CN116909165A - Real-time simulation method for physical field of large power generation equipment component - Google Patents
Real-time simulation method for physical field of large power generation equipment component Download PDFInfo
- Publication number
- CN116909165A CN116909165A CN202310337264.5A CN202310337264A CN116909165A CN 116909165 A CN116909165 A CN 116909165A CN 202310337264 A CN202310337264 A CN 202310337264A CN 116909165 A CN116909165 A CN 116909165A
- Authority
- CN
- China
- Prior art keywords
- physical field
- power generation
- generation equipment
- simulation
- real
- 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
- 238000004088 simulation Methods 0.000 title claims abstract description 94
- 238000010248 power generation Methods 0.000 title claims abstract description 73
- 238000000034 method Methods 0.000 title claims abstract description 31
- 238000012549 training Methods 0.000 claims abstract description 35
- 238000012360 testing method Methods 0.000 claims abstract description 31
- 238000005094 computer simulation Methods 0.000 claims abstract description 6
- 238000013461 design Methods 0.000 claims abstract description 6
- 238000009826 distribution Methods 0.000 claims abstract description 4
- 238000005259 measurement Methods 0.000 claims description 31
- 238000013528 artificial neural network Methods 0.000 claims description 10
- 238000013507 mapping Methods 0.000 claims description 6
- 230000005672 electromagnetic field Effects 0.000 claims description 5
- 239000012530 fluid Substances 0.000 claims description 4
- 230000008878 coupling Effects 0.000 claims description 2
- 238000010168 coupling process Methods 0.000 claims description 2
- 238000005859 coupling reaction Methods 0.000 claims description 2
- 238000004519 manufacturing process Methods 0.000 claims description 2
- 238000012795 verification Methods 0.000 claims description 2
- 238000012544 monitoring process Methods 0.000 abstract description 12
- 238000004364 calculation method Methods 0.000 abstract description 11
- 238000012423 maintenance Methods 0.000 description 4
- 238000013473 artificial intelligence Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 239000003245 coal Substances 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 238000010200 validation analysis Methods 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B17/00—Systems involving the use of models or simulators of said systems
- G05B17/02—Systems involving the use of models or simulators of said systems electric
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/23—Design optimisation, verification or simulation using finite element methods [FEM] or finite difference methods [FDM]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- 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
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
-
- 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
- G06T17/20—Finite element generation, e.g. wire-frame surface description, tesselation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2111/00—Details relating to CAD techniques
- G06F2111/10—Numerical modelling
-
- 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
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- 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
- G06N3/045—Combinations of networks
-
- 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
- G06N3/0464—Convolutional networks [CNN, ConvNet]
-
- 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
- G06N3/0499—Feedforward networks
-
- 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
- Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
- Y02E60/00—Enabling technologies; Technologies with a potential or indirect contribution to GHG emissions mitigation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- Evolutionary Computation (AREA)
- Geometry (AREA)
- Software Systems (AREA)
- Economics (AREA)
- General Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Computer Hardware Design (AREA)
- Water Supply & Treatment (AREA)
- Human Resources & Organizations (AREA)
- Automation & Control Theory (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Artificial Intelligence (AREA)
- Public Health (AREA)
- Computer Graphics (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- Strategic Management (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a real-time simulation method of a physical field of a large power generation equipment part, which is based on design data of the large power generation equipment part, carries out finite element simulation modeling on a real running environment of the power generation equipment to obtain a finite element simulation model of the physical field of the large power generation equipment part, and generates a physical field simulation data set in the simulation environment; training and testing the accuracy of the AI proxy model according to the physical field simulation data set to obtain a physical field AI proxy model meeting expected requirements; and driving the physical field AI proxy model to perform physical field real-time simulation by utilizing the measured data of the equipment working condition measuring sensor collected regularly. In the simulation realization process, the AI model is used for replacing the finite element simulation model to predict the physical field distribution in the deployment stage, the invention can realize the full-equipment full-state monitoring, avoid a large amount of grid calculation in the finite element simulation, reduce the operation complexity, promote the simulation speed and perform online real-time physical field simulation.
Description
Technical Field
The invention relates to the technical field of intelligent operation and maintenance, artificial intelligence and finite element simulation of power generation equipment, in particular to a real-time simulation method of a physical field of a large-scale power generation equipment component.
Background
The stable and reliable operation of the large-scale power generation equipment is a precondition for guaranteeing the stable supply of the power, and is related to national life. In the industry, the health state of the equipment can be confirmed by directly or indirectly monitoring physical field information such as stress field, temperature field, electromagnetic field, sound field, fluid field and the like of the equipment, and key information is provided for the operation and maintenance of the equipment.
The state monitoring of the current power generation equipment mainly uses a traditional limited measuring point monitoring mode. If such conventional monitoring means are employed, there are a number of disadvantages: (1) Because of the huge size of the power generation equipment, the measurement range of a single measuring point is limited, the cost is huge, and the power generation equipment cannot be realized; (2) The power generation equipment has a complex structure, and partial physical fields, such as a temperature field and an electromagnetic field, are distributed in the equipment and cannot be measured by arranging a sensor; (3) If the measuring points are added, the problem of high cost exists, and only a small number of measuring points can be arranged on the part of the power generation equipment, so that obvious monitoring blind areas exist, the health state of the equipment can not be comprehensively estimated, and potential safety hazards exist.
Due to the lack of comprehensive status information of the power generation equipment, traditional equipment predictive maintenance often only predicts what time the equipment is bad, but cannot predict what critical parts of the equipment are problematic. In addition, in the traditional monitoring process, the equipment is required to stop running when monitoring key parts of some power generation equipment, then a professional temporarily arranges sensors on site to measure related parameters, and after the measurement is completed, the temporarily arranged sensors are removed to allow the equipment to run again. The method not only affects the power generation continuity of the power station, but also has high labor cost and low efficiency.
Disclosure of Invention
In order to solve the problems of insufficient measuring points, blind areas in monitoring and the like in the current power generation equipment state monitoring, the invention provides a real-time simulation method for the physical field of a large power generation equipment part, and the rapid and complete virtual measurement of the equipment state can be realized based on the combination of finite element simulation and artificial intelligence.
The technical scheme of the invention is as follows:
a real-time simulation method for the physical field of a large-scale power generation equipment component comprises the following steps:
1) Finite element high-precision modeling: based on design data of large-scale power generation equipment parts, finite element simulation modeling is carried out on the real operation environment of the power generation equipment, possible operation working conditions of the large-scale power generation equipment are set in the simulation environment, distribution conditions of various physical fields are simulated, and a physical field finite element simulation model of the large-scale power generation equipment parts is built;
the finite element simulation modeling is to perform full-size simulation modeling of 1 to 1 according to design parameters of power generation equipment. The method comprises the steps of carrying out fine mesh subdivision on power generation equipment, wherein the finer the mesh subdivision is, the more accurate the finite element simulation calculation is, and then solving each finite element unit formed after subdivision so as to meet the high-precision modeling requirement;
the simulation calculation of the finite element on the physical field is essentially to solve differential equations of the endpoints after the physical entity mesh of the power generation equipment is split, and boundary conditions are required to be specified in order to solve the equations. Boundary refers to the boundary between an object and an external environment, for example, the boundary between the device and the environment needs to be set firstly for analyzing the structural stability of the power generation device; such as which positions are stationary after installation, and the like. The boundary condition is also called a definite solution condition, and the unknown number in the differential equation solution, namely the data of the physical field, can be solved only by a given initial value;
2) Generating a physical field simulation data set in a simulation environment based on a physical field finite element simulation model of the large power generation equipment component;
3) Splitting a physical field simulation data set into a training set I and a testing set I, wherein the training set I is used for training an AI proxy model, and the testing set I is used for carrying out precision testing on the AI proxy model trained by the training set I so as to enable the AI proxy model to learn the mapping relation between the input and the output of finite element simulation; for the AI proxy model which meets the expected requirement after the precision test, the AI proxy model is used as a physical field AI proxy model for real-time simulation;
4) For the physical field AI proxy model available for real-time simulation: and driving the physical field AI proxy model to perform physical field real-time simulation by utilizing the actual measurement data of the equipment working condition measurement sensor periodically collected from the power station site of the large-scale power generation equipment.
Further, the large-scale power generation equipment refers to various power generation equipment for providing electric energy, and covers wind power, solar energy, hydropower, nuclear power, a gas turbine and coal power, and typical equipment comprises a wind power unit, solar power generation equipment, a hydropower unit, a nuclear power unit, a thermal power unit and the like.
The power generation equipment can generate different physical fields in different working environments, for example, each part of the equipment can be stressed to generate certain deformation in the running process of the water turbine, and all the deformation are combined together to form a stress field. In the same way, each part of the equipment generates a certain temperature in the working process, and all the equipment is stably combined together to form a temperature field. Electromagnetic fields, acoustic fields, fluid fields, and the like. The physical fields in the simulation step include a single physical field and two or more multiple physical coupling fields, and specifically include at least one or more of a stress field, a temperature field, an electromagnetic field, a sound field, a fluid field, and the like.
Further, the structure of the AI proxy model comprises a feedforward neural network, a convolution neural network and a circulation neural network, and the specific structure can be formed by any combination of the three neural networks.
Further, step 1) after a physical field finite element simulation model of a large-scale power generation equipment component is built, the accuracy of the physical field finite element simulation model can be corrected through real data measured by a local physical field measurement sensor arranged on a power station site of the large-scale power generation equipment; thus, a complete digital twin body of a physical field of a large-scale power generation equipment component is constructed in a simulation environment, and complete mapping of equipment states is realized.
Further, the accuracy verification in step 3) is to evaluate the AI proxy model accuracy by mean absolute error (MAE, mean absolute error) and standard mean absolute error (NMAE, normalized mean absolute error):
wherein, the sample number of the test set, () is a finite element simulation value, and' () is an AI proxy model predicted value; if NMAE is in the error accuracy receiving range, the accuracy of the AI proxy model is considered to meet the expected requirement; if NMAE is not in the error accuracy acceptance range, the accuracy of the AI proxy model is not considered to meet the expected requirement.
And for the AI proxy model which does not meet the expected requirement, training is continued through the training set until the accuracy meets the expected requirement.
Further, the actual measurement data of the local physical field sensor is collected periodically from the power station site of the large power generation equipment, a physical field actual measurement data set is manufactured and split into a physical field actual measurement training set and a physical field actual measurement test set, a new training set II and a new test set II are formed, an AI proxy model is trained through the training set II, and the accuracy of the AI proxy model is tested through the test set II until the accuracy of the AI proxy model meets the expected requirement.
When the method is used for multi-physical-field real-time simulation, a physical-field AI proxy model which can be used for real-time simulation is deployed in a power station test, and the physical-field real-time simulation is performed by using equipment working condition measurement sensor data arranged on site to drive the AI model. In the deployment stage, an AI model is used for "proxy" of the finite element simulation model of the physical field, and the segmentation of millions or even tens of millions of finite element grids and the calculation of finite element numerical values can be avoided, so that the calculation complexity is greatly reduced, the calculation speed is greatly improved, and the real-time physical field simulation is realized.
The technical scheme of the invention has the following beneficial effects:
(1) Building a high-precision physical field finite element simulation model based on large-scale power generation equipment design data, correcting finite element modeling precision by matching with local physical field actual measurement data, realizing full-equipment full-state monitoring, and completely solving the problems of insufficient measuring points and blind areas in monitoring existing in the operation and maintenance of the traditional power generation equipment;
(2) The invention trains the AI proxy model of the physical field by utilizing the data set; in the actual deployment stage, the AI model is used for replacing the finite element simulation model, so that millions of grid calculation and even tens of millions of grid calculation caused by large volume and complex structure of large power generation equipment are avoided, the calculation amount is greatly reduced, the simulation calculation time of the original single working condition can be shortened from hours to seconds, and online real-time physical field simulation can be performed.
Drawings
FIG. 1 is a schematic diagram of a real-time simulation flow of the present invention.
Fig. 2 is a schematic diagram of the physical field AI proxy model training process of the present invention.
Detailed Description
The technical scheme of the invention is further elaborated below in conjunction with the description and drawings.
As shown in FIG. 1, the simulation method of the physical field of the large-scale power generation equipment component comprises the following simulation steps:
1) Finite element high-precision modeling: based on design data of large-scale power generation equipment parts, high-precision full-size finite element simulation modeling is carried out on the real operation environment of the power generation equipment, possible operation working conditions of the large-scale power generation equipment are set in the simulation environment, distribution conditions of various physical fields are simulated, and a physical field finite element simulation model for forming the large-scale power generation equipment parts is built.
After a physical field finite element simulation model of a large power generation equipment part is built, the accuracy of the physical field finite element simulation model can be corrected through real data measured by a local physical field measurement sensor arranged on a power station site of the large power generation equipment; thus, a complete digital twin body of a physical field of a large-scale power generation equipment component is constructed in a simulation environment, and complete mapping of equipment states is realized.
2) Based on the physical field finite element simulation model of the large power generation equipment component, a physical field simulation data set is generated in a simulation environment.
3) As shown in fig. 2, the physical field simulation data set is split into a training set one and a testing set one, the training set one is used for training the AI proxy model, and the testing set one is used for performing precision test on the AI proxy model trained by the training set one so that the AI proxy model learns the mapping relation between the input and the output of the finite element simulation; for the AI proxy model which meets the expected requirement after the precision test, the AI proxy model is used as a physical field AI proxy model for real-time simulation;
the AI proxy model comprises a feedforward neural network, a convolution neural network and a circulation neural network.
Specific accuracy validation is to evaluate AI proxy model accuracy by mean absolute error (MAE, mean absolute error) and standard mean absolute error (NMAE, normalizedmean absolute error):
wherein, the sample number of the test set, () is a finite element simulation value, and' () is an AI proxy model predicted value; if NMAE is in the error accuracy receiving range, the accuracy of the AI proxy model is considered to meet the expected requirement; if NMAE is not in the error accuracy acceptance range, the accuracy of the AI proxy model is not considered to meet the expected requirement.
And for the AI proxy model which does not meet the expected requirement, training is continued through the training set until the accuracy meets the expected requirement.
Optimization for AI proxy model: and periodically collecting actual measurement data of a local physical field sensor from a power station site of the large power generation equipment, manufacturing a physical field actual measurement data set, splitting the actual measurement data set into a physical field actual measurement training set and a physical field actual measurement test set, forming a new training set II and a new test set II, training an AI proxy model through the training set II, and testing the accuracy of the AI proxy model through the test set II until the accuracy of the AI proxy model meets the expected requirement.
4) For the physical field AI proxy model available for real-time simulation: and driving the physical field AI proxy model to perform physical field real-time simulation by utilizing the actual measurement data of the equipment working condition measurement sensor periodically collected from the power station site of the large-scale power generation equipment.
When the simulation method is used for multi-physical-field real-time simulation, taking real-time simulation of the stress field of the bearing rack under the hydroelectric generating set of the hydroelectric generating station as an example, a physical-field AI proxy model for real-time simulation is deployed in the power station test, and the AI model is driven to perform real-time simulation by using the on-site arranged equipment working condition measurement sensor data.
According to the actual condition of the machine set site, a measuring sensor can be arranged at the key part of the lower frame. The sensor mainly measures two types of data, one type is used for measuring working condition data of the unit, such as the load born by the frame; one type is used to measure stress values at critical locations under current conditions (i.e., load).
The correction process comprises the following steps: firstly, training of an AI proxy model depends on a finite element simulation calculation result, and the initial finite element simulation engineering construction is completely dependent on a theoretical value and has a certain deviation from the actual situation of the site, so that the modeling precision of the finite element simulation engineering needs to be corrected by using a stress measurement sensor of the site, the error between the stress value simulated by the finite element and the stress measurement sensor value of the site is ensured to be within an allowable range, and further a training set I and a testing set I for training the stress real-time simulation AI proxy model are obtained through simulation.
The optimization process comprises the following steps: after the AI proxy model is deployed on the site of the power station, a new training data set II and a new testing data set II are formed by continuously collecting working condition data and local stress measurement data measured by the site sensor, and the AI proxy model is subjected to fine tuning, so that simulation errors of the AI proxy model are always kept in an error operation range.
Claims (8)
1. A real-time simulation method for the physical field of a large-scale power generation equipment part is characterized by comprising the following steps:
1) Based on design data of large-scale power generation equipment parts, finite element simulation modeling is carried out on the real operation environment of the power generation equipment, possible operation working conditions of the large-scale power generation equipment are set in the simulation environment, distribution conditions of various physical fields are simulated, and a physical field finite element simulation model of the large-scale power generation equipment parts is built;
2) Generating a physical field simulation data set in a simulation environment based on a physical field finite element simulation model of the large power generation equipment component;
3) Splitting a physical field simulation data set into a training set I and a testing set I, wherein the training set I is used for training an AI proxy model, and the testing set I is used for carrying out precision testing on the AI proxy model trained by the training set I so as to enable the AI proxy model to learn the mapping relation between the input and the output of finite element simulation; for the AI proxy model which meets the expected requirement after the precision test, the AI proxy model is used as a physical field AI proxy model for real-time simulation;
4) For the physical field AI proxy model available for real-time simulation: and driving the physical field AI proxy model to perform physical field real-time simulation by utilizing the actual measurement data of the equipment working condition measurement sensor periodically collected from the power station site of the large-scale power generation equipment.
2. The real-time simulation method for the physical field of the large power generation equipment component according to claim 1, wherein the method comprises the following steps: the large-scale power generation equipment refers to various power generation equipment for providing electric power energy, and at least comprises a wind turbine generator, solar power generation equipment, a hydroelectric turbine generator, a nuclear power turbine generator and a thermal power turbine generator.
3. The real-time simulation method for the physical field of the large power generation equipment component according to claim 1, wherein the method comprises the following steps: the physical fields comprise a single physical field and two or more multi-physical coupling fields, and specifically at least comprise one or more of stress fields, temperature fields, electromagnetic fields, sound fields, fluid fields and the like.
4. The real-time simulation method for the physical field of the large power generation equipment component according to claim 1, wherein the method comprises the following steps: the AI proxy model has a structure formed by any combination of a feedforward neural network, a convolution neural network and a circulation neural network.
5. The real-time simulation method for the physical field of the large power generation equipment component according to claim 1, wherein the method comprises the following steps: step 1), after a physical field finite element simulation model of a large power generation equipment part is built, correcting the precision of the physical field finite element simulation model through real data measured by a local physical field measurement sensor arranged on a power station site of the large power generation equipment; thus, a complete digital twin body of a physical field of a large-scale power generation equipment component is constructed in a simulation environment, and complete mapping of equipment states is realized.
6. The real-time simulation method for the physical field of the large power generation equipment component according to claim 1, wherein the method comprises the following steps: the accuracy verification in step 3) is to evaluate the AI proxy model accuracy by mean absolute error (MAE, mean absolute error) and standard mean absolute error (NMAE, normalized mean absolute error):
wherein, the sample number of the test set, () is a finite element simulation value, and' () is an AI proxy model predicted value; if NMAE is in the error accuracy receiving range, the accuracy of the AI proxy model is considered to meet the expected requirement; if NMAE is not in the error accuracy acceptance range, the accuracy of the AI proxy model is not considered to meet the expected requirement.
7. The real-time simulation method for the physical field of the large power generation equipment component according to claim 6, wherein the method comprises the following steps: and for the AI proxy model which does not meet the expected requirement, training is continued through the training set until the accuracy meets the expected requirement.
8. The real-time simulation method for the physical field of the large power generation equipment component according to claim 1 or 7, wherein: and periodically collecting actual measurement data of a local physical field sensor from a power station site of the large power generation equipment, manufacturing a physical field actual measurement data set, splitting the actual measurement data set into a physical field actual measurement training set and a physical field actual measurement test set, forming a new training set II and a new test set II, training an AI proxy model through the training set II, and testing the accuracy of the AI proxy model through the test set II until the accuracy of the AI proxy model meets the expected requirement.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310337264.5A CN116909165A (en) | 2023-03-31 | 2023-03-31 | Real-time simulation method for physical field of large power generation equipment component |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310337264.5A CN116909165A (en) | 2023-03-31 | 2023-03-31 | Real-time simulation method for physical field of large power generation equipment component |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116909165A true CN116909165A (en) | 2023-10-20 |
Family
ID=88363493
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310337264.5A Pending CN116909165A (en) | 2023-03-31 | 2023-03-31 | Real-time simulation method for physical field of large power generation equipment component |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116909165A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117973236A (en) * | 2024-04-01 | 2024-05-03 | 北京适创科技有限公司 | Correction method, system, equipment and storage medium for simulation result |
-
2023
- 2023-03-31 CN CN202310337264.5A patent/CN116909165A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117973236A (en) * | 2024-04-01 | 2024-05-03 | 北京适创科技有限公司 | Correction method, system, equipment and storage medium for simulation result |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110259646B (en) | Wind generating set component state early warning method based on historical data | |
| Liu et al. | Wind power plant prediction by using neural networks | |
| US8639454B2 (en) | Method of efficacy anticipation and failure examination for an apparatus | |
| CN110363334B (en) | Grid line loss prediction method of photovoltaic grid connection based on gray neural network model | |
| CN102052233A (en) | Water turbine regulating system module used for stability analysis of power system | |
| WO2024040608A1 (en) | Model training method for energy management system, apparatus, and storage medium | |
| CN116909165A (en) | Real-time simulation method for physical field of large power generation equipment component | |
| CN115549187A (en) | Stability evaluation method for wind power integration | |
| Xu et al. | Short-term wind speed prediction based on GRU | |
| Bai et al. | Automatic modeling and optimization for the digital twin of a regional multi-energy system | |
| CN107221933B (en) | Probabilistic load flow calculation method | |
| Zhao et al. | Gas turbine exhaust system health management based on recurrent neural networks | |
| CN108233356B (en) | Consistency evaluation method and evaluation platform for photovoltaic inverter controller | |
| CN117272110A (en) | Running state fault diagnosis and prediction method for fan generator and gearbox | |
| Li et al. | An overview of digital twin concept for key components of renewable energy systems | |
| Yusupov et al. | Development of a simulation model for assessing the technical condition of transformers exploited in hydroelectric stations | |
| CN116629406A (en) | Power consumption carbon emission prediction method based on improved Prophet model | |
| Devarajan et al. | Modeling and Simulation of Smart Power Systems Using HIL | |
| Jiangxia et al. | A combination prediction model for wind farm output power | |
| Liu et al. | Short-Term Wind Speed and Wind Power Prediction Based on Meteorological Model Modification | |
| Sun et al. | Reliability prediction of distribution network based on PCA-GA-BP neural network | |
| CN106782703A (en) | A kind of flexible measurement method of one loop of nuclear power station cooling agent leakage | |
| Barrios Gómez et al. | System identification applied to a single area electric power system under frequency response | |
| CN115374957B (en) | Radial flow type small hydropower station multiscale missing measurement data reconstruction method | |
| CN113468843B (en) | Construction method of equivalent model of active power distribution 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 |