CN117194900A - Equipment operation lightweight monitoring method and system based on self-adaptive sensing - Google Patents
Equipment operation lightweight monitoring method and system based on self-adaptive sensing Download PDFInfo
- Publication number
- CN117194900A CN117194900A CN202311248570.8A CN202311248570A CN117194900A CN 117194900 A CN117194900 A CN 117194900A CN 202311248570 A CN202311248570 A CN 202311248570A CN 117194900 A CN117194900 A CN 117194900A
- Authority
- CN
- China
- Prior art keywords
- equipment operation
- data
- operation data
- equipment
- item
- 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
- 238000012544 monitoring process Methods 0.000 title claims abstract description 63
- 238000000034 method Methods 0.000 title claims abstract description 39
- 238000013507 mapping Methods 0.000 claims abstract description 77
- 238000013528 artificial neural network Methods 0.000 claims abstract description 40
- 238000012545 processing Methods 0.000 claims description 28
- 239000011159 matrix material Substances 0.000 claims description 24
- 230000003044 adaptive effect Effects 0.000 claims description 12
- 238000013527 convolutional neural network Methods 0.000 claims description 12
- 230000006870 function Effects 0.000 claims description 12
- 230000008569 process Effects 0.000 claims description 12
- 238000012549 training Methods 0.000 claims description 8
- 230000001133 acceleration Effects 0.000 claims description 6
- 238000000354 decomposition reaction Methods 0.000 claims description 6
- 238000000605 extraction Methods 0.000 claims description 6
- 238000007781 pre-processing Methods 0.000 claims description 6
- 238000013139 quantization Methods 0.000 claims description 6
- 238000004891 communication Methods 0.000 claims description 3
- 238000005315 distribution function Methods 0.000 claims description 3
- 238000003860 storage Methods 0.000 claims description 3
- 239000013585 weight reducing agent Substances 0.000 claims description 3
- 230000008447 perception Effects 0.000 claims description 2
- 230000005540 biological transmission Effects 0.000 abstract description 9
- 239000002699 waste material Substances 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 230000036541 health Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
Classifications
-
- 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
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/02—Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS]
Landscapes
- Testing And Monitoring For Control Systems (AREA)
Abstract
The application discloses a device operation lightweight monitoring method and system based on self-adaptive sensing, comprising the following steps: according to the first equipment operation data and the first equipment operation state, a first target mapping relation is learned by using a neural network; noise adding and orthogonal initializing are carried out on the first equipment operation data, and the processed first equipment operation data is lightened through a first target mapping relation to obtain second equipment operation data; and according to the first equipment operation data and the second equipment operation data, learning a second target mapping relation by using the neural network. The application realizes quick online transmission and reduces the resource consumption of computers and mobile equipment; and learning a second target mapping relation by using the neural network, wherein the second target mapping relation is used for obtaining second equipment operation data according to the first equipment operation data, so that self-adaption light weight is realized.
Description
Technical Field
The application relates to the technical field of equipment monitoring light weight, in particular to an equipment operation light weight monitoring method and system based on self-adaptive sensing.
Background
The device monitoring refers to monitoring and detecting the running state of the device in real time or periodically to acquire data such as performance indexes and fault information of the device, and analyzing, processing and explaining the data so as to discover the health condition of the device in time. The equipment monitoring is usually realized by technical means such as a sensor, a monitoring system, computer software and the like, so that the reliability, availability and efficiency of the equipment are improved, the failure rate and maintenance cost of the equipment are reduced, and the life cycle value of the equipment is improved. The equipment monitoring is widely applied in the fields of manufacturing industry, energy, traffic, construction, environmental protection and the like.
The equipment operation monitoring generally monitors all working condition information of equipment, so that the comprehensiveness of monitoring is guaranteed, the accuracy of equipment operation monitoring can be guaranteed through comprehensive monitoring, but the comprehensive monitoring can lead to heavy informatization burden such as information on-line transmission and operation consumption, equipment resource waste is easy to cause, and the monitoring process is complex.
Disclosure of Invention
The application aims to provide a device operation lightweight monitoring method and system based on self-adaptive sensing, which are used for solving the technical problems that in the prior art, comprehensive monitoring can lead to heavy informatization burden such as information on-line transmission, operation consumption and the like, equipment resource waste is easy to cause, and the monitoring process is complicated.
In order to solve the technical problems, the application specifically provides the following technical scheme:
an equipment operation lightweight monitoring method based on self-adaptive sensing comprises the following steps:
acquiring first equipment operation data and a first equipment operation state, wherein the first equipment operation state and the first equipment operation data have a time sequence corresponding relation, the first equipment operation data comprises a plurality of first data items, the plurality of first data items correspond to a plurality of working condition information of target equipment, and the first equipment operation state corresponds to a working condition category of the target equipment;
according to the first equipment operation data and the first equipment operation state, a neural network is utilized to learn a first target mapping relation, and the first target mapping relation is used for obtaining the first equipment operation state according to the first equipment operation data;
noise adding and orthogonal initializing are carried out on the first equipment operation data, the processed first equipment operation data is lightened through a first target mapping relation, and second equipment operation data is obtained and used for representing the first equipment operation data after the weight reduction;
and according to the first equipment operation data and the second equipment operation data, a second target mapping relation is learned by using a neural network, and the second target mapping relation is used for obtaining the second equipment operation data according to the first equipment operation data.
As a preferred embodiment of the present application, the obtaining of the first target mapping relationship includes:
taking the first equipment operation data as a first input item of the neural network, taking the first equipment operation state as a first output item of the neural network, and learning and training the neural network based on the first input item and the first output item to obtain a first target mapping relation;
the expression of the first target mapping relation is as follows:
Label＝CNN(S1)；
in the formula, label is a first equipment operation state, S1 is first equipment operation data, and CNN is a neural network.
As a preferred embodiment of the present application, the noise adding and orthogonal initializing process for the first device operation data includes:
carrying out data quantization on each first data item in the first equipment operation data by utilizing a spearman algorithm, summing the data quantization results of the association relation among each first data item to obtain the data symmetry of the first equipment operation data,
if the data symmetry of the first equipment operation data is larger than a first preset value, gaussian noise is added to each first data item in the first equipment operation data, and orthogonal initialization processing is carried out on each first data item added with the Gaussian noise;
and if the data symmetry of the first equipment operation data is smaller than or equal to a first preset value, carrying out orthogonal initialization processing on each first data item in the first equipment operation data.
As a preferred embodiment of the present application, the adding gaussian noise to each first data item includes:
and adding Gaussian noise to each first data item in sequence, wherein the adding expression of the Gaussian noise is as follows:
w l _add＝w l +N(0,βσ(r(w l )))；
wherein w is l Add is the first data item, w, after adding Gaussian noise l For the first data item of the first, N (0, βσ (r (w l ) And) is a gaussian distribution function, σ (r (w) l ) Is r (w) l ) Beta is the data symmetry of the first device operation data, r (w l ) Is the association relationship between the first data items, wherein r (w l )∈P K×K ，P K×K And K is a matrix specification of the incidence relation matrix among the first data items, and l is a counting variable.
As a preferred embodiment of the present application, the orthogonal initialization process includes:
each first data item added with Gaussian noise is subjected to orthogonal initialization processing, and the method comprises the following steps:
performing QR decomposition on each first data item added with Gaussian noise to obtain a quadrature matrix Q and an upper triangular matrix R;
carrying out orthogonal initialization operation on each first data item added with Gaussian noise to obtain a new first data item, wherein the function expression of the new data item is as follows:
w l _add_new＝Q⊙sign(diag(R))；
wherein w is l The add_new is the orthogonal initialization result of the first data item after Gaussian noise is added, diag (R) is a diagonal matrix of R, sign (diag (R)) is a sign function of diag (R);
performing orthogonal initialization processing on each first data item in the first device operation data, including:
performing QR decomposition on each first data item of the first equipment operation data to obtain a quadrature matrix Q and an upper triangular matrix R;
carrying out orthogonal initialization operation on each first data item of the first equipment operation data to obtain a new first data item, wherein the function expression of the new data item is as follows:
w l _new＝Q⊙sign(diag(R))；
wherein w is l The_new is the orthogonal initialization result of the first data item in the first equipment operation data, the diag (R) is the diagonal matrix of R, and sign (diag (R)) is the sign function of diag (R)。
As a preferred solution of the present application, the light-weighting of the processed first device operation data according to the first target mapping relationship includes:
respectively and sequentially replacing corresponding first data items in the first equipment operation data by each first data item after orthogonal initialization to obtain a group of third equipment operation data, wherein only one first data item in the third equipment operation data and the first equipment operation data is different;
obtaining the second equipment operation state through the first target mapping relation according to the third equipment operation data;
comparing the second device operating state with the first device operating state according to the corresponding relation between the third device operating data and the first device operating data, wherein,
if the first equipment operation state is the same as the second equipment operation state, marking the first data item corresponding to the second equipment operation state as an unnecessary data item;
if the first equipment operation state is different from the second equipment operation state, marking the first data item corresponding to the second equipment operation state as a necessary data item;
and eliminating unnecessary data items in the first equipment operation data to obtain second equipment operation data.
As a preferred embodiment of the present application, the obtaining of the second target mapping relationship includes:
taking the first equipment operation data as a second input item of the neural network, taking the second equipment operation data as a second output item of the neural network, and learning and training the neural network based on the second input item and the second output item to obtain a second target mapping relation;
the expression of the second target mapping relation is as follows:
S2＝CNN(S1)；
in the formula, S2 is second equipment operation data, S1 is first equipment operation data, and CNN is a neural network.
As a preferred embodiment of the present application, the first device operation data is at least one of image data, text data, or audio data.
As a preferred solution of the present application, the present application provides a lightweight monitoring system applying the apparatus operation lightweight monitoring method based on adaptive sensing, including:
the data acquisition module comprises at least one of an instrument, a remote terminal unit and a programmable logic control unit and is used for acquiring a plurality of working condition information of target equipment;
the hardware acceleration image preprocessing module is used for processing a plurality of working condition information of the target equipment into first equipment operation data;
the feature extraction module is used for learning a first target mapping relation by utilizing a neural network according to the first equipment operation data and the first equipment operation state;
the feature recognition result is used for carrying out noise addition and orthogonal initialization processing on the first equipment operation data, and carrying out light weight on the processed first equipment operation data through a first target mapping relation to obtain second equipment operation data;
the communication module is used for transmitting the first equipment operation data of the hardware acceleration image preprocessing module to the feature extraction module and the feature recognition result, and transmitting the second equipment operation data of the feature recognition result to the lower computer module;
the upper computer data display and storage module is used for storing a first target mapping relation and storing a second target mapping relation.
As a preferred scheme of the application, the first equipment operation state has a time sequence corresponding relation with first equipment operation data, the first equipment operation data comprises a plurality of first data items, the plurality of first data items correspond to a plurality of working condition information of target equipment, and the first equipment operation state corresponds to a working condition category of the target equipment;
the first target mapping relation is used for obtaining a first equipment operation state according to first equipment operation data;
the second equipment operation data are used for representing the first equipment operation data after light weight;
and the second target mapping relation is used for obtaining second equipment operation data according to the first equipment operation data.
Compared with the prior art, the application has the following beneficial effects:
according to the application, noise addition and orthogonal initialization processing are carried out on the first equipment operation data, the processed first equipment operation data is light-weighted through the first target mapping relation, so that quick online transmission is realized, and the resource consumption of a computer and mobile equipment is reduced; and learning a second target mapping relation by using the neural network, wherein the second target mapping relation is used for obtaining second equipment operation data according to the first equipment operation data, so that self-adaption light weight is realized.
Drawings
In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below. It will be apparent to those of ordinary skill in the art that the drawings in the following description are exemplary only and that other implementations can be obtained from the extensions of the drawings provided without inventive effort.
FIG. 1 is a flowchart of a method for monitoring the operation light weight of a device with adaptive sensing according to an embodiment of the present application;
fig. 2 is a block diagram of a lightweight monitoring system according to an embodiment of the present application.
Detailed Description
The following description of the embodiments of the present application will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present application, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are intended to be within the scope of the application.
The equipment operation monitoring generally monitors all working condition information of equipment, so that the comprehensiveness of monitoring is guaranteed, the accuracy of equipment operation monitoring can be guaranteed through comprehensive monitoring, but the comprehensive monitoring can lead to heavy informatization burden such as information on-line transmission and operation consumption, equipment resource waste is easy to cause, and the monitoring process is complex. Therefore, the application provides a device operation light-weight monitoring method based on self-adaptive perception, which obtains second device operation data according to first device operation data, realizes self-adaptive light-weight of device monitoring, realizes quick online transmission, and reduces resource consumption of computers and mobile devices.
As shown in fig. 1, the application provides a device operation lightweight monitoring method based on adaptive sensing, which comprises the following steps:
acquiring first equipment operation data and a first equipment operation state, wherein the first equipment operation state and the first equipment operation data have a time sequence corresponding relation, the first equipment operation data comprises a plurality of first data items, the plurality of first data items correspond to a plurality of working condition information of target equipment, and the first equipment operation state corresponds to a working condition category of the target equipment;
according to the first equipment operation data and the first equipment operation state, a neural network is utilized to learn a first target mapping relation, and the first target mapping relation is used for obtaining the first equipment operation state according to the first equipment operation data;
noise adding and orthogonal initializing are carried out on the first equipment operation data, the processed first equipment operation data is lightened through a first target mapping relation, second equipment operation data is obtained, and the second equipment operation data is used for representing the first equipment operation data after the weight reduction;
and according to the first equipment operation data and the second equipment operation data, learning a second target mapping relation by using the neural network, wherein the second target mapping relation is used for obtaining the second equipment operation data according to the first equipment operation data.
In order to determine the mapping relation between the equipment operation state and the equipment operation data, the equipment operation state can be obtained by learning and training through the neural network to obtain the equipment operation data, and the first equipment operation data characterizes the monitoring of all the working condition information of the equipment, so that the equipment operation state is accurately identified through the first equipment operation data obtained through comprehensive monitoring, and the accuracy of equipment operation monitoring can be ensured through comprehensive monitoring.
In order to lighten the monitoring of the equipment, the application carries out the light-weight processing on all the working condition information of the equipment, namely, the equipment running state is obtained through the monitoring of less working condition information, the light-weight processing is obtained according to the first equipment running data and the first target mapping relation, so that the light-weight equipment running data (second equipment running data) inherits the accuracy of equipment running state identification in the first target mapping relation, namely, the first equipment running state can be accurately obtained by utilizing the second equipment running data, and compared with the light-weight of the first equipment running data, the second equipment running data has quick online transmission, and the resource consumption of computers and mobile equipment (monitoring sensors of the equipment running data) is reduced.
In order to further apply the light weight of the equipment monitoring, the light weight process is packaged into the second target mapping relation, so that the light weight self-adaption can be realized, namely, the second equipment operation data is directly obtained through the first equipment operation data, a complicated light weight analysis process is not needed, and the light weight process is simpler and more convenient.
In order to determine the mapping relation between the equipment operation state and the equipment operation data, the application uses the neural network to perform learning training to obtain the equipment operation state directly through the equipment operation data, and the first equipment operation data characterizes the monitoring of all the working condition information of the equipment, so that the equipment operation state is accurately identified through the first equipment operation data obtained through comprehensive monitoring, the accuracy of the equipment operation monitoring can be ensured through the comprehensive monitoring, and the application comprises the following specific steps:
the obtaining of the first target mapping relation comprises the following steps:
taking the first equipment operation data as a first input item of the neural network, taking the first equipment operation state as a first output item of the neural network, and learning and training the neural network based on the first input item and the first output item to obtain a first target mapping relation;
the expression of the first target mapping relation is as follows:
Label＝CNN(S1)；
in the formula, label is a first equipment operation state, S1 is first equipment operation data, and CNN is a neural network.
According to the application, through analyzing the data symmetry of the first equipment operation data, the higher the data symmetry is, the more the data items which are related to each other and exist in the first equipment operation data are indicated, and the more the redundant data items are in the operation state identification process, so that the data symmetry in the first equipment operation data is destroyed through noise addition and orthogonal initialization processing, the data items which are added with noise but do not influence the state identification result in the first equipment operation data can be identified by utilizing the difference between the identification results of the equipment operation states generated by the first equipment operation data before and after the processing, the noise is added to the data items but do not influence the state identification result, the lower the necessity of the data items in the state identification is indicated, or the existence of the data items is suggested, and the light weight of the first equipment operation data is realized.
And the light weight of the first equipment operation data depends on the first target mapping relation, so that the second equipment operation data after the light weight of the first equipment operation data is used for maintaining the state identification accuracy of comprehensive monitoring when the state equipment is operated.
Noise adding and orthogonal initializing processing are carried out on the first equipment operation data, and the method comprises the following steps:
carrying out data quantization on each first data item in the first equipment operation data by utilizing a spearman algorithm, summing the data quantization results of the association relation among each first data item to obtain the data symmetry of the first equipment operation data,
if the data symmetry of the first equipment operation data is larger than a first preset value, gaussian noise is added to each first data item in the first equipment operation data, and orthogonal initialization processing is carried out on each first data item added with the Gaussian noise;
and if the data symmetry of the first equipment operation data is smaller than or equal to a first preset value, carrying out orthogonal initialization processing on each first data item in the first equipment operation data.
Adding gaussian noise to each first data item, comprising:
and adding Gaussian noise to each first data item in sequence, wherein the adding expression of the Gaussian noise is as follows:
w l _add＝w l +N(0,βσ(r(w l )))；
wherein w is l Add is the first data item, w, after adding Gaussian noise l For the first data item of the first, N (0, βσ (r (w l ) And) is a gaussian distribution function, σ (r (w) l ) Is r (w) l ) Beta is the data symmetry of the first device operation data, r (w l ) Is the association relationship between the first data items, wherein r (w l )∈P K×K ，P K×K And K is a matrix specification of the incidence relation matrix among the first data items, and l is a counting variable.
An orthogonal initialization process comprising:
each first data item added with Gaussian noise is subjected to orthogonal initialization processing, and the method comprises the following steps:
performing QR decomposition on each first data item added with Gaussian noise to obtain a quadrature matrix Q and an upper triangular matrix R;
carrying out orthogonal initialization operation on each first data item added with Gaussian noise to obtain a new first data item, wherein the function expression of the new data item is as follows:
w l _add_new＝Q⊙sign(diag(R))；
wherein w is l The add_new is the orthogonal initialization result of the first data item after Gaussian noise is added, diag (R) is a diagonal matrix of R, sign (diag (R)) is a sign function of diag (R);
performing orthogonal initialization processing on each first data item in the first device operation data, including:
performing QR decomposition on each first data item of the first equipment operation data to obtain a quadrature matrix Q and an upper triangular matrix R;
carrying out orthogonal initialization operation on each first data item of the first equipment operation data to obtain a new first data item, wherein the function expression of the new data item is as follows:
w l _new＝Q⊙sign(diag(R))；
wherein w is l And_new is the orthogonal initialization result of the first data item in the first equipment operation data, diag (R) is a diagonal matrix of R, and sign (diag (R)) is a sign function of diag (R).
The light weight of the processed first equipment operation data is carried out through the first target mapping relation, and the method comprises the following steps:
respectively and sequentially replacing corresponding first data items in the first equipment operation data by each first data item after orthogonal initialization to obtain a group of third equipment operation data, wherein only one first data item in the third equipment operation data and the first equipment operation data is different;
obtaining the second equipment operation state through the first target mapping relation according to the third equipment operation data;
comparing the second device operating state with the first device operating state according to the corresponding relation between the third device operating data and the first device operating data, wherein,
if the first equipment operation state is the same as the second equipment operation state, marking the first data item corresponding to the second equipment operation state as an unnecessary data item;
if the first equipment operation state is different from the second equipment operation state, marking the first data item corresponding to the second equipment operation state as a necessary data item;
and eliminating unnecessary data items in the first equipment operation data to obtain second equipment operation data.
In order to lighten the monitoring of the equipment, the application carries out the light-weight processing on all the working condition information of the equipment, namely, the equipment running state is obtained through the monitoring of less working condition information, the light-weight processing is obtained according to the first equipment running data and the first target mapping relation, so that the light-weight equipment running data (second equipment running data) inherits the accuracy of equipment running state identification in the first target mapping relation, namely, the first equipment running state can be accurately obtained by utilizing the second equipment running data, and compared with the light-weight of the first equipment running data, the second equipment running data has quick online transmission, and the resource consumption of computers and mobile equipment (monitoring sensors of the equipment running data) is reduced.
In order to further apply the light weight of the equipment monitoring, the light weight process is packaged into the second target mapping relation, so that the light weight self-adaption can be realized, namely, the second equipment operation data is directly obtained through the first equipment operation data, and a complicated light weight analysis process is not needed, so that the light weight process is simpler and more convenient, and the method specifically comprises the following steps:
the obtaining of the second target mapping relation comprises the following steps:
taking the first equipment operation data as a second input item of the neural network, taking the second equipment operation data as a second output item of the neural network, and learning and training the neural network based on the second input item and the second output item to obtain a second target mapping relation;
the expression of the second target mapping relation is as follows:
S2＝CNN(S1)；
in the formula, S2 is second equipment operation data, S1 is first equipment operation data, and CNN is a neural network.
The first equipment operation data is at least one of image data, text data or audio data, and the text data comprises parameters of operation time, operation speed, load condition, temperature, pressure and the like of equipment recorded in a text, parameters of current, voltage, power, resistance and the like of the equipment, and parameters of vibration amplitude, frequency, phase and the like of the equipment. The audio data includes parameters of the device such as sound frequency, sound intensity, sound characteristics, etc. The image data includes photographs, videos, infrared images, etc. of the device.
As shown in fig. 2, the present application provides a lightweight monitoring system for an application of an apparatus operation lightweight monitoring method based on adaptive sensing, comprising:
the data acquisition module comprises at least one of an instrument, a remote terminal unit and a programmable logic control unit and is used for acquiring a plurality of working condition information of target equipment;
the hardware acceleration image preprocessing module is used for processing a plurality of working condition information of the target equipment into first equipment operation data;
the feature extraction module is used for learning a first target mapping relation by utilizing a neural network according to the first equipment operation data and the first equipment operation state;
the feature recognition result is used for carrying out noise addition and orthogonal initialization processing on the first equipment operation data, and carrying out light weight on the processed first equipment operation data through a first target mapping relation to obtain second equipment operation data;
the communication module is used for transmitting the first equipment operation data of the hardware acceleration image preprocessing module to the feature extraction module and the feature recognition result, and transmitting the second equipment operation data of the feature recognition result to the lower computer module;
the upper computer data display and storage module is used for storing a first target mapping relation and storing a second target mapping relation.
The first equipment operation state has a time sequence corresponding relation with first equipment operation data, the first equipment operation data comprises a plurality of first data items, the plurality of first data items correspond to a plurality of working condition information of target equipment, and the first equipment operation state corresponds to a working condition category of the target equipment;
the first target mapping relation is used for obtaining a first equipment operation state according to first equipment operation data;
the second equipment operation data is used for representing the first equipment operation data after light weight;
the second target mapping relationship is used for obtaining second equipment operation data according to the first equipment operation data.
According to the application, noise addition and orthogonal initialization processing are carried out on the first equipment operation data, the processed first equipment operation data is light-weighted through the first target mapping relation, so that quick online transmission is realized, and the resource consumption of a computer and mobile equipment is reduced; and learning a second target mapping relation by using the neural network, wherein the second target mapping relation is used for obtaining second equipment operation data according to the first equipment operation data, so that self-adaption light weight is realized.
The above embodiments are only exemplary embodiments of the present application and are not intended to limit the present application, the scope of which is defined by the claims. Various modifications and equivalent arrangements of this application will occur to those skilled in the art, and are intended to be within the spirit and scope of the application.
Claims (10)
1. An equipment operation lightweight monitoring method based on self-adaptive sensing is characterized in that: the method comprises the following steps:
acquiring first equipment operation data and a first equipment operation state, wherein the first equipment operation state and the first equipment operation data have a time sequence corresponding relation, the first equipment operation data comprises a plurality of first data items, the plurality of first data items correspond to a plurality of working condition information of target equipment, and the first equipment operation state corresponds to a working condition category of the target equipment;
according to the first equipment operation data and the first equipment operation state, a neural network is utilized to learn a first target mapping relation, and the first target mapping relation is used for obtaining the first equipment operation state according to the first equipment operation data;
noise adding and orthogonal initializing are carried out on the first equipment operation data, the processed first equipment operation data is lightened through a first target mapping relation, and second equipment operation data is obtained and used for representing the first equipment operation data after the weight reduction;
and according to the first equipment operation data and the second equipment operation data, a second target mapping relation is learned by using a neural network, and the second target mapping relation is used for obtaining the second equipment operation data according to the first equipment operation data.
2. The device operation lightweight monitoring method based on adaptive sensing according to claim 1, wherein: the obtaining of the first target mapping relation includes:
taking the first equipment operation data as a first input item of the neural network, taking the first equipment operation state as a first output item of the neural network, and learning and training the neural network based on the first input item and the first output item to obtain a first target mapping relation;
the expression of the first target mapping relation is as follows:
Label＝CNN(S1)；
in the formula, label is a first equipment operation state, S1 is first equipment operation data, and CNN is a neural network.
3. The device operation lightweight monitoring method based on adaptive sensing according to claim 2, wherein: the noise adding and orthogonal initializing process for the first device operation data includes:
carrying out data quantization on each first data item in the first equipment operation data by utilizing a spearman algorithm, summing the data quantization results of the association relation among each first data item to obtain the data symmetry of the first equipment operation data,
if the data symmetry of the first equipment operation data is larger than a first preset value, gaussian noise is added to each first data item in the first equipment operation data, and orthogonal initialization processing is carried out on each first data item added with the Gaussian noise;
and if the data symmetry of the first equipment operation data is smaller than or equal to a first preset value, carrying out orthogonal initialization processing on each first data item in the first equipment operation data.
4. A device operation lightweight monitoring method based on adaptive sensing according to claim 3, characterized in that: the adding gaussian noise to each first data item comprises:
and adding Gaussian noise to each first data item in sequence, wherein the adding expression of the Gaussian noise is as follows:
w l _add＝w l +N(0，βσ(r(w l )))；
wherein w is l Add is the first data item, w, after adding Gaussian noise l For the first data item of the first, N (0, βσ (r (w l ) And) is a gaussian distribution function, σ (r (w) l ) Is r (w) l ) Beta is the data symmetry of the first device operation data, r (w l ) Is the association relationship between the first data items, wherein r (w l )∈P K×K ，P K×K And K is a matrix specification of the incidence relation matrix among the first data items, and l is a counting variable.
5. The device operation lightweight monitoring method based on adaptive sensing according to claim 4, wherein: the orthogonal initialization process includes:
each first data item added with Gaussian noise is subjected to orthogonal initialization processing, and the method comprises the following steps:
performing QR decomposition on each first data item added with Gaussian noise to obtain a quadrature matrix Q and an upper triangular matrix R;
carrying out orthogonal initialization operation on each first data item added with Gaussian noise to obtain a new first data item, wherein the function expression of the new data item is as follows:
w l _add_new＝Q⊙sign(diag(R))；
wherein w is l The add_new is the orthogonal initialization result of the first data item after Gaussian noise is added, diag (R) is a diagonal matrix of R, sign (diag (R)) is a sign function of diag (R);
performing orthogonal initialization processing on each first data item in the first device operation data, including:
performing QR decomposition on each first data item of the first equipment operation data to obtain a quadrature matrix Q and an upper triangular matrix R;
carrying out orthogonal initialization operation on each first data item of the first equipment operation data to obtain a new first data item, wherein the function expression of the new data item is as follows:
w l _new＝Q⊙sign(diag(R))；
wherein w is l And_new is the orthogonal initialization result of the first data item in the first equipment operation data, diag (R) is a diagonal matrix of R, and sign (diag (R)) is a sign function of diag (R).
6. The device operation lightweight monitoring method based on adaptive sensing according to claim 5, wherein: the light weight of the processed first device operation data through the first target mapping relation includes:
respectively and sequentially replacing corresponding first data items in the first equipment operation data by each first data item after orthogonal initialization to obtain a group of third equipment operation data, wherein only one first data item in the third equipment operation data and the first equipment operation data is different;
obtaining the second equipment operation state through the first target mapping relation according to the third equipment operation data;
comparing the second device operating state with the first device operating state according to the corresponding relation between the third device operating data and the first device operating data, wherein,
if the first equipment operation state is the same as the second equipment operation state, marking the first data item corresponding to the second equipment operation state as an unnecessary data item;
if the first equipment operation state is different from the second equipment operation state, marking the first data item corresponding to the second equipment operation state as a necessary data item;
and eliminating unnecessary data items in the first equipment operation data to obtain second equipment operation data.
7. The device operation lightweight monitoring method based on adaptive sensing according to claim 6, wherein: the obtaining of the second target mapping relation includes:
taking the first equipment operation data as a second input item of the neural network, taking the second equipment operation data as a second output item of the neural network, and learning and training the neural network based on the second input item and the second output item to obtain a second target mapping relation;
the expression of the second target mapping relation is as follows:
S2＝CNN(S1)；
in the formula, S2 is second equipment operation data, S1 is first equipment operation data, and CNN is a neural network.
8. The device operation lightweight monitoring method based on adaptive sensing according to claim 1, wherein: the first device operational data is at least one of image data, text data, or audio data.
9. A lightweight monitoring system applying the adaptive perception-based device operation lightweight monitoring method of any of claims 1-8, comprising:
the data acquisition module comprises at least one of an instrument, a remote terminal unit and a programmable logic control unit and is used for acquiring a plurality of working condition information of target equipment;
the hardware acceleration image preprocessing module is used for processing a plurality of working condition information of the target equipment into first equipment operation data;
the feature extraction module is used for learning a first target mapping relation by utilizing a neural network according to the first equipment operation data and the first equipment operation state;
the feature recognition result is used for carrying out noise addition and orthogonal initialization processing on the first equipment operation data, and carrying out light weight on the processed first equipment operation data through a first target mapping relation to obtain second equipment operation data;
the communication module is used for transmitting the first equipment operation data of the hardware acceleration image preprocessing module to the feature extraction module and the feature recognition result, and transmitting the second equipment operation data of the feature recognition result to the lower computer module;
the upper computer data display and storage module is used for storing a first target mapping relation and storing a second target mapping relation.
10. A lightweight monitoring system as in claim 9, wherein: the first equipment operation state has a time sequence corresponding relation with first equipment operation data, the first equipment operation data comprises a plurality of first data items, the plurality of first data items correspond to a plurality of working condition information of target equipment, and the first equipment operation state corresponds to a working condition category of the target equipment;
the first target mapping relation is used for obtaining a first equipment operation state according to first equipment operation data;
the second equipment operation data are used for representing the first equipment operation data after light weight;
and the second target mapping relation is used for obtaining second equipment operation data according to the first equipment operation data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311248570.8A CN117194900A (en) | 2023-09-25 | 2023-09-25 | Equipment operation lightweight monitoring method and system based on self-adaptive sensing |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311248570.8A CN117194900A (en) | 2023-09-25 | 2023-09-25 | Equipment operation lightweight monitoring method and system based on self-adaptive sensing |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117194900A true CN117194900A (en) | 2023-12-08 |
Family
ID=88992341
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311248570.8A Pending CN117194900A (en) | 2023-09-25 | 2023-09-25 | Equipment operation lightweight monitoring method and system based on self-adaptive sensing |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117194900A (en) |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103124179A (en) * | 2013-01-14 | 2013-05-29 | 江苏大学 | Electric power system data reconfiguration decompressing method based on orthogonal matching pursuit |
| CN106022972A (en) * | 2016-06-30 | 2016-10-12 | 中国电力科学研究院 | State matrix symmetry-based power distribution network abnormal data identification method |
| CN108804715A (en) * | 2018-07-09 | 2018-11-13 | 北京邮电大学 | Merge multitask coordinated recognition methods and the system of audiovisual perception |
| CN112991223A (en) * | 2021-04-06 | 2021-06-18 | 深圳棱镜空间智能科技有限公司 | Image enhancement method, device, equipment and medium based on reversible neural network |
| CN113159072A (en) * | 2021-04-22 | 2021-07-23 | 中国人民解放军国防科技大学 | Online ultralimit learning machine target identification method and system based on uniform regularization |
| CN114238368A (en) * | 2021-12-07 | 2022-03-25 | 广东电网有限责任公司广州供电局 | Transformer fault diagnosis method and device, computer equipment and readable storage medium |
| US20220288783A1 (en) * | 2021-03-10 | 2022-09-15 | Nvidia Corporation | Machine learning of grasp poses in a cluttered environment |
| CN115293212A (en) * | 2022-08-15 | 2022-11-04 | 西安欧亚学院 | Equipment running state monitoring method based on audio perception and digital twins |
| CN116506043A (en) * | 2023-03-18 | 2023-07-28 | 西安电子科技大学 | Method, system, equipment and medium for recognizing disturbed state of PSK digital modulation communication signal |
-
2023
- 2023-09-25 CN CN202311248570.8A patent/CN117194900A/en active Pending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103124179A (en) * | 2013-01-14 | 2013-05-29 | 江苏大学 | Electric power system data reconfiguration decompressing method based on orthogonal matching pursuit |
| CN106022972A (en) * | 2016-06-30 | 2016-10-12 | 中国电力科学研究院 | State matrix symmetry-based power distribution network abnormal data identification method |
| CN108804715A (en) * | 2018-07-09 | 2018-11-13 | 北京邮电大学 | Merge multitask coordinated recognition methods and the system of audiovisual perception |
| US20220288783A1 (en) * | 2021-03-10 | 2022-09-15 | Nvidia Corporation | Machine learning of grasp poses in a cluttered environment |
| CN112991223A (en) * | 2021-04-06 | 2021-06-18 | 深圳棱镜空间智能科技有限公司 | Image enhancement method, device, equipment and medium based on reversible neural network |
| CN113159072A (en) * | 2021-04-22 | 2021-07-23 | 中国人民解放军国防科技大学 | Online ultralimit learning machine target identification method and system based on uniform regularization |
| CN114238368A (en) * | 2021-12-07 | 2022-03-25 | 广东电网有限责任公司广州供电局 | Transformer fault diagnosis method and device, computer equipment and readable storage medium |
| CN115293212A (en) * | 2022-08-15 | 2022-11-04 | 西安欧亚学院 | Equipment running state monitoring method based on audio perception and digital twins |
| CN116506043A (en) * | 2023-03-18 | 2023-07-28 | 西安电子科技大学 | Method, system, equipment and medium for recognizing disturbed state of PSK digital modulation communication signal |
Non-Patent Citations (7)
| Title |
|---|
| DAN LING ET AL.: "Control performance monitoring for EWMA-based run-to-run control in semiconductor manufacturing processes", 《PROCEEDINGS OF THE 33RD CHINESE CONTROL CONFERENCE》, 15 September 2014 (2014-09-15), pages 2990 - 2994 * |
| WEI HU ET AL.: "Provable Benefit of Orthogonal Initialization in Optimizing Deep Linear Networks", 《ARXIV》, 16 January 2020 (2020-01-16), pages 1 - 16 * |
| 张丕富: "神经网络在接触网2C系统异物入侵识别中的应用", 《电气化铁道》, vol. 32, no. 1, 30 December 2021 (2021-12-30), pages 159 - 161 * |
| 张琼等: "一种新的基于粗糙集的leader聚类算法", 《计算机科学》, no. 03, 25 March 2008 (2008-03-25), pages 177 - 179 * |
| 汪滢等: "移动信息节点变化下数据挖掘方法研究与仿真", 《计算机仿真》, vol. 33, no. 03, 15 March 2016 (2016-03-15), pages 278 - 281 * |
| 王邦祥等: "神经网络近似模型在液压支架顶梁轻量化设计中的应用", 《轻工学报》, vol. 33, no. 02, 12 April 2018 (2018-04-12), pages 87 - 94 * |
| 胡念苏等: "汽轮发电机组经济性监测诊断系统中失效参数的处理", 《武汉大学学报(工学版)》, no. 06, 30 December 2003 (2003-12-30), pages 91 - 95 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10699195B2 (en) | Training of artificial neural networks using safe mutations based on output gradients | |
| Lv et al. | A deep convolution generative adversarial networks based fuzzing framework for industry control protocols | |
| US11500370B2 (en) | System for predictive maintenance using generative adversarial networks for failure prediction | |
| CN109871002B (en) | Concurrent abnormal state identification and positioning system based on tensor label learning | |
| CN112508078A (en) | Image multitask multi-label identification method, system, equipment and medium | |
| CN114662006A (en) | End cloud collaborative recommendation system and method and electronic equipment | |
| CN114398466A (en) | Complaint analysis method and device based on semantic recognition, computer equipment and medium | |
| Kayode et al. | Lirul: A lightweight lstm based model for remaining useful life estimation at the edge | |
| CN117194900A (en) | Equipment operation lightweight monitoring method and system based on self-adaptive sensing | |
| CN117036732A (en) | Electromechanical equipment detection system, method and equipment based on fusion model | |
| CN116739154A (en) | Fault prediction method and related equipment thereof | |
| CN116843395A (en) | Alarm classification method, device, equipment and storage medium of service system | |
| CN116244640A (en) | Unbalanced bearing fault diagnosis method and system based on federal learning | |
| CN110751055A (en) | Intelligent manufacturing system | |
| Noje et al. | Automatic System Based on Riesz MV-algebras, for Predictive Maintenance of Bearings of Industrial Equipment Using Temperature Sensors | |
| CN113886140A (en) | Artificial intelligence model output data judgment system based on credibility verification | |
| CN114513355A (en) | Malicious domain name detection method, device, equipment and storage medium | |
| CN114461499A (en) | Abnormal information detection model construction method and gray scale environment abnormal detection method | |
| CN113822313A (en) | Method and device for detecting abnormity of graph nodes | |
| CN112597828A (en) | Training method and device of webpage recognition model and webpage recognition method | |
| CN113688924B (en) | Abnormal order detection method, device, equipment and medium | |
| CN114757541B (en) | Performance analysis method, device, equipment and medium based on training behavior data | |
| CN117057772A (en) | Real-time tracking display method and system for equipment fault detection and maintenance | |
| CN117453536A (en) | System abnormality analysis method, system abnormality analysis device, computer device and storage medium | |
| CN117851252A (en) | Interface exception handling method, device, equipment and storage medium thereof |
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 |