Heat management and service life prediction method and equipment for waste heat recovery equipment

Abstract

The invention discloses a heat management and service life prediction method and equipment for waste heat recovery equipment. The method comprises the following steps: firstly, collecting related data of the waste heat recovery equipment, integrating historical maintenance data, fault data and material attribute data of the equipment to form a multi-dimensional data set, and based on the data set, fusing a multi-physics theory and a simulation technology to construct an equipment digital twinborn body; training by utilizing actual operation data and fault data of equipment and simulation data generated by a digital twinborn body to obtain a life prediction model; then, under the safety constraint of the equipment structure, determining optimized operation parameters based on a target optimization function which aims at prolonging the residual life of the equipment and improving the waste heat recovery efficiency; inputting the optimized operation parameters into digital twinborn simulation to determine corresponding waste heat recovery efficiency, and predicting the residual life of corresponding equipment based on a simulation result through a life prediction model; and if the waste heat recovery efficiency and the residual life of the equipment corresponding to the optimized operation parameters are superior to those of the current state, the equipment is controlled to operate according to the optimized operation parameters.

Description

Technical Field

[0001] This application relates to the field of thermal power generation technology, and in particular to a method and equipment for thermal management and life prediction of waste heat recovery equipment. Background Technology

[0002] In the process of industrial energy conservation and intelligent transformation, waste heat recovery equipment (such as waste heat boilers and heat exchangers) is widely used in chemical, energy, and metallurgical fields, and is a core piece of equipment for improving energy utilization efficiency. These devices operate under complex conditions of high temperature, high pressure, and corrosive media for extended periods. Their thermal management efficiency and service life directly affect production efficiency and operational safety; therefore, related optimization and prediction have become key requirements for industrial management.

[0003] Currently, existing technologies mainly collect equipment operation data and historical maintenance records through sensors and industrial control systems, and use purely data-driven machine learning models to estimate the remaining lifespan of the equipment; thermal management optimization is mostly aimed at maximizing recycling efficiency, and operating parameters are adjusted through experience or simple algorithms.

[0004] However, purely data-driven prediction models lack physical mechanism support, are sensitive to data noise, have large prediction biases, and cannot trace the root cause of failures. Furthermore, in existing technologies, thermal management and life prediction are independent of each other, without considering structural safety and material loss constraints, and pursuing efficiency under high loads can easily accelerate equipment degradation. Summary of the Invention

[0005] To address the aforementioned issues, this application provides a method and equipment for thermal management and lifespan prediction of waste heat recovery equipment.

[0006] The embodiments of this application disclose the following technical solutions: In a first aspect, embodiments of this application provide a method for thermal management and lifespan prediction of a waste heat recovery device, the method comprising: Collect operational status data, environmental data, and structural status data of the waste heat recovery equipment, and integrate historical maintenance data, fault data, and material property data of the waste heat recovery equipment to form a multi-dimensional dataset; Based on multi-dimensional datasets, a digital twin of the waste heat recovery equipment is constructed by integrating multiphysics theory and simulation technology; A life prediction model is trained based on actual operating data, fault data, and simulation data generated by a digital twin of the waste heat recovery equipment. The actual operating data, fault data, and simulation data are all data from the same time period. The output of the life prediction model includes the remaining life of the equipment. Under the structural safety constraints of the waste heat recovery equipment, the optimal operating parameters of the waste heat recovery equipment are determined based on the objective optimization function; wherein, the objective optimization function is constructed under the structural safety constraints with the goal of improving the remaining service life of the equipment and the waste heat recovery efficiency; The optimized operating parameters are input into the digital twin for simulation to determine the waste heat recovery efficiency corresponding to the optimized operating parameters; and the remaining equipment lifespan corresponding to the optimized operating parameters is predicted based on the simulation results of the digital twin using the life prediction model. If the waste heat recovery efficiency and remaining equipment lifespan corresponding to the optimized operating parameters are both better than those of the current waste heat recovery equipment, then the waste heat recovery equipment should be controlled to operate according to the optimized operating parameters.

[0007] In one possible implementation, operational status data, environmental data, and structural status data of the waste heat recovery equipment are collected, including: Sensors are deployed in key parts of the waste heat recovery equipment to collect operational and structural status data. Key parts refer to areas that affect the waste heat recovery efficiency, structural safety, or operational stability of the waste heat recovery equipment. Key parts include at least one of the following: the heat-receiving surface, the medium flow channel, the shell, the flue gas inlet and outlet end faces, and the heat exchange actuator. Environmental data is acquired through the industrial control system interface; where environmental data refers to external correlation data that affects the waste heat recovery thermal efficiency, structural corrosion, or heat load balance of the waste heat recovery equipment.

[0008] In one possible implementation, based on the multi-dimensional dataset, a digital twin of the waste heat recovery equipment is constructed by integrating multiphysics theory and simulation technology, including: Based on the three-dimensional structural model of the waste heat recovery equipment, and by integrating multi-physics theory and simulation technology with multi-dimensional datasets, a digital twin is constructed through multi-field coupling simulation. Among them, the multiphysics theory includes at least one of heat transfer, fluid mechanics and materials mechanics, and the simulation technology includes at least one of finite element simulation technology and computational fluid dynamics technology; the digital twin is used to map the internal physical field distribution and performance degradation process of the waste heat recovery equipment in real time based on the multidimensional dataset; the internal physical field distribution includes at least one of temperature field, flow field and stress field, and the performance degradation process includes at least one of scaling, corrosion and creep damage.

[0009] In one possible implementation, after constructing the digital twin, the above method further includes: Select the actual operating parameters of the waste heat recovery equipment during a preset time period, input the actual operating parameters into the digital twin for simulation, and obtain the simulation data of the digital twin; If the deviation between the simulation data and the actual operating data (operating status data and structural status data) is greater than or equal to a preset threshold, adjust the thermal conductivity coefficient, flow field resistance coefficient, and material fatigue parameters of the digital twin. Repeated calibration is performed until the deviation between the simulation data and the actual operating data is less than a preset threshold, so that the simulation state of the digital twin is close to that of the waste heat recovery equipment.

[0010] In one possible implementation, a lifespan prediction model is trained based on actual operating data of the waste heat recovery equipment, fault data of the waste heat recovery equipment, and simulation data generated by a digital twin, including: The actual operating data, fault data and simulation data within the same time period are divided into training set, validation set and test set according to a preset ratio; A model framework for lifespan prediction is constructed by combining long short-term memory networks with attention mechanisms; Using medium temperature, pressure, flow rate, structural stress, and scale thickness as input features, and the remaining lifespan of the waste heat recovery equipment as the output label, the model is iteratively trained through backpropagation algorithm until the prediction error of the model is less than the model prediction error standard, thus obtaining the lifespan prediction model.

[0011] In one possible implementation, the objective optimization function is a weighted summation function with the dual objectives of maximizing the remaining equipment lifespan and maximizing waste heat recovery efficiency; The objective optimization function is expressed as f(X) = α·L(X) + β·E(X); Where X is the operating parameter to be optimized; α and β are weighting coefficients, α+β=1, and α, β∈(0,1); L(X) is the equipment remaining life function corresponding to the operating parameter X to be optimized, and E(X) is the waste heat recovery efficiency function corresponding to the operating parameter X to be optimized; The structural safety constraint is expressed as σ(X)≤σ0; where σ(X) is the structural stress of the equipment corresponding to the operating parameter X to be optimized, and σ0 is the preset safety threshold; the preset safety threshold is determined based on at least one of the following: equipment material properties, design standards, and industry safety specifications.

[0012] In one possible implementation, the above method also includes: Every preset period, the life prediction model is updated using the actual operating data of the waste heat recovery equipment, the fault data of the waste heat recovery equipment, and the simulation data of the digital twin within the preset period.

[0013] Secondly, embodiments of this application disclose a thermal management and lifespan prediction device for waste heat recovery equipment, the device comprising: The data acquisition module is used to collect operating status data, environmental data, and structural status data of the waste heat recovery equipment, and integrate the equipment's historical maintenance data, fault data, and material property data to form a multi-dimensional dataset. The building module is used to construct a digital twin of waste heat recovery equipment based on multi-dimensional datasets and by integrating multi-physics theory and simulation technology. The training module is used to train a life prediction model based on actual operating data, fault data, and simulation data generated from a digital twin of the waste heat recovery equipment. The actual operating data, fault data, and simulation data are data from the same time period. The output of the life prediction model includes the remaining life of the equipment. The optimization module is used to determine the optimal operating parameters of the waste heat recovery equipment based on the objective optimization function; wherein, the objective optimization function is constructed under the constraint of the structural safety of the waste heat recovery equipment with the goal of improving the remaining life of the equipment and the waste heat recovery efficiency; The simulation module is used to input optimized operating parameters into the digital twin for simulation, determine the waste heat recovery efficiency corresponding to the optimized operating parameters, and use the life prediction model to predict the remaining life of the equipment corresponding to the optimized operating parameters based on the simulation results of the digital twin. The operation module is used to optimize the waste heat recovery efficiency and remaining equipment lifespan of the operating parameters to be better than the current waste heat recovery equipment, and to control the waste heat recovery equipment to operate according to the optimized operating parameters.

[0014] In one possible implementation, the acquisition module is specifically used to deploy sensors at key locations of the waste heat recovery equipment to collect operational and structural status data. Key locations refer to areas affecting the waste heat recovery efficiency, structural safety, or operational stability of the waste heat recovery equipment, and include at least one of the following: the equipment's heating surface, medium flow channel, shell, flue gas inlet / outlet end face, and heat exchange actuator. Environmental data is acquired through an industrial control system interface. Environmental data refers to external correlation data affecting the waste heat recovery thermal efficiency, structural corrosion, or heat load balance of the waste heat recovery equipment.

[0015] In one possible implementation, the building module is specifically used to construct a digital twin based on a three-dimensional structural model of the waste heat recovery equipment, integrating multi-physics theory and simulation technology with a multi-dimensional dataset, through multi-field coupled simulation. The multi-physics theory includes at least one of heat transfer, fluid mechanics, and materials mechanics, and the simulation technology includes at least one of finite element simulation and computational fluid dynamics. The digital twin is used to map the internal physical field distribution and performance degradation process of the waste heat recovery equipment in real time based on the multi-dimensional dataset. The internal physical field distribution includes at least one of temperature field, flow field, and stress field, and the performance degradation process includes at least one of scaling, corrosion, and creep damage.

[0016] In one possible implementation, the construction module is further used to select the actual operating parameters of the waste heat recovery equipment during a preset time period, input the actual operating parameters into the digital twin for simulation, and obtain the simulation data of the digital twin; if the deviation between the simulation data and the corresponding operating state data and structural state data of the actual operating data is greater than or equal to a preset threshold, the thermal conductivity coefficient, flow field resistance coefficient and material fatigue parameters of the digital twin are adjusted, and the calibration is repeated until the deviation between the simulation data and the actual operating data is less than the preset threshold, so that the simulation state of the digital twin is close to that of the waste heat recovery equipment.

[0017] In one possible implementation, the training module is specifically used to divide the actual operating data, fault data, and simulation data within the same time period into training set, validation set, and test set according to a preset ratio; a long short-term memory network combined with an attention mechanism is used to construct the model framework of the life prediction model; with medium temperature, pressure, flow rate, structural stress, and scale thickness as input features, and the remaining life of the waste heat recovery equipment as the output label, the model is iteratively trained through the backpropagation algorithm until the prediction error of the model is less than the model prediction error standard, thus obtaining the life prediction model.

[0018] In one possible implementation, the objective optimization function is a weighted summation function with the dual objectives of maximizing the remaining equipment lifespan and maximizing waste heat recovery efficiency; The objective optimization function is expressed as f(X) = α·L(X) + β·E(X); Where X is the operating parameter to be optimized; α and β are weighting coefficients, α+β=1, and α, β∈(0,1); L(X) is the equipment remaining life function corresponding to the operating parameter X to be optimized, and E(X) is the waste heat recovery efficiency function corresponding to the operating parameter X to be optimized; The structural safety constraint is expressed as σ(X)≤σ0; where σ(X) is the structural stress of the equipment corresponding to the operating parameter X to be optimized, and σ0 is the preset safety threshold; the preset safety threshold is determined based on at least one of the following: equipment material properties, design standards, and industry safety specifications.

[0019] In one possible implementation, the building module is also used to update the life prediction model every preset period using the actual operating data of the waste heat recovery equipment, the fault data of the waste heat recovery equipment, and the simulation data of the digital twin within the preset period.

[0020] Thirdly, embodiments of this application disclose a control device, including a processor and a memory. The memory is used to store programs, instructions, or code, and the processor is used to execute the programs, instructions, or code in the memory to complete the thermal management and life prediction method of the waste heat recovery equipment as described in any of the first aspects.

[0021] Fourthly, embodiments of this application disclose a computer-readable storage medium, characterized in that it stores a computer program, which is loaded by a processor to execute the thermal management and life prediction method of the waste heat recovery equipment as described in any of the first aspects.

[0022] This application provides a method and device for thermal management and lifespan prediction of waste heat recovery equipment. The method first collects operational status data, environmental data, and structural status data of the waste heat recovery equipment, and integrates historical maintenance data, fault data, and material property data to form a multi-dimensional dataset. Based on this dataset, a digital twin of the equipment is constructed by fusing multiphysics theory and simulation technology. Then, a lifespan prediction model for the remaining lifespan of the equipment is trained using actual operational data, fault data, and simulation data generated by the digital twin within the same time period. Subsequently, under the structural safety constraints of the equipment, optimized operating parameters are determined based on a target optimization function aimed at improving the remaining lifespan of the equipment and the waste heat recovery efficiency. The optimized operating parameters are input into the digital twin simulation to determine the corresponding waste heat recovery efficiency, and the remaining lifespan of the corresponding equipment is predicted based on the simulation results using the lifespan prediction model. If both the waste heat recovery efficiency and the remaining lifespan of the equipment corresponding to the optimized operating parameters are better than the current state, the equipment is controlled to operate according to the optimized operating parameters.

[0023] This application embodiment integrates multi-dimensional data and constructs a digital twin. By combining actual operation, fault, and simulation data from the same time period to train a life prediction model, it effectively improves the accuracy of equipment remaining life prediction and avoids prediction bias caused by a single data source. An optimization function is constructed with structural safety as a constraint and improved equipment remaining life and waste heat recovery efficiency as dual objectives, achieving synergy between thermal management and life protection. This avoids the problem of simply pursuing efficiency while ignoring equipment wear and tear, and reduces the risk of equipment failure through structural safety constraints. After digital twin simulation verification and life prediction model evaluation, equipment operation is controlled only when the optimization effect is better than the current state, ensuring the effectiveness and reliability of the optimization scheme. Ultimately, a dynamic balance between the operating efficiency and service life of the waste heat recovery equipment is achieved, improving the safety and overall benefits of equipment operation. Attached Figure Description

[0024] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0025] Figure 1 A schematic flowchart illustrating a thermal management and lifespan prediction method for a waste heat recovery device provided in this application embodiment; Figure 2 This is a schematic diagram of the structure of a thermal management and life prediction device for a waste heat recovery equipment provided in an embodiment of this application. Detailed Implementation

[0026] As described earlier, the industry has developed some technical solutions for life prediction of waste heat recovery equipment. Purely data-driven life prediction models lack the support of physical mechanisms such as heat transfer, fluid mechanics, and materials mechanics. Relying solely on statistical data for prediction not only results in poor interpretability and difficulty in tracing the root cause of failures, but also makes them sensitive to sensor noise, missing data, and other issues. They are prone to prediction bias under complex operating conditions and cannot provide a reliable basis for equipment maintenance.

[0027] To address this technical problem, this application provides a method and device for thermal management and lifespan prediction of waste heat recovery equipment. The method first collects operational status data, environmental data, and structural status data of the waste heat recovery equipment, and integrates historical maintenance data, fault data, and material property data to form a multi-dimensional dataset. Based on this dataset, a digital twin of the equipment is constructed by fusing multiphysics theory and simulation technology. Then, a lifespan prediction model is trained using actual operational data, fault data, and simulation data generated by the digital twin within the same time period, outputting the remaining lifespan of the equipment. Subsequently, under the structural safety constraints of the equipment, optimized operating parameters are determined based on a target optimization function aimed at improving the remaining lifespan of the equipment and the waste heat recovery efficiency. The optimized operating parameters are input into the digital twin simulation to determine the corresponding waste heat recovery efficiency, and the lifespan prediction model predicts the corresponding remaining lifespan of the equipment based on the simulation results. If the waste heat recovery efficiency and remaining lifespan of the equipment corresponding to the optimized operating parameters are both better than the current state, the equipment is controlled to operate according to the optimized operating parameters.

[0028] This application embodiment integrates multi-dimensional data and constructs a digital twin. By combining actual operation, fault, and simulation data from the same time period to train a life prediction model, it effectively improves the accuracy of equipment remaining life prediction and avoids prediction bias caused by a single data source. An optimization function is constructed with structural safety as a constraint and improved equipment remaining life and waste heat recovery efficiency as dual objectives, achieving synergy between thermal management and life protection. This avoids the problem of simply pursuing efficiency while ignoring equipment wear and tear, and reduces the risk of equipment failure through structural safety constraints. After digital twin simulation verification and life prediction model evaluation, equipment operation is controlled only when the optimization effect is better than the current state, ensuring the effectiveness and reliability of the optimization scheme. Ultimately, a dynamic balance between the operating efficiency and service life of the waste heat recovery equipment is achieved, improving the safety and overall benefits of equipment operation.

[0029] The method provided in this application embodiment can be widely applied to core waste heat recovery equipment such as waste heat boilers and plate heat exchangers in chemical, metallurgical, and thermal power industries. Its hardware deployment and operating logic are deeply adapted to actual industrial conditions. Temperature sensors, strain gauges, electromagnetic flowmeters, and pressure sensors are deployed at key locations such as the equipment's heating surface, medium flow channel, flue gas inlet and outlet faces, and heat exchange actuators to collect real-time operating and structural status data. Simultaneously, it interfaces with the environmental monitoring module and the online flue gas monitoring system via the Digital Crossconnected System (DCS) interface or Programmable Logic Controller (PLC) interface in the industrial control system to obtain environmental data such as ambient temperature and flue gas composition. All data is aggregated to an edge server or cloud through standardized interfaces to form a multi-dimensional dataset. A digital twin is constructed based on a workstation equipped with multiphysics simulation software. A life prediction model is trained using an industrial server. After solving the objective optimization function through an intelligent optimization algorithm, the optimized operating parameters, such as water flow rate and flue gas guide plate angle, are sent to the PLC control system. This drives actuators such as variable frequency pumps, electric regulating valves, and flow guiding devices to precisely adjust the equipment's operating status. At the same time, the optimization effect is verified through real-time simulation using the digital twin. Ultimately, a dynamic balance between efficiency and lifespan is achieved for the equipment while ensuring structural safety.

[0030] Whether it's a flue gas waste heat recovery system for 300MW~1000MW thermal power units or a dust-laden flue gas heat exchange scenario in the chemical industry, this solution can adapt to complex operating conditions through hardware collaboration, effectively solving the pain point of difficulty in balancing efficiency and safety in industrial waste heat recovery equipment.

[0031] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present application.

[0032] See Figure 1 , Figure 1 This is a flowchart illustrating a method for thermal management and lifespan prediction of a waste heat recovery device provided in an embodiment of this application. The execution subject of this method can be a server, desktop computer, or other computing device capable of computation. The following description uses a computing device as the execution subject, and the method includes: S101: The computing device collects the operating status data, environmental data, and structural status data of the waste heat recovery equipment, and integrates the equipment's historical maintenance data, fault data, and material property data to form a multi-dimensional dataset.

[0033] In this embodiment, the computing device can be an industrial-grade edge server. The computing device can connect to various sensors deployed on the waste heat recovery equipment through wired communication protocols to obtain the operating status data and structural status data of the waste heat recovery equipment.

[0034] Meanwhile, the computing device can connect to the DCS interface or PLC interface through the OPC UA protocol to obtain environmental data, and retrieve historical maintenance data, fault data and material property data of the equipment through the industrial database interface.

[0035] After acquiring various types of raw data, the computing device can also preprocess the collected raw data. The computing device can use the 3σ criterion to remove abnormal data caused by sensor noise, use linear interpolation to complete missing data, perform smoothing and noise reduction processing on continuous data, and then classify and store the data according to timestamps and data types to form a standardized multi-dimensional dataset.

[0036] By collecting and integrating data from multiple channels and in all dimensions, the dataset is ensured to cover both dynamic parameters of the device's real-time operation and static data such as historical status and material characteristics. This provides comprehensive and accurate basic data support for the subsequent construction of the digital twin, avoiding distortion of the twin simulation due to missing or incomplete data.

[0037] This application embodiment achieves efficient collection of operational status data, structural status data, and environmental data through targeted hardware deployment and system integration. The specific implementation methods of various data collection methods are described in detail below with reference to specific implementation methods.

[0038] In one possible implementation, the computing device collects operational status data, environmental data, and structural status data of the waste heat recovery equipment, including: The computing device deploys sensors at key locations within the waste heat recovery equipment to collect operational and structural status data. The computing device also acquires environmental data via an industrial control system interface.

[0039] In this embodiment, critical components refer to areas that affect the waste heat recovery efficiency, structural safety, or operational stability of the waste heat recovery equipment. Critical components include at least one of the following: the heating surface, the medium flow channel, the shell, the flue gas inlet and outlet faces, and the heat exchange actuator.

[0040] Taking a flue-type waste heat boiler commonly used in the chemical industry as an example, the heating surface may include water-cooled wall tube bundles and superheater tube walls, the medium flow channel may include feedwater inlet pipe, flue gas flow channel and bend section flow channel, the shell may include boiler shell and flange connection, flue gas inlet and outlet end faces, and the heat exchange actuator may include flue gas guide plate and feedwater regulating valve installation position.

[0041] Sensors adapted to complex industrial operating conditions are strategically deployed in the aforementioned key areas. For example, fiber optic temperature sensors (measuring range -50℃ to 600℃) and strain gauges (measuring accuracy ±1MPa) are attached to the heated surfaces and shell surfaces to collect real-time structural data such as pipe wall temperature and shell thermal stress. Electromagnetic flow meters (measuring range 50m³ / h to 150m³ / h) and pressure transmitters (measuring range 0 to 2MPa) are installed in the media flow channels and at the inlet and outlet ends to capture operating status data such as feedwater flow rate and flue gas pressure. Vibration sensors (measuring range 0 to 10mm / s) are deployed near the heat exchange actuators to monitor the vibration amplitude of the mechanism during operation.

[0042] All sensors are connected to the computing device via shielded cables and transmitted to the computing device via Ethernet, ensuring the stability and real-time performance of data transmission.

[0043] This application embodiment obtains first-hand data from the core working area of ​​the equipment directly by precisely deploying points in key locations and selecting appropriate sensors, avoiding data interference from non-critical areas, while ensuring the synchronization of the acquisition of operational status data and structural status data.

[0044] In this embodiment, environmental data refers to externally related data that affects the waste heat recovery efficiency, structural corrosion, or heat load balance of the waste heat recovery equipment. Environmental data includes, for example, basic parameters of the environment in which the waste heat recovery equipment is located (such as ambient temperature and atmospheric pressure), flue gas-related parameters (such as flue gas composition and dust concentration), and environmental parameters affecting structural corrosion (such as atmospheric humidity).

[0045] The computing device can connect to the DCS system and the flue gas online monitoring system via the OPC UA protocol. The computing device can retrieve environmental data such as ambient temperature, atmospheric pressure, and atmospheric humidity from the DCS system. It can also obtain SO2 concentration and NO2 concentration at the flue gas inlet and outlet from the flue gas online monitoring system. X Concentration and dust concentration data are updated at a frequency consistent with the sensor acquisition frequency to ensure time synchronization between environmental data and operational and structural status data.

[0046] The computing device acquires environmental data through the industrial control system interface, avoiding the cost of deploying a large number of additional environmental sensors and relying on the mature monitoring capabilities of existing industrial systems to ensure the accuracy and continuity of environmental data. This supplementary data ensures that the multi-dimensional dataset not only covers the equipment's own condition but also incorporates external influencing factors. This provides crucial support for subsequent digital twin simulations of the impact of flue gas composition on scaling and corrosion, as well as heat load balance calculations, further enhancing the comprehensiveness of the dataset.

[0047] S102: The computing device is based on a multi-dimensional dataset and integrates multi-physics theory and simulation technology to construct a digital twin of the waste heat recovery equipment.

[0048] In this embodiment, the computing device first reconstructs the spatial layout of core components such as the heating surface, medium flow channel, and shell based on the structural data of the waste heat recovery equipment, thus obtaining a three-dimensional structural model of the waste heat recovery equipment. Subsequently, the computing device integrates heat transfer, fluid mechanics, and materials mechanics theories, combined with finite element method and computational fluid dynamics (CFD), to import the multi-dimensional dataset into the three-dimensional structural model of the waste heat recovery equipment, ultimately constructing a high-fidelity digital twin.

[0049] Digital twins can accurately map the distribution of temperature, flow, and stress fields inside equipment. They can also simulate performance degradation processes such as scaling and corrosion, quantify the impact of scaling and corrosion on heat exchange efficiency, and achieve state linkage between physical equipment and virtual models, providing a reliable virtual carrier for subsequent model training and parameter optimization.

[0050] The following section details the process of building a digital twin, focusing on specific implementation details: In one possible implementation, the computing device constructs a digital twin of the waste heat recovery equipment based on a multi-dimensional dataset, integrating multiphysics theory and simulation technology, including: The computing device is based on the three-dimensional structural model of the waste heat recovery equipment. It integrates multi-physics theory and simulation technology with multi-dimensional datasets and constructs a digital twin through multi-field coupling simulation.

[0051] In this embodiment, the multiphysics theory includes at least one of heat transfer, fluid mechanics, and materials mechanics. The simulation technology includes at least one of finite element simulation and computational fluid dynamics. The digital twin is used to map the internal physical field distribution and performance degradation process of the waste heat recovery equipment in real time based on a multi-dimensional dataset. The internal physical field distribution includes at least one of temperature field, flow field, and stress field. The performance degradation process includes at least one of scaling, corrosion, and creep damage.

[0052] The computing device can construct a 3D structural model of the waste heat recovery equipment based on its structural data. This 3D model can recreate the structural details and spatial positions of core components such as the boiler shell, water-cooled wall tube bundles, flue gas passages, baffles, and feedwater pipes. Simultaneously, it labels key parameters such as material properties (e.g., shell is made of Q345R steel, tube bundles are made of 20G seamless steel pipe), wall thickness, and interface dimensions of each component, ensuring complete consistency between the 3D structural model and the physical equipment in terms of structure, size, and material. For minor deformations that may occur after long-term operation, the 3D structural model can be fine-tuned using historical maintenance data to further improve the fit between the model and the physical equipment, providing a precise geometric basis for subsequent multi-field coupled simulations.

[0053] The computing device uses multi-dimensional datasets as simulation inputs for a multiphysics simulation platform, integrating three core physical field theories: heat transfer, fluid mechanics, and materials mechanics. It combines finite element simulation technology with CFD technology to carry out multi-field coupled simulation, achieving collaborative simulation and interrelation of different physical fields.

[0054] From the perspective of single-physics field simulation, the computing device, based on heat transfer theory and finite element technology, combines data such as feedwater temperature, flue gas temperature, and heat transfer efficiency to calculate the heat exchange coefficient between the pipe wall and the medium, simulate the temporal variation of the spatial distribution of the temperature field inside the boiler, and accurately capture the temperature gradient of key parts such as the heating surface and shell. Based on fluid mechanics theory and CFD technology, the computing device simulates the turbulent flow state of flue gas in the channel and the laminar flow characteristics of feedwater in the tube bundle, reconstructing the velocity distribution, pressure loss, and vortex dead zone locations of the flow field. Based on materials mechanics theory and finite element technology, the computing device combines data such as structural stress, material tensile strength, and fatigue life curves to analyze the stress field distribution of the equipment under the dual effects of temperature and pressure loads, and locate stress concentration areas.

[0055] From the perspective of multi-field coupling logic, the computing device associates three major physical fields through the coupling algorithm of the simulation platform to simulate the performance degradation process of scaling (quantifying the fouling deposition rate based on flue gas dust concentration and operating temperature data), corrosion (calculating the degree of material erosion based on flue gas composition and atmospheric humidity data), and creep damage (simulating the accumulation of material plastic deformation based on pipe wall temperature and stress data). By dynamically comparing real-time data with simulation data, parameters such as thermal conductivity coefficient and flow field resistance coefficient are adjusted to ensure the consistency of the state between the twin and the physical device.

[0056] This step constructs a digital twin by integrating multiphysics theory and simulation technology, breaking through the limitation of existing technologies that can only visualize the surface state of equipment, and realizing the virtual reconstruction of the internal physical field distribution and performance degradation process of the equipment. Based on the dual support of a three-dimensional structural model and a multi-dimensional dataset, the digital twin can accurately map the real-time state of the physical equipment, and the generated simulation data has high fidelity.

[0057] To further improve the simulation accuracy of the digital twin and ensure that its virtual state is highly consistent with the physical operating state of the waste heat recovery equipment, thus providing reliable support for subsequent life prediction model training and parameter optimization verification, this application also provides a dynamic calibration process for the digital twin. Specific implementation details are as follows: In one possible implementation, after constructing the digital twin, the computing device selects the actual operating parameters of the waste heat recovery device over a preset time period, and inputs the actual operating parameters into the digital twin for simulation to obtain the simulation data of the digital twin.

[0058] If the deviation between the simulation data and the actual operating data (operating state data and structural state data) is greater than or equal to a preset threshold, the computing device adjusts the thermal conductivity coefficient, flow field resistance coefficient, and material fatigue parameters of the digital twin. This calibration is repeated until the deviation between the simulation data and the actual operating data is less than the preset threshold, making the simulation state of the digital twin approximate that of the waste heat recovery device.

[0059] For ease of explanation, this application uses a preset time period of 24 consecutive hours as an example. This preset time period covers both rated load operation and scenarios with small load fluctuations, comprehensively reflecting the parameter change patterns under normal operating conditions and avoiding incomplete calibration due to a single operating condition.

[0060] The computing device extracts actual operating parameters of the waste heat recovery equipment within a preset time period from a multi-dimensional dataset, categorized by timestamp. These parameters include feedwater flow rate, feedwater temperature, flue gas inlet and outlet temperatures, flue gas pressure, and pipe wall thermal stress. The computing device synchronously inputs this set of actual operating parameters into a pre-constructed digital twin, controlling the twin to initiate simulations according to the actual operating sequence of the physical equipment. This simulates the equipment's operation under identical conditions and outputs simulation data for the corresponding time period. This simulation data includes the digital twin's operational status data and structural status data as it operates according to the actual operating parameters.

[0061] The computing device employs a mean squared error algorithm to calculate the deviation between the simulated operational status data and the actual operational status data dimension by dimension, and also calculates the deviation between the simulated structural status data and the actual structural status data. If both calculated deviations are less than a preset threshold, it indicates that the operational status of the digital twin is very close to that of the waste heat recovery equipment, and the digital twin can be used to simulate the operation of the waste heat recovery equipment. If either deviation is greater than or equal to the preset threshold, it indicates that there is a significant deviation between the operational status of the digital twin and the physical equipment, and the simulation parameters of the digital twin need to be adjusted accordingly.

[0062] When adjusting the simulation parameters of a digital twin, the key parameters to adjust are the thermal conductivity coefficient, flow field drag coefficient, and material fatigue parameters. Among these, the thermal conductivity coefficient primarily affects the accuracy of the temperature field simulation; if the temperature dimension deviation is too large, it can be adjusted by ±0.02 W / (m²). Fine-tune the step size of K until the deviation between the simulated and actual temperature values ​​converges. The flow field drag coefficient is related to the simulation results of flow field velocity and pressure. If the deviation of the flow field or pressure exceeds the standard, it can be adjusted in a step size gradient of ±0.05 to correct the deviation of the medium flow simulation. Material fatigue parameters directly affect the simulation effect of stress field and performance degradation. If the stress value deviation is too large, the derived parameters such as tensile strength and fatigue limit of the material can be corrected by ±3% to ensure that the structural stress simulation is consistent with reality.

[0063] After each parameter adjustment, the actual operating parameters for the same preset time period are re-input into the twin, the simulation is repeated, and the deviation is calculated to form a closed-loop iteration.

[0064] When the iterative calibration is completed until all deviations are less than the preset threshold, the simulation parameter adjustment is stopped. At this point, the simulation state of the digital twin is close to the actual operating state of the waste heat recovery equipment. It can not only replicate the real distribution of the internal temperature field, flow field and stress field of the equipment, but also the simulated performance degradation processes such as scaling, corrosion and creep damage are highly consistent with the wear law of the physical equipment.

[0065] To address issues such as material aging and operating condition drift during long-term operation, the computing equipment can also undergo periodic calibration. For example, every 72 hours, the computing equipment automatically triggers a calibration process, selecting the latest 24-hour actual operating parameters to repeat the above calibration steps, dynamically correcting the twin simulation parameters, and continuously maintaining high fidelity.

[0066] S103: The computing device trains a life prediction model based on the actual operating data of the waste heat recovery equipment, the fault data of the waste heat recovery equipment, and the simulation data generated by the digital twin.

[0067] Existing purely data-driven models lack the support of physical mechanisms. The embodiments of this application integrate actual operation and fault data and twin simulation data within the same time period, allowing the model to simultaneously take into account both data statistical patterns and equipment physical operation mechanisms, thereby improving the accuracy and reliability of remaining life prediction.

[0068] In this embodiment, the computing device first selects actual operating data, fault data, and simulation data from the multi-dimensional dataset and twin simulation results for the same time period. Then, it performs synchronous preprocessing on the three types of data, removing outliers and filling in missing values, before dividing them into training and validation sets according to a preset ratio.

[0069] The computing device can use a Long Short-Term Memory (LSTM) network to build the basic framework of the model. Combined with historical operation and maintenance records, the actual remaining lifespan of the waste heat recovery equipment is calculated. Preprocessed data of three types is used as input features, and the actual remaining lifespan of the equipment is used as the output label. The model is iteratively trained using the backpropagation algorithm, while the prediction accuracy is verified using a validation set, until the model prediction error is below a preset threshold, completing the lifespan prediction model training. The output of the lifespan prediction model includes the remaining lifespan of the equipment.

[0070] To further enhance the lifetime prediction model's ability to capture key features, improve the accuracy and generalization ability of equipment remaining lifetime prediction, and make the lifetime prediction model more closely match the time-series operating characteristics and performance degradation patterns of waste heat recovery equipment, this application embodiment further refines the training process of the lifetime prediction model. The specific implementation method is as follows: In one possible implementation, the computing device divides actual operating data, fault data, and simulation data within the same time period into training, validation, and test sets according to a preset ratio. The computing device uses a Long Short-Term Memory (LSTM) network combined with an attention mechanism to construct the model framework of the life prediction model. The computing device uses medium temperature, pressure, flow rate, structural stress, and scale thickness as input features, and the remaining lifespan of the waste heat recovery equipment as the output label. Iterative training is performed using a backpropagation algorithm until the model's prediction error is less than the model prediction error standard, thus obtaining the life prediction model.

[0071] In this embodiment, the computing device first extracts actual operating data, fault data, and simulation data within the same time period to ensure data time alignment and operating condition matching. Then, the integrated data is divided into a training set, a validation set, and a test set according to a preset ratio. The training set is used for iterative learning of model parameters, the validation set is used for real-time adjustment of model hyperparameters to avoid overfitting, and the test set is used for final evaluation of the model's generalization ability. The clear division of labor among the three sets ensures the scientific rigor and reliability of model training. Preprocessing is still required before data partitioning, using the 3σ criterion to remove outliers and linear interpolation to fill in missing values, ensuring the completeness and accuracy of the input data.

[0072] This application employs an LSTM combined with an attention mechanism to construct a lifespan prediction model framework, balancing time-series data processing capabilities with key feature focusing capabilities. The LSTM network excels at capturing dependencies in long-term time-series data, effectively uncovering the inherent patterns of equipment operating parameters and performance degradation over time, thus adapting to the long-term time-series characteristics of waste heat recovery equipment. The attention mechanism assigns weights to input features, focusing on core features that significantly impact equipment lifespan, such as medium temperature, structural stress, and scale thickness, while mitigating interference from irrelevant features and enhancing the model's sensitivity to key damage factors. Simultaneously, the model framework uses an adaptive learning rate, initially set to 0.001 and dynamically adjusted during iterations based on the loss value, thereby adapting to the complex distribution characteristics of multi-source data.

[0073] During training, the core input features are medium temperature, medium pressure, medium flow rate, structural stress, and scale thickness, covering the dimensions of waste heat recovery equipment operation status, structural safety, and performance degradation. The actual remaining lifespan of the equipment calculated based on historical operation and maintenance records and fault data is used as the output label, and the model is driven to iteratively train through the backpropagation algorithm.

[0074] After each iteration, the prediction error of the lifespan prediction model is calculated using the validation set, such as by using Mean Absolute Error (MAE). If the model's prediction error is greater than or equal to a preset standard, the model's hyperparameters are adjusted, such as the number of neurons in the LSTM hidden layer and the weight coefficients of the attention mechanism, and the training iteration process is repeated. This continues until the prediction error of the lifespan prediction model on the validation set is less than the preset standard. Then, the model's generalization ability is verified using the test set. Once it is confirmed that there are no overfitting or underfitting issues, the lifespan prediction model training is complete.

[0075] The training method for the life prediction model provided in this application not only continues the advantages of multi-source data fusion, but also makes up for the shortcomings of traditional LSTM models in paying insufficient attention to key features. This allows the trained model to not only fit the data patterns, but also accurately capture the core factors affecting the life of the equipment, further improving the accuracy and reliability of remaining life prediction and providing a more accurate decision-making basis for subsequent thermal management optimization.

[0076] S104: Under the structural safety constraints of the waste heat recovery equipment, the computing device determines the optimal operating parameters of the waste heat recovery equipment based on the objective optimization function.

[0077] Structural safety constraints refer to the parameter boundary conditions that ensure the structural stability of waste heat recovery equipment and prevent damage and failure. The objective optimization function is a mathematical function constructed under these structural safety constraints with the goal of simultaneously improving the remaining lifespan of the equipment and the waste heat recovery efficiency. Optimized operating parameters are the optimal operating parameters obtained by solving the objective optimization function, balancing equipment lifespan and operating efficiency while ensuring the structural safety of the waste heat recovery equipment.

[0078] To overcome the limitations of existing technologies that solely pursue waste heat recovery efficiency while neglecting equipment structural safety and lifespan loss, this application embodiment can achieve a synergistic balance between thermal management efficiency, equipment lifespan, and structural safety by constructing a dual-objective optimization function and binding structural safety constraints, thereby avoiding irreversible damage to the equipment caused by extreme operating conditions.

[0079] In this embodiment, the computing device constructs the objective optimization function using a weighted summation method. Simultaneously, a structural safety constraint is set: the thermal stress of the equipment pipe wall ≤ a preset stress. The preset stress can be determined based on material properties and industry standards, for example, 300 MPa. Then, the computing device solves the objective optimization function, using parameters such as water flow rate and flue gas guide plate angle as parameters to be optimized. It then selects the parameter combinations that satisfy the constraints and optimize the objective optimization function's value, using these as the optimized operating parameters.

[0080] The embodiments of this application effectively reduce the one-sidedness of single-objective optimization through the design of dual-objective optimization and structural safety constraints. This ensures both improved waste heat recovery efficiency and reduced equipment performance degradation and extended service life. At the same time, structural safety constraints solidify the bottom line for equipment operation.

[0081] To accurately quantify the dual-objective optimization requirements of equipment remaining lifespan and waste heat recovery efficiency, and to clarify the core boundaries of structural safety, making the determination of optimized operating parameters more scientific and feasible, this application further refines the construction form of the objective optimization function and the rules for setting structural safety constraints. The specific implementation is as follows: In one possible implementation, the objective optimization function is a weighted summation function with the dual objectives of maximizing the remaining equipment lifespan and maximizing waste heat recovery efficiency. The expression for the objective optimization function can be f(X) = α·L(X) + β·E(X).

[0082] Where X is the operating parameter to be optimized; α and β are weighting coefficients, α+β=1, and α, β∈(0,1); L(X) is the equipment remaining life function corresponding to the operating parameter X to be optimized, and E(X) is the waste heat recovery efficiency function corresponding to the operating parameter X to be optimized.

[0083] The structural safety constraint is expressed as σ(X)≤σ0. Where σ(X) is the structural stress of the equipment corresponding to the operating parameter X to be optimized, and σ0 is the preset safety threshold; the preset safety threshold is determined based on at least one of the following: equipment material properties, design standards, and industry safety specifications.

[0084] In this embodiment, the objective optimization function adopts a weighted summation form, balancing the dual objectives of maximizing the remaining lifespan of the equipment and maximizing waste heat recovery efficiency within the same mathematical framework. Here, X, as the operating parameter to be optimized, can include adjustable parameters such as feedwater flow rate, flue gas guide vane angle, and boiler outlet flue gas temperature. These operating parameters directly affect the equipment's heat exchange efficiency and lifespan loss. L(X) is the remaining lifespan function corresponding to operating parameter X, used to reflect the equipment's lifespan trend under different parameter combinations. E(X) is the waste heat recovery efficiency function corresponding to operating parameter X, which can be calculated based on parameters such as medium temperature and flow rate, and the heat exchange mechanism formula.

[0085] The weighting coefficients α and β are dynamically allocated according to actual production needs, satisfying the constraint that α + β = 1 and α, β ∈ (0, 1). For example, when production prioritizes ensuring the long-term stable operation of equipment, α = 0.6 and β = 0.4 can be set to focus on extending equipment life; when it is necessary to prioritize improving energy utilization and controlling energy consumption costs, α = 0.4 and β = 0.6 can be adjusted to focus on improving recycling efficiency. This weighted approach avoids the conflict between two objectives and can adapt to the priority requirements of different production scenarios, making the optimization direction more aligned with practical applications.

[0086] Structural safety constraints are explicitly expressed in the form of stress boundaries as σ(X)≤σ0. Parameter limitations prevent irreversible damage such as corrosion and cracking caused by excessive stress. Here, σ(X) represents the structural stress of the equipment corresponding to the operating parameter X to be optimized, which can be calculated using digital twin simulation, focusing on the superposition of thermal and pressure stresses in key areas such as the heating surface and shell. σ0 is a preset safety threshold. In this embodiment, the preset safety threshold can be determined by combining the material properties of the core components of the waste heat boiler (such as the tensile strength of Q345R steel), equipment design standards, and the industry safety standard "Safety Technical Supervision Regulations for Industrial Boilers." For example, the preset safety threshold σ0 is 300MPa, which provides safety redundancy while avoiding excessive constraints that lead to insufficient optimization space.

[0087] Meanwhile, the determination of the preset safety threshold can be flexibly adapted to different equipment types. If the equipment material is upgraded or the application scenario changes, the σ0 value can be adjusted based on the new material properties, design parameters and industry standards to ensure the universality and rigor of the constraints and reduce the structural safety risks that may be caused by the optimization parameters from the source.

[0088] When solving for the objective optimization function, the computing device can employ a particle swarm optimization algorithm, using X as the optimization variable and embedding σ(X)≤σ0 as a constraint condition into the solution process. During the solution process, multiple sets of operating parameter combinations are generated iteratively, and each combination is substituted into the objective optimization function to calculate the function value. Simultaneously, a digital twin is used to verify whether the structural stress corresponding to each set of parameters meets the constraints. Finally, the parameter combination that meets both structural safety requirements and maximizes the value of f(X) is selected as the final optimized operating parameters.

[0089] This application's embodiments, through explicit mathematical expressions and constraint rules, transform dual-objective optimization from qualitative requirements into quantitative calculations. This ensures the accuracy of the optimization direction while safeguarding the bottom line of equipment operation through structural safety constraints, effectively balancing the needs of production efficiency, equipment lifespan, and operational safety. It provides clear and reliable parameter basis for subsequent simulation verification and equipment control.

[0090] S105: The computing device inputs the optimized operating parameters into the digital twin for simulation to determine the waste heat recovery efficiency corresponding to the optimized operating parameters.

[0091] Step S105 is the core verification step for the optimized parameters. The computing device uses a digital twin to simulate and calculate the waste heat recovery efficiency corresponding to the optimized operating parameters, verifying the efficiency improvement effect of the optimized operating parameters in the actual operating scenario. At the same time, it provides a consistent simulation scenario support for subsequent steps to predict the remaining lifespan of the equipment under these parameters, avoiding the risks and losses caused by testing directly on the physical equipment.

[0092] In this embodiment, the computing device can first preprocess the optimized operating parameters and check whether each parameter in the optimized operating parameters is within the equipment's operating limit range, such as the water flow rate not exceeding 100m³ / h and the flue gas guide plate angle adjustment range of 0°~45°, to ensure that the parameters are actually operable and avoid exceeding the equipment's hardware bearing capacity, which could lead to simulation distortion or equipment damage.

[0093] Subsequently, the computing device synchronously inputs the preprocessed optimized operating parameters into the dynamically calibrated digital twin according to the time sequence, controls the digital twin to start the simulation, and simulates the complete process of the waste heat recovery equipment operating with optimized operating parameters.

[0094] During the simulation, the digital twin, based on the multiphysics coupling mechanism, replicates and optimizes the medium flow, heat exchange, and structural stress state corresponding to the operating parameters, and outputs core simulation data in real time. This includes indicators directly related to waste heat recovery efficiency, such as feedwater inlet and outlet temperatures, flue gas inlet and outlet temperatures, medium flow rate, and heat exchange coefficient of the heat exchange surface. The computing equipment can collect simulation data at preset time intervals and continuously simulate for a preset time length to cover the complete operating conditions after parameter stabilization, ensuring the representativeness and stability of the data.

[0095] Based on the collected simulation data, the computing equipment uses the heat balance method to calculate the waste heat recovery efficiency. First, it calculates the heat absorbed by the feedwater (i.e., the recovered waste heat) and the initial waste heat carried by the flue gas, based on the medium flow rate, specific heat capacity, and inlet / outlet temperature difference. Then, it calculates the waste heat recovery efficiency corresponding to the optimized operating parameters using the formula η = (Heat absorbed by feedwater / Initial waste heat of flue gas) × 100%. To further improve the accuracy of the results, the computing equipment can calculate the efficiency value for multiple sets of collected simulation data and take the average as the final result, effectively reducing the impact of random simulation errors on the efficiency calculation.

[0096] This application's embodiments verify the waste heat recovery efficiency of optimized operating parameters through digital twin simulation, eliminating the need for trial runs on physical equipment. This avoids the risks of equipment damage and production interruptions caused by parameter mismatches, and significantly reduces the verification cost and cycle of the optimization scheme. Relying on a calibrated digital twin, the simulation process can accurately replicate the heat transfer patterns under actual operating conditions, and the calculated waste heat recovery efficiency has high reliability, providing accurate efficiency indicators to support subsequent judgments on whether the optimization scheme is superior to the current state.

[0097] S106: The computing device uses a life prediction model to predict and optimize the remaining life of the device based on the simulation results of the digital twin.

[0098] The computing device can selectively extract and preprocess simulation results output by the digital twin, filtering out core feature data closely related to equipment lifespan. Examples include: temperature field distribution data of key equipment components under optimized operating parameters (average pipe wall temperature, temperature gradient), structural stress data (thermal stress on heated surfaces, peak shell stress), time-series data of media parameters (stability of feedwater flow rate, concentration of flue gas components), and performance degradation correlation data (simulated values ​​of scale thickness growth rate and corrosion rate). The computing device standardizes the extracted data, transforming simulation data of different dimensions and units into feature vectors that meet the input requirements of the lifespan prediction model. Simultaneously, it removes outlier data points caused by instantaneous fluctuations during the simulation process and smooths data trends using a moving average method, ensuring the stability and consistency of the input data.

[0099] Subsequently, the computing device inputs the preprocessed simulation feature vectors into the life prediction model. After the life prediction model starts its operation, it first uses an attention mechanism to assign weights to the input features, focusing on core features that significantly affect equipment life, such as structural stress, scale thickness growth rate, and wall temperature gradient, while weakening the interference of secondary features. Then, it uses LSTM network layers to mine the temporal correlation patterns of each feature over time, and combines the physical mechanisms and data statistical laws learned during model training to quantify the impact of optimized operating parameters on equipment performance degradation. For example, it predicts the heat exchange surface loss cycle based on the scale thickness growth rate, calculates the material fatigue accumulation rate based on structural stress data, and outputs the initial predicted value of the remaining equipment life corresponding to the optimized operating parameters.

[0100] To further improve the reliability of the prediction results, the computing device can also perform post-processing calibration on the initial prediction values. The computing device compares the prediction results with the actual remaining life data of the equipment under similar historical operating conditions. If the deviation is within a preset range (±10%), the prediction value is directly output. If the deviation exceeds the range, the device fine-tunes the input feature weights of the model by combining detailed data such as stress fluctuations and attenuation rates from the digital twin simulation results, and re-predicts until the result meets the standard.

[0101] This application's embodiments rely on a lifespan prediction model trained with multi-source data and combine it with simulation results from a digital twin to achieve accurate prediction of the remaining lifespan of equipment under optimized operating parameters, thus overcoming the evaluation shortcomings of traditional optimization schemes. The combination of full-dimensional feature support from simulation results and the model's attention mechanism and time-series analysis capabilities ensures that the prediction results not only closely match the actual operating mechanism of the equipment but also accurately capture the long-term impact of optimized parameters on lifespan, resulting in high reliability. Furthermore, the prediction process does not require intervention in the physical equipment's operation, avoiding the impact of lifespan assessment on production. The synergy between the model and simulation significantly improves assessment efficiency and further enhances engineering practicality.

[0102] S107: If the waste heat recovery efficiency and remaining lifespan of the optimized operating parameters are both better than the current waste heat recovery efficiency and remaining lifespan of the waste heat recovery equipment, the calculation equipment controls the waste heat recovery equipment to operate according to the optimized operating parameters.

[0103] Step S107 verifies through dual-index comparison to ensure that parameter adjustments are only performed when both efficiency and lifespan are improved, reducing the risk of optimizing one index while damaging the other, and achieving a dynamic balance between equipment heat recovery efficiency and remaining lifespan.

[0104] In this embodiment, the current waste heat recovery efficiency of the waste heat recovery equipment is obtained by collecting real-time operating data (feedwater and flue gas inlet and outlet temperatures and flow rates) and calculating it according to the heat balance method in step S105; the current remaining lifespan of the waste heat recovery equipment is obtained by inputting the current operating parameters into the lifespan prediction model. Simultaneously, quantitative judgment criteria are set: an improvement in optimized efficiency greater than or equal to a first threshold, and an extension of optimized lifespan greater than or equal to a second threshold, are considered better than the current state. The first and second thresholds can be flexibly adjusted according to production needs.

[0105] Subsequently, the computing device initiates a dual-indicator comparison. If the optimized waste heat recovery efficiency corresponding to the optimized operating parameters is greater than or equal to the first threshold compared to the current waste heat recovery efficiency of the waste heat recovery device, and the optimized remaining lifespan of the device corresponding to the optimized operating parameters is less than the second threshold compared to the current remaining lifespan of the waste heat recovery device, the current operating state of the waste heat recovery device is retained to avoid sacrificing device lifespan for efficiency improvement.

[0106] If the optimized efficiency of the waste heat recovery corresponding to the optimized operating parameters is less than the first threshold compared to the current optimized efficiency of the waste heat recovery equipment, and the optimized efficiency of the remaining lifespan of the equipment corresponding to the optimized operating parameters is greater than or equal to the second threshold compared to the current remaining lifespan of the waste heat recovery equipment, the current operating state of the waste heat recovery equipment is retained to avoid sacrificing efficiency to improve equipment lifespan.

[0107] When the waste heat recovery efficiency or remaining equipment lifespan corresponding to the optimized operating parameters fails to meet the preset improvement standards, the parameter issuance process is terminated. The weighting coefficients α and β of the target optimization function are adjusted accordingly based on the non-compliance indicators. For example, if the waste heat recovery efficiency fails to meet the standard, the β value is increased and the α value is decreased; if the remaining lifespan fails to meet the standard, the opposite is true. Then, the process returns to step S104, where the function is re-solved under the premise that the structural safety constraint σ(X)≤σ0 remains unchanged, generating new optimization parameters. The simulation and lifespan prediction process of S105-S106 is repeated, and the dual-indicator comparison of S107 is executed again until both indicators meet the standards or the preset maximum number of adjustments is reached.

[0108] If the optimized operating parameters correspond to a waste heat recovery efficiency greater than or equal to the current waste heat recovery efficiency of the waste heat recovery equipment, and the optimized operating parameters correspond to a remaining lifespan of the equipment greater than or equal to the current remaining lifespan of the waste heat recovery equipment, it indicates that if the waste heat recovery equipment operates according to the optimized operating parameters, it can simultaneously achieve efficiency improvement and lifespan extension. Calculate the equipment start-up control command issuance process.

[0109] During the control execution phase, the computing device can send optimized operating parameters to the PLC control system of the waste heat recovery equipment via the OPC UA protocol. The PLC control system adjusts the actuators of the waste heat recovery equipment according to the optimized operating parameters. For example, the variable frequency pump adjusts the feedwater flow rate to the optimized value, and the electric actuator adjusts the angle of the flue gas guide plate to the target position. Simultaneously, the temperature and pressure monitoring modules provide real-time feedback on the adjusted operating status. During the adjustment process, the computing device continuously collects the actual operating data of the waste heat recovery equipment. If parameters deviate from the optimized value or structural stress approaches the safety threshold, a fine-tuning command is immediately triggered to ensure the equipment smoothly transitions to the optimized operating condition and operates stably. If the optimized solution fails to meet the standards after comparison, the computing device executes step S104, triggering parameter re-optimization and iteration, forming a continuous optimization mechanism.

[0110] To address issues such as material aging and operating condition drift during long-term equipment operation, and to continuously ensure the accuracy of the life prediction model so that it always adapts to the actual operating state of the equipment, this application also provides a method for periodically updating the life prediction model, the specific implementation of which is as follows: In one possible implementation, after each preset period, the computing device updates the life prediction model using the actual operating data of the waste heat recovery equipment, the fault data of the waste heat recovery equipment, and the simulation data of the digital twin within the preset period.

[0111] The embodiments of this application can set a preset cycle based on the operating stability and failure frequency of the waste heat recovery equipment.

[0112] Upon reaching a cycle node, the computing device acquires the actual operating data, fault data, and simulation data synchronously generated by the digital twin of the pre-waste heat recovery equipment within a preset cycle. The computing device integrates the new data with historical data, divides it into training, validation, and test sets according to a preset ratio, and uses a model framework based on the original Long Short-Term Memory network combined with an attention mechanism. It uses data such as medium temperature, pressure, and structural stress as input features, and the actual remaining lifespan calculated within the cycle as the output label. The parameters of the lifespan prediction model are iteratively updated using the backpropagation algorithm. The validation set is used to monitor prediction errors, and once the target is met, the generalization ability is verified using the test set. The lifespan prediction model is then updated and put into subsequent prediction work.

[0113] By periodically updating the life prediction model, the model can dynamically adapt to changes in the state of the waste heat recovery equipment, continuously ensuring the accuracy of the remaining life prediction and providing reliable support for the long-term, efficient, and safe operation of the waste heat recovery equipment.

[0114] Based on the above method embodiments, this application also provides a thermal management and lifespan prediction device for waste heat recovery equipment, such as... Figure 2 As shown, the device includes: The data acquisition module 201 is used to collect the operating status data, environmental data and structural status data of the waste heat recovery equipment, and integrate the historical maintenance data, fault data and material property data of the waste heat recovery equipment to form a multi-dimensional dataset. Module 202 is used to construct a digital twin of the waste heat recovery device based on the multi-dimensional dataset and by integrating multiphysics theory and simulation technology. Training module 203 is used to train a life prediction model based on the actual operating data of the waste heat recovery equipment, the fault data of the waste heat recovery equipment, and the simulation data generated by the digital twin; wherein the actual operating data, the fault data, and the simulation data are data within the same time period; the output of the life prediction model includes the remaining life of the equipment; The optimization module 204 is used to determine the optimized operating parameters of the waste heat recovery equipment based on the objective optimization function; wherein, the objective optimization function is constructed under the constraint of the structural safety of the waste heat recovery equipment with the objective of improving the remaining life of the equipment and the waste heat recovery efficiency; The simulation module 205 is used to input the optimized operating parameters into the digital twin for simulation, determine the waste heat recovery efficiency corresponding to the optimized operating parameters, and use the life prediction model to predict the remaining life of the equipment corresponding to the optimized operating parameters based on the simulation results of the digital twin. The operation module 206 is used to control the waste heat recovery equipment to operate according to the optimized operation parameters if the waste heat recovery efficiency and remaining equipment lifespan corresponding to the optimized operation parameters are both better than the current waste heat recovery efficiency and remaining equipment lifespan of the waste heat recovery equipment.

[0115] In one possible implementation, the acquisition module 201 is specifically used to deploy sensors at key locations of the waste heat recovery equipment to collect the operating status data and the structural status data; wherein, the key locations refer to areas that affect the waste heat recovery efficiency, structural safety, or operational stability of the waste heat recovery equipment, and the key locations include at least one of the equipment's heating surface, medium flow channel, shell, flue gas inlet and outlet end faces, and heat exchange actuator; the environmental data is acquired through an industrial control system interface; wherein, the environmental data refers to external correlation data that affects the waste heat recovery thermal efficiency, structural corrosion, or heat load balance of the waste heat recovery equipment.

[0116] In one possible implementation, the construction module 202 is specifically used to construct the digital twin based on the three-dimensional structural model of the waste heat recovery equipment, and by integrating multiphysics theory and simulation technology with the multidimensional dataset, through multi-field coupling simulation; wherein the multiphysics theory includes at least any one of heat transfer, fluid mechanics, and materials mechanics, and the simulation technology includes at least any one of finite element simulation technology and computational fluid dynamics technology; the digital twin is used to map the internal physical field distribution and performance degradation process of the waste heat recovery equipment in real time based on the multidimensional dataset; the internal physical field distribution includes at least any one of temperature field, flow field, and stress field, and the performance degradation process includes at least any one of scaling, corrosion, and creep damage.

[0117] In one possible implementation, the construction module 202 is further configured to select the actual operating parameters of the waste heat recovery equipment during a preset time period, input the actual operating parameters into the digital twin for simulation, and obtain the simulation data of the digital twin; if the deviation between the simulation data and the corresponding operating state data and structural state data of the actual operating data is greater than or equal to a preset threshold, the thermal conductivity coefficient, flow field resistance coefficient and material fatigue parameters of the digital twin are adjusted, and the calibration is repeated until the deviation between the simulation data and the actual operating data is less than the preset threshold, so that the simulation state of the digital twin is close to that of the waste heat recovery equipment.

[0118] In one possible implementation, the training module 203 is specifically used to divide the actual operating data, fault data, and simulation data within the same time period into a training set, a validation set, and a test set according to a preset ratio; to construct the model framework of the life prediction model using a long short-term memory network combined with an attention mechanism; and to iteratively train the model using the medium temperature, pressure, flow rate, structural stress, and scale thickness as input features, and the remaining life of the waste heat recovery equipment as the output label, through a backpropagation algorithm, until the prediction error of the model is less than the model prediction error standard, thereby obtaining the life prediction model.

[0119] In one possible implementation, the objective optimization function is a weighted summation function with the dual objectives of maximizing the remaining lifespan of the equipment and maximizing the waste heat recovery efficiency; The expression for the objective optimization function is f(X) = α·L(X) + β·E(X); Where X is the operating parameter to be optimized; α and β are weighting coefficients, α+β=1, and α, β∈(0,1); L(X) is the equipment remaining life function corresponding to the operating parameter X to be optimized, and E(X) is the waste heat recovery efficiency function corresponding to the operating parameter X to be optimized; The structural safety constraint is expressed as σ(X)≤σ0; where σ(X) is the structural stress of the equipment corresponding to the operating parameter X to be optimized, and σ0 is a preset safety threshold; the preset safety threshold is determined based on at least one of the following: equipment material properties, design standards, and industry safety specifications.

[0120] In one possible implementation, the construction module 202 is further configured to update the life prediction model every preset period using the actual operating data of the waste heat recovery equipment, the fault data of the waste heat recovery equipment, and the simulation data of the digital twin within the preset period.

[0121] This application also provides a control device. The control device may include a memory and a processor. The processor is used to execute the thermal management and lifespan prediction methods for the waste heat recovery equipment described in any of the above embodiments. The memory may be random access memory (RAM), flash memory, read-only memory (ROM), non-volatile read-only memory (EPROM), registers, hard disk, removable disk, etc.

[0122] Memory can store computer instructions. When these instructions are executed by a processor, the processor can use them to implement thermal management and lifespan prediction methods for waste heat recovery equipment. Memory can also store data.

[0123] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented, in whole or in part, as a computer program product. A computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the flow or function according to the embodiments of this application is generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, Digital Subscriber Line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape) or a semiconductor medium (e.g., solid-state disk (SSD)).

[0124] This application also provides a readable storage medium for storing the methods provided in the above embodiments. For example, RAM, flash memory, ROM, EPROM, registers, hard disk, removable disk, or any other form of storage medium in the art.

[0125] In the embodiments of this application, the terms "first" and "second" (if they exist) are used only as name identifiers and do not represent the order of first and second.

[0126] It should be noted that the various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. Regarding the methods disclosed in the embodiments, since they correspond to the product embodiments disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to in the description of the product embodiments.

[0127] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for thermal management and lifespan prediction of waste heat recovery equipment, characterized in that, The method includes: Collect operational status data, environmental data, and structural status data of the waste heat recovery equipment, and integrate the equipment's historical maintenance data, fault data, and material property data to form a multi-dimensional dataset; Based on the multi-dimensional dataset, a digital twin of the waste heat recovery equipment is constructed by integrating multiphysics theory and simulation technology; Based on the actual operating data of the waste heat recovery equipment, the fault data of the waste heat recovery equipment, and the simulation data generated by the digital twin, a life prediction model is trained; wherein the actual operating data, the fault data, and the simulation data are data within the same time period; the output of the life prediction model includes the remaining life of the equipment; Under the structural safety constraints of the waste heat recovery equipment, the optimized operating parameters of the waste heat recovery equipment are determined based on the objective optimization function; wherein, the objective optimization function is constructed under the structural safety constraints with the goal of improving the remaining lifespan of the equipment and the waste heat recovery efficiency; The optimized operating parameters are input into the digital twin for simulation to determine the waste heat recovery efficiency corresponding to the optimized operating parameters; and the remaining equipment lifespan corresponding to the optimized operating parameters is predicted based on the simulation results of the digital twin using the lifespan prediction model. If the waste heat recovery efficiency and remaining equipment lifespan corresponding to the optimized operating parameters are both better than the current waste heat recovery efficiency and remaining equipment lifespan of the waste heat recovery equipment, the waste heat recovery equipment is controlled to operate according to the optimized operating parameters.

2. The method according to claim 1, characterized in that, The data collected on the operating status, environmental data, and structural status of the waste heat recovery equipment include: The operating status data and structural status data are collected by sensors; wherein, the sensors are deployed in key parts of the waste heat recovery equipment, and the key parts refer to the areas that affect the waste heat recovery efficiency, structural safety or operational stability of the waste heat recovery equipment, and the key parts include at least any one of the equipment's heating surface, medium flow channel, shell, flue gas inlet and outlet end face and heat exchange actuator; The environmental data is obtained through the industrial control system interface; wherein, the environmental data refers to external correlation data that affects the waste heat recovery thermal efficiency, structural corrosion, or heat load balance of the waste heat recovery equipment.

3. The method according to claim 1, characterized in that, The construction of a digital twin of the waste heat recovery equipment based on the multi-dimensional dataset, integrating multiphysics theory and simulation technology, includes: Based on the three-dimensional structural model of the waste heat recovery equipment, and by integrating multi-physics theory and simulation technology with the multi-dimensional dataset, the digital twin is constructed through multi-field coupling simulation. The multiphysics theory includes at least one of heat transfer, fluid mechanics, and materials mechanics; the simulation technology includes at least one of finite element simulation and computational fluid dynamics; the digital twin is used to map the internal physical field distribution and performance degradation process of the waste heat recovery equipment in real time based on the multidimensional dataset; the internal physical field distribution includes at least one of temperature field, flow field, and stress field; and the performance degradation process includes at least one of scaling, corrosion, and creep damage.

4. The method according to claim 1, characterized in that, After constructing the digital twin, the method further includes: Select the actual operating parameters of the waste heat recovery equipment during a preset time period, input the actual operating parameters into the digital twin for simulation, and obtain the simulation data of the digital twin; If the deviation between the simulation data and the actual operating data corresponding to the operating status data and structural status data is greater than or equal to a preset threshold, the thermal conductivity coefficient, flow field resistance coefficient, and material fatigue parameters of the digital twin are adjusted. Repeat the calibration until the deviation between the simulation data and the actual operating data is less than the preset threshold, so that the simulation state of the digital twin is close to that of the waste heat recovery device.

5. The method according to claim 1, characterized in that, The lifespan prediction model is trained based on the actual operating data of the waste heat recovery equipment, the fault data of the waste heat recovery equipment, and the simulation data generated by the digital twin, including: The actual operating data, fault data and simulation data within the same time period are divided into training set, validation set and test set according to a preset ratio; The lifespan prediction model is constructed by combining a long short-term memory network with an attention mechanism. Using medium temperature, pressure, flow rate, structural stress, and scale thickness as input features, and the remaining lifespan of the waste heat recovery equipment as the output label, the model is iteratively trained through backpropagation algorithm until the prediction error of the model is less than the model prediction error standard, thus obtaining the lifespan prediction model.

6. The method according to claim 1, characterized in that, The objective optimization function is a weighted summation function with the dual objectives of maximizing the remaining lifespan of the equipment and maximizing the waste heat recovery efficiency. The expression for the objective optimization function is f(X) = α·L(X) + β·E(X); Where X is the operating parameter to be optimized; α and β are weighting coefficients, α+β=1, and α, β∈(0,1); L(X) is the equipment remaining life function corresponding to the operating parameter X to be optimized, and E(X) is the waste heat recovery efficiency function corresponding to the operating parameter X to be optimized; The structural safety constraint is expressed as σ(X)≤σ0; where σ(X) is the structural stress of the equipment corresponding to the operating parameter X to be optimized, and σ0 is a preset safety threshold; the preset safety threshold is determined based on at least one of the following: equipment material properties, design standards, and industry safety specifications.

7. The method according to claim 1, characterized in that, The method further includes: Every preset period, the life prediction model is updated using the actual operating data of the waste heat recovery equipment, the fault data of the waste heat recovery equipment, and the simulation data of the digital twin within the preset period.

8. A thermal management and lifespan prediction device for waste heat recovery equipment, characterized in that, The device includes: The data acquisition module is used to collect operating status data, environmental data, and structural status data of the waste heat recovery equipment, and integrate the equipment's historical maintenance data, fault data, and material property data to form a multi-dimensional dataset. The construction module is used to construct a digital twin of the waste heat recovery device based on the multi-dimensional dataset and by integrating multiphysics theory and simulation technology. The training module is used to train a life prediction model based on the actual operating data of the waste heat recovery equipment, the fault data of the waste heat recovery equipment, and the simulation data generated by the digital twin; wherein the actual operating data, the fault data, and the simulation data are data within the same time period; the output of the life prediction model includes the remaining life of the equipment; An optimization module is used to determine the optimized operating parameters of the waste heat recovery equipment based on a target optimization function; wherein, the target optimization function is constructed under the constraint of the structural safety of the waste heat recovery equipment with the goal of improving the remaining lifespan of the equipment and the waste heat recovery efficiency; The simulation module is used to input the optimized operating parameters into the digital twin for simulation, determine the waste heat recovery efficiency corresponding to the optimized operating parameters, and use the life prediction model to predict the remaining life of the equipment corresponding to the optimized operating parameters based on the simulation results of the digital twin. The operation module is used to control the waste heat recovery equipment to operate according to the optimized operation parameters if the waste heat recovery efficiency and remaining equipment lifespan corresponding to the optimized operation parameters are both better than the current waste heat recovery efficiency and remaining equipment lifespan of the waste heat recovery equipment.

9. A control device, characterized in that, It includes a processor and a memory, the memory being used to store programs, instructions, or code, and the processor being used to execute the programs, instructions, or code in the memory to complete the thermal management and life prediction method of the waste heat recovery equipment as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The device contains a computer program that is loaded by a processor to execute the thermal management and life prediction method for the waste heat recovery device as described in any one of claims 1-7.

Images