Thermal power generating unit operation and maintenance decision digital twin test system and method

Abstract

The invention discloses a thermal power generating unit operation and maintenance decision digital twinning test system. The system comprises a physical layer which is composed of thermal power generating unit equipment, a sensor and edge computing equipment; the data acquisition and transmission layer comprises a plurality of sensor interfaces and transmits the acquired data to a cloud or a local server in real time; according to the data twinborn model layer, sub-models of all devices of the thermal power generating unit are built based on physical modeling, the sub-models are optimized and integrated through a machine learning algorithm, and a digital twinborn model is built; the test and optimization decision layer is used for evaluating an operation and maintenance decision according to a simulation result and selecting an optimal operation and maintenance strategy; and a man-machine interaction layer. The invention belongs to the technical field of generator set operation and maintenance, and particularly provides a method for solving the problems of high data fusion difficulty, lack of real-time performance and verification mechanism, insufficient operation and maintenance optimization capability and the like in the prior art.

Description

Technical Field

[0001] The present invention belongs to the technical field of operation and maintenance of generator sets, and specifically refers to a digital twin test system and method for operation and maintenance decision-making of thermal power generator sets. Background Art

[0002] Thermal power generator sets are an important part of modern power systems, and their operation and maintenance are directly related to the stability and economy of the power grid. Traditional operation and maintenance methods mainly rely on regular inspections and experience-based maintenance strategies, which not only make it difficult to accurately predict equipment failures, but also easily lead to problems such as high maintenance costs, shortened equipment life, and lagged fault response. With the expansion of unit scale and the complex and changeable operating environment, it is difficult to quickly and accurately predict equipment failures and optimize operation and maintenance strategies in this way.

[0003] In recent years, Digital Twin technology has received extensive attention in the field of intelligent operation and maintenance. Digital Twin can map the state of the physical system in real time and achieve data-driven intelligent decision-making. However, the existing digital twin operation and maintenance systems for thermal power units still have the following technical difficulties:

[0004] 1. In terms of data fusion, thermal power generator sets involve multiple sensor data (temperature, pressure, vibration, flow, etc.), and the formats, precisions, and acquisition frequencies of these data are different, making it difficult to fuse.

[0005] 2. The combination of traditional physical models and data-driven models is weak, making it difficult to achieve high-precision prediction and real-time optimization decisions, and there is a lack of an efficient test environment, making it difficult to verify the reliability and applicability of digital twin decisions.

[0006] 3. Existing intelligent operation and maintenance solutions are often limited to single-point equipment, and the actual operating conditions are variable, making it difficult to make optimal decisions across the entire unit. Summary of the Invention

[0007] The technical problems to be solved by the present invention are the high difficulty of data fusion, the lack of real-time performance and verification mechanism, and the insufficient operation and maintenance optimization ability in the prior art.

[0008] To solve the above problems, the technical solutions adopted by the present invention are as follows:

[0009] On the one hand, the present invention proposes a digital twin test system for operation and maintenance decision-making of thermal power generator sets, including the following architecture:

[0010] Physical layer:

[0011] It is composed of thermal power unit equipment, sensors (such as temperature sensors, pressure sensors, vibration sensors, air flow sensors), edge computing devices, etc., and is used to obtain equipment operation status data, such as temperature, pressure, vibration, load, etc.

[0012] Data acquisition and transmission layer:

[0013] It includes a variety of sensor interfaces, which are used to collect the operation data of each device and sensor of the generator set equipment, and transmit the collected data to the cloud or local server in real time;

[0014] In addition, the data is preprocessed, and the preprocessing includes data cleaning and data fusion to improve the data quality.

[0015] Digital twin model layer:

[0016] Sub-models of each device of the thermal power generator set are constructed based on physical modeling. For example, a boiler model is constructed using thermodynamic principles, and a steam turbine model is constructed using mechanical dynamics principles, etc.;

[0017] Machine learning algorithms are used to optimize and integrate the above sub-models to construct a digital twin model to reflect the interrelationships between devices and the operating characteristics of the overall system. It is trained with a large amount of historical operation data to improve the accuracy of the model.

[0018] Testing and optimization decision-making layer:

[0019] Based on the digital twin model, different fault scenarios and operation and maintenance strategies are set, and simulation calculations are run to obtain the results of the changes in the operating state of the unit under different conditions; the operation and maintenance decisions are evaluated according to the simulation results, and the optimal operation and maintenance strategy is selected. For example: simulate the impact of different maintenance time intervals on the overall operation of the unit under the condition of boiler tube rupture failure, or the impact of different load adjustment strategies on the efficiency of the generator set;

[0020] Reinforcement learning algorithms are used to optimize the operation and maintenance decisions to improve the operation efficiency of the unit; algorithm comparison experiments are designed to evaluate the reliability of different operation and maintenance strategies.

[0021] Human-computer interaction layer:

[0022] A visualization interface is provided to display the unit status and the recommended results of operation and maintenance strategies.

[0023] Furthermore, the testing and optimization decision-making layer further includes an optimization module. The optimization module regularly compares the prediction results of the digital twin model with the operation data of the actual unit, calculates error metrics such as root mean square error, etc.; updates the digital twin model according to the error situation, including adjusting model parameters, optimizing algorithms, etc., to ensure the accuracy and effectiveness of the model, and re-run the operation and maintenance decision simulation.

[0024] Preferably, the data fusion adopts the Kalman filter algorithm combined with the neural network algorithm to fuse data from different sources and in different formats to obtain a unified representation of the device operation state data.

[0025] Preferably, the data twin model layer adopts multi-level modeling, specifically including the following sub-modules:

[0026] Physical simulation modeling module: Based on the mechanism model to simulate the operation behavior of the equipment, using methods such as finite element analysis (FEA), computational fluid dynamics (CFD), or / and thermal system simulation, etc., to construct sub-models of each equipment of the thermal power unit, such as thermal system simulation, fluid dynamics analysis, etc.;

[0027] Data-driven prediction module: Using methods such as machine learning and deep learning, based on historical data to predict the equipment status and potential faults;

[0028] Model integration and decision optimization module: Using data assimilation technology to integrate the sub-models of each equipment, combined with reinforcement learning and optimization algorithms, to provide an adaptive operation and maintenance strategy optimization scheme; The sub-models include boiler model, steam turbine model, and generator model.

[0029] Real-time data fusion and self-correction module: Continuously adjust the digital twin model through real-time sensor data to improve the simulation accuracy.

[0030] On the other hand, the present invention proposes a digital twin test method for the operation and maintenance decision of a thermal power generation unit. This method is implemented based on the above system, specifically including the following steps:

[0031] Step S1: Start the sensor, obtain the operation data of each equipment of the thermal power generation unit according to the set acquisition frequency, and perform data denoising, missing value filling, and anomaly detection; Transmit the processed data to the data twin model layer;

[0032] Step S2: Construct an initial sub-model according to the physical characteristics of the equipment, and then use machine learning algorithms to train and optimize the sub-model; Integrate each sub-model into a complete digital twin model of the thermal power generation unit, and conduct a preliminary verification;

[0033] Step S3: Set different fault scenarios and operation and maintenance strategy combinations in the digital twin model, run the simulation calculation, and obtain the operation status change results of the unit under different conditions, such as power output, equipment life loss, etc.; Evaluate the operation and maintenance decision according to the simulation results, select the optimal operation and maintenance strategy and deploy it to the actual unit.

[0034] Preferably, it further includes: Continuously optimize in combination with real-time data, regularly collect the operation data of the actual unit, compare and analyze it with the predicted data of the digital twin model, adjust and update the digital twin model according to the error situation, and then re-conduct the operation and maintenance decision simulation.

[0035] Adopting the above solution, the beneficial effects obtained by the present invention are as follows:

[0036] 1. Combine physical modeling and data-driven technologies to improve the prediction speed and accuracy of the unit's operating status. At the same time, provide a virtual test environment that can quickly adapt to different operating conditions and environmental changes, and improve the reliability of operation and maintenance decisions.

[0037] 2. Through the simulation of different operation and maintenance strategies by the digital twin model, more accurately predict equipment failures and system performance, reduce problems such as equipment damage caused by decision-making errors, and problems such as reduced power generation efficiency.

[0038] 3. Reduce unnecessary equipment maintenance and replacement costs, improve the operating efficiency of the unit, and greatly reduce the power generation cost. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 It is a schematic diagram of the architecture composition of a digital twin test system for operation and maintenance decision-making of a thermal power generation unit provided by the present invention;

[0040] Figure 2 It is a flowchart of a digital twin test method for operation and maintenance decision-making of a thermal power generation unit provided in this embodiment.

[0041] The accompanying drawings are used to provide a further understanding of the present invention, and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention, and do not constitute a limitation to the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0042] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments; based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0043] Embodiment 1

[0044] As Figure 1 shown, a digital twin test system for operation and maintenance decision-making of a thermal power generation unit proposed by the present invention includes a physical layer, a data acquisition and transmission layer, a data twin model layer, a test and optimization decision layer, and a human-computer interaction layer arranged in sequence from top to bottom.

[0045] Physical layer: It is composed of thermal power generation unit equipment, sensors (such as temperature sensors, pressure sensors, vibration sensors, air flow sensors), edge computing devices, etc., and is used to obtain equipment operation status data, such as temperature, pressure, vibration, load, etc.

[0046] Data Acquisition and Transmission Layer: It includes a variety of sensor interfaces for collecting the operation data of each device and sensor of the generator set equipment, and transmitting the collected data to the cloud or local server in real time.

[0047] In addition, the data is preprocessed, and the preprocessing includes data cleaning and data fusion to improve the data quality.

[0048] Among them, a data fusion unit can also be set in this layer. The data fusion adopts the Kalman filter algorithm combined with the neural network algorithm to fuse data from different sources and in different formats to obtain a unified representation of the equipment operation status data.

[0049] Data Twin Model Layer: Sub-models of each device of the thermal power generator set are constructed based on physical modeling. For example, the boiler model is constructed using thermodynamic principles, and the steam turbine model is constructed using mechanical dynamics principles, etc.

[0050] Among them, the data twin model layer adopts multi-level modeling, specifically including the following sub-modules:

[0051] Physical Simulation Modeling Module: Based on the mechanism model, it simulates the equipment operation behavior, and uses methods such as finite element analysis (FEA), computational fluid dynamics (CFD) or / and thermal system simulation to construct sub-models of each device of the thermal power unit, such as thermal system simulation, fluid dynamics analysis, etc.;

[0052] Data-driven Prediction Module: Using methods such as machine learning and deep learning, it predicts the equipment status and potential faults based on historical data;

[0053] Model Integration and Decision Optimization Module: It uses data assimilation technology to integrate the sub-models of each device, and combines reinforcement learning and optimization algorithms to provide an optimized solution for adaptive operation and maintenance strategies; the sub-models include the boiler model, the steam turbine model, and the generator model.

[0054] Real-time Data Fusion and Self-correction Module: It continuously adjusts the digital twin model through real-time sensor data to improve the simulation accuracy.

[0055] Machine learning algorithms are used to optimize and integrate the above sub-models to construct a digital twin model to reflect the interrelationships between devices and the operating characteristics of the overall system, and it is trained with a large amount of historical operation data to improve the accuracy of the model.

[0056] Testing and Optimization Decision Layer:

[0057] Based on the digital twin model, different fault scenarios and operation and maintenance strategies are set, and simulation calculations are run to obtain the results of the changes in the operating status of the unit under different conditions; the operation and maintenance decisions are evaluated based on the simulation results, and the optimal operation and maintenance strategy is selected. For example, simulate the impact of different maintenance time intervals on the overall operation of the unit under the condition of boiler tube rupture, or the impact of different load adjustment strategies on the efficiency of the generator set;

[0058] Use the reinforcement learning algorithm to optimize the operation and maintenance decisions and improve the operating efficiency of the unit; design algorithm comparison experiments to evaluate the reliability of different operation and maintenance strategies.

[0059] The testing and optimization decision-making layer also includes an optimization module. The optimization module regularly compares the prediction results of the digital twin model with the operating data of the actual unit, calculates error metrics such as root mean square error, etc.; updates the digital twin model according to the error situation, including adjusting model parameters, optimizing algorithms, etc., to ensure the accuracy and effectiveness of the model, and re-run the operation and maintenance decision simulation.

[0060] Human-computer interaction layer:

[0061] Provide a visual interface for displaying the unit status and the recommended results of operation and maintenance strategies.

[0062] Embodiment 2

[0063] As Figure 2 shown, based on the above system, this embodiment provides a digital twin testing method for operation and maintenance decision-making of thermal power generator sets, including the following steps:

[0064] Step S1: Start the sensors and obtain the operating data of each device of the thermal power generator set according to the set acquisition frequency;

[0065] Perform data denoising, missing value filling, and anomaly detection; transmit the processed data to the data twin model layer.

[0066] Step S2: Construct an initial sub-model according to the physical characteristics of the device, and then use machine learning algorithms to train and optimize the sub-model;

[0067] Integrate each sub-model into a complete digital twin model of the thermal power generator set and conduct preliminary verification.

[0068] Step S3: Set different combinations of fault scenarios and operation and maintenance strategies in the digital twin model, run simulation calculations to obtain the results of the changes in the operating status of the unit under different conditions; evaluate the operation and maintenance decisions based on the simulation results, select the optimal operation and maintenance strategy and deploy it to the actual unit.

[0069] Step S4: Continuously optimize by combining real-time data. Regularly collect the operation data of the actual unit, compare and analyze it with the predicted data of the digital twin model, adjust and update the digital twin model according to the error situation, and then re-perform the operation and maintenance decision-making simulation.

[0070] Embodiment 3

[0071] Combined with the above system and method, an embodiment is provided:

[0072] Install a variety of sensors such as temperature sensors and pressure sensors on Unit 3 of a certain thermal power plant. The temperature sensor collects data such as the steam temperature at the boiler outlet and the temperatures before and after each stage of the steam turbine, and the pressure sensor collects data such as the boiler drum pressure and the main steam pressure. The collection frequency is set to once every 5 minutes.

[0073] The data acquisition and transmission layer filters these data to remove noise interference, and then performs data fusion through the Kalman filter algorithm combined with the neural network algorithm to obtain unified equipment operation status data.

[0074] Construct a boiler sub-model according to the thermodynamic principle of the boiler, considering processes such as fuel combustion and heat transfer. Construct a steam turbine sub-model using the mechanical dynamics principle of the steam turbine, including processes such as blade rotation and steam work.

[0075] Use the long short-term memory network algorithm in deep learning to train these sub-models, and use the historical operation data of this unit in the past year for training, so that the model can accurately reflect the operation characteristics and mutual relationships of the equipment.

[0076] Set the boiler slagging fault scenario in the digital twin model, and then simulate the impact of different soot cleaning cycles (such as once a week and once every two weeks) on the overall operation of the unit.

[0077] The results show that when soot cleaning is performed once a week, the power generation efficiency loss of the unit is relatively small, and the equipment life loss is also relatively low, so it is determined that soot cleaning once a week is a better operation and maintenance strategy.

[0078] In addition, collect the actual operation data of Unit 3 once a month, and compare it with the predicted data of the digital twin model. Calculate the root mean square error. When the error exceeds 5%, adjust the parameters of the digital twin model, such as adjusting the calculation parameters of the boiler combustion efficiency, etc., and then re-perform the operation and maintenance decision-making simulation.

[0079] The above describes the present invention and its implementation manners. Such description is not restrictive. What is shown in the drawings is only one of the implementation manners of the present invention, and the actual structure is not limited thereto. In general, if those of ordinary skill in the art are inspired by it and, without departing from the gist of the present invention, design similar structural manners and embodiments to this technical solution without creative efforts, they shall fall within the protection scope of the present invention.

Claims

1. A digital twin test system for operation and maintenance decision-making of thermal power generation units, characterized in that, It includes: Physical layer: used to obtain device operation status data; Data acquisition and transmission layer: used to collect the operation data of each device and sensor of the generator set equipment, transmit the collected data to the cloud or local server in real time; and preprocess the data; Data twin model layer: based on physical modeling, sub-models of each device of the thermal power generator set are constructed, and machine learning algorithms are used to optimize and integrate the above sub-models to construct a digital twin model; Testing and optimization decision-making layer: based on the digital twin model, different fault scenarios and operation and maintenance strategies are set, simulation calculations are run, and the operation status change results of the unit under different conditions are obtained; the operation and maintenance decisions are evaluated according to the simulation results, and the optimal operation and maintenance strategy is selected; Human-computer interaction layer: provides a visual interface for displaying the unit status and operation and maintenance strategy recommendation results.

2. The digital twin test system for operation and maintenance decision-making of a thermal power generation unit according to claim 1, wherein: The testing and optimization decision-making layer further includes an optimization module, and the optimization module regularly compares the prediction results of the digital twin model with the operation data of the actual unit to calculate the error index; The digital twin model is updated according to the error situation, including adjusting model parameters, optimizing algorithms, and re-running the operation and maintenance decision simulation.

3. The digital twin test system for operation and maintenance decision-making of a thermal power generation unit according to claim 1, wherein: The pre-processor includes data cleaning and data fusion to improve data quality.

4. The digital twin test system for operation and maintenance decision-making of a thermal power generation unit according to claim 3, characterized in that: The data fusion uses the Kalman filter algorithm combined with the neural network algorithm to fuse data from different sources and in different formats to obtain a unified representation of device operation status data.

5. A digital twin test system for operation and maintenance decision-making of thermal power generation units according to claim 1, characterized in that: The data twin model layer adopts multi-level modeling, specifically including the following sub-modules: Physical simulation modeling module: based on the mechanism model, simulate the device operation behavior, and use methods such as finite element analysis, computational fluid dynamics or / and thermal system simulation to construct sub-models of each device of the thermal power unit; Data-driven prediction module: use machine learning and deep learning methods to predict device status and potential faults based on historical data; Model integration and decision optimization module: use data assimilation technology to integrate the sub-models of each device, combine reinforcement learning and optimization algorithms to provide an adaptive operation and maintenance strategy optimization plan; Real-time data fusion and self-calibration module: continuously adjust the digital twin model through real-time sensor data to improve the simulation accuracy.

6. The digital twin test system for operation and maintenance decision-making of a thermal power generation unit according to claim 5, wherein: The sub-models include a boiler model, a steam turbine model, and a generator model.

7. A method for testing the operation and maintenance decision-making digital twin of a thermal power generation unit, which is tested by using a digital twin test system for operation and maintenance decision-making of a thermal power generation unit as described in any one of claims 1-6, characterized in that It includes the following steps: Step S1: Start the sensor, obtain the operation data of each device of the thermal power generator set according to the set acquisition frequency, and perform data denoising, missing value filling, and anomaly detection; transmit the processed data to the data twin model layer; Step S2: Construct an initial sub-model according to the physical characteristics of the device, and then use machine learning algorithms to train and optimize the sub-model; integrate each sub-model into a complete digital twin model of the thermal power generator set and conduct preliminary verification; Step S3: Set different fault scenarios and operation and maintenance strategy combinations in the digital twin model, run simulation calculations, and obtain the operation status change results of the unit under different conditions; evaluate the operation and maintenance decisions according to the simulation results, select the optimal operation and maintenance strategy and deploy it to the actual unit.

8. A method for testing a digital twin of an operation and maintenance decision-making for a thermal power generation unit according to claim 7, characterized in that: Continuously optimize in combination with real-time data, regularly collect the operation data of the actual unit, conduct comparative analysis with the predicted data of the digital twin model, adjust and update the digital twin model according to the error situation, and then re-perform the operation and maintenance decision-making simulation.

Images