Digital twinborn simulation prediction method for electro-thermal conversion equipment of integrated energy system and related device

Abstract

The invention discloses a digital twinborn simulation prediction method and a related device for electro-thermal conversion equipment of an integrated energy system, and belongs to the field of industrial automation and energy system optimizing.The method comprises the following steps that a modeling basic assumption is completed through the equipment operation condition, actual physical parameters and material physical characteristics of the electro-thermal conversion equipment; model initial conditions and model boundary conditions are introduced, unstable heat conduction modeling is carried out on the temperature of the electro-thermal conversion equipment to obtain a physical model of the electro-thermal conversion equipment, and discretization solving is carried out on the physical model of the electro-thermal conversion equipment to obtain a temperature heat conduction simulation result; designing a loss function, building a model structure, and performing model training based on a temperature heat conduction simulation result to obtain an electrothermal conversion equipment temperature prediction model so as to realize digital twinborn simulation prediction of the electrothermal conversion equipment. According to the method, the problems of poor interpretability of a single data driving model and strong dependency of pure physical mechanism model parameters in temperature modeling of the electrothermal conversion equipment can be solved.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of industrial automation and energy system optimization, and particularly relates to a digital twin simulation prediction method for an integrated energy system electric-thermal conversion device and a related device. BACKGROUND

[0002] As a key heat supply device in industrial production, the temperature control accuracy of an industrial electric boiler directly affects energy utilization efficiency, production safety and process stability. However, the internal heat conduction process of the electric boiler is complex and is affected by multiple physical factors such as medium flow, heat conduction, boundary conditions (such as inlet water temperature and heat source power), and the like. Traditional single modeling methods are difficult to balance model accuracy and practicality.

[0003] Although a single data-driven method such as a machine learning model trained on historical data can fit the temperature variation trend, it lacks physical meaning explanation and requires a large amount of labeled data. The prediction performance significantly decreases when the working condition changes or data is missing. In addition, the model generalization ability is limited, making it difficult to adapt to different devices or running conditions. For example, although a data-driven model (such as a two-dimensional convolutional neural network) can capture spatial features, it requires a large amount of two-dimensional temperature field labeled data and is difficult to explain the physical mechanism (such as Fourier's law and convection effect). When the working condition changes (such as uneven radial heat source distribution), the model generalization ability is insufficient, and the prediction may fail due to incomplete data coverage.

[0004] In addition, a mechanism model based on physical principles such as heat conduction differential equations in cylindrical coordinates and fluid mechanics has clear physical meaning, but requires accurate acquisition of radial and axial heat diffusion coefficients, convective heat transfer coefficients and other parameters. However, the radial medium flow state (such as turbulence intensity) in industrial scenes is difficult to measure in real time, and the boundary conditions (such as heat loss of the cylindrical side wall) are complex, resulting in high parameter sensitivity of the pure mechanism model and significant error under non-steady state conditions (such as obvious radial temperature stratification during the startup stage).

[0005] The operation data (such as temperature, flow, power, etc.) accumulated in the industrial field can reflect the device operation law, and the physical mechanism provides the bottom constraint of the heat conduction process. Therefore, how to organically combine the two to build a prediction model with explainability and high robustness is the key to solving the accurate temperature control of the electric boiler. The existing technology does not fully exploit the synergistic driving of physical constraints and data features, resulting in limited model performance improvement. SUMMARY

[0006] The purpose of the present application is to provide a digital twin simulation prediction method for an integrated energy system electric heat conversion device and related devices to solve the problems of poor interpretability of single data-driven models and strong parameter dependence of pure physical mechanism models in existing industrial electric boiler temperature modeling. By data and physical mechanism dual driving modeling, high-precision and interpretable prediction of the electric boiler temperature field is realized, providing support for the optimization control of industrial energy systems.

[0007] To achieve the above purpose, the present application adopts the following technical solutions:

[0008] In a first aspect, a digital twin simulation prediction method for an integrated energy system electric heat conversion device includes the following steps:

[0009] The device operating conditions of the electric heat conversion device in the actual industrial scene are analyzed, and the actual physical parameters and material physical properties of the electric heat conversion device are obtained. Based on the device operating conditions, actual physical parameters and material physical properties, modeling basic assumptions are completed. Model initial conditions and model boundary conditions are introduced based on the modeling basic assumptions, and non-steady-state heat conduction modeling of the temperature of the electric heat conversion device is performed.

[0010] Based on the modeling basic assumptions, model initial conditions, model boundary conditions and non-steady-state heat conduction modeling, an electric heat conversion device physical model is obtained. The electric heat conversion device physical model is discretized and solved to obtain temperature heat conduction simulation results.

[0011] Based on the electric heat conversion device physical model, a loss function is designed, and a model structure is built. The model structure and the loss function are combined, and a model training is performed based on the temperature heat conduction simulation results to obtain an electric heat conversion device temperature prediction model. The digital twin simulation prediction of the electric heat conversion device is realized by using the electric heat conversion device temperature prediction model.

[0012] In some embodiments, the actual physical parameters include shape, size, heating working condition, inlet and outlet water position and inlet water temperature.

[0013] The step of completing modeling basic assumptions based on the device operating conditions, actual physical parameters and material physical properties specifically includes:

[0014] One-dimensional simplification assumptions or two-dimensional simplification assumptions are made based on the actual physical parameters, as well as shape regularity assumptions. The one-dimensional simplification assumption is that the electric heat conversion device has axial heat conduction and convection, and ignores the radial temperature gradient. The two-dimensional simplification assumption is that the electric heat conversion device has axial heat conduction and convection, and introduces a radial temperature gradient. The shape regularity assumption is that the internal structure of the electric heat conversion device is regular, and the internal medium is uniformly distributed.

[0015] According to the device operating condition, the heat transfer mode assumption is that the heat conduction inside the electrothermal conversion device and the water flow in the electrothermal conversion device are natural convection; the initial condition assumption is that the temperature in the entire electrothermal conversion device is uniform at the initial moment; and the boundary condition assumption includes a first type of boundary condition, a second type of boundary condition, and a third type of boundary condition.

[0016] The first type of boundary condition is that the temperature of the heat source remains constant; the second type of boundary condition is that the temperature of the water flowing into the water inlet is regarded as constant, and the water inlet flow is stable and does not change with time; when one-dimensional simplification is performed, the third type of boundary condition is as follows:

[0017]

[0018] When two-dimensional simplification is performed, the third type of boundary condition is as follows:

[0019]

[0020] where λ is the thermal conductivity, is the partial derivative of the temperature along the x direction, is the derivative of the temperature along the outer normal direction of the boundary, h is the convective heat transfer coefficient, u(x=L x , y, t) and u are boundary temperatures obtained by different models, and U env is the medium temperature on the boundary.

[0021] According to the physical properties of the substance, the model physical parameter related assumption is that the thermal diffusivity, the thermal conductivity, and the convective heat transfer coefficient are constant during the entire heat conduction process, and the specific heat capacity and the density of water are constant values.

[0022] In some embodiments, based on the modeling basic assumption, in the step of introducing the model initial condition and the model boundary condition, the model initial condition is the preset temperature of each position of the entire electrothermal conversion device before starting, and the model boundary condition includes a top boundary, a bottom boundary, and a side boundary.

[0023] The top boundary is that the heat source is located at the top end of the electrothermal conversion device, and the first type of boundary condition is used in the action area thereof during work to maintain a constant temperature to heat the medium in the electrothermal conversion device.

[0024] The bottom boundary is that the third type of boundary condition is used to set the heat exchange condition between the bottom of the electrothermal conversion device and the external environment.

[0025] The side boundary is that the side is an adiabatic boundary, and the heat loss is not considered.

[0026] In some embodiments, the step of modeling the temperature of the electrothermal conversion device in a non-steady heat conduction mode specifically comprises: modeling the temperature of the electrothermal conversion device in a non-steady one-dimensional heat conduction mode or a non-steady two-dimensional heat conduction mode.

[0027] The one-dimensional heat conduction equation obtained by modeling the temperature of the electrothermal conversion device in a non-steady one-dimensional heat conduction mode is as follows:

[0028]

[0029] wherein, α eff = α + α conv , α is a thermal diffusion coefficient, α conv is an equivalent thermal diffusion coefficient caused by convection, u is a temperature field, x is a spatial coordinate, and t is time.

[0030] The two-dimensional heat conduction equation obtained by modeling the temperature of the electrothermal conversion device in a non-steady two-dimensional heat conduction mode is as follows:

[0031]

[0032] wherein, α eff = α + α conv , α is a thermal diffusion coefficient, α conv is an equivalent thermal diffusion coefficient caused by convection, u is a temperature field, x and y are spatial coordinates, and t is time.

[0033] In some embodiments, based on the modeling basic assumptions, the model initial conditions, the model boundary conditions, and the non-steady heat conduction modeling, a physical model of the electrothermal conversion device is obtained, and the step of discretely solving the physical model of the electrothermal conversion device specifically comprises:

[0034] The physical model of the electrothermal conversion device is discretized in space and time to divide its time coordinate and spatial coordinate into discrete grids;

[0035] The one-dimensional heat conduction equation or the two-dimensional heat conduction equation is discretized by using the explicit finite difference method to obtain an explicit difference equation;

[0036] The model initial conditions and the model boundary conditions are discretized to obtain discretized conditions;

[0037] Based on the internal nodes of the discrete grids, each time step is traversed, and the temperature field temperature of the current node is calculated by using the explicit difference equation, and the discretized results of the model initial conditions and the model boundary conditions are updated by using the discretized conditions;

[0038] After the traversal is completed, a temperature heat conduction simulation result is obtained.

[0039] In some embodiments, the loss function comprises: a physical loss, an initial condition loss, a boundary condition loss, and a data loss.

[0040] In a second aspect, a digital twin simulation prediction system of an integrated energy system electro-thermal conversion device comprises:

[0041] A modeling assumption module is configured to analyze device operating conditions of the electro-thermal conversion device in an actual industrial scene, and obtain actual physical parameters and material physical properties of the electro-thermal conversion device, complete modeling basic assumptions based on the device operating conditions, the actual physical parameters and the material physical properties, and introduce model initial conditions and model boundary conditions based on the modeling basic assumptions, and perform non-steady-state heat conduction modeling on the temperature of the electro-thermal conversion device.

[0042] A discretization solving module is configured to obtain an electro-thermal conversion device physical model based on the modeling basic assumptions, the model initial conditions, the model boundary conditions and the non-steady-state heat conduction modeling, perform discretization solving on the electro-thermal conversion device physical model, and obtain a temperature heat conduction simulation result.

[0043] A model building and training module is configured to design a loss function based on the electro-thermal conversion device physical model, build a model structure, combine the model structure and the loss function, perform model training based on the temperature heat conduction simulation result to obtain an electro-thermal conversion device temperature prediction model, and realize digital twin simulation prediction of the electro-thermal conversion device by using the electro-thermal conversion device temperature prediction model.

[0044] In a third aspect, an electronic device comprises a memory, a processor, and a computer program stored in the memory and executable in the processor, and the processor implements the steps of the digital twin simulation prediction method of the integrated energy system electro-thermal conversion device when executing the computer program.

[0045] In a fourth aspect, a computer readable storage medium stores a computer program, and the computer program implements the steps of the digital twin simulation prediction method of the integrated energy system electro-thermal conversion device when executed by a processor.

[0046] In a fifth aspect, a computer program product comprises a computer program, and the computer program implements the steps of the digital twin simulation prediction method of the integrated energy system electro-thermal conversion device when executed by a processor.

[0047] Compared with the prior art, the present application has the following beneficial effects:

[0048] The application is based on the physical model of the electro-thermal conversion equipment to design a loss function, and a model structure is built, the model structure and the loss function are combined, and a temperature prediction model of the electro-thermal conversion equipment is obtained based on the model training of the temperature heat conduction simulation result. The temperature heat conduction simulation result is used as driving data, and the physical model of the electro-thermal conversion equipment is used as a physical mechanism guide. The physical mechanism and data characteristics are combined, the physical mechanism ensures the interpretability and rationality of the model, the data driving improves the adaptability of the model to the actual working condition, solves the limitation of a single method, and realizes the quantitative cooperation of data and mechanism through a multi-dimensional loss function, instead of simple superposition, and improves the model precision. In addition, the application introduces the initial condition of the model and the boundary condition of the model on the basis of the basic assumption of the modeling, and at the same time, the temperature of the electro-thermal conversion equipment is modeled by non-steady heat conduction, which supports the temperature prediction under the non-steady working condition and is suitable for the real-time control demand of the industrial electric boiler. The temperature prediction model of the electro-thermal conversion equipment is obtained based on the model training of the temperature heat conduction simulation result, so that the model parameters can be automatically optimized through the historical data, reducing the cost of manual parameter adjustment and improving the efficiency; the model training has small data demand and strong model generalization ability, so that the equipment sensor deployment can be reduced on the basis of ensuring the model prediction accuracy, reducing the hardware cost and data storage cost. BRIEF DESCRIPTION OF DRAWINGS

[0049] Figure 1 The flowchart of the digital twin simulation prediction method of the electro-thermal conversion equipment of the comprehensive energy system provided for embodiment one is provided.

[0050] Figure 2 The flowchart of the digital twin simulation prediction method of the electro-thermal conversion equipment of the comprehensive energy system provided for embodiment two is provided.

[0051] Figure 3 The overall framework diagram of the temperature prediction model of the electro-thermal conversion equipment under one-dimensional digital twin simulation prediction provided for embodiment one is provided.

[0052] Figure 4 The specific flowchart of the digital twin simulation prediction method of the electro-thermal conversion equipment of the comprehensive energy system provided by the application is provided.

[0053] Figure 5 The structure diagram of the digital twin simulation prediction system of the electro-thermal conversion equipment of the comprehensive energy system provided for embodiment three is provided. DETAILED DESCRIPTION

[0054] In order for those skilled in the art to better understand the application scheme, the technical solutions of the application will be further described in detail below with reference to the drawings, and the content is an explanation of the application rather than a limitation.

[0055] It should be noted that the terms "comprising" and "having" and any variations thereof in the specification and claims of the present application are intended to cover not exclusively including, for example, a process, method, system, product or apparatus comprising a series of steps or units does not have to be limited to those steps or units clearly listed, but can include other steps or units that are not clearly listed or inherent to these processes, methods, systems, products or apparatuses.

[0056] Embodiment one

[0057] As shown in Figure 1 and Figure 4 , the embodiment provides a digital twin simulation prediction method of an integrated energy system electric heat conversion device, the electric heat conversion device of the embodiment takes a cylindrical electric boiler in an industrial scene as an example, and one-dimensional digital twin simulation prediction is specifically performed;

[0058] The method specifically comprises the following steps:

[0059] Step 1 analyzes the device running status in the actual industrial scene, obtains the actual physical parameters of the industrial electric boiler device, and completes the modeling basic assumptions, which are as follows:

[0060] Step 1.1 obtains the actual physical parameters of the electric boiler device through the device running status in the actual industrial scene, including the shape, size, electric boiler heating working condition, inlet and outlet water position, inlet water temperature, etc. of the device;

[0061] The shape of the electric boiler in the current industrial environment is cylindrical, with a diameter of about 200 cm and a height of about 210 cm;

[0062] The heating temperature of the electric heating pipe is 100℃, located at the top end of the electric boiler, distributed in a mesh shape, and the heating mode is direct heating type, with the pipe directly contacting the medium;

[0063] The water inlet is located at the bottom end of the electric boiler, and the water inlet temperature is 60℃, and the water outlet is located at the top end of the electric boiler.

[0064] Step 1.2 gives the geometric and structural related assumptions of the model according to the actual physical parameters of the electric boiler, including one-dimensional simplification assumption and shape rule assumption:

[0065] Step 1.2.1 makes one-dimensional simplification assumption to the model, only considers the heat conduction and convection of the electric boiler in the axial direction, and ignores the radial temperature gradient, because in the actual operation, the temperature change in the electric boiler is more significant in the axial direction, and the temperature change in the radial direction is relatively small, which can be ignored. The influence on the overall heat conduction;

[0066] Step 1.2.2 makes shape rule assumption to the model, assuming that the internal structure of the electric boiler is regular and the internal medium is uniformly distributed.

[0067] Step 1.3 According to the actual operation of the electric boiler, the heat transfer related assumptions of the electric boiler temperature are given, including heat transfer mode assumption, initial condition assumption and boundary condition assumption:

[0068] Step 1.3.1 Heat transfer mode assumption: The heat transfer mode of the model mainly considers heat conduction and natural convection. There are three types of convection in heat transfer: natural convection, forced convection and mixed convection. The relative size of Grashof number (Gr) and Reynolds number (Re) can be used to determine the type of convection. Grashof number is an important dimensionless parameter, which is mainly used to describe the relative strength of buoyancy and viscous force in natural convection process. Reynolds number is also a dimensionless parameter used to characterize the flow of fluid, which can be used to determine whether the fluid flow state belongs to laminar flow or turbulent flow. Generally, Gr / Re 2 can be used to represent the relative size of Grashof number and Reynolds number. During the operation of the electric boiler, since the design of the bottom inlet and the top outlet is consistent with the direction of natural convection, the flow of water is relatively stable, which can meet Gr / Re 2 >>10, that is, the heat transfer mode in the electric boiler is mainly heat conduction, and the convection mode of water flow is natural convection;

[0069] Step 1.3.2 The initial condition assumption of the model is to assume that the temperature in the entire electric boiler is uniform under the initial state;

[0070] Step 1.3.3 The boundary condition assumption of the model is:

[0071] Top boundary (heat source): assume that the temperature of the heat source remains constant (first type of boundary condition);

[0072] Bottom boundary (water inlet): the temperature of the water flowing into the water inlet is considered constant, and the water flow is stable and does not change with time (second type of boundary condition);

[0073] The water inlet adopts convective boundary condition (third type of boundary condition), which considers the case where the medium is placed in another medium. The temperature U of the external medium is often different from the temperature u on the surface of the medium under study. Considering the heat flowing through the surface of the medium under study, it should be determined by Fourier's law from the internal medium, that is:

[0074]

[0075] where represents the directional derivative of u along the unit outward normal x on the boundary S;

[0076] From the external aspect, it should be determined by Newton's cooling law, that is:

[0077] dQ = h(u - U)dSdt

[0078] The third type of boundary condition is obtained by combining the above:

[0079]

[0080] where λ is the thermal conductivity, is the derivative of temperature along the outward normal direction of the boundary, h is the convective heat transfer coefficient, u is the boundary temperature obtained by the model, U env is the medium temperature on the boundary, the thermal conductivity λ is 0.6 W / (m·K) in the current scenario, the convective heat transfer coefficient h is 10 W / (m 2 ·K), and the ambient temperature U env is 50℃;

[0081] Side boundary: it is assumed that the side is an adiabatic boundary, i.e. no heat exchange with the outside.

[0082] Step 1.4: According to the physical properties of the substances in the electric boiler, the related assumptions of the model physical parameters are given:

[0083] The physical parameters are constant, the thermal diffusivity α, the thermal conductivity λ, the convective heat transfer coefficient h, and other physical parameters remain constant during the entire heat conduction process, and the thermal diffusivity α is 1.4×10 -7 m 2 / s in the current scenario;

[0084] It is assumed that the specific heat capacity, density, and other physical parameters of water are constant and do not change with temperature and position.

[0085] Step 2: On the basis of the basic assumptions of modeling, the initial conditions and boundary conditions of the model are introduced, as follows:

[0086] Step 2.1: Introduce the initial conditions of the model:

[0087] At the initial time (before starting), the temperature at each position of the entire electric boiler is U0, and the initial temperature U0 is set to 60 in the current scenario;

[0088] Step 2.2: The boundary conditions of the model are introduced as follows:

[0089] Top boundary (heat source): the electric heating pipe is located at the top, and when it is working in its action area, the first type of boundary condition is used to maintain a constant temperature U heat-source to heat the medium in the electric boiler, and the heat source temperature U heat-source is set to 100 in the current scenario;

[0090] Bottom boundary (water inlet): considering the heat exchange between the bottom of the electric boiler and the outside environment, the third type of boundary condition is used;

[0091] Side boundary: side is adiabatic boundary, no heat loss is considered.

[0092] Step 3 carries out the non-steady-state one-dimensional heat conduction modeling of the electric boiler temperature, as follows:

[0093] In order to obtain the expression of the temperature field of the heat-conducting temperature object, the change relationship that the temperature field in the object should satisfy is established according to the law of conservation of energy and Fourier's law, which is called the heat conduction differential equation. The heat conduction differential equation is a general equation that all temperature fields of heat-conducting objects should satisfy.

[0094] The general form of the three-dimensional non-steady-state heat conduction differential equation is:

[0095]

[0096] Where u is the temperature field, x, y, and z are spatial coordinates, t is time, p is the density of the object, and l is the thermal conductivity.

[0097] Since only vertical heat conduction is considered, the temperature difference of the liquid in the horizontal direction (radial direction) is ignored, and it is regarded as a one-dimensional problem. The control equation is obtained as follows:

[0098]

[0099] Where is the thermal diffusion coefficient, u is the temperature field, x is the spatial coordinate, and t is the time.

[0100] On the basis of heat conduction, the influence of natural convection of the medium in the electric boiler needs to be considered, so the modified heat conduction equation is introduced. The effective thermal diffusion coefficient a is introduced in the one-dimensional heat conduction equation. eff :

[0101]

[0102] Where, a eff = a + a conv , a is the thermal diffusion coefficient, and a conv is the equivalent thermal diffusion coefficient caused by convection. In this method, a conv = 0.1·a is approximately taken.

[0103] Step 4: Discretize the obtained physical model by selecting appropriate numerical methods to solve, and obtain the simulation results of the electric boiler temperature heat conduction, as follows:

[0104] Step 4.1: First, perform spatial discretization and time discretization, and divide the time and spatial coordinates of the model into discrete grids.

[0105] The electric boiler model is spatially discretized, and the height L is divided into N xA grid with a grid spacing of The spatial nodes are numbered i = 0, 1, 2, …, N x where i = 0 is the top and i = N x is the bottom.

[0106] The electric boiler is time-discretized with a total simulation time of t total The time is divided into N t grids with a grid spacing of The time nodes are numbered j = 0, 1, 2, …, N t .

[0107] The temperature field u(x, t) is discretized into a two-dimensional matrix u[i, j], where u[i, j] represents the temperature at position i and time j, and the matrix size is N x × N t .

[0108] Step 4.2 uses explicit finite difference method to discretize the control equation, and uses central difference format for the second-order spatial derivative in the control equation, and uses forward difference format for the time derivative, to obtain an explicit difference equation. Adjust the discretization parameters to ensure that the discretization model meets the stability requirements of the explicit finite difference method, and ensure the convergence of the numerical solution.

[0109] The second-order spatial derivative in the control equation is discretized using the central difference format to obtain the discretization result as:

[0110]

[0111] where u i,j represents the temperature at time step j and spatial position i.

[0112] The time derivative is discretized using the forward difference format to obtain the discretization result as:

[0113]

[0114] Substitute the above discretization format into the control equation to obtain:

[0115]

[0116] After rearrangement, the explicit difference equation is obtained as:

[0117] u i,j+1 = u i,j + a·(u i+1,j - 2u i,j + u i-1,j )

[0118] where, is the discretization parameter, the stability requirement of explicit finite difference method needs to be met The time and space discretization grid size of the model needs to be adjusted according to the stability requirement to ensure that the numerical solution does not diverge, so the temperature field discretization grid parameters are set to L = 2m, N x = 100, t total = 100000, N t = 1000, and a = 0.154 < 0.5.

[0119] Step 4.3 Discretization of initial conditions and boundary conditions according to initial conditions and boundary conditions;

[0120] The initial condition is discretized as u[:,0] = U0, which means that at time t = 0, the temperature in the entire electric boiler is uniformly U0.

[0121] The top boundary condition is discretized as u[0,:] = U heat-source , which means that at position x = 0, the temperature is U heat-source throughout the simulation time, i.e. the heat source temperature is fixed.

[0122] The bottom boundary condition is discretized using backward difference method:

[0123]

[0124] Substituting the third type of boundary condition equation gives:

[0125]

[0126] The discretization equation is obtained as:

[0127]

[0128] According to the above discretization conditions, the initial condition and boundary condition discretization results of the model are obtained.

[0129] Step 4.4 Iterative solution, for each time step j, traverse the internal node i, and update the temperature field temperature using the explicit difference equation.

[0130] Step 5 The solution results in Step 4 are used as driving data, and the above physical model is used as physical mechanism guidance to design the loss function, including physical loss, initial condition loss, boundary condition loss, and data loss, as follows:

[0131] The loss function of the model consists of four parts:

[0132]

[0133] L i ∈{LFC ,L BC ,L IC ,L DC}

[0134]

[0135]

[0136] wherein, L FC , L BC , L IC , L DC are loss functions corresponding to physical loss, boundary condition loss, initial condition loss and data loss terms respectively; Ω, Φ, Ψ, Z respectively refer to the calculation domains obtained by sampling during training of different loss terms; N f , N b , N i , N m are the sample numbers of each calculation domain; ||·|| represents the Euclidean norm; λ f , λ bc respectively represent the input physical parameters in the physical loss and boundary condition loss; θ represents the neural network parameters; represents the temperature predicted by the algorithm; U heat-source represents the temperature at the heat source position; U0 represents the temperature field temperature under the initial condition; u real-part represents the driving data obtained by sampling from the numerical solution of physical simulation according to a certain proportion;

[0137] Step 6 builds the model structure, trains the model, evaluates the model, and realizes the steps of one-dimensional modeling simulation prediction of the electric boiler temperature driven by data and mechanism as follows:

[0138] Initialize the model, define the model input and output and the neural network structure NN(r, t; θ), wherein r is the spatial coordinate of the electric boiler input by the model, t is the corresponding time coordinate, and θ is the neural network parameter;

[0139] Optimal parameter combination and optimizer are searched by using the grid method, and are used for model training, and finally the model structure is determined as shown in Table 1, and the overall framework of the model is shown in Figure 3 ;

[0140] Table 1 Temperature prediction model structure of electric heat conversion equipment under one-dimensional digital twin simulation prediction

[0141]

[0142] The network weights are updated by back propagation to minimize the total loss;

[0143] The generalization ability of the model is evaluated by using the numerical solution not participating in the training, and the MSE error is selected as the evaluation index:

[0144]

[0145] Among them, n x is the number of samples of the test set at spatial point x, n t is the number of samples in the test set at time point t; u real-rest (x i ,t j ) is the remaining numerical solution in the physical simulation results that did not participate in the training; The temperature of the temperature field predicted by the algorithm;

[0146] In the current scenario, the proportional interval sampling method is used to determine the ratio of compensation data and test data. The sampling ratios are set to percent = 0.5, percent = 0.25, and percent = 0.1 for experiments. The appropriate sampling ratio is selected according to the experimental results to achieve the accuracy of model prediction while taking into account the characteristics of training the model with a small amount of compensation data.

[0147] This embodiment combines physical mechanisms and data characteristics. The physical mechanism ensures the interpretability and rationality of the model, and data-driven improves the model's adaptability to actual working conditions, addressing the limitations of a single method. Quantitative synergy between data and mechanisms is achieved through a multi-dimensional loss function, rather than simple superposition, thereby improving model accuracy. Temperature prediction under non-steady-state conditions is supported, which is suitable for the real-time control needs of industrial electric boilers. Model parameters can be automatically optimized through historical data, reducing manual parameter adjustment costs and improving efficiency. Model training has a small data requirement and a strong model generalization capability. Therefore, while ensuring the accuracy of model predictions, the deployment of equipment sensors can be reduced, reducing hardware costs and data storage costs. In addition, accurate temperature prediction is provided for energy-saving optimization of electric boilers, helping to reduce energy consumption. At the same time, a general framework is provided for multi-physics field modeling of complex energy systems, which can be extended to other industrial equipment (such as heat exchangers and reactors).

[0148] Example 2

[0149] like Figure 2 and Figure 4 As shown, this embodiment provides a digital twin simulation prediction method for an electric heat conversion device in an integrated energy system. The electric heat conversion device in this embodiment takes a cylindrical electric boiler in a certain industrial scenario as an example, and specifically performs two-dimensional digital twin simulation prediction;

[0150] The method specifically comprises the following steps:

[0151] Step 1: Analyze the equipment operating conditions in actual industrial scenarios, obtain the actual physical parameters of industrial electric boiler equipment, and complete the basic modeling assumptions, as follows:

[0152] Step 1.1 Obtain the actual physical parameters of the electric boiler equipment through the equipment running status in the actual industrial scene, including the shape, size of the equipment, electric boiler heating, water inlet and outlet position, water inlet temperature, etc.

[0153] The shape of the electric boiler in the current industrial environment is cylindrical, with a diameter of about 200 cm and a height of about 210 cm.

[0154] The heating temperature of the electric heating pipe is 100℃, located at the top end of the electric boiler, distributed in a mesh shape, and the heating mode is direct heating type, with the pipe directly contacting the medium.

[0155] The water inlet is located at the bottom end of the electric boiler, with a water inlet temperature of 60℃, and the water outlet is located at the top end of the electric boiler.

[0156] Step 1.2 Make geometric and structural assumptions for the model based on the actual physical parameters of the electric boiler, including two-dimensional simplification assumptions and shape regularity assumptions:

[0157] Step 1.2.1 Make two-dimensional simplification assumptions for the model, only considering axial heat conduction and convection of the electric boiler, and introducing radial temperature gradient. In actual operation, the temperature change in the vertical direction is more significant, and the temperature change in the radial direction is relatively small, which has negligible effect on overall heat conduction. Therefore, a three-dimensional model is not needed.

[0158] Step 1.2.2 Make shape regularity assumptions for the model, assuming that the internal structure of the electric boiler is regular and the internal medium is uniformly distributed.

[0159] Step 1.3 Make heat transfer related assumptions for the temperature of the electric boiler based on the actual running status of the electric boiler, including heat transfer mode assumptions, initial condition assumptions and boundary condition assumptions:

[0160] Step 1.3.1 The heat transfer mode of the model mainly considers heat conduction and natural convection. There are three types of convection in heat transfer: natural convection, forced convection and mixed convection. The relative size of Grashof number (Gr) and Reynolds number (Re) can be used to determine the type of convection. Grashof number is an important dimensionless parameter, mainly used to describe the relative strength of buoyancy and viscous force in natural convection. Reynolds number is also a dimensionless parameter used to characterize the flow of fluid, which can be used to determine whether the fluid flow is laminar or turbulent. Generally, Gr / Re 2 can be used to represent the relative size of Grashof number and Reynolds number. In the running process of the electric boiler, since the design of water inlet at the bottom and water outlet at the top is consistent with the direction of natural convection, the flow of water is relatively stable, which can satisfy Gr / Re 2>> 10, i.e. the heat transfer mode in the electric boiler is mainly heat conduction, and the flow mode of water is natural convection;

[0161] Step 1.3.2 Model initial condition assumption: Assume that the temperature in the entire electric boiler is uniform at the initial time;

[0162] Step 1.3.3 Model boundary condition assumption:

[0163] Top boundary (heat source): Assume that the temperature of the heat source remains constant (first type of boundary condition);

[0164] Bottom boundary (water inlet): The temperature of the water flowing into the water inlet is considered constant, and the water inlet flow is stable and does not change with time;

[0165] The convective boundary condition (third type of boundary condition) is used at the water inlet, which examines the case where the medium is placed in another medium. The temperature U of the external medium is often different from the temperature u on the surface of the medium under examination. The heat flowing through the surface of the medium under examination should be determined by the Fourier law from the internal medium, i.e.:

[0166]

[0167] where represents the directional derivative of u along the unit outer normal direction x on the boundary S;

[0168] From the external aspect, it should be determined by Newton's cooling law, i.e.:

[0169] dQ = h (u - U) dSdt

[0170] Combining the above, the third type of boundary condition is obtained:

[0171]

[0172] where λ is the thermal conductivity, is the partial derivative of temperature along x, h is the convective heat transfer coefficient, u (x = L x , y, t) is the boundary temperature obtained by the model, U env is the temperature of the medium on the boundary, the thermal conductivity λ in the current scenario is 0.6 W / (m·K), the convective heat transfer coefficient h is 10 W / (m 2 ·K), and the temperature of the medium on the boundary U env is 50℃;

[0173] Side boundary: Assume that the side is an adiabatic boundary, i.e. no heat exchange with the outside.

[0174] Step 1.4 According to the physical properties of the substances in the electric boiler, the physical parameter related assumptions of the model are given:

[0175] The physical parameters are constant, the thermal diffusion coefficient a, the thermal conductivity λ, the convective heat transfer coefficient h, and other physical parameters remain constant throughout the heat transfer process. The thermal diffusion coefficient a is 1.4 x 10 -7 m 2 / s in the current scenario.

[0176] It is assumed that the specific heat capacity, density, and other physical parameters of water are constant values and do not change with temperature and position.

[0177] Step 2 introduces the initial conditions and boundary conditions of the model based on the basic assumptions of modeling, as follows:

[0178] Step 2.1 introduces the initial conditions of the model:

[0179] At the initial time (before starting), the temperature at each position of the entire electric boiler is U0, and the initial temperature U0 is set to 60 in the current scenario.

[0180] Step 2.2 introduces the boundary conditions of the model as follows:

[0181] Top boundary (heat source): The electric heating pipe is located at the top, and when it is working, it adopts the first type of boundary condition in its action area, maintaining a constant temperature U heat-source The medium in the electric boiler is heated, and the heat source temperature U heat-source is set to 100 in the current scenario.

[0182] Bottom boundary (water inlet): Considering the heat exchange between the bottom of the electric boiler and the external environment, the third type of boundary condition is adopted.

[0183] Side boundary: The side is an adiabatic boundary, and no heat loss is considered.

[0184] Step 3 performs non-steady-state two-dimensional heat conduction modeling of the electric boiler temperature, as follows:

[0185] To obtain the expression of the temperature field of the heat-conducting object, the change relationship that the temperature field in the object should satisfy is established according to the law of conservation of energy and Fourier's law, which is called the heat conduction differential equation. The heat conduction differential equation is a general equation that all heat-conducting objects should satisfy.

[0186] The general form of the three-dimensional non-steady-state heat conduction differential equation is:

[0187]

[0188] Where u is the temperature field, x, y, and z are spatial coordinates, t is time, p is the density of the object, and λ is the thermal conductivity.

[0189] Since only axial heat conduction is considered, the temperature difference of liquid in the horizontal direction (radial direction) is ignored, and it is regarded as a two-dimensional problem, and the control equation is obtained as follows:

[0190]

[0191] wherein is the thermal diffusion coefficient, u is the temperature field, x and y are spatial coordinates, and t is time;

[0192] On the basis of heat conduction, the influence of natural convection of medium in the electric boiler needs to be considered, so the modified heat conduction equation is introduced, and the effective thermal diffusion coefficient α is introduced in the two-dimensional heat conduction equation eff :

[0193]

[0194] wherein, α eff = α + α conv , α is the thermal diffusion coefficient, and α conv is the equivalent thermal diffusion coefficient caused by convection, and the method approximately takes α conv = 0.1·α.

[0195] Step 4: Discretize the obtained physical model by selecting appropriate numerical methods to solve, and obtain the temperature heat conduction simulation result of the electric boiler, as follows:

[0196] Step 4.1: First, spatial discretization and time discretization are performed, and the time and spatial coordinates of the model are divided into discrete grids;

[0197] The electric boiler model is spatially discretized, and the height L x is divided into N x grids, and the grid spacing is The spatial node number is i = 0, 1, 2, …, N x , and the x-direction coordinate system is vertically downward and perpendicular to the bottom surface; the electric boiler model is spatially discretized, and the width L y is divided into N y grids, and the grid spacing is The spatial node number is j = 0, 1, 2, …, N y , and the y-direction coordinate system is horizontal to the bottom surface from left to right;

[0198] wherein i = 0, j = 0 is the coordinate origin position, located at the left upper corner of the electric boiler, i = N x , j = N y is the coordinate vertex position, located at the right lower corner of the electric boiler;

[0199] The electric boiler is time-discretized, and the total simulation time t total is set, and the time is divided into Nt a grid with grid spacing of The time node number is k = 0, 1, 2, …, N t ;

[0200] The temperature field u(x, y, t) is discretized into a three-dimensional matrix u[i, j, t], representing the temperature at position (i, j) and time t, and the matrix size is N x ×N y ×N t .

[0201] Step 4.2 uses explicit finite difference method to discretize the control equation, and for the second-order spatial derivative in the control equation, the central difference format is used, and the forward difference format is used for the time derivative, to obtain the explicit difference equation, adjust the discretization parameters to ensure that the discretization model meets the stability requirements of the explicit finite difference method, and ensure the convergence of the numerical solution;

[0202] For the second-order spatial derivative of x direction in the control equation The central difference format is used to obtain the discretization result as:

[0203]

[0204] For the second-order spatial derivative of y direction in the control equation The central difference format is used to obtain the discretization result as:

[0205]

[0206] Where, 2u i,j,k represents the temperature at time step k, spatial position (i, j);

[0207] For the time derivative The forward difference format is used to obtain the discretization result as:

[0208]

[0209] Substitute the above discretization format into the control equation to obtain:

[0210]

[0211] After arrangement, the explicit difference equation is obtained:

[0212] u i,j,k+1 = u i,j,k + a x ·(u i+1,j,k - 2u i,j,k + u i-1,j,k ) + a y ·(u i,j+1,k - 2u i,j,k+u i,j-1,k )

[0213] where, is the discretization parameter, the stability requirement of explicit finite difference method requires that a x +a y ≤0.5, the time and space discretization grid size of the model needs to be adjusted according to the stability requirement, so the temperature field discretization grid parameter L x = L y = 2m, N x = N y = 100, t total = 100000, N t = 1000, a x +a y = 0.308 < 0.5.

[0214] Step 4.3 Discretize the initial conditions and boundary conditions according to the initial conditions and boundary conditions;

[0215] The initial condition is discretized as u[:,:,0] = U0, which means that at time t = 0, the temperature in the entire electric boiler is uniformly U0.

[0216] The top boundary condition is discretized as u[0,:,:] = U heat-source , which means that at position x = 0, i.e. the topmost end of the two-dimensional plane of the electric boiler (j ∈ [0, N y -1]), the temperature is U heat-source , i.e. the temperature of the heat source is fixed, throughout the simulation time.

[0217] The bottom boundary condition is discretized using backward difference method:

[0218]

[0219] Substituting the third type of boundary condition equation gives:

[0220]

[0221] The discretization equation is obtained by rearranging:

[0222]

[0223] This formula is used for each grid node in the y direction on the lower boundary x = L x temperature update;

[0224] The side boundary condition is adiabatic, i.e. no heat loss at the boundary, and the left side boundary condition is discretized as:

[0225] u[:,0,n+1] = u[:,1,n+1]

[0226] denotes that all temperature values at the left boundary y = 0 at the n+1 time step in the iteration process are equal to the temperature values at its adjacent position y = 1;

[0227] The right boundary condition is discretized as:

[0228] u[:,-1,n+1] = u[:,-2,n+1]

[0229] denotes that all temperature values at the right boundary y = N y -1 at the n+1 time step in the iteration process are equal to the temperature values at its adjacent position y = N y -2;

[0230] The initial condition and boundary condition discretization results of the model are obtained according to the above discretization conditions.

[0231] Step 4.4 Iterative solution, for each time step j, traverse the internal node (i, j), and update the temperature field temperature using the explicit difference equation.

[0232] Step 5 Take the solution of Step 4 as driving data, take the physical model obtained above as the physical mechanism guide, design the loss function, including the physical loss, the initial condition loss, the boundary condition loss, and the data loss, as follows:

[0233] The loss function of the model is composed of four parts:

[0234]

[0235] L i ∈{L FC ,L BC ,L IC ,L DC}

[0236]

[0237] Where L FC , L BC , L IC , L DC are the loss functions corresponding to the physical loss, the boundary condition loss, the initial condition loss, and the data loss term respectively; Ω, Φ, Ψ, Z respectively indicate the calculation domain obtained by sampling when training different loss terms; N f , N b , N i , N m are the sample numbers of each calculation domain; ||·|| represents the Euclidean norm; λ f , λ bcrepresents the physical parameters of the input in the physical loss and the boundary condition loss, respectively; θ represents the neural network parameters; represents the temperature predicted by the algorithm; U heat-source represents the temperature of the heat source position; U0 represents the temperature of the temperature field under the initial condition; u real-part represents the driving data obtained by sampling the numerical solution of the physical simulation according to a certain proportion;

[0238] Step 6 builds the model structure, trains the model, evaluates the model, and realizes the steps of the data-mechanism dual-driven two-dimensional modeling simulation prediction of the electric boiler temperature as follows:

[0239] Model initialization is performed, and the model input and output and neural network structure NN(r, t; θ) are defined, where r is the spatial coordinate of the electric boiler input of the model, t is the corresponding time coordinate, and θ is the neural network parameter;

[0240] The grid method is used to search for the optimal parameter combination and the optimizer, and is used for model training, and finally the model structure is determined as shown in Table 3, and the overall framework of the model is shown in Figure 3 ;

[0241] Table 3 Temperature prediction model structure of electric heat conversion equipment under two-dimensional digital twin simulation prediction

[0242]

[0243]

[0244] The network weights are updated through back propagation to minimize the total loss;

[0245] The generalization ability of the model is evaluated using the numerical solution that did not participate in the training, and the MSE error is selected as the evaluation index:

[0246]

[0247] where n x is the sampling number of the test set at the spatial point x, n y is the sampling number of the test set at the spatial point y, and n t is the sampling number of the test set at the time point; u real-rest (x i ,y j ,t k ) is the remaining numerical solution in the physical simulation result that did not participate in the training; is the temperature field temperature predicted by the algorithm;

[0248] In the current scenario, the method of equal proportion interval sampling is selected to determine the proportion of compensation data and test data, and the sampling proportions are set as percent=0.5, percent=0.25, and percent=0.1 for experiments. According to the experimental results, a suitable sampling proportion is selected to achieve the accuracy of model prediction while taking into account the characteristics of training the model with a small amount of compensation data.

[0249] On the basis of the one-dimensional digital twin simulation prediction method of embodiment one, the two-dimensional modeling further solves the complex scenario demand of the radial and axial heat conduction coupling of the industrial electric boiler. The physical mechanism and data characteristics are combined. The physical mechanism ensures the interpretability and rationality of the model, and the data driving improves the adaptability of the model to the actual working condition, solving the limitations of a single method. Through a multi-dimensional loss function, the data and mechanism are quantitatively coordinated rather than simply superimposed, improving the model precision. It supports diversified boundary conditions in a two-dimensional scenario (such as the convective heat dissipation of the cylindrical outer wall and the two-dimensional convective boundary of the water inlet at the bottom of the flat plate), avoiding the physical distortion caused by the simplification of the boundary of the one-dimensional model. It supports temperature prediction under non-steady-state working conditions and is suitable for real-time control requirements of industrial electric boilers. Model parameters can be automatically optimized through historical data, reducing the cost of manual parameter adjustment and improving efficiency. The model training has a small data demand and strong model generalization ability, so it can reduce the deployment of equipment sensors, reduce hardware costs and data storage costs while ensuring the accuracy of model prediction. The embodiment provides accurate temperature prediction for energy-saving optimization of electric boilers, helping to reduce energy consumption. At the same time, it provides a general framework for multi-physical field modeling of complex energy systems, which can be extended to other industrial equipment (such as heat exchangers and reaction kettles).

[0250] Embodiment three

[0251] As shown in Figure 5 , the embodiment provides a digital twin simulation prediction system for an electric heat conversion device of a comprehensive energy system, comprising:

[0252] A modeling assumption module is configured to analyze the device operating conditions of the electric heat conversion device in an actual industrial scenario, and obtain the actual physical parameters and material physical characteristics of the electric heat conversion device. Based on the device operating conditions, actual physical parameters and material physical characteristics, modeling basic assumptions are completed. On the basis of the modeling basic assumptions, model initial conditions and model boundary conditions are introduced, and non-steady-state heat conduction modeling is performed on the temperature of the electric heat conversion device.

[0253] A discretization solving module is configured to obtain an electric heat conversion device physical model based on the modeling basic assumptions, model initial conditions, model boundary conditions and non-steady-state heat conduction modeling, and perform discretization solving on the electric heat conversion device physical model to obtain a temperature heat conduction simulation result.

[0254] The model building and training module is configured to design a loss function based on the physical model of the electrothermal conversion device, build a model structure, combine the model structure and the loss function, and train the model based on the temperature heat conduction simulation result to obtain an electrothermal conversion device temperature prediction model, and use the electrothermal conversion device temperature prediction model to realize digital twin simulation prediction of the electrothermal conversion device.

[0255] The division of the modules in the embodiments of the present application is illustrative, and is only a logical function division. In actual implementation, another division manner can be used. In addition, the function modules in each embodiment of the present application can be integrated in one processor, or can be physically separated, or two or more modules can be integrated in one module. The integrated module can be realized in the form of hardware or in the form of a software function module.

[0256] The embodiment also provides a computer device including a processor and a memory. The memory is configured to store a computer program (the computer program includes a calculation component and an iteration component, and can perform model calculation and model updating). The computer program includes program instructions. The processor is configured to execute the program instructions stored in the computer storage medium. The processor can be a central processing unit (CPU), and can also be another general-purpose processor, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or another programmable logic device, a discrete gate or transistor logic device, a discrete hardware component, etc. The processor is a computing core and a control core of the terminal, and is suitable for implementing one or more instructions. Specifically, the processor is suitable for loading and executing one or more instructions in the computer storage medium to implement a corresponding method flow or a corresponding function. The processor in the embodiments of the present application can be used for the operation of the digital twin simulation prediction method of the comprehensive energy system electrothermal conversion device.

[0257] The embodiment also provides a storage medium, specifically a computer readable storage medium (Memory), which is a memory device in a computer device and is used to store programs and data. It can be understood that the computer readable storage medium herein can include an internal storage medium in the computer device, and of course can also include an extended storage medium supported by the computer device. The computer readable storage medium provides a storage space, which stores an operating system of the terminal. In addition, one or more instructions suitable for being loaded and executed by the processor are also stored in the storage space, and the instructions can be one or more computer programs (including program codes). It should be noted that the computer readable storage medium herein can be a high-speed RAM memory or a non-volatile memory such as at least one disk memory. One or more instructions stored in the computer readable storage medium can be loaded and executed by the processor to realize the corresponding steps of the digital twin simulation prediction method of the comprehensive energy system electric-thermal conversion device in the above embodiment.

[0258] The embodiment also provides a computer program product, which includes a computer program. When the computer program is executed by the processor, the corresponding steps of the digital twin simulation prediction method of the comprehensive energy system electric-thermal conversion device in the above embodiment are realized.

[0259] Those skilled in the art will understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROMs, optical storage, etc.) containing computer-usable program code.

[0260] The present application is described with reference to flowcharts and / or block diagrams of the method, device (system), and computer program product according to the embodiments of the present application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and the combination of the flows and / or blocks in the flowcharts and / or block diagrams can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device produce a device that implements the functions specified in the flowcharts and / or block diagrams. Figure 1 The function of the device specified in one flow or multiple flows and / or blocks Figure 1 The function of the device specified in one flow or multiple flows and / or blocks

[0261] These computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer readable memory produce an article of manufacture including instructions which implement the flow Figure 1 one or more flows and / or blocks Figure 1 one or more blocks or multiple blocks.

[0262] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions executed on the computer or other programmable apparatus provide steps for implementing the flow Figure 1 one or more flows and / or blocks Figure 1 one or more blocks or multiple blocks.

[0263] Finally, it should be noted that the above-mentioned embodiments are merely used to illustrate the technical solutions of the present application, rather than limiting the same. Even though the present application has been described in detail with reference to the above-mentioned embodiments, those skilled in the art should understand that the specific embodiments of the present application can be modified or replaced equivalently, and any modification or replacement without departing from the spirit and scope of the present application should be covered within the protection scope of the claims of the present application.

Claims

1. A digital twin simulation prediction method for electric heat conversion equipment in an integrated energy system, characterized in that: The following steps are involved: Analyze the operating conditions of electrothermal conversion equipment in actual industrial scenarios, and obtain the actual physical parameters and material physical properties of the electrothermal conversion equipment. Based on the operating conditions, actual physical parameters, and material physical properties, complete basic modeling assumptions. Based on the basic modeling assumptions, introduce model initial conditions and model boundary conditions, and simultaneously perform unsteady-state heat conduction modeling on the temperature of the electrothermal conversion equipment. Based on the basic modeling assumptions, model initial conditions, model boundary conditions and non-steady-state heat conduction modeling, a physical model of the electrothermal conversion device is obtained, and the physical model of the electrothermal conversion device is discretized and solved to obtain temperature heat conduction simulation results; Based on the physical model of the electrothermal conversion device, a loss function is designed, and a model structure is constructed at the same time. The model structure and the loss function are combined, and model training is performed based on the temperature heat conduction simulation results to obtain a temperature prediction model of the electrothermal conversion device. The temperature prediction model of the electrothermal conversion device is used to realize the digital twin simulation prediction of the electrothermal conversion device.

2. The digital twin simulation prediction method for electric heat conversion equipment in a comprehensive energy system according to claim 1 is characterized in that: The actual physical parameters include: shape, size, heating conditions, water inlet and outlet positions and water inlet temperature; The steps to complete the basic modeling assumptions based on the equipment operating conditions, actual physical parameters and material physical properties include: A one-dimensional simplified assumption or a two-dimensional simplified assumption, as well as a shape regularity assumption, are made based on the actual physical parameters. The one-dimensional simplified assumption is that the electrothermal conversion device has heat conduction and convection in the axial direction, and the radial temperature gradient is ignored. The two-dimensional simplified assumption is that the electrothermal conversion device has heat conduction and convection in the axial direction, and a radial temperature gradient is introduced. The shape regularity assumption is that the internal structure of the electrothermal conversion device is regular, and the internal medium is evenly distributed. Based on the operating conditions of the device, a heat transfer mode assumption, an initial condition assumption, and a boundary condition assumption are made. The heat transfer mode assumption is that heat conduction occurs inside the electrothermal conversion device, and natural convection occurs in the water flow of the electrothermal conversion device. The initial condition assumption is that the temperature inside the entire electrothermal conversion device is uniform at the initial moment. The boundary condition assumptions include first-class boundary conditions, second-class boundary conditions, and third-class boundary conditions. The first type of boundary condition is that the temperature of the heat source remains constant; the second type of boundary condition is that the temperature of the water flowing into the water inlet is considered constant, and the water flow rate is stable and does not change with time; when making a one-dimensional simplified assumption, the third type of boundary condition is as follows: When making a two-dimensional simplified assumption, the third type of boundary condition is as follows: Where λ is the thermal conductivity, is the partial derivative of temperature along the x direction, is the derivative of temperature along the normal direction outside the boundary, h is the convection heat transfer coefficient, u(x=L x ,y,t) and u(x=L x ,t) is the boundary temperature obtained from different models, U env is the medium temperature on the boundary; Based on the physical properties of the material, assumptions related to the physical parameters of the model are made. The assumptions related to the physical parameters of the model are: the thermal diffusivity, thermal conductivity and convective heat transfer coefficient are constant throughout the entire heat conduction process, and the specific heat capacity and density of water are constant values.

3. The digital twin simulation prediction method for electric-thermal conversion equipment in an integrated energy system according to claim 2, characterized in that: In the step of introducing the model initial conditions and model boundary conditions based on the basic modeling assumptions, the model initial conditions are: the preset temperatures of various positions of the entire electrothermal conversion device before startup, and the model boundary conditions include the top boundary, the bottom boundary, and the side boundaries; The top boundary is: the heat source is located at the top of the electrothermal conversion device, and the first type of boundary conditions are used in its active area during operation to maintain a constant temperature to heat the medium in the electrothermal conversion device; The bottom boundary is: the heat exchange between the bottom of the electrothermal conversion device and the external environment is set by using the third type of boundary conditions; The side boundary is: the side is an adiabatic boundary and heat loss is not considered.

4. The digital twin simulation prediction method for electric-thermal conversion equipment in an integrated energy system according to claim 1, characterized in that: The step of performing unsteady heat conduction modeling on the temperature of the electrothermal conversion device specifically includes: performing unsteady one-dimensional heat conduction modeling or unsteady two-dimensional heat conduction modeling on the temperature of the electrothermal conversion device; The one-dimensional heat conduction equation is obtained by performing non-steady-state one-dimensional heat conduction modeling on the temperature of the electrothermal conversion device, as shown in the following formula: Among them, α eff =α+α conv , α is the thermal diffusion coefficient, α conv is the equivalent thermal diffusion coefficient caused by convection, u is the temperature field, x is the spatial coordinate, and t is the time; The two-dimensional heat conduction equation is obtained by performing non-steady-state two-dimensional heat conduction modeling on the temperature of the electrothermal conversion device, as shown in the following formula: Among them, α eff =α+α conv , α is the thermal diffusion coefficient, α conv is the equivalent thermal diffusion coefficient caused by convection, u is the temperature field, x and y are the spatial coordinates, and t is the time.

5. The digital twin simulation prediction method for electric heat conversion equipment in an integrated energy system according to claim 4 is characterized in that: Based on the basic modeling assumptions, model initial conditions, model boundary conditions, and unsteady-state heat conduction modeling, a physical model of the electrothermal conversion device is obtained, and the steps of discretizing and solving the physical model of the electrothermal conversion device specifically include: The physical model of the electrothermal conversion device divides its own time coordinates and space coordinates into discrete grids through spatial discretization and time discretization; Discretizing the one-dimensional heat conduction equation or the two-dimensional heat conduction equation using an explicit finite difference method to obtain an explicit difference equation; Discretizing the model initial conditions and the model boundary conditions to obtain discretized conditions; Traversing the nodes of each time step based on the internal nodes of the discrete grid, and calculating the temperature field temperature of the current node through the explicit differential equation, while updating the discretization results of the model initial conditions and model boundary conditions through the discretization conditions; After the traversal is completed, the temperature heat conduction simulation result is obtained.

6. The digital twin simulation prediction method for electric heat conversion equipment in an integrated energy system according to claim 1, characterized in that: The loss function includes: physical loss, initial condition loss, boundary condition loss and data loss.

7. A digital twin simulation prediction system for electric heat conversion equipment in an integrated energy system, characterized in that: include: A modeling hypothesis module is used to analyze the operating conditions of electrothermal conversion equipment in actual industrial scenarios, obtain the actual physical parameters and material physical properties of the electrothermal conversion equipment, complete basic modeling assumptions based on the equipment operating conditions, actual physical parameters, and material physical properties, introduce model initial conditions and model boundary conditions based on the basic modeling assumptions, and simultaneously perform unsteady-state heat conduction modeling on the temperature of the electrothermal conversion equipment; A discretization solution module is used to obtain a physical model of the electrothermal conversion device based on the basic modeling assumptions, model initial conditions, model boundary conditions and non-steady-state heat conduction modeling, and to perform discretization solution on the physical model of the electrothermal conversion device to obtain temperature heat conduction simulation results; A model building and training module is used to design a loss function based on the physical model of the electrothermal conversion device, and at the same time build a model structure, combine the model structure and loss function, and perform model training based on the temperature heat conduction simulation results to obtain a temperature prediction model of the electrothermal conversion device, and use the temperature prediction model of the electrothermal conversion device to realize digital twin simulation prediction of the electrothermal conversion device.

8. An electronic device, characterized in that: It includes a memory, a processor, and a computer program stored in the memory and executable in the processor. When the processor executes the computer program, it implements the steps of the digital twin simulation prediction method of the electrothermal conversion equipment of the integrated energy system as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of a digital twin simulation prediction method for an electrothermal conversion device of an integrated energy system as described in any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that: When the computer program is executed by a processor, the steps of the digital twin simulation prediction method of the electric-thermal conversion equipment of the integrated energy system according to any one of claims 1 to 6 are implemented.

Images