Thermal power plant intelligent control system and method based on NPU

Abstract

The invention relates to an NUP-based intelligent control system for a thermal power plant, which combines a DCS with an NPU to construct a novel thermal power control system comprising a control layer, a calculation layer, a communication layer and a management layer. In a calculation layer, traditional control logic is executed in a DCS, and advanced algorithms such as artificial intelligence and the like are executed by an NPU acceleration module; a general compiler is adopted to support a mainstream programming language to construct an algorithm library, and a deep learning framework is combined to improve a function; the data management uses a lightweight database; and the communication layer adopts a high-speed industrial Ethernet and is compatible with a mainstream protocol to realize data interaction. According to the scheme, the defects of insufficient response speed, algorithm processing capability and parallel operation capability when a traditional DCS deals with the problems of new energy fluctuation and the like are solved, and the effects of improving the flexibility and stability of a unit, improving the calculation speed of an AI algorithm, reducing the data query and extraction speed, realizing self-adaptive adjustment of control parameters of the unit and iterative optimization of performance are achieved. And the system has good compatibility and expansibility, and can be seamlessly connected with the existing DCS system.

Description

Technical Field

[0001] The present application relates to the field of power plant operation and maintenance control, and in particular to an NPU-based intelligent control system, method, computer equipment, and computer-readable storage medium for thermal power plants. Background Art

[0002] In modern power systems, the flexibility of thermal power units has become a key issue in adapting to the large-scale grid connection of renewable energy. Although traditional DCS (Distributed Control System) has high integration and stability, its response speed and algorithm processing capabilities are insufficient when faced with the volatility, randomness, and uncertainty of renewable energy.

[0003] With the large-scale application of artificial intelligence (AI) and big data analysis technologies in power plants, the traditional DCS method of processing and executing algorithms through the CPU is unable to guarantee its parallel computing capabilities and response requirements. Summary of the Invention

[0004] The embodiments of the present application provide a NUP-based intelligent control system, method, computer device, and computer-readable storage medium for a thermal power plant, to at least address the problem in related technologies that traditional DCS methods are difficult to ensure parallel computing and response requirements of power plants.

[0005] In a first aspect, an embodiment of the present application provides a NPU-based intelligent control system for a thermal power plant, characterized in that it is used to combine a DCS with an NPU to implement intelligent control of a thermal power plant. The system includes: a control layer, a computing layer, and a communication layer, wherein:

[0006] The control layer is used to collect the operation data of the thermal power plant and, through the DCS control mode, perform the bottom-level operation monitoring and bottom-level operation control of the power plant according to the operation data;

[0007] The computing layer is used to perform the computational processing of conventional control tasks through the DCS control mode based on the operating data, and to accelerate the computational processing of intelligent control tasks through the NPU control mode;

[0008] The communication layer is used to implement data interaction between the DCS control mode and the NPU control mode using high-speed industrial Ethernet.

[0009] In some embodiments, the system further includes a management layer, which is used to store control algorithms related to the intelligent control tasks and call corresponding control algorithms through a standardized interface according to task requirements.

[0010] In some embodiments, when the control layer performs the underlying operation control, the computing layer is configured to obtain the operation result of the intelligent control task through the NPU control mode based on the operation data, and send the operation result to the control layer through the communication layer;

[0011] The control layer is used to adjust and optimize control parameters according to the calculation results of the NPU through a programmable logic controller, obtain optimized control parameters, and execute the underlying operation control through the optimized control parameters.

[0012] In some embodiments, the conventional control tasks include: sequential control tasks and preset control logic configuration tasks, and the intelligent control tasks include: artificial intelligence tasks, deep learning tasks and optimization algorithm modeling tasks and matrix operation tasks.

[0013] In some embodiments, the system further includes a pre-built module, which is used to define a class function through a compiler and build an algorithm library related to the intelligent control task based on the class function. The NPU control mode calls the corresponding control algorithm from the algorithm library according to task requirements to execute the intelligent control task.

[0014] Furthermore, a deep learning model related to the artificial intelligence task and the deep learning task is constructed based on the class function, and after completing the training of the deep learning model, the deep learning model is deployed to the computing layer for NPU control mode.

[0015] In some embodiments, the system also includes a data management module, which is used to store and manage the operation data and transit data through a lightweight database, and to call class functions and the lightweight database through an API interface of a preset programming language.

[0016] In some embodiments, the NPU control mode is used to perform control tasks related to the intelligent control, obtain vector results, and transmit the vector results to the DPU module;

[0017] The DPU module is used to output a control signal based on the vector result in combination with the measurement quantity and communication configuration information through a built-in custom algorithm black box module.

[0018] In a second aspect, an embodiment of the present application provides a method for intelligent control of a thermal power plant based on a NPU, characterized in that it is used to combine a DCS with an NPU to implement intelligent control of a thermal power plant, and the method includes:

[0019] Collecting the operating data of the thermal power plant through the control layer, and performing the bottom-level operation monitoring and bottom-level operation control of the power plant based on the operating data through the DCS control mode;

[0020] Through the computing layer, based on the operating data, the conventional control tasks are executed through the DCS control mode, and the intelligent control tasks are accelerated through the NPU control mode;

[0021] Through the communication layer, high-speed industrial Ethernet is adopted to realize data interaction between the DCS control mode and the NPU control mode.

[0022] In a third aspect, an embodiment of the present application provides a computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method described in the first aspect above when executing the computer program.

[0023] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described in the first aspect above.

[0024] Compared with related technologies, the NUP-based intelligent control system for thermal power plants provided in the embodiment of the present application combines the DCS system with the NPU to construct a new thermal power control system including a control layer, a computing layer, a communication layer, and a management layer. At the computing layer, traditional control logic is executed in the DCS, and advanced algorithms such as artificial intelligence are executed by the NPU acceleration module; in terms of algorithm implementation, a general compiler is used to support mainstream programming languages ​​to build an algorithm library, and the deep learning framework is combined to improve functionality; data management relies on a lightweight database; model deployment utilizes SSH transmission and is authenticated through a key pair; the communication layer uses high-speed industrial Ethernet and compatible mainstream protocols to achieve data interaction. This solution solves the shortcomings of traditional DCS in terms of insufficient response speed, algorithm processing capability, and parallel computing capability when dealing with problems such as the volatility of new energy, and achieves the goal of improving the flexibility and stability of the unit, increasing the calculation speed of the AI ​​algorithm, reducing the speed of data query and extraction, and realizing adaptive adjustment of unit control parameters, iterative performance optimization, and having good compatibility and scalability, and can be seamlessly connected with the existing DCS system. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0026] Figure 1 This is a structural block diagram of a thermal power plant intelligent control system based on NUP according to an embodiment of the present application;

[0027] Figure 2 This is a schematic diagram of a connection architecture between a host computer and an NPU according to an embodiment of the present application;

[0028] Figure 3 is a flow chart of a thermal power plant intelligent control method based on NUP according to an embodiment of the present application;

[0029] Figure 4 Schematic diagram of the internal structure of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0030] In order to make the purpose, technical solutions and advantages of this application more clearly understood, the present application is described and illustrated below in conjunction with the accompanying drawings and examples. It should be understood that the specific embodiments described herein are merely used to explain this application and are not intended to limit this application. Based on the embodiments provided in this application, all other embodiments obtained by those of ordinary skill in the art without making any creative efforts are within the scope of protection of this application.

[0031] Obviously, the drawings described below are merely examples or embodiments of the present application. Those skilled in the art can, without inventive effort, apply the present application to other similar scenarios based on these drawings. Furthermore, it is also understood that, although the effort involved in such a development process may be complex and lengthy, for those skilled in the art related to the content disclosed in this application, changes in design, manufacturing, or production based on the technical content disclosed in this application are merely conventional technical means and should not be construed as an insufficiency of the content disclosed in this application.

[0032] References to "embodiments" in this application mean that a particular feature, structure, or characteristic described in connection with the embodiment may be included in at least one embodiment of the application. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor does it refer to independent or alternative embodiments that are mutually exclusive of other embodiments. It is understood, both explicitly and implicitly, by those skilled in the art that the embodiments described in this application may be combined with other embodiments unless there is a conflict.

[0033] Unless otherwise defined, the technical or scientific terms used in this application should have the ordinary meaning understood by a person of ordinary skill in the technical field to which this application belongs. The words "one", "a", "the" and the like used in this application do not indicate a limit on quantity and may indicate the singular or plural. The terms "include", "comprise", "have" and any variations thereof used in this application are intended to cover non-exclusive inclusions; for example, a process, method, system, product or device that includes a series of steps or modules (units) is not limited to the listed steps or units, but may also include steps or units that are not listed, or may also include other steps or units that are inherent to these processes, methods, products or devices. The words "connect", "connected", "coupled" and the like used in this application are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. The word "multiple" used in this application refers to two or more. "And / or" describes the association relationship of associated objects, indicating that three relationships can exist. For example, "A and / or B" can mean: A exists alone, A and B exist at the same time, and B exists alone. The character " / " generally indicates that the objects before and after are in an "or" relationship. The terms "first", "second", "third", etc. involved in this application are only used to distinguish similar objects and do not represent a specific order for the objects.

[0034] Against the backdrop of the "dual carbon" goals and the construction of a new power system, the large-scale grid integration of renewable energy sources, such as wind and solar, has brought new opportunities and challenges to power system operations. Compared with traditional power generation models, renewable energy generation is characterized by randomness, volatility, and intermittency, posing severe challenges to grid stability. To address this issue, the flexibility and deep peak-shaving capabilities of thermal power units, as conventional power sources, have become key factors in ensuring the safe and stable operation of the power grid. By introducing technologies such as artificial intelligence (AI) and big data analysis, intelligent modeling and optimized control of the dynamic characteristics of thermal power units can significantly improve the flexibility and economic efficiency of the units. This intelligent control system can not only autonomously adapt to load changes but also achieve optimized scheduling and efficient operation under the conditions of renewable energy fluctuations.

[0035] In modern power systems, the flexible regulation of thermal power units has become a key issue in adapting to the large-scale integration of renewable energy. While traditional DCS systems offer high integration and stability, they lack the response speed and algorithm processing capabilities to address the volatility, randomness, and uncertainty of renewable energy. Furthermore, with the widespread adoption of artificial intelligence (AI) and big data analytics in power plants, traditional DCS systems, which rely on CPUs for algorithm processing and execution, are unable to maintain the parallel computing capabilities and responsiveness required.

[0036] In light of this, this application proposes a novel thermal power control system that combines a DCS system with a Network Processing Unit (NPU). By incorporating the NPU's efficient parallel computing capabilities and deep learning acceleration features, it can better adapt to the complex and changing operating environment of modern power grids and enhance the flexibility and stability of thermal power units.

[0037] Figure 1 This is a structural block diagram of a thermal power plant intelligent control system based on NUP according to an embodiment of the present application, such as Figure 1 As shown, the system includes a control layer 10, a computing layer 11, a communication layer 12 and a management layer 13, wherein:

[0038] The control layer 10 is used to collect the operation data of the thermal power plant and, through the DCS control mode, perform the bottom-level operation monitoring and bottom-level operation control of the power plant based on the operation data;

[0039] The control layer 10 plays a key role in the overall thermal power control system, performing low-level monitoring, control, and data collection for the unit. Specifically, low-level monitoring involves obtaining real-time information on the operating status and parameters of each unit component, such as temperature, pressure, and flow. Low-level control involves operating and controlling the unit equipment based on the monitored data and pre-set rules to ensure safe and stable operation.

[0040] In this embodiment, the traditional DCS (distributed control system) is a mature technology in the industrial control field, with a stable and reliable control model. The control layer 10 of this system retains this model to ensure the stability and continuity of the system's basic control functions, avoiding the risks and uncertainties associated with adopting a completely new control model.

[0041] Furthermore, parameter tuning and optimization are to adjust and optimize the control parameters in order to achieve the best performance of the control system. In this system, parameter tuning and optimization are not performed independently by the control layer 10. The NPU (network processing unit) uses its computing power and advanced algorithms to tune and optimize the control parameters. When the control layer 10 executes the underlying control logic, the computing layer 11 obtains the calculation results through the NPU system and sends the calculation results to the control layer 10 through the communication surface layer;

[0042] Finally, the PLC (Programmable Logic Controller) is the executive component in the control layer 10. It receives the calculation results of the NPU from the communication layer and adjusts the control parameters according to the established logic and rules, thereby realizing parameter setting and optimization, and ultimately accurately controlling the unit equipment.

[0043] It should be noted that the NPU in the computing layer possesses strong computing capabilities. It is used to execute artificial intelligence, deep learning, optimization algorithm modeling, and advanced control algorithms involving matrix operations for parameter tuning and optimization tasks in the control layer 10. These algorithms can analyze and process large amounts of data, find the optimal solution for control parameters, perform optimization calculations, and generate corresponding calculation results.

[0044] The computing layer 11 is used to perform the computational processing of conventional control tasks through the DCS control mode, and to accelerate the computational processing of intelligent control tasks through the NPU control mode;

[0045] The computing layer 11 in this system consists of two main parts: It processes conventional control tasks such as traditional control logic configuration and sequential control in DCS control mode. It accelerates intelligent control tasks such as artificial intelligence, deep learning, and optimization algorithm modeling, as well as matrix operations, through the NPU, enabling high-speed data processing and intelligent optimization algorithm execution.

[0046] Specifically, sequential control tasks execute a series of steps in a pre-defined order, with each step strictly prioritized to ensure smooth system operation. For example, when starting a thermal power plant, various devices must be turned on in a specific order. Preset control logic configuration tasks configure parameters and set functions for each link in the control system based on established logical relationships, laying the foundation for stable system operation.

[0047] Traditional control logic configuration and sequential control tasks are characterized by clear rules and fixed processes. Distributed Control Systems (DCSs), with their long-standing industrial control applications and mature technology, are capable of stably and reliably executing these traditional control tasks. Implementing these tasks within a DCS leverages its established control architecture and extensive application experience to ensure the proper functioning of the system's basic control functions.

[0048] Within intelligent control tasks, artificial intelligence (AI) tasks utilize intelligent algorithms to simulate human intelligence, enabling identification, prediction, and decision-making regarding the complex operating states of thermal power units. Deep learning tasks leverage deep neural networks to automatically extract key features from large amounts of data, address complex nonlinear problems, and enhance the system's adaptability to the complex operating conditions introduced by renewable energy grid integration. Optimization algorithm modeling tasks construct mathematical models to identify optimal system operating parameters and strategies, improving the unit's operating efficiency and economics. Matrix operations tasks provide efficient data processing and computational support for these tasks, accelerating algorithm execution. Intelligent control tasks rely on powerful computing capabilities to better address the volatility, randomness, and uncertainty inherent in renewable energy.

[0049] Understandably, artificial intelligence, deep learning, optimization algorithm modeling, and advanced control algorithms involving matrix operations typically require processing large amounts of data and complex calculations. The NPU (Network Processing Unit) is specifically designed to accelerate these calculations, with powerful parallel computing capabilities and efficient deep learning processing capabilities. Executing these advanced control algorithms through the NPU acceleration module significantly increases data processing speed, enabling rapid execution of intelligent optimization algorithms and enabling the system to better cope with the complex operating conditions and uncertainties brought about by the integration of new energy sources into the grid.

[0050] Figure 2 This is a schematic diagram of a connection architecture between a host computer and an NPU according to an embodiment of the present application. Figure 2 As shown in the figure, the NPU is equipped with a dedicated NPU host computer for algorithm development and call-up. Developers can write and debug algorithms here and then transfer them to the NPU for execution. The DPU is connected to multiple host computers, which are responsible for system configuration, monitoring, and other functions. Operators can use them to operate the DPU, which processes control signals and other functions within the system. The NPU, DPU, and historian are all connected via a bus connected via high-speed industrial Ethernet. The historian can store various data during the operation of the NPU and DPU, providing data support for system operation analysis, fault diagnosis, and other aspects. This connection architecture enables efficient data exchange between components, contributing to the coordinated operation and intelligent control of the entire system.

[0051] More specifically, the system also includes a pre-built module, which is interconnected with the computing layer. Using a compiler, the pre-built module defines class functions, which are code encapsulations with specific functionality. Based on these class functions, an algorithm library matching the intelligent control task is constructed and deployed in the computing module.

[0052] The algorithm library serves as a "toolbox" that contains various tools (algorithms) required for processing tasks such as artificial intelligence, deep learning, optimization algorithm modeling, and matrix operations. When faced with different intelligent control tasks, the NPU module can select the appropriate algorithm from this "toolbox" to perform the task according to specific needs.

[0053] Furthermore, the pre-built modules also construct algorithm models related to AI and deep learning tasks based on class functions. For example, a neural network model might be built to predict unit operating status. After training, these algorithm models are deployed to the NPU module. The NPU module can directly use these trained models to process actual tasks, improving processing efficiency and accuracy.

[0054] It's also worth noting that the algorithm development involved in this system utilizes a universal compiler and supports mainstream programming languages ​​such as Python, C, and C++. This multi-language support allows developers to choose the appropriate language for development based on their specific task requirements and personal programming habits. The host computer compiler directly synchronizes with the NPU, ensuring that algorithms written and compiled on the host computer can run quickly and accurately on the NPU module.

[0055] Furthermore, the algorithm library is implemented through self-developed function classes. These function classes are built on deep learning frameworks such as PyTorch and TensorFlow. These deep learning frameworks provide a rich set of tools and models, making the algorithm library powerful. This library supports the rapid deployment and updating of adaptive optimization and control strategies. The system can dynamically adjust algorithms and control strategies based on actual operating conditions to adapt to different operating conditions.

[0056] This system integrates key modeling and prediction models from PyTorch and TensorFlow, such as linear programming models, neural network models, and support vector regression models, into its self-deployed function classes. These models can be directly called, allowing developers to quickly apply mature models to different tasks, reducing development time and workload. The NPU in the intelligent control system possesses multi-threaded execution capabilities and can handle multiple tasks simultaneously, such as executing large model operations like DeepSeek. Furthermore, it can simultaneously support the storage and analysis of heterogeneous data streams, such as images and sounds, enabling the system to handle more complex and diverse data, further enhancing the system's intelligence and application scope.

[0057] In summary, the pre-built modules provide the foundation and tools for the NPU to perform intelligent control tasks by building an algorithm library and deploying algorithm models, while the specific algorithm implementation method ensures the flexibility and efficiency of algorithm development as well as the intelligence and adaptability of the system.

[0058] The system also includes a data management module, which uses the lightweight database SQLite to store and manage data, and implements calls between class functions and the database through an API interface in Python. For model management, PKL files are stored as binary files and called.

[0059] Regarding model deployment, this system first constructs a Python algorithm model within a defined class function. After model training is complete, the model is deployed to the slave computer. Due to the large-scale localization of DCS systems currently being implemented in domestic power plants, the systems used are mostly Linux distributions, and the slave computers also use Linux. Therefore, model deployment is performed via SSH between the slave and host computers. Given the closed network environment of power plants, SSH key pairs are used for authentication to prevent repeated verification during transmission and ensure network security. A public key is generated on the master computer, and a private key is added to the slave computer to ensure the legitimacy of the transmitted model.

[0060] Furthermore, the control algorithm integrates model predictive control, generalized predictive control, self-tuning PID control, and neural network control. Control algorithms are compiled and downloaded after configuration on the host computer. Vector calculations for AI-based control algorithms are performed by the NPU, and the results are transmitted to the DPU. Custom algorithm black box modules are built within the DPU configuration. Input pins are for measurement and communication-related configurations, while output pins are for calculated control signals.

[0061] It should be noted that this system integrates model predictive control, generalized predictive control, self-tuning PID control, and neural network control. Model predictive control uses the system's prediction model and optimization objective to continuously optimize control inputs; generalized predictive control is suitable for systems with long time lags, time-varying properties, and nonlinearities; self-tuning PID control automatically adjusts PID controller parameters; and neural network control leverages its powerful self-learning and nonlinear mapping capabilities to handle complex control problems. The synergy of these multiple algorithms allows the system to flexibly select the optimal control strategy for varying operating conditions.

[0062] Among them, the control algorithm is first compiled on the host computer, and the code is converted into a form that the machine can recognize. Then, the parameters, operating modes, etc. are configured according to the actual control needs. After completion, it is downloaded to the system-related equipment to prepare for actual operation. The AI-based control algorithm involves vector calculations, which are handled by the NPU. The NPU's powerful parallel computing and deep learning acceleration capabilities enable it to complete vector calculations quickly, and the calculation results are transmitted to the DPU to provide data support for subsequent control signal generation. A custom algorithm black box module is built in the DPU configuration. The input pin receives the measured quantity (such as real-time data such as unit temperature and pressure) and communication-related configuration information. After internal processing, the output pin outputs the calculated control signal. The "black box" feature simplifies the system design. The outside world does not need to understand the details of the internal calculations and only focuses on the input-output relationship, which enhances the modularity and versatility of the system.

[0063] The communication layer 12 is used to implement data interaction between the DCS control mode and the NPU control mode using high-speed industrial Ethernet.

[0064] Communication Layer 12 acts as a bridge between the DCS system and the NPU system (lower computer), enabling data exchange between the two. To ensure fast and accurate data transmission, high-speed industrial Ethernet is used to meet real-time data transmission requirements and ensure timely system response. Furthermore, the system's communication methods are compatible with Modbus and OPC UA, two mainstream industrial control protocols, enabling the system to better interface with diverse devices and systems.

[0065] Furthermore, the NPU and DPU are connected to the historical station and DCS via a bus, also using high-speed industrial Ethernet. This connection architecture enables efficient data transmission between components. Data processed by the NPU is promptly transmitted to the DPU, historical station, and DCS. DPU feedback data, stored data from the historical station, and control instructions from the DCS are also smoothly transmitted to the NPU, ensuring coordinated operation of all system components.

[0066] The management layer 13 is used to store various control algorithm files and call corresponding control algorithms through standardized interfaces according to task requirements.

[0067] Management layer 13 primarily focuses on control algorithm files. In terms of "synchronization," it ensures that control algorithm file versions are consistent across different devices and locations, preventing system anomalies caused by version differences. The "scheduling" function rationally arranges the execution order and execution time of each control algorithm based on the system's real-time operating conditions and task requirements, ensuring efficient utilization of system resources. "Management" encompasses operations such as storing, updating, and deleting algorithm files, ensuring orderly storage and convenient access to algorithm files. Providing standardized interfaces facilitates the rapid access and use of different algorithms, reduces the difficulty of algorithm development and integration, and improves system scalability. For example, when a new optimization algorithm needs to be introduced, it can be quickly integrated into the system through a standardized interface without the need for complex adaptation work.

[0068] Through the above system, the DCS system is combined with the NPU to construct a new thermal power control system consisting of a control layer 10, a computing layer 11, a communication layer 12, and a management layer 13. In the computing layer 11, traditional control logic is executed in the DCS, while advanced algorithms such as artificial intelligence are executed by the NPU acceleration module. For algorithm implementation, a general-purpose compiler is used to support mainstream programming languages ​​to build an algorithm library, integrating deep learning frameworks to enhance functionality. Data management utilizes a lightweight database. Model deployment utilizes SSH transmission and key pair authentication. The communication layer 12 uses high-speed industrial Ethernet and compatible mainstream protocols for data exchange. This solution addresses the shortcomings of traditional DCS systems in terms of response speed, algorithm processing capabilities, and parallel computing power when dealing with issues such as renewable energy volatility. It achieves the goals of improving unit flexibility and stability, increasing the calculation speed of AI algorithms, reducing data query and retrieval speeds, enabling adaptive adjustment of unit control parameters and iterative performance optimization, and possessing excellent compatibility and scalability, enabling seamless integration with existing DCS systems.

[0069] On the other hand, the embodiment of the present application also provides a thermal power plant intelligent control method based on NPU, which is used to combine DCS with NPU to realize thermal power plant intelligent control. Figure 3 is a flow chart of a thermal power plant intelligent control method based on NUP according to an embodiment of the present application, such as Figure 3 As shown, the method includes the following steps:

[0070] S301, collecting the operation data of the thermal power plant through the control layer, and performing bottom-level operation monitoring and bottom-level operation control of the power plant based on the operation data through the DCS control mode;

[0071] S302, through the computing layer, based on the operating data, the DCS control mode is used to perform the computational processing of conventional control tasks, and the NPU control mode is used to accelerate the computational processing of intelligent control tasks;

[0072] S303, through the communication layer, adopts high-speed industrial Ethernet to realize data interaction between DCS control mode and NPU control mode.

[0073] Through the above steps S101 to S103, the DCS system is combined with the NPU. At the computing layer, traditional control logic is executed in the DCS, while advanced algorithms such as artificial intelligence are executed by the NPU acceleration module. In terms of algorithm implementation, a general-purpose compiler is used to support mainstream programming languages ​​to build an algorithm library, and deep learning frameworks are combined to enhance functionality. Data management utilizes a lightweight database. Model deployment utilizes SSH transmission and key pair authentication. The communication layer uses high-speed industrial Ethernet and compatible mainstream protocols to achieve data interaction. This solution addresses the shortcomings of traditional DCS in terms of response speed, algorithm processing power, and parallel computing capabilities when dealing with issues such as the volatility of new energy. It achieves the goal of improving unit flexibility and stability, increasing the calculation speed of AI algorithms, reducing data query and extraction speed, achieving adaptive adjustment of unit control parameters, iterative performance optimization, and possessing good compatibility and scalability, allowing for seamless integration with existing DCS systems.

[0074] In one embodiment, Figure 4 is a schematic diagram of the internal structure of an electronic device according to an embodiment of the present application, such as Figure 4 As shown, an electronic device is provided, which may be a server, and its internal structure diagram may be as shown in FIG. Figure 4 As shown. This electronic device includes a processor, a network interface, internal memory, and non-volatile memory connected via an internal bus. The non-volatile memory stores an operating system, a computer program, and a database. The processor provides computing and control capabilities, the network interface communicates with external terminals via a network connection, the internal memory provides an environment for the operating system and computer program. When executed by the processor, the computer program implements a NUP-based intelligent control method for thermal power plants. The database stores data.

[0075] Those skilled in the art will understand that Figure 4 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the electronic device to which the solution of the present application is applied. The specific electronic device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0076] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing the relevant hardware through a computer program, which can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0077] The above embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that a person skilled in the art could make various modifications and improvements without departing from the spirit of the present application, all of which fall within the scope of protection of the present application. Therefore, the scope of protection of the present patent application shall be determined by the appended claims.

Claims

1. A thermal power plant intelligent control system based on NUP, characterized in that: The system is used to combine DCS with NPU to realize intelligent control of thermal power plants. The system includes: a control layer, a computing layer, and a communication layer, wherein: The control layer is used to collect the operation data of the thermal power plant and, through the DCS control mode, perform the bottom-level operation monitoring and bottom-level operation control of the power plant according to the operation data; The computing layer is used to perform the computational processing of conventional control tasks through the DCS control mode based on the operating data, and to accelerate the computational processing of intelligent control tasks through the NPU control mode; The communication layer is used to use high-speed industrial Ethernet to achieve data interaction between the DCS and the NPU control mode.

2. The system according to claim 1, wherein: The system further comprises a management layer, which is used to store control algorithms related to the intelligent control tasks and call corresponding control algorithms through a standardized interface according to task requirements.

3. The system according to claim 1, wherein: During the process of the control layer executing the bottom layer operation control, the computing layer is used to obtain the operation result of the intelligent control task through the NPU control mode based on the operation data, and send the operation result to the control layer through the communication layer; The control layer is used to adjust and optimize control parameters according to the calculation results of the NPU through a programmable logic controller, obtain optimized control parameters, and execute the underlying operation control through the optimized control parameters.

4. The system according to claim 3, characterized in that The conventional control tasks include: sequential control tasks and preset control logic configuration tasks, and the intelligent control tasks include: artificial intelligence tasks, deep learning tasks, optimization algorithm modeling tasks and matrix operation tasks.

5. The system according to claim 1, wherein: The system also includes a pre-built module, which is used to define a class function through a compiler and build an algorithm library related to the intelligent control task based on the class function. The NPU control mode calls the corresponding control algorithm from the algorithm library according to the task requirements to execute the intelligent control task. Furthermore, a deep learning model related to the artificial intelligence task and the deep learning task is constructed based on the class function, and after completing the training of the deep learning model, the deep learning model is deployed to the computing layer for NPU control mode.

6. The system according to claim 5, characterized in that The system also includes a data management module, which is used to store and manage the operating data and transit data through a lightweight database, and to call class functions and the lightweight database through an API interface of a preset programming language.

7. The system according to claim 1, wherein: The NPU control mode is used to execute control tasks related to the intelligent control, obtain vector results, and transmit the vector results to the DPU module; The DPU module is used to output a control signal based on the vector result in combination with the measurement quantity and communication configuration information through a built-in custom algorithm black box module.

8. A thermal power plant intelligent control method based on NUP, characterized in that: The method for combining a DCS with an NPU to realize intelligent control of a thermal power plant includes: Collecting the operating data of the thermal power plant through the control layer, and performing the bottom-level operation monitoring and bottom-level operation control of the power plant based on the operating data through the DCS control mode; Through the computing layer, based on the operating data, the conventional control tasks are executed through the DCS control mode, and the intelligent control tasks are accelerated through the NPU control mode; Through the communication layer, high-speed industrial Ethernet is adopted to realize data interaction between the DCS control mode and the NPU control mode.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the method according to claim 8 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to claim 8 is implemented.

Images