CN115184563A - Chemical workshop field data acquisition method based on digital twinning - Google Patents
Chemical workshop field data acquisition method based on digital twinning Download PDFInfo
- Publication number
- CN115184563A CN115184563A CN202211095433.0A CN202211095433A CN115184563A CN 115184563 A CN115184563 A CN 115184563A CN 202211095433 A CN202211095433 A CN 202211095433A CN 115184563 A CN115184563 A CN 115184563A
- Authority
- CN
- China
- Prior art keywords
- gas concentration
- risk
- hidden layer
- virtual space
- representing
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 239000000126 substance Substances 0.000 title claims abstract description 37
- 238000000034 method Methods 0.000 title claims abstract description 34
- 238000013507 mapping Methods 0.000 claims abstract description 31
- 230000006870 function Effects 0.000 claims abstract description 16
- 238000002372 labelling Methods 0.000 claims abstract description 4
- 239000013598 vector Substances 0.000 claims description 23
- 239000011159 matrix material Substances 0.000 claims description 22
- 238000003384 imaging method Methods 0.000 claims description 7
- 238000005070 sampling Methods 0.000 claims description 5
- 238000012886 linear function Methods 0.000 claims description 4
- 238000006386 neutralization reaction Methods 0.000 claims description 4
- 238000013528 artificial neural network Methods 0.000 abstract description 4
- 238000013473 artificial intelligence Methods 0.000 abstract description 2
- 238000012544 monitoring process Methods 0.000 description 8
- 230000000007 visual effect Effects 0.000 description 6
- 238000010586 diagram Methods 0.000 description 5
- 238000012824 chemical production Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 4
- 230000003287 optical effect Effects 0.000 description 4
- 238000004519 manufacturing process Methods 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 238000013480 data collection Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000003889 chemical engineering Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000009792 diffusion process Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000011958 production data acquisition Methods 0.000 description 1
- 238000003860 storage Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
- 238000012549 training Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0635—Risk analysis of enterprise or organisation activities
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N33/00—Investigating or analysing materials by specific methods not covered by groups G01N1/00 - G01N31/00
- G01N33/0004—Gaseous mixtures, e.g. polluted air
- G01N33/0009—General constructional details of gas analysers, e.g. portable test equipment
- G01N33/0062—General constructional details of gas analysers, e.g. portable test equipment concerning the measuring method or the display, e.g. intermittent measurement or digital display
- G01N33/0067—General constructional details of gas analysers, e.g. portable test equipment concerning the measuring method or the display, e.g. intermittent measurement or digital display by measuring the rate of variation of the concentration
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T15/00—3D [Three Dimensional] image rendering
- G06T15/04—Texture mapping
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/70—Determining position or orientation of objects or cameras
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/80—Analysis of captured images to determine intrinsic or extrinsic camera parameters, i.e. camera calibration
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Strategic Management (AREA)
- Software Systems (AREA)
- Entrepreneurship & Innovation (AREA)
- Economics (AREA)
- Computer Graphics (AREA)
- General Health & Medical Sciences (AREA)
- Game Theory and Decision Science (AREA)
- Quality & Reliability (AREA)
- Immunology (AREA)
- Geometry (AREA)
- Biochemistry (AREA)
- Development Economics (AREA)
- Analytical Chemistry (AREA)
- Educational Administration (AREA)
- Medicinal Chemistry (AREA)
- Food Science & Technology (AREA)
- Combustion & Propulsion (AREA)
- Marketing (AREA)
- Operations Research (AREA)
- Pathology (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
Abstract
The invention discloses a digital twin-based chemical workshop field data acquisition method, which relates to the technical field of combination of artificial intelligence, virtual reality and the Internet of things, and comprises the following steps: establishing a mapping relation between a real space and a virtual space; arranging a gas concentration sensor in a chemical workshop, and collecting a time sequence numerical value of gas concentration; dividing a virtual space into a plurality of subspaces; establishing a risk model of gas concentration; obtaining a cost function through learning; iteratively calculating a risk model to obtain a learned risk model; calculating the probability of risk on the subspace; and when the probability of the risk existing on a certain subspace is greater than the probability of the risk existing on the rest subspaces, taking the subspace as a risk source, and labeling the subspace. The method comprises the steps of establishing a virtual space according to field space data and gas concentration data, establishing a corresponding risk model, analyzing the probability of risk sources at different positions on the field by adopting a neural network method, and marking and warning high-probability risk points in the virtual space.
Description
Technical Field
The invention relates to the technical field of combination of artificial intelligence, virtual reality and the Internet of things, in particular to a chemical workshop field data acquisition method based on digital twins.
Background
The digital twin is a new technology for building a model in a virtual domain to correspond to an entity in a physical domain, thereby simulating and describing the state of the entity in the physical domain, and is being widely used for intelligent manufacturing in various industries. The digital twin is based on data acquisition, information related to production in a real environment is converted into electric signals through acquisition equipment such as a sensor, and the electric signals are further sampled into data for analysis and storage, so that the processes of virtualization and digitization of the objective physical world are realized. The data acquisition is a digital twin keystone, and the integrity, accuracy and real-time performance of the acquisition determine the application effect of the digital twin.
Production data acquisition based on digital twins has gained wide attention in recent years and has been researched and applied in key industries such as chemical industry, machinery, electric power, traffic and the like. The key to digital twin-based production data collection is collection, processing, and virtualization. In the aspect of acquisition, on-site data are acquired based on a large-scale sensor, and efficient acquisition of a data source is realized; in the aspect of processing, a data processing model is established, large-scale data is modeled, and an abstract entity with practical significance is generated; in the aspect of virtualization, abstract entities are displayed by adopting related technologies, so that a user can intuitively feel the real situation of a physical domain to make a decision.
The monitoring of safety is a basic requirement of chemical production. The safety monitoring of the chemical production is carried out through the digital twin, abstract monitoring contents and monitoring objects can be specified, monitoring personnel can know the conditions of the chemical production site more visually, and the efficiency and the effect of the safety monitoring are improved. The sensors for collecting gas concentration and the like are distributed and installed on the site of a chemical workshop, and the data collected by each sensor reflects the state of the workshop site in a small range. However, the data generated by the sensors are isolated from each other, and data from different sensors must be integrated in a certain way to comprehensively reflect the field state. Still other acquisition methods using a camera, for example, patent CN110334701A, provide a data acquisition method based on deep learning and multi-vision in a digital twin environment, which can capture a wider range of field states, but cannot capture such invisible critical data as gas concentration.
Disclosure of Invention
The invention provides a chemical workshop field data acquisition method based on digital twinning, which is used for overcoming at least one technical problem in the prior art.
The embodiment of the invention provides a chemical workshop field data acquisition method based on digital twinning, which comprises the following steps:
establishing a mapping relation between a real space and a virtual space of a chemical workshop;
arranging a plurality of gas concentration sensors at different positions of a chemical workshop site;
collecting time series values of gas concentration at a predetermined sampling rate using the gas concentration sensor
；
Dividing the virtual space into a plurality of subspaces to obtain linear correlation between the real coordinate of the gas concentration sensor and the subspaces; each of the subspaces has a unique identifier
，
Representing the serial numbers of different subspaces on three coordinate axes;
defining an input using the real space coordinates of the gas concentration sensor and the time series values of the gas concentration as inputsGoes out to
Establishing a risk model of gas concentration; wherein,
representing a node; the risk model of the gas concentration comprises a first hidden layer, a second hidden layer, a third hidden layer, a fourth hidden layer and a fifth hidden layer, wherein the first hidden layer is
，
The nodes that represent the hidden layer(s),
representing a shared linear weight connection,
for the linear bias parameter, s represents the sensor number,
the dimensions of the vector are represented as,
represents a non-linear function, defined as
，
As the parameter(s) is (are),
representing a real number domain; the second hidden layer is
，
Indicating neutralization of sensors in a first hidden layer
The node of the corresponding node is a node of the corresponding node,
for the corresponding linear weight connection, t represents time,
in order to be a linear bias parameter,
representing a vector dimension; the third hidden layer is
，
Indicating sensor
Is a node in the input layer,
、
、
in order to be connected correspondingly,
is a linear bias parameter; the fourth hidden layer is
，
、
Is that
And node
、
The connection of (a) to (b),
in order to be a linear bias parameter,
representing a vector dimension; the fifth hidden layer is
，
Is a node
And hidden layer node
The connection of (a) to (b),
is a linear bias parameter;
obtaining a cost function of the risk model by learning the risk model
；
Iteratively calculating the risk model by using the cost function to obtain a learned risk model;
the gas concentration sensor acquires a time sequence value of gas concentration, the time sequence value and the real space coordinate of the gas concentration sensor are input into the learned risk model, and the learned risk model sequentially calculates the probability of risks existing in the subspace corresponding to the gas concentration sensor;
and when the probability of the risk existing on one subspace is greater than the probability of the risk existing on the rest subspace, taking the subspace as a risk source, and labeling the subspace in the virtual space.
Optionally, the establishing of the mapping relationship between the real space and the virtual space of the chemical plant specifically includes:
arranging a camera in a real space of the chemical workshop, acquiring an image by using the camera, and acquiring coordinates of a target point in the real space according to the image;
establishing a virtual space through a computer, and setting mapping points corresponding to the target points in the virtual space;
and obtaining the mapping relation between the virtual space and the real space according to the coordinates of the mapping point in the virtual space and the coordinates of the target point in the real space.
Optionally, obtaining, according to the image, coordinates of a target point in the real space, specifically:
calibrating the internal parameter matrix and the external parameter matrix of the camera according to the linear model of the camera imaging
Obtaining the coordinates of the target point in the real space; wherein,
which represents the homogeneous coordinates of the image,
representing world homogeneous coordinates of the target point,
a matrix of intrinsic parameters representing the camera is shown,
a matrix of extrinsic parameters representing the camera is shown,
is a linear calibration parameter.
The innovation points of the embodiment of the invention comprise:
1. in the embodiment, an optical sensor carried by a camera is used for collecting a field scene, a three-dimensional reconstruction technology is used for establishing mapping of the field scene in a virtual space, and key information such as the position and texture of a field target is recovered and used as a basis for constructing a virtual space of a chemical workshop, and the method is one of innovation points of the embodiment of the invention.
2. In the embodiment, data are acquired by a plurality of gas concentration sensors arranged at different positions on the site, an independent sensor for sensing the specific gas concentration is arranged at each position, the position coordinates of the sensor in the real space are recorded for implementing positioning in the real space, and the risk point location is displayed in the virtual space by establishing a large-scale sensor set, so that the method is one of the innovation points of the embodiment of the invention.
3. In the embodiment, a virtual space is established according to field space data and gas concentration data, a corresponding risk model is established, the probability of risk sources existing at different positions on the field is analyzed by adopting a neural network method, and high-probability risk points are marked and warned in the virtual space, which is one of innovation points of the embodiment of the invention.
4. In this embodiment, defining an output layer and each hidden layer, and establishing a structure of an obtained risk model through a neural network is one of innovative points of the embodiment of the present invention.
Drawings
In order to more clearly illustrate the embodiments of the present invention 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 is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
FIG. 1 is a flow chart of a data collection method according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of a risk model provided by an embodiment of the present invention;
FIG. 3 is a schematic diagram of a virtual space and a subspace provided in accordance with an embodiment of the present invention;
fig. 4 is a flowchart of a mapping relationship between a real space and a virtual space according to an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be obtained by a person skilled in the art without inventive effort based on the embodiments of the present invention, are within the scope of the present invention.
It is to be noted that the terms "comprises" and "comprising" and any variations thereof in the embodiments and drawings of the present invention are intended to cover non-exclusive inclusions. For example, a process, method, system, article, or apparatus that comprises a list of steps or elements is not limited to only those steps or elements listed, but may alternatively include other steps or elements not listed, or inherent to such process, method, article, or apparatus.
The embodiment of the invention discloses a chemical workshop field data acquisition method based on digital twinning. The following are detailed below.
Fig. 1 is a flowchart of a data acquisition method according to an embodiment of the present invention, fig. 2 is a schematic view of a risk model according to an embodiment of the present invention, fig. 3 is a schematic view of a virtual space and a subspace according to an embodiment of the present invention, and referring to fig. 1 to fig. 3, the method for acquiring data of a chemical workshop site based on a digital twin according to the embodiment includes:
step 1: establishing a mapping relation between a real space and a virtual space of a chemical workshop;
step 2: arranging a plurality of gas concentration sensors at different positions of a chemical workshop site;
and 3, step 3: collecting time series values of gas concentration using a gas concentration sensor at a predetermined sampling rate
；
And 4, step 4: dividing the virtual space into a plurality of subspaces to obtain the linear correlation between the real coordinate of the gas concentration sensor and the subspaces; each subspace has a unique identifier
，
Representing the serial numbers of different subspaces on three coordinate axes;
and 5: the real space coordinates of the gas concentration sensor and the time sequence value of the gas concentration are used as input, and the defined output is
Establishing a risk model of gas concentration; wherein,
representing a node; the risk model of the gas concentration comprises a first hidden layer, a second hidden layer, a third hidden layer, a fourth hidden layer and a fifth hidden layer, wherein the first hidden layer is
，
The nodes that represent the hidden layer(s),
representing a shared linear weight connection,
for the linear bias parameter, s represents the sensor number,
the dimensions of the vector are represented in the representation,
represents a non-linear function, defined as
，
As a function of the parameters of the system,
representing a real number domain; the second hidden layer is
，
Indicating neutralization of sensors in a first hidden layer
The corresponding node is connected with the corresponding node,
for the corresponding linear weight connection, t represents time,
in order to be a linear bias parameter,
representing a vector dimension; the third hidden layer is
，
Indicating sensor
Is a node in the input layer,
、
、
in order to be connected correspondingly,
is a linear bias parameter; the fourth hidden layer is
，
、
Is that
And node
、
The connection of (a) to (b),
for linearly biasing a parameterThe number of the first and second groups is counted,
representing a vector dimension; the fifth hidden layer is
，
Is a node
And hidden layer node
The connection of (a) to (b),
is a linear bias parameter;
and 6: obtaining a cost function of the risk model by learning the risk model
；
And 7: iteratively calculating a risk model by using a cost function to obtain a learned risk model;
and 8: the method comprises the steps that a gas concentration sensor collects a time sequence value of gas concentration, the time sequence value and real space coordinates of the gas concentration sensor are input into a learned risk model, and the learned risk model sequentially calculates the probability of risks existing in subspaces corresponding to the gas concentration sensor;
and step 9: and when the probability of the risk existing on a certain subspace is greater than the probability of the risk existing on the rest subspaces, taking the subspace as a risk source, and labeling the subspace in the virtual space.
Specifically, referring to fig. 1, in the method for acquiring field data of a chemical workshop based on digital twins according to this embodiment, a mapping relationship between a real space and a virtual space of the chemical workshop is first established in step 1, and when the mapping relationship is established, a camera needs to be used for capturing imagesThe on-board optical sensor collects the field scene of the chemical workshop, then the mapping of the field scene in the virtual space is established by utilizing the three-dimensional reconstruction technology, and the position, texture and other related information of the field target are recovered. The virtual space is a three-dimensional space, the value of each coordinate in the virtual space is a vector, and the coordinate is the position coordinate in the real space
Linear mapping of (2)
The vector represents texture information of the visual object at a corresponding position in the real space. The virtual space is used as the linear scale reconstruction of the real space, the texture of the visual target in the real space reappears in the virtual space according to the consistent proportion, the visual reconstruction of the field space is realized, and a user can intuitively feel the field state by observing the virtual space on a computer.
In order to collect the gas concentration, a plurality of gas concentration sensors are arranged on the site of the chemical engineering workshop through step 2, the gas concentration sensors are located at different positions, independent gas concentration sensors for sensing specific gas concentration are arranged at different positions, and the position coordinates of the gas concentration sensors in real space are recorded
And the positioning of the gas concentration sensor in the real space is realized. Then in step 3, the gas concentration sensor is enabled to collect the gas concentration at a preset sampling rate, and a group of time sequence numerical values of the gas concentration changing along with the time are obtained
Wherein s represents a sensor number, t is a time stamp, and the acquisition time stamps of all gas concentration sensors are controlled by a unique central control program on the same site to realize synchronous sampling.
Obtaining a time series value of the concentration of the collected gas
And then, establishing a risk model according to the gas concentration data, and analyzing the probability of risk sources existing in different positions on the site. When the risk model is established, in order to present the risk points more intuitively in the virtual space, the virtual space is divided into a plurality of subspaces through step 4, please refer to fig. 3, in order to make each subspace have a unique identifier, the subspace is divided according to three orthogonal coordinate axes of the three-dimensional virtual space
、
、
The virtual space is divided into several subspaces. Each obtained subspace is a three-dimensional cube, namely the boundary of the subspace is parallel to the coordinate axis; no intersection exists between any two subspaces, and the set of all the subspaces is equal to the original space; any face of a subspace cube only intersects with a unique face of another subspace cube; the position coordinates of any gas concentration sensor necessarily belong to a certain subspace, and no more than one gas concentration sensor is included in each subspace. When the above conditions are satisfied, the obtained subspace has a unique identifier
，
Is a three-dimensional matrix of which the matrix is,
is an element of the matrix that is,
indicating a matrix element index, i.e. subspace edge
Relative order of the coordinate axes. By the subspace identification method, linear correlation between the subspace and the spatial position of the sensor can be established.
And 5, taking the real space coordinate of the gas concentration sensor and the time sequence numerical value of the gas concentration as input, taking the mark of the virtual space subspace as output, and establishing a risk model of the gas concentration through a neural network. In this embodiment, the output is defined as
Wherein
representing a node.
Referring to FIG. 2, the risk model includes a first hidden layer defined as
，
The nodes in the hidden layer are represented as,
representing a shared linear weight connection,
1 node representing a hidden layer
And
the 5 nodes shown are connected together and,
is the time series value of the gas concentration collected in step 3.
For the linear bias parameter, s represents the sensor number, the set of nodes represented by the first hidden layer
Contains S independent vectors, S is the total number of sensors.
And representing vector dimensions, wherein the dimension of each vector is T, and T is the total length of acquisition time. By acquiring the characteristic of the gas concentration on the independent sensor changing along with the time, the model can identify the diffusion stage of the gas concentration, and the identification precision of the gas concentration distribution is optimized.
Referring to FIG. 2, the risk model further includes a second hidden layer, which is defined as
，
Representing a first secretIn-reservoir neutralization sensor
The corresponding node is connected with the corresponding node,
for the corresponding linear weight connection, t represents a timestamp,
in order to be a linear bias parameter,
representing vector dimensions, i.e.
In the form of a vector, the vector,
refers to the second dimension of the vector. Set of nodes of a second hidden layer
The correlation between the collected values of different sensor gas concentration changes is described.
Referring to FIG. 2, the risk model further includes a third hidden layer, which is defined as
，
、
、
Indicating sensor
Real space coordinates of (are) inputThe nodes in the entry layer are connected to each other,
、
、
in order to be connected correspondingly,
is a linear bias parameter. Set of nodes of third hidden layer
The correlation between the spatial coordinates of different sensors is described. It should be noted that a and B in the second hidden layer and the third hidden layer are merely identifiers for distinguishing different outputs, and have no practical meaning.
Referring to FIG. 2, the risk model further includes a fourth hidden layer, which is defined as
，
、
Is that
And node
、
The connection of (a) to (b),
for linearly biasing a parameterThe number of the first and second groups is counted,
the dimension of the vector is represented as,
is a 1024-dimensional vector, in which the parameter k represents the dimension of the vector. Set of nodes of fourth hidden layer
A correlation between sensor spatial coordinates and gas concentration acquisition values is described.
Referring to FIG. 2, the risk model further includes a fifth hidden layer, which is
， 
Is a node
And hidden layer node
The connection of (a) to (b),
is a linear bias parameter.
In the risk model diagram shown in fig. 2, the respective parameters are not represented by the superscripts and subscripts, but the meanings thereof are the same as those of the parameters represented by the superscripts and subscripts, and gs (t) in fig. 2 corresponds to, for example, the above-mentioned gs (t)
H3 (k) corresponds to the above
And other parameters are not listed here.
In the embodiment, a risk model among sensor collected data, a real space and a virtual space is established through a defined output layer and a defined hidden layer
。
Each element value of (1) represents the probability of the risk source of the gas leakage in the virtual subspace, and the value range is 0-1; when the value is 0, the risk does not exist, and when the value is 1, the risk is determined to exist. Using risk models
Before monitoring, a training sample set is prepared, a probability matrix output by a model is artificially marked, the output value of a virtual subspace where a gas leakage source is located is marked as 1, and the rest are marked as 0. Marking of the sample
Obtaining model output from input calculation of samples
。
Then, through step 6, learning the risk model is performed, and a cost function of the risk model can be obtained
，
The function of a natural index is represented,
and the natural logarithmic function is expressed, and the identification performance under small-scale sample data can be improved by adopting an exponential-logarithmic function combination. In step 7, a risk model is iteratively calculated by using a cost function to obtain parameters of each hidden layer, so that the parameters of each hidden layer can be obtainedTo the learned risk model. During iterative computation, an extreme value of the cost function can be solved by adopting a backward propagation method, so that the learning of each parameter in the risk model is completed.
In step 8, the gas concentration sensor collects a time series value of the gas concentration, it should be noted that the time series value of the gas concentration collected in step 8 may refer to step 3, which is not described herein again. And then inputting the time sequence value and the real space coordinate of the gas concentration sensor into the learned risk model, wherein the learned risk model can calculate the probability of risks in each subspace. Then, step 9, the subspace corresponding to the maximum risk probability is taken as the risk source, and the subspace is labeled in the virtual space by adopting a special display mode.
The digital twin-based chemical workshop field data acquisition method provided by the invention combines the advantages of the gas concentration sensor and the camera optical sensor, efficiently and quickly acquires the data of the chemical production field, monitors the gas leakage risk, and carries out modeling and virtual display on the potential risk points in the form of digital twin, so that the potential risk positions can be quickly positioned in the chemical workshop in a large range, the risk points can be visually displayed in the virtual space, and monitoring personnel can be helped to quickly find the risk.
Optionally, fig. 4 is a flowchart of a mapping relationship between a real space and a virtual space provided in an embodiment of the present invention, please refer to fig. 4, and in step 1, the mapping relationship between the real space and the virtual space of the chemical plant is established, specifically: step 11: arranging a camera in a real space of a chemical workshop, acquiring an image by using the camera, and acquiring the coordinates of a target point in the real space according to the image; step 12: establishing a virtual space through a computer, and setting a mapping point corresponding to the target point in the virtual space; step 13: and obtaining the mapping relation between the virtual space and the real space according to the coordinates of the mapping point in the virtual space and the coordinates of the target point in the real space.
Specifically, referring to fig. 4, when a mapping relationship between a real space and a virtual space of a chemical plant is established, firstly, a camera is set in the real space of the chemical plant in step 11The camera is used for acquiring images, and then a field scene is acquired from a plurality of images acquired from the same scene, so that coordinates of a target point in a real space, such as information of spatial positions, textures and the like of various targets in the field scene, are acquired. In step 12, a three-dimensional virtual space is established by the computer, and a mapping point corresponding to the target point is set in the virtual space, that is, a mapping point of the position and texture information of the on-site target is set in the virtual space, the value of each coordinate in the virtual space is a vector, and the coordinate is a position coordinate in the real space
Linear mapping of
The vector represents texture information of the visual object at the corresponding position in the real space.
After the mapping point is obtained, in step 13, the mapping relationship between the virtual space and the real space can be obtained according to the coordinate of the mapping point in the virtual space and the coordinate of the target point in the real space. The virtual space is used as the linear scale reconstruction of the real space by establishing the mapping of the real space and the virtual space, the texture of the visual target in the real space reappears in the virtual space according to a consistent proportion, the visual reconstruction of the field space is realized, and a user can intuitively feel the field state by observing the virtual space on a computer.
Optionally, referring to fig. 4, in step 11, obtaining coordinates of the target point in the real space according to the image, specifically: calibrating the internal parameter matrix and the external parameter matrix of the camera according to the linear model of the camera imaging
Obtaining the coordinates of the target point in the real space; wherein,
which represents the homogeneous coordinates of the image,
representing world homogeneous coordinates of the target point,
a matrix of intrinsic parameters representing the camera is shown,
a matrix of extrinsic parameters representing the camera is shown,
are linear calibration parameters.
Specifically, referring to fig. 4, image pixels are projections of objects in a scene on an imaging plane of a camera, each image pixel corresponds to an object in the scene, coordinates of the image pixels are related to real world coordinates of the object, and values of the image pixels are related to real texture of the object. The intrinsic parameter matrix of the camera represents the basic parameters of the camera imaging, and is related to the properties of the imaging element of the camera, such as pixel size, focal length and the like, and is unrelated to the position of the camera in the field space; the external parameter matrix of the camera represents the relative position of the camera in the field space, including the optical axis direction, three-dimensional space offset and the like. Therefore, when the coordinates of the target point in the real space are acquired, the internal and external parameter matrixes of the camera are calibrated, and then the linear model imaged by the camera is used
The coordinates of the target point in the real space can be obtained, and therefore the position and texture information of the field target can be recovered from the image pixels. Linear model of camera imaging
In (1),
which represents the homogeneous coordinates of the image,
representing world homogeneous coordinates of the target point,
a matrix of intrinsic parameters representing the camera is shown,
a matrix of extrinsic parameters representing the camera is shown,
are linear calibration parameters.
Besides the above mode, the internal and external parameter matrixes of the camera can be calibrated by directly utilizing the position and texture information of the field target. Further, combining the internal and external parameter matrices can obtain:
wherein, in the process,
。
being a 3-by-4 matrix comprising 12 free variables, theoretically measuring at least 6 on-site targets can solve for since each image pixel homogeneous coordinate comprises 2 free variables
Is determined. Measuring more field targets and solving by adopting linear optimization method
The solution accuracy can be improved.
Based on the method provided by the invention, the inventor conducts experimental verification aiming at chemical workshops of different scales to obtain experimental results shown in table 1, and the table 1 is an experimental result schematic diagram.
TABLE 1
Referring to the table 1, experimental results show that the chemical workshop field data acquisition method based on the digital twinning, provided by the invention, is suitable for chemical workshops of different scales, has high identification and positioning accuracy and quick positioning, can effectively position the point position of the gas leakage risk, and improves the risk discovery efficiency.
Those of ordinary skill in the art will understand that: the figures are merely schematic representations of one embodiment, and the blocks or flow diagrams in the figures are not necessarily required to practice the present invention.
Those of ordinary skill in the art will understand that: modules in the devices in the embodiments may be distributed in the devices in the embodiments according to the description of the embodiments, or may be located in one or more devices different from the embodiments with corresponding changes. The modules of the above embodiments may be combined into one module, or further split into multiple sub-modules.
Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (3)
1. A chemical workshop field data acquisition method based on digital twinning is characterized by comprising the following steps:
establishing a mapping relation between a real space and a virtual space of a chemical workshop;
arranging a plurality of gas concentration sensors at different positions of a chemical workshop site;
collecting time series values of gas concentration at a predetermined sampling rate using the gas concentration sensor
；
Dividing the virtual space into a plurality of subspaces to obtain linear correlation between the real coordinate of the gas concentration sensor and the subspaces; each of the subspaces having a unique identifier
，
Representing the serial numbers of different subspaces on three coordinate axes;
the real space coordinates of the gas concentration sensor and the time sequence numerical value of the gas concentration are used as input, and the defined output is
Establishing a risk model of gas concentration; wherein,
representing a node; the risk model of the gas concentration comprises a first hidden layer, a second hidden layer, a third hidden layer, a fourth hidden layer and a fifth hidden layer, wherein the first hidden layer is
，
A node representing a hidden layer of the image,
representing a shared linear weight connection,
for the linear bias parameter, s represents the sensor number,
the dimensions of the vector are represented in the representation,
represents a non-linear function, defined as
，
As the parameter(s) is (are),
representing a real number domain; the second hidden layer is
，
Indicating neutralization of a first hidden layer
The node of the corresponding node is a node of the corresponding node,
for the corresponding linear weight connection, t represents time,
in order to be a linear bias parameter,
representing a vector dimension; the third hidden layer is
，
Indicating sensor
Is a node in the input layer,
、
、
in order to be connected correspondingly,
is a linear bias parameter; the fourth hidden layer is
，
、
Is that
And node
、
The connection of (a) to (b),
in order to be a linear bias parameter,
representing a vector dimension; the fifth hidden layer is
，
Is a node
And hidden layer node
The connection of (a) to (b),
is a linear bias parameter;
obtaining a cost function of the risk model by learning the risk model
；
Iteratively calculating the risk model by using the cost function to obtain a learned risk model;
the gas concentration sensor acquires a time sequence value of gas concentration, the time sequence value and the real space coordinate of the gas concentration sensor are input into the learned risk model, and the learned risk model sequentially calculates the probability of risks existing in the subspace corresponding to the gas concentration sensor;
and when the probability of the risk existing on one subspace is greater than the probability of the risk existing on the rest subspaces, taking the subspace as a risk source, and labeling the subspace in the virtual space.
2. The digital twin-based chemical plant site data acquisition method according to claim 1, wherein the establishing of the mapping relationship between the real space and the virtual space of the chemical plant specifically comprises:
arranging a camera in a real space of the chemical workshop, acquiring an image by using the camera, and acquiring coordinates of a target point in the real space according to the image;
establishing a virtual space through a computer, and setting mapping points corresponding to the target points in the virtual space;
and obtaining the mapping relation between the virtual space and the real space according to the coordinates of the mapping point in the virtual space and the coordinates of the target point in the real space.
3. The digital twin-based chemical plant field data acquisition method according to claim 2, wherein coordinates of a target point in the real space are acquired according to the image, specifically:
calibrating the internal parameter matrix and the external parameter matrix of the camera according to the linear model of the camera imaging
Obtaining the coordinates of the target point in the real space; wherein,
which represents the homogeneous coordinates of the image,
representing the world homogeneous coordinates of the target point,
a matrix of intrinsic parameters representing the camera is shown,
a matrix of extrinsic parameters representing the camera is shown,
is a linear calibration parameter.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211095433.0A CN115184563B (en) | 2022-09-08 | 2022-09-08 | Chemical workshop field data acquisition method based on digital twinning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211095433.0A CN115184563B (en) | 2022-09-08 | 2022-09-08 | Chemical workshop field data acquisition method based on digital twinning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115184563A true CN115184563A (en) | 2022-10-14 |
| CN115184563B CN115184563B (en) | 2022-12-02 |
Family
ID=83522337
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211095433.0A Active CN115184563B (en) | 2022-09-08 | 2022-09-08 | Chemical workshop field data acquisition method based on digital twinning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115184563B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116188714A (en) * | 2023-04-27 | 2023-05-30 | 山东翰林科技有限公司 | Geospatial coordinate processing method based on illusion engine |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110334701A (en) * | 2019-07-11 | 2019-10-15 | 郑州轻工业学院 | Collecting method based on deep learning and multi-vision visual under the twin environment of number |
| CN111967202A (en) * | 2020-08-08 | 2020-11-20 | 西北工业大学 | Artificial intelligence-based aircraft engine extreme speed performance digital twinning method |
| CN112070279A (en) * | 2020-08-19 | 2020-12-11 | 浙江工业大学 | Product processing control method based on digital twinning technology |
| CN112947368A (en) * | 2021-02-02 | 2021-06-11 | 安徽理工大学 | Fault diagnosis device and method for three-phase asynchronous motor |
| CN114000907A (en) * | 2021-12-10 | 2022-02-01 | 重庆邮电大学 | Mine ventilation equipment intelligent regulation and control system based on digital twin technology |
| WO2022121923A1 (en) * | 2020-12-10 | 2022-06-16 | 东北大学 | Smart modelling method and apparatus of complex industrial process digital twin system, device, and storage medium |
| EP4040343A1 (en) * | 2021-02-05 | 2022-08-10 | Tata Consultancy Services Limited | Method and system for building digital twin by leveraging prior knowledge |
-
2022
- 2022-09-08 CN CN202211095433.0A patent/CN115184563B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110334701A (en) * | 2019-07-11 | 2019-10-15 | 郑州轻工业学院 | Collecting method based on deep learning and multi-vision visual under the twin environment of number |
| CN111967202A (en) * | 2020-08-08 | 2020-11-20 | 西北工业大学 | Artificial intelligence-based aircraft engine extreme speed performance digital twinning method |
| CN112070279A (en) * | 2020-08-19 | 2020-12-11 | 浙江工业大学 | Product processing control method based on digital twinning technology |
| WO2022121923A1 (en) * | 2020-12-10 | 2022-06-16 | 东北大学 | Smart modelling method and apparatus of complex industrial process digital twin system, device, and storage medium |
| CN112947368A (en) * | 2021-02-02 | 2021-06-11 | 安徽理工大学 | Fault diagnosis device and method for three-phase asynchronous motor |
| EP4040343A1 (en) * | 2021-02-05 | 2022-08-10 | Tata Consultancy Services Limited | Method and system for building digital twin by leveraging prior knowledge |
| CN114000907A (en) * | 2021-12-10 | 2022-02-01 | 重庆邮电大学 | Mine ventilation equipment intelligent regulation and control system based on digital twin technology |
Non-Patent Citations (1)
| Title |
|---|
| 王忠强 等: "基于数字孪生技术的选煤厂智能管控系统", 《智能矿山》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116188714A (en) * | 2023-04-27 | 2023-05-30 | 山东翰林科技有限公司 | Geospatial coordinate processing method based on illusion engine |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115184563B (en) | 2022-12-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN117575550B (en) | BIM technology-based three-dimensional visual management system for wind farm data | |
| CN116229019A (en) | Digital twinning-oriented large-scene fusion three-dimensional reconstruction method and system | |
| CN113408034B (en) | Engineering construction intelligent cloud management platform based on BIM building three-dimensional model technology | |
| CN111912937B (en) | Intelligent real-time tracing method and system for indoor sudden pollutants | |
| EP4336385A1 (en) | Method and apparatus for updating target detection model | |
| CN112146761A (en) | Human body temperature measurement compensation method based on machine learning | |
| CN116993681A (en) | Substation inspection defect detection method and system | |
| CN108959713A (en) | Target range and face positional shift measurement method based on convolutional neural networks | |
| CN115184563B (en) | Chemical workshop field data acquisition method based on digital twinning | |
| CN116933626A (en) | Data monitoring method and device based on digital twinning | |
| CN115858940A (en) | Steel structure welding process quality management recommendation method based on big data processing | |
| Zdziebko et al. | Synthetic image generation using the finite element method and blender graphics program for modeling of vision-based measurement systems | |
| Mattioli et al. | 2D–3D spatial registration for remote inspection of power substations | |
| CN116308185A (en) | Visual progress input and early warning method for constructional engineering | |
| Patel et al. | Deep Learning-Based Plant Organ Segmentation and Phenotyping of Sorghum Plants Using LiDAR Point Cloud | |
| CN115585731A (en) | Space-air-ground integrated hydropower station space state intelligent monitoring management system and method thereof | |
| CN115130369A (en) | Transformer substation equipment comprehensive state evaluation and active early warning method based on digital twinning | |
| CN116821409B (en) | Production safety management method and system based on image data processing | |
| Yoon et al. | UAV-based automated 3D modeling framework using deep learning for building energy modeling | |
| CN117291042A (en) | Underwater three-dimensional temperature field reconstruction method and device based on peak surface feature consideration | |
| CN117688313A (en) | Method for cooperatively processing mapping data by network | |
| CN115661446A (en) | Pointer instrument indication automatic reading system and method based on deep learning | |
| CN115761265A (en) | Method and device for extracting substation equipment in laser radar point cloud | |
| CN116297620B (en) | Magnetic variable measurement method and system for nuclear magnetic resonance apparatus | |
| CN109934333A (en) | Object localization method based on full Connection Neural Network |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |