Energy-saving optimization method and device for thermal power auxiliary engine system, electronic equipment and storage medium

Abstract

The invention provides an energy-saving optimization method and device for a thermal power auxiliary engine system, electronic equipment and a storage medium, relates to the technical field of data processing, and aims to realize multi-physical-scale and multi-time-scale state estimation of the thermal power auxiliary engine system by constructing a hybrid dynamic model fusing a mechanism model and a time sequence prediction model. Meanwhile, on-line self-adaptive identification and correction are carried out on key operation parameters in the hybrid dynamic model based on the state estimation result so as to track the performance change of the auxiliary machine equipment, and a collaborative optimization control strategy of the system is generated by using the corrected hybrid dynamic model and adopting a multi-target dynamic optimization algorithm under the condition that a preset operation constraint condition is met. And the collaborative optimization control strategy is converted into an executable control instruction through a hierarchical coordination control mechanism, and the executable control instruction is issued to a field controller of each auxiliary machine equipment, so that the problem of relatively large deviation of an optimization result effect caused by limited processing capacity of a thermal power auxiliary machine system under a dynamic working condition and difficulty in self-adaptive updating of model parameters can be solved.

Description

Technical Field

[0001] This disclosure relates to the field of data processing technology, and in particular to a method and apparatus for energy-saving optimization of thermal power auxiliary equipment systems, electronic equipment, and storage media. Background Technology

[0002] Thermal power generation, as a crucial component of the global energy supply system, is widely used in critical scenarios such as grid peak shaving and baseload power supply. With the continuous demand for operational flexibility and energy efficiency improvements in power systems, collaborative energy-saving optimization technology for thermal power auxiliary systems has become a core direction for improving unit economics. Existing technologies utilize mathematical optimization methods such as linear programming, nonlinear programming, and dynamic programming, combined with intelligent algorithms such as genetic algorithms and neural networks, to construct a collaborative optimization system aimed at minimizing the total energy consumption of auxiliary systems. Specifically, this system covers key aspects from fuel transportation, pulverization, flue gas systems to desulfurization and denitrification. The mechanistic model is based on physical conservation laws and equipment characteristic curves, while the data-driven model employs deep learning time-series prediction networks (such as LSTM / GRU), using Kalman filtering or extended Kalman filtering algorithms to achieve hybrid dynamic modeling. However, existing methods have significant limitations under dynamic operating conditions: directly employing static mathematical models and fixed-parameter optimization strategies fails to effectively integrate the time-varying characteristics of equipment performance (such as decreased fan efficiency η_fan and increased heat exchanger fouling thermal resistance), leading to accumulated model prediction errors. Therefore, when units face rapid load changes (such as the transition from 500MW to 600MW) or fluctuations in fuel quality, traditional optimization schemes struggle to balance energy minimization with process parameter stability. Furthermore, the lack of online parameter identification mechanisms means that efficiency degradation due to equipment aging significantly exacerbates control deviations. This contradiction between model inaccuracies and control disconnect not only reduces the energy efficiency of auxiliary systems during dynamic processes but also risks exceeding equipment operating limits, ultimately impacting the economic viability, safety, and environmental compliance of thermal power plants. Summary of the Invention

[0003] This disclosure provides a method and apparatus for energy-saving optimization of thermal power auxiliary equipment systems, as well as electronic equipment and storage media. Its main objective is to at least partially solve one of the technical problems in related technologies.

[0004] According to a first aspect of this disclosure, an energy-saving optimization method for a thermal power auxiliary equipment system is provided, comprising: Based on a hybrid dynamic model, the state estimation of thermal power auxiliary equipment system is performed at multiple physical and time scales. The hybrid dynamic model is obtained by fusing a mechanistic model and a time-series prediction model. Based on the state estimation results, key operating parameters in the hybrid dynamic model are adaptively identified and corrected online to track performance changes of auxiliary equipment. Using the modified hybrid dynamic model and a multi-objective dynamic optimization algorithm, a collaborative optimization control strategy for the thermal power auxiliary equipment system is generated under the condition of satisfying preset operating constraints. The hierarchical coordination control mechanism transforms the collaborative optimization control strategy into executable control commands, which are then sent to the field controllers of each auxiliary device to achieve collaborative operation of the system.

[0005] Optionally, the mechanism model describes the mass balance, energy balance, momentum balance and thermodynamic balance relationships of the thermal power auxiliary equipment system through a set of differential equations; The time-series prediction model employs a recurrent neural network structure to capture the nonlinear dynamic characteristics and time-delay effects of the system. The mechanistic model and the time series prediction model are fused by a state estimation algorithm to combine the physical consistency of the mechanistic model with the nonlinear mapping capability of the data-driven model.

[0006] Optionally, the online adaptive identification and correction of key operating parameters in the hybrid dynamic model based on the state estimation results includes: A recursive estimation algorithm is used to identify the key operating parameters in real time, and to achieve joint estimation of state and parameters in a nonlinear system. The identified parameters are updated to the hybrid dynamic model in real time, and the model correction results are continuously verified. A parameter library is established based on historical identification results to predict the performance degradation trend of the auxiliary equipment.

[0007] Optionally, the step of employing a multi-objective dynamic optimization algorithm to generate a collaborative optimization control strategy for the thermal power auxiliary equipment system under preset operating constraints includes: Construct a multi-objective optimization function, which includes a total system energy consumption term, an operational stability penalty term, a pollutant emission penalty term, and an operational flexibility penalty term; Set operational constraints, which include physical limitations of equipment, limitations of process parameters, limitations of safety protection, limitations of environmental emission standards, and limitations of the rate of change of control variables; A rolling optimization and feedback correction mechanism is adopted to solve the multi-objective optimization function in the prediction time domain, generate the optimal control sequence, and generate the cooperative optimization control strategy based on the optimal control sequence.

[0008] Optionally, the hierarchical coordination control mechanism includes: The higher-level coordination control is used to decompose the global optimization goal into the collaborative operation goals of each auxiliary equipment subsystem; Lower-level local control is implemented by the field control loops of each auxiliary device, used to track the setpoints issued by the upper-level coordinated control and suppress local disturbances; and The parameters of the lower-level local controller are adaptively adjusted online to adapt to changes in the operating conditions of the thermal power auxiliary equipment system.

[0009] Optional, also includes: Based on the auxiliary decision-making mechanism, when the multi-objective dynamic optimization algorithm is inefficient, a fast control suggestion is provided. The auxiliary decision-making mechanism is trained offline and deployed online through machine learning algorithms. Its reward function is associated with the multi-objective optimization function and includes a penalty term for violating security constraints.

[0010] According to a second aspect of this disclosure, an energy-saving optimization device for a thermal power auxiliary system is provided, comprising: An estimation unit is used to perform state estimation of a thermal power auxiliary system at multiple physical and time scales based on a hybrid dynamic model, wherein the hybrid dynamic model is obtained by fusing a mechanistic model and a time-series prediction model. The correction unit is used to perform online adaptive identification and correction of key operating parameters in the hybrid dynamic model based on the state estimation results, so as to track the performance changes of auxiliary equipment. The generation unit is used to generate a collaborative optimization control strategy for the thermal power auxiliary equipment system by using the modified hybrid dynamic model and a multi-objective dynamic optimization algorithm, under the condition of satisfying preset operating constraints. The conversion unit is used to convert the collaborative optimization control strategy into executable control commands through a hierarchical coordination control mechanism, and then send them to the field controllers of each auxiliary equipment to realize the collaborative operation of the system.

[0011] Optionally, the mechanism model describes the mass balance, energy balance, momentum balance and thermodynamic balance relationships of the thermal power auxiliary equipment system through a set of differential equations; The time-series prediction model employs a recurrent neural network structure to capture the nonlinear dynamic characteristics and time-delay effects of the system. The mechanistic model and the time series prediction model are fused by a state estimation algorithm to combine the physical consistency of the mechanistic model with the nonlinear mapping capability of the data-driven model.

[0012] Optionally, the correction unit is also used for: A recursive estimation algorithm is used to identify the key operating parameters in real time, and to achieve joint estimation of state and parameters in a nonlinear system. The identified parameters are updated to the hybrid dynamic model in real time, and the model correction results are continuously verified. A parameter library is established based on historical identification results to predict the performance degradation trend of the auxiliary equipment.

[0013] Optionally, the generating unit is also used for: Construct a multi-objective optimization function, which includes a total system energy consumption term, an operational stability penalty term, a pollutant emission penalty term, and an operational flexibility penalty term; Set operational constraints, which include physical limitations of equipment, limitations of process parameters, limitations of safety protection, limitations of environmental emission standards, and limitations of the rate of change of control variables; A rolling optimization and feedback correction mechanism is adopted to solve the multi-objective optimization function in the prediction time domain, generate the optimal control sequence, and generate the cooperative optimization control strategy based on the optimal control sequence.

[0014] Optionally, the hierarchical coordination control mechanism includes: The higher-level coordination control is used to decompose the global optimization goal into the collaborative operation goals of each auxiliary equipment subsystem; Lower-level local control is implemented by the field control loops of each auxiliary device, used to track the setpoints issued by the upper-level coordinated control and suppress local disturbances; and The parameters of the lower-level local controller are adaptively adjusted online to adapt to changes in the operating conditions of the thermal power auxiliary equipment system.

[0015] Optional, also includes: An auxiliary decision-making unit is used to provide rapid control suggestions based on an auxiliary decision-making mechanism when the multi-objective dynamic optimization algorithm is inefficient. The auxiliary decision-making mechanism is trained offline and deployed online using a machine learning algorithm. Its reward function is associated with the multi-objective optimization function and includes a penalty term for violating safety constraints.

[0016] According to a third aspect of this disclosure, an electronic device is provided, comprising: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method described in the first aspect above.

[0017] According to a fourth aspect of this disclosure, a non-transitory computer-readable storage medium is provided storing computer instructions, wherein the computer instructions are configured to cause the computer to perform the method described in the first aspect above.

[0018] According to a fifth aspect of this disclosure, a computer program product is provided, comprising a computer program that, when executed by a processor, implements the method described in the first aspect above.

[0019] The energy-saving optimization method, device, electronic equipment, and storage medium for thermal power auxiliary equipment systems disclosed herein achieve state estimation of the thermal power auxiliary equipment system across multiple physical and time scales by constructing a hybrid dynamic model that integrates a mechanistic model and a time-series prediction model. Based on the state estimation results, key operating parameters in the hybrid dynamic model are adaptively identified and corrected online to track performance changes in the auxiliary equipment. Furthermore, using the corrected hybrid dynamic model and a multi-objective dynamic optimization algorithm, a collaborative optimization control strategy for the system is generated under preset operating constraints. This collaborative optimization control strategy is then transformed into executable control commands and distributed to the field controllers of each auxiliary equipment through a hierarchical coordination control mechanism. Therefore, this method addresses the limitations of existing technologies in handling nonlinear coupling, significant time delays, and parameter evolution characteristics under dynamic operating conditions of thermal power auxiliary equipment systems. It also overcomes the difficulty in adaptively updating model parameters with equipment performance changes, leading to significant deviations between optimization results and actual operating effects, and hindering the achievement of full-condition, adaptive, and collaborative energy-saving optimization. This method significantly improves the energy-saving effect of thermal power auxiliary equipment systems under dynamic operating conditions, enables adaptive and intelligent system operation, and enhances system stability, safety, and operational flexibility.

[0020] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of this disclosure, nor is it intended to limit the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description

[0021] The accompanying drawings are provided to better understand this solution and do not constitute a limitation of this disclosure. Wherein: Figure 1 A flowchart illustrating an energy-saving optimization method for a thermal power auxiliary equipment system provided in this embodiment of the disclosure; Figure 2 A schematic diagram of the structure of an energy-saving optimization device for a thermal power auxiliary system provided in this embodiment of the present disclosure; Figure 3 A schematic block diagram of an example electronic device provided for embodiments of this disclosure. Detailed Implementation

[0022] The exemplary embodiments of this disclosure are described below with reference to the accompanying drawings, including various details of the embodiments to aid understanding, and should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this disclosure. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0023] The following description, with reference to the accompanying drawings, outlines an energy-saving optimization method and apparatus for thermal power auxiliary systems, as well as electronic devices and storage media, representing embodiments of this disclosure.

[0024] Figure 1 This is a flowchart illustrating an energy-saving optimization method for a thermal power auxiliary system provided in an embodiment of this disclosure.

[0025] like Figure 1 As shown, the method includes the following steps: Step 101: Based on the hybrid dynamic model, perform state estimation of the thermal power auxiliary equipment system at multiple physical and time scales, wherein the hybrid dynamic model is obtained by fusing the mechanistic model and the time-series prediction model.

[0026] In the embodiments of this disclosure, to address the technical problem of insufficient system state estimation accuracy in existing thermal power auxiliary systems due to the coexistence of multiple physical scales (such as the coupling relationship between the operating characteristics of individual equipment and the overall system) and multiple time scales (such as instantaneous parameter fluctuations and long-term performance evolution), and the difficulty of traditional single models in simultaneously taking into account the consistency of physical laws and the ability to capture complex nonlinear and time-delay effects, this technical solution first constructs a hybrid dynamic model. This model is formed by fusing a mechanistic model and a time-series prediction model. By leveraging the ability of the mechanistic model to characterize the core physical processes of the system, and the ability of the time-series prediction model to capture complex dynamic characteristics that are difficult to model mechanistically, a comprehensive state estimation of the thermal power auxiliary system across multiple physical and time scales can be achieved, providing accurate system state basis for subsequent optimized control. As one implementation method, the mechanistic model can be established based on physical conservation laws to describe the balance relationship of system mass, energy, etc., while the time-series prediction model can be a model with time-series data processing capabilities. The fusion of the two can be achieved through filtering algorithms to combine the advantages of both and improve model accuracy.

[0027] By using a hybrid dynamic model that integrates mechanistic and temporal prediction models to achieve multi-scale state estimation, the physical rationality of the state estimation results is ensured, and the complex dynamic characteristics of the system are accurately captured. This effectively improves the accuracy and reliability of the state estimation of thermal power auxiliary equipment systems, laying a solid foundation for the formulation and execution of subsequent collaborative optimization control strategies.

[0028] Step 102: Based on the state estimation results, the key operating parameters in the hybrid dynamic model are adaptively identified and corrected online to track the performance changes of auxiliary equipment.

[0029] In the embodiments of this disclosure, to address the technical problem that auxiliary equipment in thermal power plants is prone to performance degradation, wear, or ash accumulation during long-term operation, causing the key operating parameters initially set in the hybrid dynamic model to gradually deviate from the actual operating characteristics of the equipment, thereby affecting the accuracy of the model's description of the system state, this technical solution, based on the aforementioned multi-physical-scale and multi-time-scale state estimation results, adopts an online adaptive approach to identify and correct key operating parameters in the hybrid dynamic model. By tracking parameter changes in real time to match the performance evolution of the auxiliary equipment, it ensures that the hybrid dynamic model can always accurately reflect the actual operating state of the system. As one implementation method, the key operating parameters to be identified can be determined first, and the equipment input and output data can be continuously monitored. A parameter identification algorithm suitable for linear or nonlinear systems can be used to achieve online parameter updates. A parameter library can also be established to store historical identification results to help predict the future performance degradation trend of the equipment.

[0030] By online adaptively identifying and correcting key operating parameters of the hybrid dynamic model, the problem of disconnection between the model parameters and the actual performance of the equipment is effectively avoided, ensuring the long-term effectiveness and accuracy of the hybrid dynamic model and providing reliable model support for the subsequent generation of precise collaborative optimization control strategies.

[0031] Step 103: Using the modified hybrid dynamic model and a multi-objective dynamic optimization algorithm, a collaborative optimization control strategy for the thermal power auxiliary equipment system is generated under the condition of satisfying preset operating constraints.

[0032] In the embodiments of this disclosure, to address the technical problem that the optimization process of thermal power auxiliary equipment systems needs to simultaneously meet multiple operational objectives (such as energy saving, stability, and environmental protection), and that traditional optimization methods are prone to generating control strategies that fail to meet the requirements of multiple objectives or do not fully consider operational constraints due to the model not matching the actual system state in real time or not fully considering operational constraints, this technical solution is based on a hybrid dynamic model that has been corrected online and can accurately reflect the current performance of auxiliary equipment and the actual state of the system. It adopts a multi-objective dynamic optimization algorithm that can coordinate the priorities and coupling relationships of multiple objectives. Under the preset operational constraints such as ensuring equipment safety, process compliance, and environmental compliance, it calculates and generates a collaborative optimization control strategy that can coordinate the operating states of each auxiliary equipment, ensuring that the strategy meets the overall optimization objectives of the system without exceeding the operational boundaries. As one implementation method, the multiple objectives may include minimizing the total energy consumption of the auxiliary equipment system and ensuring operational stability. The optimization algorithm may adopt an algorithm based on the principle of model predictive control, and the preset operational constraints may cover equipment physical limitations, environmental emission standard limitations, etc.

[0033] Based on the corrected and accurate hybrid dynamic model and multi-objective dynamic optimization algorithm, combined with preset operating constraints, the generated collaborative optimization control strategy can effectively balance multiple operating objectives of the system, while avoiding exceeding the safety and compliance boundaries, significantly improving the scientificity and practicality of the optimization control of thermal power auxiliary equipment systems.

[0034] Step 104: The collaborative optimization control strategy is transformed into executable control commands through a hierarchical coordination control mechanism and sent to the field controllers of each auxiliary equipment to achieve collaborative operation of the system.

[0035] In the embodiments of this disclosure, to address the technical problem that thermal power auxiliary systems contain various types of equipment with different functions, and that without a targeted instruction conversion and hierarchical coordination mechanism, the collaborative optimization control strategy is prone to mismatch between instructions and equipment control requirements, and asynchronous actions between equipment, making it difficult to effectively implement and thus preventing the overall collaborative operation of the system, this technical solution uses a hierarchical coordination control mechanism. First, the aforementioned generated collaborative optimization control strategy is converted into executable control instructions that meet the control logic and parameter requirements of each auxiliary equipment. Then, these instructions are precisely sent to the field controllers of the corresponding auxiliary equipment. By using a hierarchical architecture to divide and coordinate control tasks at different levels, it ensures that each auxiliary equipment operates synchronously according to a unified optimization goal, ultimately achieving collaborative operation of the system. As one implementation method, the hierarchical coordination control mechanism may include a higher-level coordination control responsible for the decomposition and allocation of global optimization goals, and a lower-level local control responsible for quickly tracking setpoints and handling local disturbances. Control instructions can be sent to the field controllers through an adapted communication protocol.

[0036] By implementing a hierarchical coordination control mechanism, the transformation and distribution of collaborative optimization control strategies are realized, effectively solving the problems of instruction adaptation and equipment coordination when the strategy is implemented. This ensures the consistency and accuracy of the actions of each auxiliary machine, provides a reliable guarantee for the overall collaborative operation of the thermal power auxiliary machine system, and improves the stability and optimization effect of the system operation.

[0037] The energy-saving optimization method for thermal power auxiliary equipment systems disclosed herein constructs a hybrid dynamic model that integrates a mechanistic model and a time-series prediction model to achieve state estimation of the thermal power auxiliary equipment system across multiple physical and time scales. Based on the state estimation results, key operating parameters in the hybrid dynamic model are adaptively identified and corrected online to track performance changes in the auxiliary equipment. Furthermore, using the corrected hybrid dynamic model and a multi-objective dynamic optimization algorithm, a collaborative optimization control strategy for the system is generated under preset operating constraints. This collaborative optimization control strategy is then transformed into executable control commands and distributed to the field controllers of each auxiliary equipment through a hierarchical coordinated control mechanism. Therefore, this method addresses the limitations of existing technologies in handling nonlinear coupling, significant time delays, and parameter evolution characteristics under dynamic operating conditions of thermal power auxiliary equipment systems. It also overcomes the difficulty in adaptively updating model parameters with equipment performance changes, leading to significant deviations between optimization results and actual operating effects, hindering the achievement of full-condition, adaptive, and collaborative energy-saving optimization. This method significantly improves the energy-saving effect of thermal power auxiliary equipment systems under dynamic operating conditions, enables adaptive and intelligent system operation, and enhances system stability, safety, and operational flexibility.

[0038] As a specific embodiment of this disclosure, based on the aforementioned basic scheme, the mechanism model is further defined to describe the mass balance, energy balance, momentum balance, and thermodynamic balance relationships of the thermal power auxiliary system through a set of differential equations; the time-series prediction model adopts a recurrent neural network structure to capture the nonlinear dynamic characteristics and time delay effects of the system; and the mechanism model and the time-series prediction model are fused through a state estimation algorithm to combine the physical consistency of the mechanism model with the nonlinear mapping capability of the data-driven model.

[0039] Specifically, in this embodiment, the mechanistic model uses a set of ordinary differential equations to characterize the mass balance, energy balance, momentum balance, and thermodynamic balance relationships of the thermal power auxiliary equipment system. For example, for the flue gas system, the equations can describe the momentum balance relationship between the blower outlet air pressure and air volume, and the thermodynamic balance relationship of heat exchange between flue gas and air in the air preheater. For the feedwater system, the equations can describe the mass balance relationship between the feedwater pump flow rate and the water volume of the water-cooled wall, and the energy balance relationship between the economizer heat absorption and the steam temperature. These ordinary differential equations are further discretized using the fourth-order Runge-Kutta method or the Euler method to adapt to subsequent numerical calculation requirements. The recurrent neural network structure used in the time series prediction model is specifically a Long Short-Term Memory (LSTM) network or a gated recurrent unit (GRU). The network contains 2-4 recurrent layers, with 64 cells per layer. The system employs 256 neurons, trained by inputting time-series data of historical operating parameters from the thermal power auxiliary equipment system (such as blower speed, furnace negative pressure, and main steam temperature). This allows the system to capture nonlinear dynamic characteristics (such as the nonlinear relationship between coal mill output and power consumption) and time-lag effects (such as the lag effect of induced draft fan adjustment on furnace negative pressure) that are difficult for mechanistic models to accurately describe. The state estimation algorithm used in model fusion is Kalman filtering or extended Kalman filtering. The mechanistic model outputs the predicted system state value, while the time-series prediction model calculates the prediction residual of the mechanistic model and provides a compensation signal. The algorithm obtains a preliminary state estimate by combining the outputs of both through a prediction update step, and then corrects the preliminary estimate using real-time operating data through a measurement update step. Ultimately, this achieves a combination of the physical consistency of the mechanistic model and the nonlinear mapping capability of the data-driven model.

[0040] Differential equations can accurately quantify the physical equilibrium relationship of the system, providing a reliable physical basis for the model; recurrent neural networks can effectively capture complex nonlinear and time-delay characteristics, making up for the modeling shortcomings of mechanistic models; state estimation algorithms achieve efficient integration of the two, significantly improving the accuracy of the hybrid dynamic model in describing the state of thermal power auxiliary equipment systems, and providing more accurate data support for subsequent state estimation.

[0041] As a specific implementation of this disclosure, based on the aforementioned basic scheme, the online adaptive identification and correction of key operating parameters in the hybrid dynamic model based on state estimation results is further defined, including: using a recursive estimation algorithm to identify the key operating parameters in real time, and realizing joint estimation of state and parameters in a nonlinear system; updating the identified parameters to the hybrid dynamic model in real time, and continuously verifying the model correction results; and establishing a parameter library based on historical identification results to predict the performance degradation trend of the auxiliary equipment.

[0042] Specifically, in this embodiment, the online adaptive identification and correction of key operating parameters of the hybrid dynamic model is performed according to the following process: First, for the identification of key operating parameters (such as fan efficiency coefficient, pipe resistance coefficient, heat exchanger fouling thermal resistance, etc.), a recursive estimation algorithm is used to achieve real-time processing. For linear systems where parameters change slowly over time, the recursive least squares (RLS) method is selected, and the forgetting factor is set to 0.95-0.999 to prioritize the use of the latest data. Parameter identification is achieved by iteratively calculating the gain vector, updating the parameter estimates and covariance matrix. For nonlinear systems, the extended Kalman filter (EKF) algorithm is used to incorporate the parameters to be identified into the system state vector. By performing local linearization processing on the nonlinear model, the state and parameters at the next time step are estimated in the prediction step, and the estimation results are corrected in the update step by combining real-time measurements to achieve joint estimation of state and parameters. Subsequently, the identified updated parameters (such as a 2% decrease in the fan efficiency coefficient) are injected in real time into the corresponding module of the hybrid dynamic model (such as the fan characteristic curve model). The model correction effect is continuously verified by calculating indicators such as mean squared error (MSE) and mean absolute percentage error (MAPE) through historical data backtesting and comparison with real-time operating data. If the prediction error exceeds a preset threshold (such as a furnace negative pressure prediction error greater than ±5Pa) within 10 consecutive sampling periods, model retraining or system diagnosis is triggered. Finally, a parameter library is established based on a time-series database (such as InfluxDB) to store timestamped historical identification results. By analyzing parameter change trends (such as the monthly decay rate of the fan efficiency coefficient), the future performance degradation rate of auxiliary equipment is predicted, providing data support for predictive maintenance of equipment.

[0043] By selectively employing recursive estimation algorithms, the real-time performance and accuracy of key parameter identification under different system characteristics are ensured; the model correction and verification mechanism guarantees the consistency between the hybrid dynamic model and the actual system; and the parameter library and performance degradation prediction provide forward-looking guidance for equipment maintenance, effectively improving the long-term reliability of the hybrid dynamic model and the efficiency of system operation and maintenance.

[0044] As a specific implementation of this disclosure, based on the aforementioned basic scheme, the method further specifies the use of a multi-objective dynamic optimization algorithm to generate a collaborative optimization control strategy for the thermal power auxiliary system under preset operating constraints. This includes: constructing a multi-objective optimization function, which includes a total system energy consumption term, an operational stability penalty term, a pollutant emission penalty term, and an operational flexibility penalty term; setting operational constraints, which include equipment physical limitations, process parameter limitations, safety protection limitations, environmental emission standard limitations, and control variable change rate limitations; and employing a rolling optimization and feedback correction mechanism to solve the multi-objective optimization function in the prediction time domain, generating an optimal control sequence, and generating the collaborative optimization control strategy based on the optimal control sequence.

[0045] Specifically, in this embodiment, the process of generating a collaborative optimization control strategy using a multi-objective dynamic optimization algorithm is as follows: First, a multi-objective optimization function is constructed, where the total system energy consumption term is the sum of the power consumption of all auxiliary equipment such as blowers, induced draft fans, and feedwater pumps within the prediction time domain, and needs to be calculated in conjunction with the real-time operating parameters of each auxiliary equipment (such as speed, flow rate, and efficiency); the operational stability penalty term is determined by quantifying the degree or fluctuation range of key process parameters such as furnace negative pressure and main steam pressure deviating from the set value (e.g., when the furnace negative pressure fluctuation exceeds ±5Pa, it is included in the penalty); the pollutant emission penalty term targets SO2 and NO. x If the emission concentration exceeds the preset environmental protection standard limit, a corresponding penalty will be imposed. The operational flexibility penalty item is related to the unit's response speed to load changes and the frequency of control variable adjustments (e.g., a penalty is triggered when the fan speed change rate exceeds ±100 rpm / min), and the priorities of each target are coordinated through dynamically adjustable weight coefficients. Secondly, operational constraints are set. Equipment physical limitations include the inherent capacity boundaries of the equipment, such as fan speed (e.g., 20%-100% rated speed) and minimum water pump flow rate; process parameter limitations cover the process requirement range, such as furnace negative pressure (e.g., -45Pa to -15Pa) and main steam temperature (e.g., 535℃ to 545℃); safety protection limitations are used to prevent dangerous conditions such as boiler overheating and overpressure; environmental emission standard limitations must comply with national or local emission requirements for flue gas pollutants; and control variable change rate limitations (e.g., valve opening change rate not exceeding ±5% / s) are used to avoid excessive control actions. Finally, a rolling optimization and feedback correction mechanism is adopted. In each control cycle, the system state of the next Np sampling steps (prediction time domain) is predicted based on the modified hybrid dynamic model. The multi-objective optimization function is solved by algorithms such as interior point method or sequential quadratic programming to obtain the optimal control sequence of the next Nc control steps (control time domain). Only the first control variable is taken to generate a collaborative optimization control strategy. In the next cycle, the optimization is repeated by combining the latest state estimation results to achieve feedback correction.

[0046] By comprehensively covering the core operational requirements of the system through multi-objective optimization functions, the clear constraints ensure the safety and compliance of the control strategy. The rolling optimization and feedback correction mechanism can adapt to the dynamic changes of the system in real time, effectively improving the accuracy and robustness of the collaborative optimization control strategy, and ensuring that the thermal power auxiliary equipment system can balance energy saving, stability, environmental protection and flexible operation under dynamic conditions.

[0047] As a specific implementation of this disclosure, based on the aforementioned basic scheme, the hierarchical coordination control mechanism is further defined as follows: upper-level coordination control, used to decompose the global optimization target into the collaborative operation target of each auxiliary equipment subsystem; lower-level local control, implemented by the field control loop of each auxiliary equipment, used to track the set values ​​issued by the upper-level coordination control and suppress local disturbances; and online adaptive adjustment of the parameters of the lower-level local controller to adapt to the changes in the operating conditions of the thermal power auxiliary equipment system.

[0048] Specifically, in this embodiment, the hierarchical coordination control mechanism is implemented as follows: The upper-level coordination control is based on the global optimization goals of the thermal power auxiliary equipment system (such as minimizing total energy consumption and ensuring overall operational stability). By analyzing the operational coupling relationships and dynamic characteristics of each auxiliary subsystem (flue gas system, feedwater system, pulverizing system, etc.), the global goal is decomposed into the collaborative operational goals of each subsystem. For example, when the boiler load increases from 60% MCR to 80% MCR, the plant-wide energy-saving goal is decomposed into the forced draft fan airflow adjustment goal, the induced draft fan negative pressure control goal, and the feedwater pump flow rate adaptation goal, ensuring that the action direction of each subsystem is consistent with the global goal. The lower-level local control is specifically implemented by the field control loops inherent in each auxiliary equipment. These loops include PID controllers or ratio control loops, such as using a PID controller for the forced draft fan. The induced draft fan tracks the setpoint airflow from the upper control unit and maintains the setpoint negative pressure in the furnace via a PID controller. When local disturbances occur, such as air leakage in the duct or instantaneous fluctuations in feedwater flow, the lower-level local controller can respond quickly, adjusting the control output within milliseconds to seconds to suppress the disturbance and prevent the spread of local fluctuations from affecting the overall system. At the same time, the parameters of the lower-level local controller (such as the proportional gain Kp, integral time Ti, and derivative time Td of the PID controller) are adaptively adjusted online. When the system operating conditions change (such as the boiler load decreasing from high load to low load), the PID parameters are dynamically corrected according to the degree of deviation from the operating conditions through fuzzy logic, or the parameter set is switched according to key variables such as boiler load using gain scheduling, to ensure that the lower-level controller can maintain good control performance and accurately track the setpoint under different operating conditions.

[0049] The upper-level coordination control enables the reasonable decomposition of global objectives, ensuring the coordinated operation of various auxiliary subsystems; the lower-level local control can quickly suppress local disturbances and improve the system's anti-interference capability; the online adaptive adjustment of lower-level controller parameters adapts to changes in operating conditions and further improves control accuracy. The combination of these three ensures the efficient operation of the hierarchical coordination control mechanism, realizing the refined and coordinated operation of the thermal power auxiliary system.

[0050] For specific application scenarios, this application provides an optimized implementation method, which further includes: based on an auxiliary decision-making mechanism, when the multi-objective dynamic optimization algorithm is inefficient, providing a fast control suggestion, wherein the auxiliary decision-making mechanism is trained offline and deployed online through a machine learning algorithm, its reward function is associated with the multi-objective optimization function, and includes a penalty term for violating security constraints.

[0051] Specifically, in this embodiment, the implementation process of the auxiliary decision-making mechanism is as follows: First, machine learning algorithms suitable for continuous action spaces, such as Deep Deterministic Policy Gradient (DDPG), Soft Actor-Commentary (SAC), or Proximal Policy Optimization (PPO), are selected to construct the auxiliary decision-making model. During the offline training phase, the thermal power auxiliary equipment system and its modified hybrid dynamic model are used as the model training environment. The environment includes the current key operating parameters of the auxiliary equipment (such as the blower speed, induced draft fan pressure, and feedwater pump flow), the real-time boiler load, the disturbance prediction for the next hour (such as load change trends and ambient temperature fluctuations), and the currently identified online equipment parameters (such as the fan efficiency coefficient and heat exchanger fouling thermal resistance). The action space is defined as the control input increment of each auxiliary equipment (such as the change in blower speed ±50 rpm and the change in induced draft fan damper opening ±3%). The reward function is designed in conjunction with the multi-objective optimization function. The negative value of the calculated result of the multi-objective optimization function (including total energy consumption, stability, emissions, and flexibility terms) is used as the base reward. An additional penalty term is set for violating safety constraints—when the model's output control action causes the system state to exceed the safety boundary (e.g., furnace negative pressure below -50 Pa or above -5 Pa, main steam temperature exceeding 545℃), a huge penalty of 10 times the absolute value of the base reward is applied to ensure the model learns safe strategies. The model is trained using millions of steps of offline simulation data (covering load increases and decreases, coal quality fluctuations, etc.), with the training process accelerated by GPU, until the model's reward values ​​converge to a stable range under various typical operating conditions. During online deployment, the solution time of multi-objective dynamic optimization algorithms (such as MPC-based optimization algorithms) is monitored in real time. When the solution time exceeds a preset threshold (such as 30 seconds) or the algorithm has difficulty converging (it still fails to meet the accuracy requirements after 100 iterations), the auxiliary decision-making mechanism is automatically activated. The current system status data is read and input into the trained model. The model outputs rapid control suggestions within 1 second (such as adjusting the blower speed to 420 rpm and maintaining the induced draft fan damper opening at 65%), which participate in the system control as temporary control commands.

[0052] By constructing an auxiliary decision-making mechanism using an offline-trained machine learning model, the problem of insufficient solution efficiency of multi-objective dynamic optimization algorithms is effectively solved, ensuring that the system can still quickly obtain control commands under extreme or complex dynamic conditions. The setting of safety constraint penalty terms in the reward function avoids the auxiliary decision-making output of dangerous control actions, further ensuring the safety and stability of the thermal power auxiliary equipment system.

[0053] It should be noted that the embodiments of this disclosure may include multiple steps. For ease of description, these steps are numbered, but these numbers are not a limitation on the execution time slots or execution order between the steps; these steps can be implemented in any order, and the embodiments of this disclosure do not limit this.

[0054] Corresponding to the aforementioned energy-saving optimization method for thermal power auxiliary equipment systems, this disclosure also proposes an energy-saving optimization device for thermal power auxiliary equipment systems. Since the device embodiments of this disclosure correspond to the aforementioned method embodiments, details not disclosed in the device embodiments can be referred to the aforementioned method embodiments, and will not be repeated here.

[0055] Figure 2 This is a schematic diagram of the structure of an energy-saving optimization device for a thermal power auxiliary system provided in an embodiment of this disclosure, as shown below. Figure 2 As shown, it includes: The estimation unit 21 is used to perform state estimation of the thermal power auxiliary equipment system at multiple physical scales and multiple time scales based on a hybrid dynamic model, wherein the hybrid dynamic model is obtained by fusing a mechanistic model and a time-series prediction model. The correction unit 22 is used to perform online adaptive identification and correction of key operating parameters in the hybrid dynamic model based on the state estimation results, so as to track the performance changes of auxiliary equipment. The generation unit 23 is used to generate a collaborative optimization control strategy for the thermal power auxiliary equipment system by using the modified hybrid dynamic model and a multi-objective dynamic optimization algorithm, under the condition of satisfying the preset operating constraints. The conversion unit 24 is used to convert the collaborative optimization control strategy into executable control commands through a hierarchical coordination control mechanism, and send them to the field controllers of each auxiliary equipment to realize the collaborative operation of the system.

[0056] The energy-saving optimization device for thermal power auxiliary equipment systems disclosed herein constructs a hybrid dynamic model that integrates a mechanistic model and a time-series prediction model to achieve state estimation of the thermal power auxiliary equipment system across multiple physical and time scales. Based on this state estimation result, key operating parameters in the hybrid dynamic model are adaptively identified and corrected online to track performance changes in the auxiliary equipment. Furthermore, using the corrected hybrid dynamic model and a multi-objective dynamic optimization algorithm, a collaborative optimization control strategy for the system is generated under preset operating constraints. This collaborative optimization control strategy is then transformed into executable control commands and distributed to the field controllers of each auxiliary equipment through a hierarchical coordination control mechanism. Therefore, this device can solve the problems in existing technologies where the limitations in handling nonlinear coupling, significant time delays, and parameter evolution characteristics under dynamic operating conditions of thermal power auxiliary equipment systems, and the difficulty in adaptively updating model parameters with equipment performance changes, lead to significant deviations between optimization results and actual operating effects, making it difficult to achieve full-condition, adaptive, and collaborative energy-saving optimization. This device significantly improves the energy-saving effect of thermal power auxiliary equipment systems under dynamic operating conditions, achieves adaptive and intelligent system operation, and enhances system stability, safety, and operational flexibility.

[0057] Furthermore, in one possible implementation of this embodiment, the mechanism model describes the mass balance, energy balance, momentum balance, and thermodynamic balance relationships of the thermal power auxiliary equipment system through a set of differential equations; The time-series prediction model employs a recurrent neural network structure to capture the nonlinear dynamic characteristics and time-delay effects of the system. The mechanistic model and the time series prediction model are fused by a state estimation algorithm to combine the physical consistency of the mechanistic model with the nonlinear mapping capability of the data-driven model.

[0058] Furthermore, in one possible implementation of this embodiment, the correction unit 22 is also used for: A recursive estimation algorithm is used to identify the key operating parameters in real time, and to achieve joint estimation of state and parameters in a nonlinear system. The identified parameters are updated to the hybrid dynamic model in real time, and the model correction results are continuously verified. A parameter library is established based on historical identification results to predict the performance degradation trend of the auxiliary equipment.

[0059] Furthermore, in one possible implementation of this embodiment, the generation unit 23 is also used for: Construct a multi-objective optimization function, which includes a total system energy consumption term, an operational stability penalty term, a pollutant emission penalty term, and an operational flexibility penalty term; Set operational constraints, which include physical limitations of equipment, limitations of process parameters, limitations of safety protection, limitations of environmental emission standards, and limitations of the rate of change of control variables; A rolling optimization and feedback correction mechanism is adopted to solve the multi-objective optimization function in the prediction time domain, generate the optimal control sequence, and generate the cooperative optimization control strategy based on the optimal control sequence.

[0060] Furthermore, in one possible implementation of this embodiment, the hierarchical coordination control mechanism includes: The higher-level coordination control is used to decompose the global optimization goal into the collaborative operation goals of each auxiliary equipment subsystem; Lower-level local control is implemented by the field control loops of each auxiliary device, used to track the setpoints issued by the upper-level coordinated control and suppress local disturbances; and The parameters of the lower-level local controller are adaptively adjusted online to adapt to changes in the operating conditions of the thermal power auxiliary equipment system.

[0061] Furthermore, in one possible implementation of this embodiment, such as Figure 2 As shown, it also includes: The auxiliary decision-making unit 25 is used to provide rapid control suggestions based on the auxiliary decision-making mechanism when the multi-objective dynamic optimization algorithm is inefficient. The auxiliary decision-making mechanism is trained offline and deployed online through machine learning algorithms. Its reward function is associated with the multi-objective optimization function and includes a penalty term for violating safety constraints.

[0062] It should be noted that the foregoing explanation of the method embodiments also applies to the apparatus of this embodiment, and the principle is the same, so it is not limited in this embodiment.

[0063] According to embodiments of this disclosure, this disclosure also provides an electronic device, a readable storage medium, and a computer program product.

[0064] Figure 3 A schematic block diagram of an example electronic device 300 that can be used to implement embodiments of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0065] like Figure 3As shown, the electronic device 300 includes a computing unit 301, which can perform various appropriate actions and processes based on a computer program stored in ROM (Read-Only Memory) 302 or a computer program loaded from storage unit 308 into RAM (Random Access Memory) 303. The RAM 303 may also store various programs and data required for the operation of the electronic device 300. The computing unit 301, ROM 302, and RAM 303 are interconnected via a bus 304. An I / O (Input / Output) interface 305 is also connected to the bus 304.

[0066] Multiple components in electronic device 300 are connected to I / O interface 305, including: input unit 306, such as keyboard, mouse, etc.; output unit 307, such as various types of displays, speakers, etc.; storage unit 308, such as disk, optical disk, etc.; and communication unit 309, such as network card, modem, wireless transceiver, etc. Communication unit 309 allows electronic device 300 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0067] The computing unit 301 can be various general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 301 include, but are not limited to, CPUs (Central Processing Units), GPUs (Graphics Processing Units), various special-purpose AI (Artificial Intelligence) computing chips, various computing units running machine learning model algorithms, DSPs (Digital Signal Processors), and any suitable processor, controller, microcontroller, etc. The computing unit 301 performs the various methods and processes described above, such as the energy-saving optimization method for thermal power auxiliary equipment systems. For example, in some embodiments, the energy-saving optimization method for thermal power auxiliary equipment systems can be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as storage unit 308. In some embodiments, part or all of the computer program can be loaded and / or installed on the electronic device 300 via ROM 302 and / or communication unit 309. When the computer program is loaded into RAM 303 and executed by the computing unit 301, one or more steps of the methods described above can be performed. Alternatively, in other embodiments, the computing unit 301 may be configured to perform the aforementioned energy-saving optimization method for thermal power auxiliary equipment systems by any other suitable means (e.g., by means of firmware).

[0068] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, FPGAs (Field Programmable Gate Arrays), ASICs (Application-Specific Integrated Circuits), ASSPs (Application-Specific Standard Products), SOCs (System-on-Chips), CPLDs (Complex Programmable Logic Devices), computer hardware, firmware, software, and / or combinations thereof. These various implementations may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0069] The program code used to implement the methods of this disclosure may be written in any combination of one or more programming languages. This program code may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code may be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0070] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, RAM, ROM, EPROM (Electrically Programmable Read-Only Memory) or flash memory, optical fiber, CD-ROM (Compact Disc Read-Only Memory), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0071] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a CRT (Cathode-Ray Tube) or LCD (Liquid Crystal Display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0072] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or computing systems that include middleware components (e.g., application servers), or computing systems that include frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include LANs (Local Area Networks), WANs (Wide Area Networks), the Internet, and blockchain networks.

[0073] Computer systems can include clients and servers. Clients and servers are generally geographically separated and typically interact via communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. A server can be a cloud server, also known as a cloud computing server or cloud host, a hosting product within the cloud computing service system that addresses the shortcomings of traditional physical hosts and VPS (Virtual Private Server) services, such as high management difficulty and weak business scalability. Servers can also be servers for distributed systems or servers incorporating blockchain technology.

[0074] It's important to note that artificial intelligence (AI) is the study of enabling computers to simulate certain human thought processes and intelligent behaviors (such as learning, reasoning, thinking, and planning). It encompasses both hardware and software technologies. AI hardware technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, and big data processing. AI software technologies primarily include computer vision, speech recognition, natural language processing, machine learning / deep learning, big data processing, and knowledge graph technologies.

[0075] The various numerical designations such as "first," "second," etc., used in this disclosure are merely for ease of description and are not intended to limit the scope of the embodiments of this disclosure, nor do they indicate a sequential order.

[0076] At least one of the features described in this disclosure can also be described as one or more, and multiple features can be two, three, four or more, and this disclosure does not impose any limitations. In the embodiments of this disclosure, for a technical feature, the technical features in that technical feature are distinguished by "first", "second", "third", "A", "B", "C" and "D", etc., and there is no sequential order or size order among the technical features described by "first", "second", "third", "A", "B", "C" and "D".

[0077] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this disclosure can be achieved, and this is not limited herein.

[0078] The specific embodiments described above do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. A method for energy-saving optimization of a thermal power auxiliary equipment system, characterized in that, include: Based on a hybrid dynamic model, the state estimation of thermal power auxiliary equipment system is performed at multiple physical and time scales. The hybrid dynamic model is obtained by fusing a mechanistic model and a time-series prediction model. Based on the state estimation results, key operating parameters in the hybrid dynamic model are adaptively identified and corrected online to track performance changes of auxiliary equipment. Using the modified hybrid dynamic model and a multi-objective dynamic optimization algorithm, a collaborative optimization control strategy for the thermal power auxiliary equipment system is generated under the condition of satisfying preset operating constraints. The hierarchical coordination control mechanism transforms the collaborative optimization control strategy into executable control commands, which are then sent to the field controllers of each auxiliary device to achieve collaborative operation of the system.

2. The method according to claim 1, characterized in that, The mechanistic model describes the mass balance, energy balance, momentum balance, and thermodynamic balance relationships of the thermal power auxiliary equipment system through a set of differential equations. The time-series prediction model employs a recurrent neural network structure to capture the nonlinear dynamic characteristics and time-delay effects of the system. The mechanistic model and the time series prediction model are fused by a state estimation algorithm to combine the physical consistency of the mechanistic model with the nonlinear mapping capability of the data-driven model.

3. The method according to claim 2, characterized in that, The online adaptive identification and correction of key operating parameters in the hybrid dynamic model based on the state estimation results includes: A recursive estimation algorithm is used to identify the key operating parameters in real time, and to achieve joint estimation of state and parameters in a nonlinear system. The identified parameters are updated to the hybrid dynamic model in real time, and the model correction results are continuously verified. A parameter library is established based on historical identification results to predict the performance degradation trend of the auxiliary equipment.

4. The method according to claim 1, characterized in that, The method employs a multi-objective dynamic optimization algorithm to generate a collaborative optimization control strategy for the thermal power auxiliary equipment system under preset operational constraints, including: Construct a multi-objective optimization function, which includes a total system energy consumption term, an operational stability penalty term, a pollutant emission penalty term, and an operational flexibility penalty term; Set operational constraints, which include physical limitations of equipment, limitations of process parameters, limitations of safety protection, limitations of environmental emission standards, and limitations of the rate of change of control variables; A rolling optimization and feedback correction mechanism is adopted to solve the multi-objective optimization function in the prediction time domain, generate the optimal control sequence, and generate the cooperative optimization control strategy based on the optimal control sequence.

5. The method according to claim 1, characterized in that, The hierarchical coordination and control mechanism includes: The higher-level coordination control is used to decompose the global optimization goal into the collaborative operation goals of each auxiliary equipment subsystem; Lower-level local control is implemented by the field control loops of each auxiliary device, used to track the setpoints issued by the upper-level coordinated control and suppress local disturbances; and The parameters of the lower-level local controller are adaptively adjusted online to adapt to changes in the operating conditions of the thermal power auxiliary equipment system.

6. The method according to claim 1, characterized in that, Also includes: Based on the auxiliary decision-making mechanism, when the multi-objective dynamic optimization algorithm is inefficient, a fast control suggestion is provided. The auxiliary decision-making mechanism is trained offline and deployed online through machine learning algorithms. Its reward function is associated with the multi-objective optimization function and includes a penalty term for violating security constraints.

7. An energy-saving optimization device for a thermal power auxiliary equipment system, characterized in that, include: An estimation unit is used to perform state estimation of a thermal power auxiliary system at multiple physical and time scales based on a hybrid dynamic model, wherein the hybrid dynamic model is obtained by fusing a mechanistic model and a time-series prediction model. The correction unit is used to perform online adaptive identification and correction of key operating parameters in the hybrid dynamic model based on the state estimation results, so as to track the performance changes of auxiliary equipment. The generation unit is used to generate a collaborative optimization control strategy for the thermal power auxiliary equipment system by using the modified hybrid dynamic model and a multi-objective dynamic optimization algorithm, under the condition of satisfying preset operating constraints. The conversion unit is used to convert the collaborative optimization control strategy into executable control commands through a hierarchical coordination control mechanism, and then send them to the field controllers of each auxiliary equipment to realize the collaborative operation of the system.

8. An electronic device, characterized in that, include: At least one processor; as well as A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-6.

9. A non-transitory computer-readable storage medium storing computer instructions, characterized in that, The computer instructions are used to cause the computer to perform the method according to any one of claims 1-6.

10. A computer program product, characterized in that, Includes a computer program that, when executed by a processor, implements the method according to any one of claims 1-6.

Images