no influence of shelters and light rays and the like. The invention adopts a density clustering mode to automatically identify the number of the life cycles of the equipment, and can overcome the problem of insufficient flexibility of the fixed cycle number.
Description
Technical Field
The invention relates to a device life cycle judging technology, in particular to a method and a system for identifying a device life cycle by adopting sound based on density clustering.
Background
The life cycle of the device is the time which elapses from the start of the use of the device until the use of the device is technically or economically unfavorable. The life cycle of the device is also called the natural life. It refers to the time that the equipment has elapsed from being put into use until it is scrapped due to physical wear causing the equipment to lose its value of use altogether.
The lifecycle is an important factor in managing and predicting device failure. The existing equipment life cycle identification method mainly depends on equipment operation time length, online data (SCADA, DCS, MES and the like), various sensor monitoring data (vibration, temperature and the like), video monitoring data and the like to judge the life cycle of the equipment. However, in practical applications, there are many limitations that online data cannot be acquired, a (contact) sensor cannot be installed, video occlusion, light intensity, and the like, and the life cycle of the device cannot be accurately obtained.
In the modern production and manufacturing process, the operation states of various devices in a production line need to be monitored, the device faults need to be identified or predicted, the faulty devices need to be maintained and replaced in time, and normal production is guaranteed. If the identification or prediction algorithm of the equipment fault can be combined with the life cycle information of the equipment, the correctness of the algorithm can be greatly improved. Meanwhile, the life cycle information of the equipment can be utilized to carry out more precise state marking on the monitoring data of the equipment, and the quality of the monitoring data of the equipment is improved.
In the existing method, the life cycle of the equipment is mainly determined as a running-in period, a healthy period, a decline period, a dangerous period, a scrapping period and the like, and the identification of the life cycle comprises 2 types: (1) estimating according to the running time of the equipment by combining with expert experience; (2) according to online data (SCADA, DCS, MES and the like), various sensor monitoring data (vibration, temperature and the like), video monitoring data and the like, machine learning methods such as ARIMA, a gray model, fuzzy logic, a neural network and the like are utilized to identify the life cycle.
The existing equipment state monitoring technology mainly depends on various contact sensors and video monitoring of SCADA/DCS and MES systems to judge the state of equipment. But monitoring based on vibration sensing is greatly affected due to mounting difficulties; the problem that online data cannot be acquired; with video image data, the state of the internal device cannot be detected due to the influence and limitation of video occlusion, light intensity, and the like.
Disclosure of Invention
The invention provides a method and a system for recognizing the life cycle of equipment by adopting sound based on density clustering to avoid the defects in the prior art, so that the life cycle of the equipment is recognized by clustering and modeling the sound information by using a density clustering algorithm.
The invention adopts the following technical scheme to solve the technical problem.
The method for identifying the life cycle of equipment by adopting the sound based on the density clustering is characterized by comprising the following steps:
step 1: collecting a sound signal; acquiring sound information of equipment operation, and performing primary processing on the sound information to obtain an original sound data set of the equipment to be detected;
step 2: a pretreatment step; denoising, framing and windowing the original sound data set;
and 3, step 3: a step of feature extraction; extracting time domain, frequency domain and time-frequency domain characteristics from the original sound data set processed in the step 2 to obtain a standard sample set formed by characteristic value vectors;
and 4, step 4: constructing a KD tree; constructing a KD tree, and dividing the standard sample space in the step 2 into l different areas;
and 5: training a clustering model; clustering the standard sample set by combining a KD tree fast search algorithm and a density clustering algorithm to obtain a clustering model, training the clustering algorithm model and obtaining the trained clustering model;
step 6: fuzzy matching identification life cycle; acquiring real-time sound information of the operation of equipment to be detected, preprocessing the real-time sound information, extracting characteristics, and inputting the real-time sound information into the clustering model trained in the step 5; and (5) carrying out fuzzy matching by using the trained clustering model in the step 5, and identifying the life cycle of the equipment to be detected.
The method for identifying the life cycle of the equipment by adopting the voice based on the density clustering is also characterized in that:
further, in the step 1: the sound information is subjected to preliminary processing including but not limited to consistency check and missing value, outlier processing.
Further, in the step 3: after the characteristics of the sample data are extracted, the sound analog signals of the original sound data set are converted into dimensionless pure numerical characteristic information.
Further, the step 4 comprises the following steps:
step 41: constructing a root node, determining a segmentation hyperplane, and segmenting a sample space into two sub-regions;
step 42: repeatedly cutting the sample space;
step 43: repeating steps 41 and 42 until there are no instances in the two sub-regions, or the number of partitions reaches l, stopping, forming the partition of the KD tree.
Constructing the KD-tree, dividing the standard sample space into i different regions includes the above 3 steps 41-43. The process of constructing a KD tree is illustrated below by a simple example.
For example, for a set of twenty elements Q { (35,1), (26,15), (17,22), (4,23), (26,4), (12,3), (22,8), (26,36), (10,30), (19,2), (5,23), (26,25), (23,16), (10,19), (6,2), (25,23), (4,26), (22,10), (7,12), (22,4) }, its KD tree and its region partitions are as in fig. 3 and fig. 4.
Further, in the step 5, clustering the standard sample set includes the following steps:
step 51: initializing; setting a core object set H, a category set C and a noise set N as empty sets, and setting an unvisited set T as all input samples;
step 52: randomly selecting a sample X in the unaccessed set TiComputing sample X using a KD TreeiThe number n of samples in the neighborhood with eps as the radius;
step 53: judging whether n is larger than a preset density threshold value or not; if yes, the sample X is processediPutting the core object set H; if not, the sample X is processediPutting the noise set N and returning to the step 52;
step 54: establishing a new class CjClass Cj(ii) a Creating candidate set S, the initial element of candidate set S being core object X generated in step 53i；
Step 55: taking any sample X from the candidate set Sp；
Step 56: if the candidate set S is not empty, return to step 55 to execute;
and step 57: c is to bejAdding a category set C;
step 58: if the unaccessed set T is not empty, the procedure returns to step 52
Step 59: output class set C ═ C1，C2，…，Ck}。
Further, in step 52, calculating X using KD treeiThe process of the number n of samples in the neighborhood with eps as the radius comprises the following steps:
step 521: searching a binary tree; searching KD tree from root point and arriving node K in turn1,K2,…,Km，KmFor the last arriving node, at KmCalculating sample points with the distance smaller than eps in the area; m is the total number of nodes;
step 522: backtracking and searching; go back to the previous level in sequencem-1,Km-2,…,K1Judging whether there is a sample in eps neighborhood in other child node space of the father node, and taking K as the samplemThe node is used as the circle center, the eps is used as the radius to draw a circle, the circle and the area separation hyperplane intersection of the layer are judged, if the circle and the area separation hyperplane intersection are intersected, whether samples in the eps neighborhood exist in other child node spaces is calculated, and then the previous level father node is traced back; if the node does not want to be handed over, the node directly backtracks to the parent node at the upper level.
The invention also provides a system adopting the method for recognizing the life cycle of the equipment based on density clustering, which comprises a sound signal acquisition module, a preprocessing module, a feature extraction module, a KD tree construction module, a fuzzy pattern matching module and a life cycle recognition module;
the sound signal acquisition module is used for acquiring sound information of equipment operation, and performing primary processing on the sound information to obtain an original sound data set of the equipment to be detected;
the preprocessing module is used for denoising, framing and windowing the original sound data set;
the characteristic extraction module is used for extracting time domain, frequency domain and time-frequency domain characteristics from the processed original sound data set to obtain a standard sample set composed of characteristic value vectors;
the KD tree constructing module is used for constructing a KD tree and dividing the sample space of the standard sample set into l different areas;
the clustering model training module is used for combining a KD tree fast search algorithm and a density clustering algorithm to perform fast clustering on the standard sample set to obtain a clustering model;
the fuzzy matching module is used for acquiring real-time sound information of the operation of the equipment to be detected, preprocessing the real-time sound information, extracting characteristics of the real-time sound information and inputting the preprocessed real-time sound information to the trained clustering model; carrying out fuzzy pattern matching by using the trained clustering model, and identifying the life cycle of the equipment to be detected by using real-time sound information;
the life cycle identification module is used for acquiring the life cycle of the equipment to be detected according to the life cycle type.
The present invention also provides an electronic device, comprising: at least one processor and a memory communicatively coupled to the at least one processor; wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method for recognizing a device lifecycle using density-based clustering.
The present invention also provides a computer readable storage medium storing a computer program which, when executed by a processor, implements the method for density clustering based lifecycle adaptation for a voice recognition device.
Compared with the prior art, the invention has the beneficial effects that:
the invention discloses a method and a system for adopting a life cycle of sound recognition equipment based on density clustering, which comprises a sound signal acquisition step, a data preprocessing step, a feature extraction step, a KD tree construction step, a clustering algorithm model training step, a fuzzy pattern matching step and a life cycle recognition step; and acquiring the life cycle of the equipment to be detected according to the life cycle type.
1. The life cycle of the equipment is identified by utilizing the sound information, and the method has the advantages of non-contact, no stop, no influence of shelters and light rays and the like
2. The method utilizes density clustering to perform clustering modeling on the sound information to obtain different life cycle models of the equipment, automatically identifies the number of different life cycles of the equipment, and can overcome the problem of insufficient flexibility of fixed cycle number
3. The KD tree is combined to realize the fast calculation of the field samples of density clustering, the time complexity is reduced from the traditional O (n) to the O (log n) of tree searching, the calculation time is shortened, and the calculation efficiency is greatly improved.
The method and the system for recognizing the life cycle of the equipment by adopting the sound based on the density clustering have the advantages of non-contact, non-stop, no influence of shelters and light rays and the like, can automatically recognize the number of different life cycles of the equipment, and can overcome the problem of insufficient flexibility of the fixed number of cycles.
Drawings
FIG. 1 is a flow chart of a method for using a lifecycle of a voice recognition device based on density clustering in accordance with the present invention.
Fig. 2 is a schematic diagram of a KD-tree of an example of a method for recognizing device life cycle using voice based on density clustering according to the present invention.
FIG. 3 is a plot of the area divisions of the example of FIG. 2, where the bolder representation of the line is the first division, and so on, and the width of the line is reduced.
Fig. 4 is a flow chart of the clustering method of step 5 of the method for using the life cycle of the voice recognition device based on density clustering according to the present invention, wherein the minimum density threshold is set empirically or obtained by using some optimization algorithm.
FIG. 5 is an exemplary diagram of the clustering results of the clustering method of the present invention, wherein different gray levels represent different categories.
Fig. 6 is an example of searching for a sample within the eps neighborhood for a certain sample point (13, 8).
The present invention will be further described with reference to the following detailed description and accompanying drawings.
Detailed Description
Referring to fig. 1-6, the method for recognizing a life cycle of a device by using voice based on density clustering according to the present invention comprises the following 6 steps:
step 1: collecting a sound signal; acquiring sound information of equipment operation, and performing primary processing on the sound information to obtain an original sound data set of the equipment to be detected;
step 2: a pretreatment step; denoising, framing and windowing the original sound data set;
and 3, step 3: a step of feature extraction; extracting time domain, frequency domain and time-frequency domain characteristics from the original sound data set processed in the step 2 to obtain a standard sample set formed by characteristic value vectors;
and 4, step 4: constructing a KD tree; constructing a KD tree, and dividing the standard sample space in the step 2 into l different areas;
and 5: training a clustering model; clustering the standard sample set by combining a KD tree fast search algorithm and a density clustering algorithm to obtain a clustering model, training the clustering algorithm model and obtaining the trained clustering model;
step 6: fuzzy matching and life cycle identification; acquiring real-time sound information of the operation of equipment to be detected, preprocessing the real-time sound information, extracting characteristics, and inputting the real-time sound information into the clustering model trained in the step 5; and (5) carrying out fuzzy matching by using the trained clustering model in the step 5, and identifying the life cycle of the equipment to be detected.
The KD tree fast search algorithm is combined with the density clustering algorithm, the time complexity of the density clustering algorithm in calculating the distance between the neighborhood samples is large by adopting the KD tree fast search algorithm, the execution efficiency is improved under the condition of meeting the requirement of calculating the concentration of the samples in the clustering algorithm, the training cost of the concentration clustering algorithm is reduced, the training and identification requirements can be met on low-configuration computing resources when the number of the samples is large, and the computing speed and the computing efficiency are improved.
Further, in the step 1: the sound information is subjected to preliminary processing, including but not limited to consistency check and missing value, outlier processing.
By carrying out consistency check and missing value and abnormal value processing on the sample data set of the sound information, the accuracy rate of sample data classification can be improved, and the execution efficiency can be improved.
Further, in the step 3: after the characteristics of the sample data are extracted, the sound analog signals of the original sound data set are converted into dimensionless pure numerical characteristic information.
Further, the step 4 comprises the following steps:
step 41: constructing a root node, determining a segmentation hyperplane, and segmenting a sample space into two sub-regions;
constructing a root node, selecting a first feature t1Is a coordinate axis; by the characteristics t of all samples1The median of the values of (a) is the root node, will pass through the root node and is associated with the coordinate axis t1The vertical hyperplane is used as a segmentation hyperplane, and the sample space is segmented into two sub-regions: a left region and a right region; the left area and the right area respectively correspond to a left sub-tree and a right sub-tree; the left region corresponds to the feature t1Is smaller than the sub-region of the segmentation point, the right region corresponding to the feature t1Is greater than the sub-region of the cut point;
step 42: repeatedly cutting the sample space;
selecting a segmentation coordinate axis ti+1＝(ti+ 1)% k (k being the dimension of the sample space), repeating S301 for the left and right regions, respectively, i.e. with the feature ti+1The median which is the coordinate axis is a dividing point and is stored as a root node of a subtree, and then the subtree is divided into two subtrees;
step 43: repeating steps 41 and 42 until there are no instances in the two sub-regions, or the number of partitions reaches l, stopping, forming the partition of the KD tree.
Constructing the KD-tree, dividing the standard sample space into i different regions includes the above 3 steps 41-43. The process of constructing a KD tree is illustrated below by a simple example.
For example, for a set of twenty elements Q { (35,1), (26,15), (17,22), (4,23), (26,4), (12,3), (22,8), (26,36), (10,30), (19,2), (5,23), (26,25), (23,16), (10,19), (6,2), (25,23), (4,26), (22,10), (7,12), (22,4) }, its KD tree and its region partitions are as in fig. 3 and fig. 4.
Further, in the step 5, clustering the standard sample set includes the following steps:
step 51: initializing; setting a core object set H, a category set C and a noise set N as empty sets, and setting an unvisited set T as all input samples;
step 52: randomly selecting a sample X in the unaccessed set TiComputing sample X using a KD TreeiThe number n of samples in the neighborhood with eps as the radius;
step 53: judging whether n is larger than a preset density threshold value or not; if yes, the sample X is processediPutting the core object set H; if not, the sample X is processediPutting the noise set N and returning to execute the step 52;
step 54: establishing a new class CjClass Cj(ii) a Creating candidate set S, the initial element of candidate set S being core object X generated in step 53i；
Step 55: taking any sample X from the candidate set Sp；
If XpIf not, adding the new class Cj(ii) a If XpIs a coreSelecting X if the heart object belongs to the core object set HpSamples in the eps neighborhood of (C), not in set CjX in (1)pAdding samples in the eps neighborhood of the cell to a candidate set S;
if XpIn the noise set N, removing the noise from the set N;
if XpIn the unaccessed set T, removing the unaccessed set T from the set T;
step 56: if the candidate set S is not empty, return to step 55 to execute;
and step 57: will CjAdding a category set C;
step 58: if the unaccessed set T is not empty, the procedure returns to step 52
Step 59: output class set C ═ C1，C2，…，Ck}。
Fig. 4 is a flowchart of the clustering method in step 5. Fig. 5 is a diagram showing the clustering result of the 20-element set Q.
Further, in step 52, calculating X using KD treeiThe process of the number n of samples in the neighborhood with eps as the radius comprises the following steps:
step 521: searching a binary tree; searching KD tree from root point, and sequentially reaching node K1,K2,…,Km，KmFor the last arriving node, at KmCalculating sample points with the distance smaller than eps in the area; m is the total number of nodes;
step 522: backtracking and searching; go back to the previous level in sequencem-1,Km-2,…,K1Judging whether there is a sample in eps neighborhood in other child node space of the father node, and taking K as the samplemThe node is used as the circle center, the eps is used as the radius to draw a circle, the circle and the area separation hyperplane intersection of the layer are judged, if the circle and the area separation hyperplane intersection are intersected, whether samples in the eps neighborhood exist in other child node spaces is calculated, and then the previous level father node is traced back; if the intersection is not wanted, directly backtracking to the parent node at the upper level.
As fig. 6 is an example of searching for samples in the eps neighborhood for a certain sample point (13,8), all regions have 5 samples in the eps neighborhood.
The invention also provides a system adopting the method for recognizing the life cycle of the equipment based on density clustering, which comprises a sound signal acquisition module, a preprocessing module, a feature extraction module, a KD tree construction module, a fuzzy pattern matching module and a life cycle recognition module;
the sound signal acquisition module is used for acquiring sound information of equipment operation, and performing primary processing on the sound information to obtain an original sound data set of the equipment to be detected in operation;
the preprocessing module is used for denoising, framing and windowing the original sound data set;
the characteristic extraction module is used for extracting time domain, frequency domain and time-frequency domain characteristics from the processed original sound data set to obtain a standard sample set composed of characteristic value vectors;
the KD tree constructing module is used for constructing a KD tree and dividing the sample space of the standard sample set into l different areas;
the clustering model training module is used for combining a KD tree fast search algorithm and a density clustering algorithm to perform fast clustering on the standard sample set to obtain a clustering model;
the fuzzy matching module is used for acquiring real-time sound information of the operation of the equipment to be detected, preprocessing the real-time sound information, extracting characteristics of the real-time sound information and inputting the preprocessed real-time sound information to the trained clustering model; carrying out fuzzy pattern matching by using the trained clustering model, and identifying the life cycle of the equipment to be detected by using real-time sound information;
the life cycle identification module is used for acquiring the life cycle of the equipment to be detected according to the life cycle type.
The present invention also provides an electronic device, comprising: at least one processor and a memory communicatively coupled to the at least one processor; wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method for employing a voice recognition device lifecycle based on density clustering.
The invention also provides a computer-readable storage medium storing a computer program which, when executed by a processor, implements the method for recognizing a lifecycle of a device using sound based on density clustering.
The method aims to solve the problem that in the prior art, when equipment fault state monitoring is carried out by using sound information of equipment operation, the fault identification or prediction accuracy is low due to the fact that most of data which can be collected by a production line are normal state data and the number of fault or abnormal state data is very small. The invention provides a method and a system for recognizing a life cycle of equipment by using sound based on density clustering.
This patent judges the life cycle of equipment based on sound information, has advantages such as non-contact, do not shut down, do not receive shelter from thing and light influence. The invention automatically identifies the number of different life cycles of the equipment by a density clustering mode, and can overcome the problem of insufficient flexibility of the fixed cycle number.
It should be noted that, in the foregoing embodiments, the descriptions of the respective embodiments have respective emphasis, and reference may be made to relevant descriptions of other embodiments for parts that are not described in detail in a certain embodiment.
As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
While preferred embodiments of the present invention have been described, additional variations and modifications in those embodiments may occur to those skilled in the art once they learn of the basic inventive concepts. Therefore, it is intended that the appended claims be interpreted as including preferred embodiments and all such alterations and modifications as fall within the scope of the invention.
It will be evident to those skilled in the art that the invention is not limited to the details of the foregoing illustrative embodiments, and that the present invention may be embodied in other specific forms without departing from the spirit or essential attributes thereof. The present embodiments are therefore to be considered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein. Any reference sign in a claim should not be construed as limiting the claim concerned.
Furthermore, it should be understood that although the present description refers to embodiments, not every embodiment may contain only a single embodiment, and such description is for clarity only, and those skilled in the art should integrate the description, and the embodiments may be combined as appropriate to form other embodiments understood by those skilled in the art. It will be apparent to those skilled in the art that various changes and modifications may be made in the present invention without departing from the spirit and scope of the invention. Thus, if such modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include such modifications and variations.
Claims (9)
1. The method for identifying the life cycle of equipment by adopting the voice based on density clustering is characterized by comprising the following steps:
step 1: collecting a sound signal; acquiring sound information of equipment operation, and performing primary processing on the sound information to obtain an original sound data set of the equipment to be detected in operation;
step 2: a pretreatment step; denoising, framing and windowing the original sound data set;
and step 3: a step of feature extraction; extracting time domain, frequency domain and time-frequency domain characteristics from the original sound data set processed in the step 2 to obtain a standard sample set formed by characteristic value vectors;
and 4, step 4: constructing a KD tree; constructing a KD tree, and dividing the standard sample space in the step 2 into l different areas;
and 5: training a clustering model; clustering the standard sample set by combining a KD tree fast search algorithm and a density clustering algorithm to obtain a clustering model, training the clustering algorithm model and obtaining the trained clustering model;
step 6: fuzzy matching and life cycle identification; acquiring real-time sound information of the operation of equipment to be detected, preprocessing the real-time sound information, extracting characteristics, and inputting the real-time sound information into the clustering model trained in the step 5; and (5) carrying out fuzzy matching by using the trained clustering model in the step 5, and identifying the life cycle of the equipment to be detected.
2. The method for adopting voice recognition device life cycle based on density clustering as claimed in claim 1, wherein in the step 1: the sound information is subjected to preliminary processing including but not limited to consistency check and missing value, outlier processing.
3. The method for adopting voice recognition device life cycle based on density clustering as claimed in claim 1, wherein in the step 3: after the characteristics of the sample data are extracted, the sound analog signals of the original sound data set are converted into dimensionless pure numerical characteristic information.
4. The method for adopting voice recognition device life cycle based on density clustering as claimed in claim 1, wherein said step 4 comprises the steps of:
step 41: constructing a root node, determining a segmentation hyperplane, and segmenting a sample space into two sub-regions;
step 42: repeatedly cutting the sample space;
step 43: repeating steps 41 and 42 until there are no instances in the two sub-regions, or the number of partitions reaches l, stopping, forming the partition of the KD tree.
Constructing the KD-tree, dividing the standard sample space into i different regions includes the above 3 steps 41-43. The process of constructing a KD tree is illustrated below by a simple example.
For example, for a set of twenty elements Q { (35,1), (26,15), (17,22), (4,23), (26,4), (12,3), (22,8), (26,36), (10,30), (19,2), (5,23), (26,25), (23,16), (10,19), (6,2), (25,23), (4,26), (22,10), (7,12), (22,4) }, its KD tree and its region partitions are as in fig. 3 and fig. 4.
5. The method for adopting voice recognition device life cycle based on density clustering as claimed in claim 1, wherein the clustering the standard sample set in the step 5 comprises the following steps:
step 51: initializing; setting a core object set H, a category set C and a noise set N as empty sets, and setting an unvisited set T as all input samples;
step 52: randomly selecting a sample X in the unaccessed set TiComputing sample X using a KD TreeiThe number n of samples in the neighborhood with eps as the radius;
step 53: judging whether n is larger than a preset density threshold value or not; if yes, the sample X is processediPutting the core object set H; if not, the sample X is processediPutting the noise set N and returning to the step 52;
step 54: establishing a new class CjClass Cj(ii) a Creating candidate set S, the initial element of candidate set S being core object X generated in step 53i；
Step 55: taking any sample X from the candidate set Sp；
Step 56: if the candidate set S is not empty, return to step 55 to execute;
and 57: c is to bejAdding a category set C;
step 58: if the unaccessed set T is not empty, the procedure returns to step 52
Step 59: output class set C ═ C1，C2，…，Ck}。
6. The method of claim 5, wherein in step 52, X is calculated using a KD treeiThe process of the number n of samples in the neighborhood with eps as the radius comprises the following steps:
step 521: searching a binary tree; searching KD tree from root point and arriving node K in turn1,K2,…,Km，KmIs arrived at lastNode at KmCalculating sample points with the distance smaller than eps in the area; m is the total number of nodes;
step 522: backtracking and searching; go back to the previous level in sequencem-1,Km-2,…,K1Judging whether there is a sample in eps neighborhood in other child node space of the father node, and taking K as the samplemThe node is used as the circle center, the eps is used as the radius to draw a circle, the circle and the area separation hyperplane intersection of the layer are judged, if the circle and the area separation hyperplane intersection are intersected, whether samples in the eps neighborhood exist in other child node spaces is calculated, and then the previous level father node is traced back; if the node does not want to be handed over, the node directly backtracks to the parent node at the upper level.
7. A system based on density clustering and adopting a method of life cycle of voice recognition equipment is characterized by comprising a voice signal acquisition module, a preprocessing module, a feature extraction module, a KD tree construction module, a fuzzy pattern matching module and a life cycle recognition module;
the sound signal acquisition module is used for acquiring sound information of equipment operation, and performing primary processing on the sound information to obtain an original sound data set of the equipment to be detected;
the preprocessing module is used for denoising, framing and windowing the original sound data set;
the characteristic extraction module is used for extracting time domain, frequency domain and time-frequency domain characteristics from the processed original sound data set to obtain a standard sample set composed of characteristic value vectors;
the KD tree constructing module is used for constructing a KD tree and dividing the sample space of the standard sample set into l different areas;
the clustering model training module is used for combining a KD tree fast search algorithm and a density clustering algorithm to perform fast clustering on the standard sample set to obtain a clustering model;
the fuzzy matching module is used for acquiring real-time sound information of the operation of the equipment to be detected, preprocessing the real-time sound information, extracting characteristics of the real-time sound information and inputting the preprocessed real-time sound information to the trained clustering model; carrying out fuzzy pattern matching by using the trained clustering model, and identifying the life cycle of the equipment to be detected by using real-time sound information;
the life cycle identification module is used for acquiring the life cycle of the equipment to be detected according to the life cycle type.
8. An electronic device, comprising: at least one processor and a memory communicatively coupled to the at least one processor; wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method for employing a voice recognition device lifecycle based on density clustering of any of claims 1 to 6.
9. A computer-readable storage medium, in which a computer program is stored which, when being executed by a processor, carries out the method for density clustering based adoption of a voice recognition device lifecycle of any one of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111679584.6A CN114530163A (en) | 2021-12-31 | 2021-12-31 | Method and system for recognizing life cycle of equipment by adopting voice based on density clustering |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111679584.6A CN114530163A (en) | 2021-12-31 | 2021-12-31 | Method and system for recognizing life cycle of equipment by adopting voice based on density clustering |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114530163A true CN114530163A (en) | 2022-05-24 |
Family
ID=81621694
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111679584.6A Pending CN114530163A (en) | 2021-12-31 | 2021-12-31 | Method and system for recognizing life cycle of equipment by adopting voice based on density clustering |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114530163A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114827864A (en) * | 2022-06-28 | 2022-07-29 | 武汉左点科技有限公司 | Bone conduction hearing aid sound signal matching gain compensation method and device |
-
2021
- 2021-12-31 CN CN202111679584.6A patent/CN114530163A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114827864A (en) * | 2022-06-28 | 2022-07-29 | 武汉左点科技有限公司 | Bone conduction hearing aid sound signal matching gain compensation method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111914644B (en) | Dual-mode cooperation based weak supervision time sequence action positioning method and system | |
| CN108985380B (en) | Point switch fault identification method based on cluster integration | |
| CN111709465B (en) | Intelligent identification method for rough difference of dam safety monitoring data | |
| CN113688665A (en) | Remote sensing image target detection method and system based on semi-supervised iterative learning | |
| CN110704616B (en) | Equipment alarm work order identification method and device | |
| CN109543693A (en) | Weak labeling data noise reduction method based on regularization label propagation | |
| CN112905380A (en) | System anomaly detection method based on automatic monitoring log | |
| CN112905412A (en) | Method and device for detecting abnormity of key performance index data | |
| CN111160959A (en) | User click conversion estimation method and device | |
| CN112183906A (en) | Machine room environment prediction method and system based on multi-model combined model | |
| CN112308148A (en) | Defect category identification and twin neural network training method, device and storage medium | |
| CN114530163A (en) | Method and system for recognizing life cycle of equipment by adopting voice based on density clustering | |
| CN113670611A (en) | Bearing early degradation evaluation method, system, medium and electronic equipment | |
| CN113159441A (en) | Prediction method and device for implementation condition of banking business project | |
| CN109635008B (en) | Equipment fault detection method based on machine learning | |
| CN116720079A (en) | Wind driven generator fault mode identification method and system based on multi-feature fusion | |
| CN115758086A (en) | Method, device and equipment for detecting faults of cigarette cut-tobacco drier and readable storage medium | |
| CN115600116A (en) | Dynamic detection method, system, storage medium and terminal for time series abnormity | |
| CN113696454A (en) | Artificial intelligence-based extrusion molding equipment fault early warning method and system | |
| CN113139332A (en) | Automatic model construction method, device and equipment | |
| KR20230063742A (en) | Method for detecting defect of product using hierarchical CNN in smart factory, and recording medium thereof | |
| CN113869194A (en) | Variable parameter milling process signal marking method and system based on deep learning | |
| CN110728310A (en) | Target detection model fusion method and system based on hyper-parameter optimization | |
| CN117648890B (en) | Semiconductor device modeling method and system based on artificial intelligence | |
| CN113033694B (en) | Data cleaning method based on deep learning |
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 |