A peak shaving method for a thermal power plant using high-temperature thermal energy storage

Abstract

The application provides a peak shaving method of a thermal power plant using high-temperature thermal energy storage, and belongs to the technical field of peak shaving of thermal power plants.The application establishes a temperature field mathematical model of a heat storage medium based on a Fourier heat conduction differential equation, uses a temperature sensor array to collect real-time temperature data and establishes a temperature change rate mathematical model, establishes a heat storage capacity demand prediction mathematical model based on power grid load prediction data, calculates heat storage medium heating and cooling power demand using a thermal energy storage peak shaving optimization model, calculates thermal power plant main steam flow regulation parameters and steam turbine load regulation parameters through a heat storage medium power matching relationship model, calculates heat storage depth and heat release depth parameters according to heat storage medium optimal operation temperature interval parameters, and finally outputs heat storage parameters and heat release parameters of the thermal power plant peak shaving operation, thereby solving the technical problem of low thermal power plant peak shaving operation efficiency caused by insufficient heat storage medium temperature field distribution prediction accuracy.

Description

Technical Field

[0001] This invention belongs to the field of peak shaving technology for thermal power plants, and more specifically, relates to a peak shaving method for thermal power plants that utilizes high-temperature thermal energy storage. Background Technology

[0002] Peak shaving operation of thermal power plants is a crucial means of power system flexibility regulation. Traditional peak shaving methods mainly rely on changing fuel supply, adjusting turbine load, and employing thermal storage technology to adapt to dynamic changes in grid load. In current power system operation, thermal storage technology, by storing excess heat in a thermal storage medium and releasing it when needed to meet load demands, has become a key technology for improving the operational flexibility of thermal power plants. Existing peak shaving methods for thermal storage systems typically use empirical models based on historical data or simple linear prediction models to predict the temperature distribution of the thermal storage medium. These methods have limited accuracy when dealing with complex nonlinear heat conduction processes. Traditional technologies often neglect the spatiotemporal coupling characteristics of the internal temperature distribution of the thermal storage medium when modeling its temperature field, making it impossible to accurately predict the temperature change patterns of the thermal storage medium under different operating conditions. Especially when facing rapidly changing grid load demands, existing technologies struggle to achieve accurate prediction and optimized control of the thermal storage medium's temperature field. Due to insufficient accuracy in predicting the temperature field distribution of the thermal storage medium in existing technologies, the thermal storage system cannot accurately control heat storage and release according to actual needs, thus affecting the response speed and peak shaving depth of thermal power plant peak shaving operations. In other words, existing technologies suffer from insufficient accuracy in predicting the temperature field distribution of thermal storage media, leading to low peak-shaving efficiency in thermal power plants. Summary of the Invention

[0003] In view of this, the present invention provides a peak-shaving method for thermal power plants that utilizes high-temperature thermal energy storage, which can solve the technical problem of low peak-shaving operation efficiency of thermal power plants due to insufficient prediction accuracy of temperature field distribution of thermal storage medium in the prior art.

[0004] This invention is implemented as follows: This invention provides a peak-shaving method for thermal power plants utilizing high-temperature thermal energy storage, comprising: establishing a mathematical model of the temperature field of the thermal energy storage medium in a high-temperature thermal energy storage system based on the steam temperature and steam pressure parameters of the boiler system of the thermal power plant; the mathematical model of the temperature field of the thermal energy storage medium uses the Fourier thermal conductivity differential equation to describe the temperature distribution law inside the thermal energy storage medium; calculating the temperature distribution field data and heat flux density distribution field data of the thermal energy storage medium using the density parameters, specific heat capacity parameters, and thermal conductivity parameters of the thermal energy storage medium; collecting real-time temperature data of the thermal energy storage medium using a temperature sensor array arranged inside the thermal energy storage medium; establishing a mathematical model of the temperature change rate of the thermal energy storage medium through time series analysis of the temperature data; and establishing a thermal energy storage capacity... A mathematical model for demand forecasting is developed. The target temperature distribution parameters and storage time window parameters of the thermal storage medium are input into the thermal energy storage peak-shaving optimization model to calculate the heating and cooling power demand parameters of the thermal storage medium. Based on the heating and cooling duration parameters of the thermal storage medium, the main steam flow regulation parameters and turbine load regulation parameters of the thermal power plant are calculated. Using the power matching relationship model of the thermal storage medium, the heat release rate and heat absorption rate parameters of the thermal storage medium are calculated, and the optimal operating temperature range parameters of the thermal storage medium are output. Based on the optimal operating temperature range parameters of the thermal storage medium and the real-time load demand data of the thermal power plant, the thermal storage depth and heat release depth parameters of the thermal storage medium are calculated. Operators adjust the corresponding equipment parameters according to the calculation results to achieve peak shaving.

[0005] The thermal storage capacity prediction function is used to calculate the thermal storage capacity demand value for future periods based on the grid load forecast data and the thermal storage medium temperature change rate parameter. The inputs include grid load forecast time series data, thermal storage medium temperature change rate parameter, thermal storage medium heat capacity parameter, thermal storage medium thermal conductivity parameter, and thermal storage medium density parameter. The outputs are the thermal storage capacity demand value and thermal storage time parameter.

[0006] The steps for establishing the training dataset for the thermal energy storage peak-shaving optimization model specifically include collecting thermal power plant operation data under different seasons and load conditions, including thermal power plant main steam temperature parameters, thermal power plant main steam pressure parameters, turbine load parameters, thermal storage medium temperature distribution field data, and grid load demand data. The collected data is then standardized according to time series to establish a mapping relationship between input feature vectors and target output vectors.

[0007] The density of the temperature sensor array is determined based on the rate of change of the temperature gradient parameters and heat flux density distribution field data of the thermal storage medium. When the temperature gradient parameter of the thermal storage medium is greater than 10℃ / m, the temperature sensor array density is 16 monitoring points per cubic meter; when the temperature gradient parameter is less than 10℃ / m, the temperature sensor array density is 8 monitoring points per cubic meter. This technical solution achieves accurate monitoring and control of the internal temperature distribution of the thermal storage medium by establishing a dynamic correlation mechanism between the temperature sensor array density and the temperature gradient parameters of the thermal storage medium. Traditional temperature monitoring systems typically use a fixed density of sensors, which cannot adaptively adjust according to the actual heat conduction state inside the thermal storage medium. This results in insufficient monitoring accuracy in areas with drastic temperature changes, while wasting resources in areas with gentle temperature changes. This technical solution analyzes the temperature gradient parameters of the thermal storage medium in real time. When the temperature gradient is large, it automatically increases the sensor density to 16 monitoring points per cubic meter, and appropriately reduces the density to 8 monitoring points per cubic meter when the temperature gradient is small. This adaptive adjustment mechanism not only ensures sufficient temperature data support in critical areas but also effectively reduces the system's hardware costs and maintenance complexity. Compared to the traditional uniform arrangement method, this solution significantly improves the accuracy of temperature field reconstruction and the predictive ability of the thermal storage system's operating status, providing a more reliable data foundation for the peak-shaving operation of thermal power plants.

[0008] The thermal storage media are categorized into three types: high-temperature molten salt thermal storage media, high-temperature ceramic particle thermal storage media, and high-temperature metal alloy thermal storage media. The high-temperature molten salt thermal storage media classification is based on the condition that the melting point of the thermal storage medium is below 600℃ and its specific heat capacity is greater than 1500 J / kg·K. The high-temperature ceramic particle thermal storage media classification is based on the condition that the melting point of the thermal storage medium is above 1000℃ and its density is greater than 3000 kg / kJ. The conditions for obtaining.

[0009] Among them, the thermal storage depth parameter of the thermal storage medium is determined by multiplying the thermal storage medium temperature change amplitude parameter and the thermal storage medium heat capacity parameter. The calculation formula for the thermal storage depth parameter is: the thermal storage depth parameter is equal to the thermal storage medium density parameter multiplied by the thermal storage medium volume parameter multiplied by the thermal storage medium specific heat capacity parameter multiplied by the thermal storage medium temperature change amplitude parameter.

[0010] Among them, the heat release depth parameter of the heat storage medium is determined by multiplying the temperature reduction parameter of the heat storage medium and the specific heat capacity of the heat storage medium. The calculation formula of the heat release depth parameter of the heat storage medium is equal to the density parameter of the heat storage medium multiplied by the volume parameter of the heat storage medium multiplied by the specific heat capacity parameter of the heat storage medium multiplied by the temperature reduction parameter of the heat storage medium.

[0011] The thermal response time parameter of the thermal storage medium is determined based on the thermal diffusivity parameter and the geometric characteristic dimension parameter of the thermal storage medium. The calculation formula for the thermal response time parameter of the thermal storage medium is equal to the square of the characteristic length parameter of the thermal storage medium divided by the thermal diffusivity parameter of the thermal storage medium.

[0012] The mathematical model of the temperature field of the thermal storage medium is established using the Fourier thermal conductivity differential equation. It includes numerical calculations of the temperature distribution field data and heat flux density distribution field data of the thermal storage medium. The numerical simulation of the heat transfer process of the thermal storage medium is achieved through grid division and boundary condition setting.

[0013] Among them, the thermal energy storage peak shaving optimization model is a time-series prediction model based on the Transformer-XL architecture, which includes a multi-layer attention mechanism and a memory mechanism. The context length of long sequence processing is controlled by the memory segment number parameter adjustment function. The memory segment number parameter adjustment function is used to dynamically adjust the memory segment number parameter of the thermal energy storage peak shaving optimization model according to the temperature change characteristics of the thermal storage medium. The memory segment number parameter adjustment function is calculated based on the thermal storage medium temperature change period parameter, thermal storage medium thermal response time parameter, thermal power plant load change amplitude parameter, and thermal inertia coefficient parameter of the thermal storage medium to obtain a memory segment number adjustment coefficient.

[0014] Among them, the mathematical model of the temperature change rate of the thermal storage medium is established by time series analysis of real-time temperature data of the thermal storage medium. The least squares method is used to fit the temperature change law of the thermal storage medium over time, and the parameters of the temperature change rate and the temperature gradient of the thermal storage medium are output.

[0015] Among them, the power matching relationship model of thermal storage medium establishes the power balance relationship between thermal storage medium and the main equipment of thermal power plant, and calculates the main steam flow regulation parameters and turbine load regulation parameters of thermal power plant through the heating power demand parameters and cooling power demand parameters of thermal storage medium.

[0016] The specific heat capacity parameter of the thermal storage medium is calculated by multiplying the density parameter, volume parameter, and specific heat capacity parameter of the thermal storage medium. The thermal diffusivity parameter of the thermal storage medium is calculated by dividing the thermal conductivity parameter by the product of the density parameter and specific heat capacity parameter of the thermal storage medium.

[0017] Among them, the periodic parameters of the thermal storage medium temperature change are determined by analyzing the periodic variation law of the thermal storage medium temperature data, and the main periodic components of the thermal storage medium temperature change are extracted by using the fast Fourier transform method. The periodic parameters of the thermal power plant load change are determined by analyzing the periodic variation law of the historical load data of the thermal power plant.

[0018] The current temperature distribution parameters of the thermal storage medium are obtained through real-time measurement data from the temperature sensor array. The geometric characteristic dimensional parameters of the thermal storage medium include the length, width, and height of the thermal storage medium container. The characteristic length parameter of the thermal storage medium is determined by the minimum dimension among the geometric characteristic dimensional parameters of the thermal storage medium.

[0019] This invention establishes a mathematical model of the thermal storage medium temperature field based on the Fourier thermal conductivity differential equation, combines real-time monitoring data from a temperature sensor array with a time-series prediction model based on the Transformer-XL architecture, and achieves high-precision prediction and optimized control of the thermal storage medium temperature distribution. This invention employs a multi-layer attention mechanism and a memory mechanism to process long-term temperature change data, dynamically adjusting the number of memory segments to adapt to the temperature change characteristics of the thermal storage medium under different operating conditions, effectively improving the accuracy of temperature field prediction and the real-time performance of system response. The thermal storage capacity demand prediction mathematical model established in this invention can accurately calculate the thermal storage demand for future periods based on grid load prediction data and the thermal storage medium temperature change rate parameter. Through a thermal storage medium power matching relationship model, it achieves coordinated and optimized control between the thermal storage system and the main equipment of the thermal power plant, thereby significantly improving the efficiency and stability of the thermal power plant's peak-shaving operation. In summary, this invention solves the technical problem mentioned in the background art of insufficient accuracy in predicting the thermal storage medium temperature field distribution, leading to low peak-shaving efficiency in thermal power plants. Attached Figure Description

[0020] Figure 1 This is a flowchart of the method of the present invention.

[0021] Figure 2 This is a schematic diagram of the deployment of the high-temperature thermal energy storage peak-shaving system in a thermal power plant in Example 2.

[0022] Figure 3 This is a graph showing the temperature distribution of the heat storage medium over time in Example 2.

[0023] Figure 4 This is a diagram showing the relationship between power grid load forecasting and thermal storage capacity demand in Example 2.

[0024] Figure 5 This is a diagram showing the power matching relationship of the thermal storage medium in Example 2. Detailed Implementation

[0025] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings.

[0026] like Figure 1 The diagram shown is a flowchart of a peak-shaving method for thermal power plants utilizing high-temperature thermal energy storage, provided by this invention. This method includes the following steps: S01. Based on the steam temperature and steam pressure parameters of the boiler system of the thermal power plant, establish a mathematical model of the temperature field of the heat storage medium in the high-temperature thermal energy storage system. The mathematical model of the temperature field of the heat storage medium uses the Fourier thermal conductivity differential equation to describe the temperature distribution law inside the heat storage medium. The temperature distribution field data and heat flux density distribution field data of the heat storage medium are calculated through the density parameter, specific heat capacity parameter and thermal conductivity parameter of the heat storage medium. S02. Real-time temperature data of the thermal storage medium is collected using a temperature sensor array arranged inside the thermal storage medium. The temperature sensor array contains uniformly distributed temperature measurement points. A mathematical model of the thermal storage medium temperature change rate is established through time series analysis of the temperature data. The mathematical model of the thermal storage medium temperature change rate outputs the thermal storage medium temperature change rate parameter and the thermal storage medium temperature gradient parameter. S03. Based on the power grid load forecast data and the temperature change rate parameter of the thermal storage medium, a mathematical model for predicting thermal storage capacity demand is established. The mathematical model for predicting thermal storage capacity demand calculates the thermal storage capacity demand value for future periods through the thermal storage capacity prediction function, and outputs the target temperature distribution parameters of the thermal storage medium and the thermal storage time window parameters. S04. Input the target temperature distribution parameters and thermal storage time window parameters of the thermal storage medium into the thermal energy storage peak shaving optimization model. The thermal energy storage peak shaving optimization model calculates the heating power demand parameters and cooling power demand parameters of the thermal storage medium based on the difference between the target temperature distribution parameters and the current temperature distribution parameters of the thermal storage medium, and outputs the heating duration parameters and cooling duration parameters of the thermal storage medium. S05. Based on the heating duration parameters and cooling duration parameters of the thermal storage medium, calculate the main steam flow regulation parameters and turbine load regulation parameters of the thermal power plant. The main steam flow regulation parameters of the thermal power plant include the main steam flow rate change parameters, and the turbine load regulation parameters include the turbine load change rate parameters. Establish a power matching relationship model for the thermal storage medium. S06. Using the power matching relationship model of the thermal storage medium, calculate the heat release rate parameter and the heat absorption rate parameter of the thermal storage medium. The heat release rate parameter and the heat absorption rate parameter of the thermal storage medium are obtained by multiplying the temperature gradient parameter and the heat flux density distribution field data of the thermal storage medium, and output the optimal operating temperature range parameter of the thermal storage medium. S07. Based on the optimal operating temperature range parameters of the thermal storage medium and the real-time load demand data of the thermal power plant, calculate the thermal storage depth parameters and the thermal release depth parameters of the thermal storage medium. The thermal storage depth parameters and the thermal release depth parameters of the thermal storage medium are determined by multiplying the thermal storage medium temperature change amplitude parameters and the thermal storage medium specific heat capacity parameters. Output the thermal storage parameters and the thermal release parameters of the thermal power plant for peak shaving operation. The operators adjust the corresponding equipment parameters according to the output results to achieve peak shaving.

[0027] The thermal storage capacity prediction function is used to calculate the thermal storage capacity demand value for future periods based on the power grid load prediction data and the thermal storage medium temperature change rate parameter. The inputs include the power grid load prediction time series data, the thermal storage medium temperature change rate parameter, the thermal storage medium heat capacity parameter, the thermal storage medium thermal conductivity parameter, and the thermal storage medium density parameter. The outputs are the thermal storage capacity demand value and the thermal storage time window parameter.

[0028] The specific structure of the thermal energy storage peak-shaving optimization model is a time-series prediction model based on the Transformer-XL architecture, which includes multi-layer attention and memory mechanisms. The context length of long sequence processing is controlled by the number of memory segments, which is determined based on the thermal storage medium temperature change cycle parameter, the thermal power plant load change cycle parameter, and the thermal storage medium thermal response time parameter. The steps for establishing the training dataset for the thermal energy storage peak-shaving optimization model specifically include collecting thermal power plant operation data under different seasons and load conditions, including thermal power plant main steam temperature parameters, thermal power plant main steam pressure parameters, turbine load parameters, thermal storage medium temperature distribution field data, and grid load demand data. The collected data is standardized according to time series, and a mapping relationship between input feature vectors and target output vectors is established. The input feature vectors include thermal storage medium temperature distribution field data, heat flux density distribution field data, current thermal power plant load parameters, and grid load prediction data. The target output vectors include thermal storage medium heating power demand parameters and thermal storage medium cooling power demand parameters. The specific steps for training the thermal energy storage peak shaving optimization model include dividing the training dataset into a training set and a validation set using a time-series segmentation method, using a weighted combination of cross-entropy loss function and mean squared error loss function as the model's loss function, optimizing model parameters through backpropagation algorithm, employing a learning rate decay strategy and an early stopping mechanism to prevent model overfitting, and dynamically adjusting the number of memory segments during training to adapt to time-series data of different lengths.

[0029] The memory segment quantity parameter adjustment function is used to dynamically adjust the memory segment quantity parameter of the thermal energy storage peak shaving optimization model according to the temperature change characteristics of the thermal storage medium. The memory segment quantity parameter adjustment function is calculated based on the thermal storage medium temperature change period parameter, thermal storage medium thermal response time parameter, thermal power plant load change amplitude parameter, and thermal inertia coefficient parameter of the thermal storage medium to obtain a memory segment quantity adjustment coefficient. When the memory segment quantity adjustment coefficient is in the range of 0 to 0.3, it is used to adjust the memory segment quantity parameter of the thermal energy storage peak shaving optimization model using the minimum memory segment quantity mode. When the memory segment quantity adjustment coefficient is in the range of 0.3 to 0.7, it is used to adjust the memory segment quantity parameter of the thermal energy storage peak shaving optimization model using the medium memory segment quantity mode. When the memory segment quantity adjustment coefficient is in the range of 0.7 to 1.0, it is used to adjust the memory segment quantity parameter of the thermal energy storage peak shaving optimization model using the maximum memory segment quantity mode.

[0030] The heat storage medium is classified into three categories: high-temperature molten salt heat storage medium, high-temperature ceramic particle heat storage medium, and high-temperature metal alloy heat storage medium. The high-temperature molten salt heat storage medium is classified based on the following conditions: its melting point is below 600℃ and its specific heat capacity is greater than 1500 J / kg·K. The high-temperature ceramic particle heat storage medium is classified based on the following conditions: its melting point is above 1000℃ and its density is greater than 3000 kg / kJ. The conditions for obtaining the classification of high-temperature metal alloy thermal storage media were determined by the following criteria: the melting point of the thermal storage medium must be between 600℃ and 1000℃, and the thermal conductivity parameter of the thermal storage medium must be greater than 50 W / m·K. The classification of high-temperature molten salt thermal storage media resulted in rapid heat storage and rapid heat release, contributing 75% to the overall thermal storage efficiency and exhibiting a 15-minute thermal response delay. The classification of high-temperature ceramic particle thermal storage media resulted in long-term heat storage and stable heat release, contributing 90% to the overall thermal storage efficiency and exhibiting a 45-minute thermal response delay. The classification of high-temperature metal alloy thermal storage media resulted in moderate heat storage and uniform heat release, contributing 80% to the overall thermal storage efficiency and exhibiting a 25-minute thermal response delay. The proportions of different types of thermal storage media in the peak-shaving operation of thermal power plants are as follows: 30% high-temperature molten salt thermal storage media, 40% high-temperature ceramic particle thermal storage media, and 30% high-temperature metal alloy thermal storage media. This proportion directly affects the overall thermal response speed and thermal storage capacity of the thermal storage system, and has a positive impact on the response time and peak-shaving depth of the thermal power plant's peak-shaving operation, including shortening the peak-shaving response time and increasing the peak-shaving capacity. However, an excessively high proportion of high-temperature ceramic particle thermal storage media will lead to negative impacts such as increased initial investment costs and increased maintenance complexity of the thermal storage system. These negative impacts can be mitigated by optimizing the proportion of thermal storage media and adopting a layered thermal storage structure.

[0031] The thermal conductivity parameter of the heat storage medium is determined based on the nonlinear relationship between the temperature distribution field data and the density parameter of the heat storage medium. When the temperature of the heat storage medium is below 500℃, the thermal conductivity parameter is calculated as 0.8 times the basic thermal conductivity. When the temperature of the heat storage medium is between 500℃ and 800℃, the thermal conductivity parameter is calculated as 1.0 times the basic thermal conductivity. When the temperature of the heat storage medium is above 800℃, the thermal conductivity parameter is calculated as 1.2 times the basic thermal conductivity.

[0032] The thermal storage depth parameter of the thermal storage medium is determined by multiplying the thermal storage medium temperature change amplitude parameter and the thermal storage medium specific heat capacity parameter. The calculation formula for the thermal storage depth parameter is: the thermal storage depth parameter is equal to the thermal storage medium density parameter multiplied by the thermal storage medium volume parameter multiplied by the thermal storage medium specific heat capacity parameter multiplied by the thermal storage medium temperature change amplitude parameter.

[0033] The heat release depth parameter of the heat storage medium is determined by multiplying the temperature reduction parameter of the heat storage medium and the specific heat capacity parameter of the heat storage medium. The calculation formula of the heat release depth parameter of the heat storage medium is: the heat release depth parameter of the heat storage medium is equal to the density parameter of the heat storage medium multiplied by the volume parameter of the heat storage medium multiplied by the specific heat capacity parameter of the heat storage medium multiplied by the temperature reduction parameter of the heat storage medium.

[0034] The thermal response time parameter of the thermal storage medium is determined based on the thermal diffusivity parameter and the geometric characteristic dimension parameter of the thermal storage medium. The calculation formula of the thermal response time parameter of the thermal storage medium is equal to the square of the characteristic length parameter of the thermal storage medium divided by the thermal diffusivity parameter of the thermal storage medium.

[0035] The mathematical model of the temperature field of the thermal storage medium is established using the Fourier thermal conductivity differential equation. It includes numerical calculations of the temperature distribution field data and heat flux density distribution field data of the thermal storage medium. The numerical simulation of the heat transfer process of the thermal storage medium is achieved through grid division and boundary condition setting.

[0036] The arrangement density of the temperature sensor array is determined based on the rate of change of the temperature gradient parameter and heat flux density distribution field data of the thermal storage medium. When the temperature gradient parameter of the thermal storage medium is greater than 10℃ / m, the temperature sensor array density is arranged with 16 monitoring points per cubic meter. When the temperature gradient parameter of the thermal storage medium is less than 10℃ / m, the temperature sensor array density is arranged with 8 monitoring points per cubic meter.

[0037] The mathematical model for the temperature change rate of the thermal storage medium is established through time series analysis of real-time temperature data of the thermal storage medium. The least squares method is used to fit the temperature change law of the thermal storage medium over time, and the parameters of the temperature change rate and the temperature gradient of the thermal storage medium are output.

[0038] The power matching relationship model of the thermal storage medium establishes the power balance relationship between the thermal storage medium and the main equipment of the thermal power plant. The main steam flow regulation parameters and turbine load regulation parameters of the thermal power plant are calculated by the heating power demand parameters and cooling power demand parameters of the thermal storage medium.

[0039] The thermal capacity parameter of the thermal storage medium is calculated by multiplying its density parameter, volume parameter, and specific heat capacity parameter. The thermal diffusivity parameter is calculated by dividing the thermal conductivity parameter by the product of the density parameter and specific heat capacity parameter. The thermal inertia parameter is calculated by multiplying the density parameter and specific heat capacity parameter.

[0040] The periodic parameter of the thermal storage medium temperature change is determined by analyzing the periodic variation pattern of the thermal storage medium temperature data, and the main periodic components of the thermal storage medium temperature change are extracted using the Fast Fourier Transform (FFT) method. The periodic parameter of the thermal power plant load change is determined by analyzing the periodic variation pattern of the thermal power plant's historical load data, and the main periodic components of the thermal power plant load change are extracted using the FFT method. The parameter of the thermal power plant load change amplitude is determined by calculating the difference between the maximum and minimum values ​​of the thermal power plant load data.

[0041] The current temperature distribution parameters of the thermal storage medium are obtained through real-time measurement data from a temperature sensor array. The geometric dimensional parameters of the thermal storage medium include the length, width, and height of the container. The characteristic length parameter is determined by the smallest dimension among the geometric dimensional parameters. The volume parameter is calculated by multiplying the length, width, and height of the geometric dimensional parameters. The temperature change range parameter is calculated by the difference between the target temperature distribution parameter and the current temperature distribution parameter. The temperature decrease range parameter is calculated by the difference between the initial temperature and the final temperature of the thermal storage medium.

[0042] The main steam temperature and pressure parameters of the thermal power plant are obtained through the steam parameter measuring device of the boiler system. The grid load forecast data is obtained through the load forecast module of the grid dispatching system. The real-time load demand data of the thermal power plant is obtained through the load monitoring system of the thermal power plant. The current load parameters of the thermal power plant are obtained through the generator unit power measuring device of the thermal power plant.

[0043] The specific implementation methods of the above steps are described in detail below. Step S01 is implemented by establishing a mathematical model of the temperature field of the heat storage medium based on the Fourier thermal conductivity differential equation to describe the temperature distribution law inside the heat storage medium. First, the steam temperature and steam pressure parameters of the thermal power plant boiler system are obtained as boundary condition inputs. The steam temperature parameters are typically in the range of 500℃ to 650℃, and the steam pressure parameters are typically in the range of 10MPa to 25MPa. Then, based on the physical property parameters of the heat storage medium, including density, specific heat capacity, and thermal conductivity, the space of the heat storage medium is meshed using the finite element method. The mesh size is determined according to the geometric characteristics of the heat storage medium, usually set to one-tenth of the characteristic length of the heat storage medium. Next, the heat transfer control equation of the heat storage medium is established. This equation is based on Fourier's law of thermal conductivity and the principle of energy conservation. The temperature distribution field data and heat flux density distribution field data of the heat storage medium are calculated through numerical solution methods. The numerical solution process adopts an implicit difference scheme, with a time step set to 0.1 seconds to 1 second, and the spatial step size determined according to the mesh size. The purpose of this step is to provide a theoretical basis and numerical calculation support for subsequent temperature monitoring and control.

[0044] The specific implementation of step S02 involves collecting real-time temperature data of the thermal storage medium using a uniformly distributed array of temperature sensors and establishing a mathematical model of the thermal storage medium's temperature change rate. First, the density of the temperature sensor array is determined based on the magnitude of the thermal storage medium's temperature gradient parameter. When the thermal storage medium's temperature gradient parameter is greater than 10℃ / m, the temperature sensor array density is 16 monitoring points per cubic meter; when the thermal storage medium's temperature gradient parameter is less than 10℃ / m, the temperature sensor array density is 8 monitoring points per cubic meter. The temperature sensors are required to achieve a measurement accuracy of ±0.5℃ and a response time of less than 1 second. Then, time series analysis is performed on the collected temperature data. The least squares method is used to fit the temperature change pattern of the thermal storage medium over time, establishing a mathematical model of the thermal storage medium's temperature change rate. This model processes real-time temperature data using a sliding window method, with a window length set to 10 to 30 minutes, calculating the temperature change rate parameter and temperature gradient parameter of the thermal storage medium at different locations. The purpose of this step is to monitor the temperature status of the thermal storage medium in real time, providing accurate temperature change information for peak-shaving control decisions.

[0045] The specific implementation of step S03 involves establishing a mathematical model for predicting thermal storage capacity demand based on grid load forecast data and the thermal storage medium temperature change rate parameter. First, grid load forecast data for the next 24 hours is obtained from the grid dispatch system, with a data sampling interval of 15 minutes. Then, combined with the thermal storage medium temperature change rate parameter, a thermal storage capacity prediction function is used to calculate the thermal storage capacity demand value for the future period. This prediction function is based on the principle of minimizing load forecast error, and its input parameters include grid load forecast time series data, thermal storage medium temperature change rate parameter, thermal storage medium specific heat capacity parameter, thermal storage medium thermal conductivity parameter, and thermal storage medium density parameter. The prediction function uses an autoregressive moving average model for time series forecasting, and the model parameters are determined through training with historical data. Next, based on the prediction results, the target temperature distribution parameters of the thermal storage medium and the thermal storage time window parameters are output. The thermal storage time window parameters are typically set to 2 to 8 hours. The purpose of this step is to predict the thermal storage capacity required for peak-shaving operation of thermal power plants, providing a basis for formulating thermal storage strategies.

[0046] The specific implementation of step S04 involves inputting the target temperature distribution parameters and thermal storage time window parameters of the thermal storage medium into the thermal energy storage peak-shaving optimization model to calculate the heating and cooling power demand parameters of the thermal storage medium. This optimization model employs a time-series prediction model based on the Transformer-XL architecture, incorporating multi-layer attention and memory mechanisms. First, the difference between the target temperature distribution parameters and the current temperature distribution parameters of the thermal storage medium is calculated; this difference reflects the required temperature adjustment range for the thermal storage medium. Then, based on the thermophysical and heat transfer characteristics of the thermal storage medium, the heating and cooling power demand parameters are calculated. The heating power demand parameters are calculated based on the energy balance of the heating process of the thermal storage medium, while the cooling power demand parameters are calculated based on the energy balance of the heat dissipation process of the thermal storage medium. Finally, the heating and cooling duration parameters of the thermal storage medium are output. The heating duration parameters are typically in the range of 1 to 6 hours, and the cooling duration parameters are typically in the range of 2 to 10 hours. The purpose of this step is to determine the operating strategy of the thermal storage medium and achieve coordinated optimization between the thermal storage system and the thermal power plant operation.

[0047] The specific implementation of step S05 involves calculating the main steam flow regulation parameters and turbine load regulation parameters of the thermal power plant based on the heating and cooling time parameters of the thermal storage medium. First, a power matching relationship model for the thermal storage medium is established, describing the power balance relationship between the thermal storage medium and the main equipment of the thermal power plant. This model is based on the principle of energy conservation and the first law of thermodynamics, calculating the main steam flow regulation parameters and turbine load regulation parameters using the heating and cooling power demand parameters of the thermal storage medium. The main steam flow regulation parameters include the main steam flow rate change parameter, which typically ranges from 10% to 90% of the rated flow. The turbine load regulation parameters include the turbine load change rate parameter, which typically ranges from 20% to 100% of the rated load. The power matching relationship model is solved using an iterative calculation method, with a convergence accuracy set to 0.1%. The purpose of this step is to establish a coordinated control relationship between the thermal storage system and the operating equipment of the thermal power plant, ensuring power balance between the thermal storage process and the power generation process.

[0048] The specific implementation of step S06 involves calculating the heat release rate parameter and heat absorption rate parameter of the thermal storage medium using a power matching relationship model. First, the heat release rate parameter is calculated by multiplying the temperature gradient parameter and heat flux density distribution field data of the thermal storage medium. This parameter reflects the rate at which the thermal storage medium releases heat to the outside. The heat absorption rate parameter is obtained using the same calculation method; this parameter reflects the rate at which the thermal storage medium absorbs heat from the outside. The calculation of the heat release rate and heat absorption rate is based on Fourier's law of thermal conduction and the principle of convective heat transfer. Then, based on the thermophysical properties and operating conditions of the thermal storage medium, the optimal operating temperature range parameter is determined. For high-temperature molten salt thermal storage media, the optimal operating temperature range is typically 280℃ to 560℃. For high-temperature ceramic particle thermal storage media, the optimal operating temperature range is typically 800℃ to 1200℃. For high-temperature metal alloy thermal storage media, the optimal operating temperature range is typically 600℃ to 900℃. The purpose of this step is to determine the optimal operating conditions of the thermal storage medium and improve the operating efficiency and stability of the thermal storage system.

[0049] The specific implementation of step S07 involves calculating the thermal storage depth parameter and the thermal release depth parameter of the thermal storage medium based on the optimal operating temperature range parameters of the thermal storage medium and the real-time load demand data of the thermal power plant. First, the real-time load demand data of the thermal power plant is acquired. This data is collected in real-time through the power plant's load monitoring system, with a sampling frequency of once per second. Then, the thermal storage depth parameter is determined by multiplying the thermal storage medium's temperature change amplitude parameter and its specific heat capacity parameter. This parameter reflects the thermal storage medium's ability to store heat. The thermal storage depth parameter is calculated by multiplying the thermal storage medium's density parameter by its volume parameter, its specific heat capacity parameter, and its temperature change amplitude parameter. The thermal release depth parameter is determined by multiplying the thermal storage medium's temperature decrease amplitude parameter and its specific heat capacity parameter. This parameter reflects the thermal storage medium's ability to release heat. The thermal release depth parameter is calculated by multiplying the thermal storage medium's density parameter by its volume parameter, its specific heat capacity parameter, and its temperature decrease amplitude parameter. Finally, the thermal storage parameters and heat release parameters for peak-shaving operation of the thermal power plant are output. Operators adjust the corresponding equipment parameters based on the output results to achieve peak shaving. The purpose of this step is to quantify the thermal storage and heat release capacity of the thermal storage system, providing precise control parameters for the peak-shaving operation of the thermal power plant.

[0050] The thermal energy storage peak-shaving optimization model adopts a deep learning time-series prediction model based on the Transformer-XL architecture, which has powerful long-sequence modeling capabilities and a memory mechanism. The detailed structure of the model includes an input embedding layer, a multi-layer attention mechanism, a memory mechanism, and an output layer. The input embedding layer is responsible for converting the input feature vector into a high-dimensional vector representation. The input feature vector contains thermal storage medium temperature distribution field data, heat flux density distribution field data, current load parameters of the thermal power plant, and grid load prediction data. The multi-layer attention mechanism includes a self-attention layer and a cross-attention layer. The self-attention layer is used to capture long-term dependencies within the sequence, while the cross-attention layer is used to establish correlations between different input features. The memory mechanism stores historical information through memory segments. The number of memory segments is determined based on the thermal storage medium temperature change cycle parameter, the thermal power plant load change cycle parameter, and the thermal storage medium thermal response time parameter. When the memory segment number adjustment coefficient is in the range of 0 to 0.3, the minimum memory segment number mode is adopted, with 32 memory segments. When the memory segment number adjustment coefficient is in the range of 0.3 to 0.7, a medium memory segment number mode is adopted, with 64 memory segments. When the memory fragment number adjustment factor is between 0.7 and 1.0, the maximum memory fragment number mode is adopted, and the memory fragment number is set to 128. The output layer adopts a fully connected layer structure, outputting the heating power requirement parameters and cooling power requirement parameters of the thermal storage medium.

[0051] The detailed steps for establishing the training dataset are as follows: First, operational data of thermal power plants under different seasons and load conditions were collected over a continuous three-year period, with a sampling frequency of once per minute. The collected data included main steam temperature parameters, main steam pressure parameters, turbine load parameters, thermal storage medium temperature distribution data, and grid load demand data. In the data preprocessing stage, the collected raw data was cleaned and standardized to remove outliers and missing values. Outliers were identified using the 3-standard-deviation principle, and missing values ​​were handled using linear interpolation. Data standardization employed the min-max normalization method, normalizing all feature data to the range of 0 to 1. In the dataset partitioning stage, a temporal partitioning method was used to divide the training dataset into a training set and a validation set, with the training set comprising 80% of the total data and the validation set comprising 20%. In the feature engineering phase, a mapping relationship is established between the input feature vector and the target output vector. The input feature vector includes data on the temperature distribution field of the thermal storage medium, the heat flux density distribution field, the current load parameters of the thermal power plant, and the grid load forecast data. The target output vector includes the heating power demand parameters and cooling power demand parameters of the thermal storage medium. In the model training phase, a weighted combination of the cross-entropy loss function and the mean squared error loss function is used as the model's loss function, with weight coefficients set to 0.3 and 0.7, respectively. The optimization algorithm uses the Adam optimizer, with an initial learning rate of 0.001 and an exponential decay strategy with a decay rate of 0.95. An early stopping mechanism is employed during training to prevent overfitting, monitoring the validation set loss function value. Training stops when the validation set loss function value does not decrease for 10 consecutive periods. In the model evaluation phase, root mean square error (RMSE) and mean absolute error (MAE) are used as evaluation metrics. The model performance requirements are an RMSE of less than 5% and an MAE of less than 3%.

[0052] The key technical ideas of this invention are mainly reflected in the following aspects. The first key technical idea is the design of a thermal energy storage peak-shaving optimization model based on the Transformer-XL architecture. This approach has significant advantages over traditional rule-based or simple neural network-based peak-shaving control methods. The Transformer-XL architecture has powerful long-sequence modeling capabilities and a memory mechanism, effectively capturing the complex nonlinear relationship between changes in thermal storage medium temperature and grid load. Traditional methods often can only process data within a short time window, making it difficult to consider the impact of long-term historical information on current decisions. The Transformer-XL model, through its memory mechanism, can utilize longer historical information, improving prediction accuracy and control performance. The second key technical idea is the synergistic optimization configuration of multiple types of thermal storage media. This approach achieves performance optimization of the thermal storage system by rationally proportioning high-temperature molten salt thermal storage media, high-temperature ceramic particle thermal storage media, and high-temperature metal alloy thermal storage media. Compared to traditional solutions using a single type of thermal storage medium, configuring multiple types of thermal storage media can fully leverage the advantages of different materials. High-temperature molten salt thermal storage media features rapid heat storage and release; high-temperature ceramic particle thermal storage media offers long-term heat storage and stable heat release; and high-temperature metal alloy thermal storage media provides moderate heat storage and uniform heat release. The synergistic effect of these three materials enables the thermal storage system to achieve optimal performance under various operating conditions. The third key technological approach is accurate temperature distribution prediction and control based on a mathematical model of the temperature field. This approach uses the Fourier thermal conductivity differential equation to establish a mathematical model of the thermal storage medium's temperature field, and employs numerical calculation methods to accurately predict the internal temperature distribution of the thermal storage medium. Compared to traditional temperature prediction methods based on empirical formulas or simplified models, this approach can more accurately describe the heat transfer process of the thermal storage medium, considering the influence of the medium's physical properties and boundary conditions, thus improving the accuracy and reliability of temperature prediction.

[0053] The synergistic effect of these key technological approaches offers significant advantages over existing technologies. The combination of the Transformer-XL architecture's optimized model and multi-type thermal storage medium configuration enables intelligent and adaptive control of the thermal storage system. The model can dynamically adjust the control strategy based on the characteristic parameters of different thermal storage media, optimizing the thermal storage and release processes. The synergistic effect of the temperature field mathematical model and the Transformer-XL optimized model achieves a seamless connection from accurate temperature distribution prediction to intelligent control decisions, improving the control accuracy and response speed of the entire peak-shaving system. The combination of multi-type thermal storage medium configuration and the temperature field mathematical model fully considers the differences in heat transfer characteristics of different thermal storage media in different temperature ranges, enabling optimized design and operation of the thermal storage system. The synergistic effect of these three technological approaches gives the entire thermal power plant's peak-shaving system a higher peak-shaving depth, faster response speed, and better economic efficiency, significantly improving the flexibility and grid adaptability of thermal power plants compared to traditional peak-shaving methods.

[0054] It should be noted that this invention also solves the following technical problems. First, the technical problem of inaccurate prediction of the thermal response time of the thermal storage medium: In existing technologies, the prediction of the thermal response time of thermal storage systems is often based on theoretical calculations or empirical estimations, which cannot accurately reflect the thermal response characteristics of different thermal storage media under actual operating conditions, resulting in the thermal storage system's inability to respond promptly to changes in grid load. This invention establishes a calculation model for the thermal response time parameters of the thermal storage medium, accurately calculates the thermal response time based on the thermal diffusivity coefficient and geometric characteristic size parameters of the thermal storage medium, and combines the thermal response delay rate characteristics of three different types of thermal storage media to achieve accurate prediction of the thermal response time of the thermal storage system, providing an accurate time reference for the real-time control of the thermal storage system. Second, the technical problem of difficulty in optimizing the thermal storage medium ratio: Traditional thermal storage systems typically use a single thermal storage medium or a simple multi-medium mixing method, which cannot optimize the ratio according to different operating conditions, resulting in low thermal storage efficiency and difficulty in cost control. This invention establishes characteristic parameter models for three categories of thermal storage media: high-temperature molten salt, high-temperature ceramic particle, and high-temperature metal alloy. It clarifies the applicable conditions and performance characteristics of different thermal storage media and achieves coordinated optimization of the overall thermal response speed and storage capacity of the thermal storage system through optimized proportions of 30%, 40%, and 30%. This ensures both the rapid response capability of the thermal storage system and the stability and economy of long-term thermal storage. Thirdly, it addresses the technical problem of insufficient precision in the coordinated control between the thermal storage system and thermal power plant equipment: Existing technologies lack a precise coordinated control mechanism between the thermal storage system and the main equipment of the thermal power plant, failing to achieve power matching and collaborative optimization between the thermal storage system and equipment such as boilers and turbines, thus affecting the overall system's operating efficiency. This invention establishes a power matching relationship model for the thermal storage media, establishing a precise mathematical relationship between the heat release rate and heat absorption rate parameters of the thermal storage media and the main steam flow regulation parameters and turbine load regulation parameters of the thermal power plant. This achieves coordinated control between the thermal storage system and the thermal power plant equipment, ensuring that the thermal storage system can perform precise heat storage and release operations according to the actual operating needs of the thermal power plant.

[0055] Specifically, the principle of this invention is as follows: This invention solves the technical problem of insufficient accuracy in predicting the temperature field distribution of thermal storage media. Its fundamental principle lies in employing a multi-level temperature field modeling and prediction optimization method. First, this invention establishes a mathematical model of the thermal storage media temperature field based on the Fourier thermal conductivity differential equation. This model can accurately describe the temperature distribution and heat conduction process within the thermal storage media, and compared to traditional empirical models or linear prediction models, it can more accurately reflect the actual thermodynamic characteristics of the thermal storage media. Second, this invention uses a temperature sensor array arranged inside the thermal storage media for real-time temperature monitoring. A mathematical model of the thermal storage media temperature change rate is established through time series analysis, providing a reliable data foundation for temperature field prediction. More importantly, this invention introduces a time-series prediction model based on the Transformer-XL architecture as the core thermal energy storage peak-shaving optimization model. This model has multi-layered attention and memory mechanisms, enabling it to process long-term temperature data and capture long-term dependencies in temperature changes. By dynamically adjusting the number of memory segments, the model can adaptively adjust the prediction window according to the thermal storage media temperature change cycle, thermal response time, and the load change characteristics of the thermal power plant, thereby significantly improving the accuracy of temperature field prediction. Furthermore, the mathematical model for predicting thermal storage capacity demand established in this invention can comprehensively consider grid load forecast data and the temperature change rate parameter of the thermal storage medium, achieving accurate prediction of future thermal storage demand and providing a scientific basis for the optimized control of the thermal storage system. The technical solution of this invention is logically sound because it forms a complete technical chain from physical modeling of the temperature field, real-time data acquisition, intelligent prediction algorithms to system optimization control. Each link cooperates with the others to ensure high accuracy in predicting the temperature field of the thermal storage medium and high efficiency in peak-shaving operation of thermal power plants.

[0056] The following provides a specific embodiment 1 of the present invention, and the specific implementation of each step in this embodiment 1 is described in detail below.

[0057] The specific implementation of step S01 involves establishing a mathematical model of the temperature field of the thermal storage medium based on the Fourier thermal conductivity differential equation. This model describes the temperature distribution within the thermal storage medium. The Fourier thermal conductivity differential equation of the thermal storage medium temperature field mathematical model is specifically expressed as follows: ; In the formula, This represents the temperature distribution field of the thermal storage medium, expressed in K. The variable is time, and the unit is seconds (s). This is the thermal diffusivity parameter for the thermal storage medium. The default value range is [0.2, 1], or you can directly use... The value, in units of ; For the Laplace operator; For the internal heat source of the thermal storage medium, the unit is... ; This refers to the density parameter of the thermal storage medium, in units of... ; This refers to the specific heat capacity parameter of the thermal storage medium, in units of... The method for determining the density of the temperature sensor array is as follows: ; In the formula, This refers to the density of the temperature sensor array, measured in units per... The density parameter of the thermal storage medium was obtained through experimental measurement, including the following steps: Step 1: preparing the thermal storage medium sample into a regular geometric shape; Step 2: measuring the volume of the thermal storage medium sample using the water displacement method or the gas displacement method; Step 3: measuring the mass of the thermal storage medium sample using a precision balance; and Step 4: calculating the density parameter of the thermal storage medium. The specific heat capacity parameter of the thermal storage medium was obtained through differential scanning calorimetry.

[0058] The specific implementation of step S02 involves establishing a mathematical model of the temperature change rate of the thermal storage medium and using the least squares method to fit the temperature change law of the thermal storage medium over time. The mathematical model of the temperature change rate of the thermal storage medium is specifically expressed as follows: ; In the formula, For the first The temperature of the heat storage medium at each temperature measurement point is expressed in K. For the first The coefficient of the first-order term of the temperature change rate at each measurement point, in units of... ; For the first The coefficient of the constant term for the rate of temperature change at each measurement point, in units of... ; For the first The random error term for each measurement point, in units of The calculation method for the temperature gradient parameters of the heat storage medium is as follows: ; In the formula, These are the spatial coordinates of the thermal storage medium, in meters (m). Parameters and The least squares method was used for fitting, including step 1: collecting temperature data of the thermal storage medium within a time window; step 2: establishing a linear regression equation between temperature and time; and step 3: optimizing the regression coefficients using the least squares method. Random error term. The range is -0.1 to 0.1.

[0059] The specific implementation of step S03 involves establishing a mathematical model for predicting thermal storage capacity demand. This model is based on grid load forecast data and the temperature change rate parameter of the thermal storage medium. The thermal storage capacity prediction function is specifically expressed as follows: ; In the formula, For future time periods The required thermal storage capacity is expressed in MJ. The forecast time window is in hours. Weighting coefficients for power grid load forecasting data, dimensionless; for Forecast data of power grid load at any given time, in MW; This is the historical time step for the load data; the default value is 24. , is a dimensionless weighting coefficient for the rate of change of temperature of the thermal storage medium; This represents the number of temperature measurement points. This represents the model bias term, in milligrams (MJ). Weight coefficients. and The model is trained using historical data, including: Step 1: collecting historical load data and thermal storage capacity data; Step 2: establishing a multiple linear regression model; Step 3: optimizing the weight coefficients using gradient descent. Bias term. The range is -50 to 50.

[0060] The specific implementation of step S04 involves establishing a dynamic adjustment mechanism for the number of memory segments in the thermal energy storage peak-shaving optimization model. The calculation method for the memory segment number adjustment coefficient is as follows: ; In the formula, A dimensionless adjustment coefficient for the number of memory segments; This is a periodic parameter representing the temperature change period of the thermal storage medium, in seconds. This is a parameter representing the load variation of thermal power plants, in MW. The reference temperature period is set to 86400s; The reference load variation range is set at 100MW; For reference density, the value is 1000. ; For reference volume, the value is 1. This formula achieves dimensionless measurement by introducing reference density and reference volume, making the adjustment coefficient a dimensionless parameter. The method for determining the number of memory segments is as follows: ; In the formula, The parameter for the number of memory segments is dimensionless. The temperature change period parameter of the thermal storage medium is obtained through Fast Fourier Transform (FFT), including step 1: collecting historical temperature data of the thermal storage medium; step 2: extracting the main periodic components using FFT; and step 3: determining the dominant period as the temperature change period parameter. The load change amplitude parameter of the thermal power plant is determined by calculating the difference between the maximum and minimum values ​​of the thermal power plant load data.

[0061] The specific implementation of step S05 involves establishing a power matching relationship model for the thermal storage medium. This model describes the power balance relationship between the thermal storage medium and the main equipment of the thermal power plant. The power matching relationship model for the thermal storage medium is specifically represented as follows: ; In the formula, This refers to the heating power requirement parameter for the thermal storage medium, in MW. The cooling power requirement parameter for the thermal storage medium is expressed in MW. The turbine efficiency coefficient is dimensionless and ranges from 0.85 to 0.95. This represents the change in turbine load, expressed in MW. The efficiency coefficient of the steam system is dimensionless and ranges from 0.90 to 0.98. This refers to the change in main steam power, expressed in MW. The calculation method for the rate of change of main steam flow in a thermal power plant is as follows: ; In the formula, Main steam flow rate, unit: ; Main steam enthalpy, unit: The turbine load change rate parameter was obtained through experimental measurement, including step 1: monitoring real-time turbine load data; step 2: calculating the load change gradient; and step 3: establishing the relationship between the load change rate and time.

[0062] The specific implementation of step S06 involves calculating the heat release rate parameter and the heat absorption rate parameter of the heat storage medium. The calculation method for the heat release rate parameter of the heat storage medium is as follows: ; In the formula, This is a parameter representing the heat release rate of the thermal storage medium, in units of... ; This is data on heat flux density distribution field, in units of The calculation method for the heat absorption rate parameter of the heat storage medium is as follows: ; In the formula, This is a parameter representing the heat absorption rate of the thermal storage medium, in units of... The heat flux density distribution field data was obtained through experimental measurements, including step 1: placing heat flux sensors on the surface of the heat storage medium; step 2: acquiring heat flux density data in real time; and step 3: establishing a spatial interpolation model to obtain the distribution field data.

[0063] The specific implementation of step S07 involves calculating the thermal storage depth parameter and the thermal release depth parameter of the thermal storage medium. The calculation method for the thermal storage depth parameter is as follows: ; In the formula, This refers to the thermal storage depth parameter of the thermal storage medium, in J. These are the volume parameters of the thermal storage medium, in units of... ; This parameter represents the temperature variation range of the thermal storage medium, expressed in Kelvin (K). During peak-shaving operation in thermal power plants, the temperature variation range of the thermal storage medium is relatively limited (typically within a few hundred degrees Celsius), and the change in specific heat capacity is relatively small within this range. When the temperature variation range is not significant... The linear approximation error is acceptable, especially in control systems where response speed is more important than absolute accuracy.

[0064] The calculation method for the heat release depth parameter of the thermal storage medium is as follows: ; In the formula, This is a parameter representing the heat release depth of the thermal storage medium, in J. This parameter represents the temperature drop of the thermal storage medium, expressed in Kelvin (K). The calculation method for the thermal response time parameter of the thermal storage medium is as follows: ; In the formula, This is the thermal response time parameter of the thermal storage medium, in seconds. This represents the characteristic length parameter of the thermal storage medium, in meters (m). The formula is essentially the reciprocal of the Fourier number. The Fourier number is defined as... ,when When the system reaches a quasi-steady state, therefore the characteristic response time .

[0065] The calculation method for the heat capacity parameters of the thermal storage medium is as follows: ; In the formula, This refers to the heat capacity parameter of the thermal storage medium, in units of... The calculation method for the thermal inertia coefficient parameter of the heat storage medium is as follows: ; In the formula, This refers to the thermal inertia coefficient parameter of the heat storage medium, with units of... The calculation method for the volume parameters of the thermal storage medium is as follows: ; In the formula, The length of the thermal storage medium container is in meters. Width of the thermal storage medium container, in meters; The height of the thermal storage medium container is expressed in meters (m). The method for determining the characteristic length parameter of the thermal storage medium is as follows: ; The calculation method for the temperature change range parameter of the heat storage medium is as follows: ; In the formula, This refers to the target temperature distribution parameter of the thermal storage medium, in K. This represents the current temperature distribution parameter of the thermal storage medium, in Kelvin (K). The calculation method for the parameter representing the temperature decrease of the thermal storage medium is as follows: ; In the formula, The initial temperature of the thermal storage medium, in K; The final temperature of the thermal storage medium is expressed in Kelvin (K). The geometric dimensions of the thermal storage medium are obtained through geometric measurements, including: Step 1: measuring the length, width, and height of the thermal storage medium container using a laser rangefinder; Step 2: calculating the geometric volume of the thermal storage medium; Step 3: subtracting the volume occupied by the internal structure to obtain the effective volume. The target temperature distribution parameters and the current temperature distribution parameters of the thermal storage medium are obtained through real-time measurement data from a temperature sensor array.

[0066] Fourier thermal conductivity differential equation Based on the fundamental physical principles of heat conduction, this paper describes the variation of internal temperature of a thermal storage medium with time and space. The equation considers the thermal diffusion characteristics of the storage medium and the influence of internal heat sources. The time partial derivative term in the equation... The Laplace operator term reflects the rate of change of temperature over time. This reflects the spatial gradient effect of temperature, heat source term It reflects the influence of external energy input on temperature distribution. Compared with traditional empirical formulas or simplified models, this equation can accurately describe the three-dimensional temperature distribution of the thermal storage medium, providing a theoretical basis for the precise control of thermal storage systems.

[0067] Mathematical model of temperature change rate of heat storage medium The least squares method is used to fit the relationship between temperature and time. This model is based on the principle of linear regression, and determines the optimal fitting parameters by minimizing the squared error between the observed values ​​and the fitted values. The coefficients of the linear term in the model are... The constant term coefficient reflects the acceleration characteristics of temperature change. It reflects the initial rate of temperature change, and the error term. Taking into account measurement uncertainties and biases caused by model simplification, this model can capture the temporal characteristics of temperature changes in the thermal storage medium compared to traditional single-point temperature monitoring methods, thus improving the accuracy and reliability of temperature prediction.

[0068] Thermal storage capacity prediction function Based on the principle of multiple linear regression, and comprehensively considering the influence of historical power grid load data and the temperature change rate of thermal storage medium, this function integrates multiple influencing factors into a predictive model through weighted summation, with weight coefficients... and This reflects the importance of each influencing factor; the bias term This function is used to correct systematic biases in the model. Compared with traditional prediction methods based on a single factor, it can comprehensively consider multiple influencing factors, thus improving the accuracy and adaptability of thermal storage capacity demand prediction.

[0069] Formula for calculating the adjustment coefficient of the number of memory fragments Based on the dynamic characteristics of the thermal storage medium and the load variation characteristics of the thermal power plant, this formula combines the time-scale characteristics of the thermal storage medium with the operating characteristics of the thermal power plant, and achieves adaptive adjustment under different operating conditions through normalization. The numerator term... This reflects the inherent time constant of the thermal storage system and the magnitude of external load disturbances; the denominator term... It provides a standardized benchmark. Compared with the traditional fixed parameter setting method, this formula can dynamically adjust the model parameters according to the actual operating conditions, which improves the adaptability and control accuracy of the peak shaving optimization model.

[0070] Thermal storage medium power matching relationship model Based on the principles of energy conservation and power balance, a mathematical relationship was established between the thermal storage medium and the main equipment of the thermal power plant. This model, through efficiency coefficients... and By taking into account the losses during the energy conversion process, this model ensures power coordination between the thermal storage process and the power generation process. Compared with the traditional independent control method, this model realizes the integrated optimization of the thermal storage system and the operation of the thermal power plant, improving the overall system's operating efficiency and stability.

[0071] Formula for calculating the heat release rate of the heat storage medium And absorption rate calculation formula Based on Fourier's law of thermal conductivity, the heat transfer rate is calculated by multiplying the temperature gradient and the heat flux density. This calculation method takes into account the spatial temperature distribution characteristics of the thermal storage medium and can accurately reflect the heat exchange process between the thermal storage medium and the outside world. Compared with the traditional average temperature calculation method, this formula can describe the local heat transfer characteristics of the thermal storage medium and provide data support for the precise control of the thermal storage system.

[0072] Formula for calculating thermal storage depth And the formula for calculating the heat release depth Based on the first law of thermodynamics and the definition of specific heat capacity, the energy storage and release capacity of the thermal storage medium is calculated by multiplying density, volume, specific heat capacity and temperature change. This calculation method directly reflects the thermal capacity characteristics of the thermal storage medium. Compared with the traditional qualitative evaluation method, this formula can quantitatively calculate the energy storage depth of the thermal storage medium, providing accurate parameter basis for peak-shaving operation of thermal power plants.

[0073] Formula for calculating thermal response time parameters of thermal storage medium Based on the theory of thermal diffusion, the thermal response characteristics of the thermal storage medium are calculated by the ratio of the square of the characteristic length to the thermal diffusivity. This formula reflects the time required for the thermal storage medium to reach thermal equilibrium, where the characteristic length term... The influence of heat transfer distance, the thermal diffusivity term The formula determines the influence of heat transfer rate. Compared with traditional empirical estimation methods, this formula can accurately predict the dynamic response characteristics of the thermal storage medium, providing theoretical guidance for the time-series control of thermal storage systems.

[0074] It should be noted that the variables involved in this invention are explained in detail in Table 1 below.

[0075] Table 1. Variable Explanation Table

[0076] To better understand and implement this invention, the following is a specific application scenario example 2: A technical team is implementing a high-temperature thermal energy storage peak-shaving retrofit project in a 300MW coal-fired power plant. The main steam temperature of the power plant's boiler system is 565℃, and the main steam pressure is 16.67MPa, facing a peak-to-valley load difference of 150MW in the power grid. The technical team designed and installed a 2000... This is a composite high-temperature thermal energy storage system. The thermal storage medium is a combination of high-temperature molten salt thermal storage medium, high-temperature ceramic particle thermal storage medium, and high-temperature metal alloy thermal storage medium, in proportions of 30%, 40%, and 30%, respectively. The system operates on the principle that coal-fired power plants generate high-temperature flue gas from coal combustion. During off-peak periods, a portion of this flue gas is extracted through a flue gas diversion pipeline and transported to a thermal storage tank for heat exchange with the thermal storage medium. During peak periods, the heat released from the storage tank heats feedwater through a heat exchanger to generate steam. This high-temperature, high-pressure steam drives an auxiliary turbine to generate electricity, providing auxiliary power to the grid. The high-temperature thermal energy storage peak-shaving system deployed in this embodiment is as follows: Figure 2 As shown.

[0077] The technical team first established a mathematical model of the temperature field of the thermal storage medium, and determined the physical property parameters of the thermal storage medium based on the steam and flue gas parameters of the boiler system in the thermal power plant. The flue gas temperature generated by the boiler burning coal is 650℃, and the flue gas composition mainly includes... , , and a small amount The smoke density is 0.68. Specific heat capacity is 1150 The technical team designed a flue gas diversion pipeline system with an inner diameter of 1.2m. The pipeline is made of high-temperature resistant stainless steel and includes flue gas flow regulating valves and temperature monitoring devices. The high-temperature molten salt heat storage medium has a density of 1800. Specific heat capacity is 1650 The thermal conductivity is 0.85. The density of the high-temperature ceramic particle heat storage medium is 3200. Specific heat capacity is 1100 Its thermal conductivity is 2.1. The density of the high-temperature metal alloy heat storage medium is 7200. Specific heat capacity is 520 The thermal conductivity is 68. The technical team used the finite element method to mesh the thermal storage medium space, setting the mesh size to 0.5m, and established a three-dimensional temperature distribution field numerical calculation model. The thermal storage medium container has dimensions of 20m in length, 15m in width, and 6.67m in height, with a characteristic length of 6.67m.

[0078] An array of temperature sensors was deployed inside the thermal storage medium. Based on the expected temperature gradient distribution, the technical team arranged sensors at a density of 16 monitoring points per cubic meter in areas with large temperature gradients and at a density of 8 monitoring points per cubic meter in areas with small temperature gradients, for a total of 2400 temperature sensors. The temperature sensors are platinum resistance thermometers with a measurement accuracy of ±0.3℃ and a response time of 0.8 seconds. The technical team established a mathematical model of the temperature change rate and used a 30-minute sliding window to perform least squares fitting on the real-time temperature data to obtain the temperature change rate parameters and temperature gradient parameters for each measurement point.

[0079] The technical team obtained grid load forecast data for the next 24 hours from the grid dispatch center, with a data sampling interval of 15 minutes, totaling 96 data points. As shown in Table 2, the grid load forecast data displays the typical daily load variation pattern.

[0080] Table 2 Typical Daily Power Grid Load Forecast Data

[0081] Based on grid load forecast data and the temperature change rate parameter of the thermal storage medium, the technical team established a mathematical model for predicting thermal storage capacity demand. The model uses load data over 24 historical time steps and temperature change rate data from 48 temperature measurement points. Weighting coefficients and bias terms were determined through training. The weighting coefficients for the grid load forecast data range from 0.05 to 0.15, and the weighting coefficients for the thermal storage medium temperature change rate range from 0.8 to 1.2. The model bias term is 38 MJ. The model outputs the thermal storage capacity demand value for the next 8 hours and the thermal storage time window parameters.

[0082] The technical team designed a thermal energy storage peak-shaving optimization model based on the Transformer-XL architecture. This model includes 12 layers of attention and memory mechanisms. The model is based on parameters such as the thermal storage medium temperature change period (86400 seconds), thermal response time (1350 seconds), power plant load variation (150MW), and thermal inertia coefficient (2.8×). The calculated adjustment factor for the number of memory segments is 0.45. The system adopts a medium memory segment count mode, with the memory segment count parameter set to 64. Figure 3 As shown, the temperature distribution field of the thermal storage medium exhibits a clear periodic pattern over time, with the temperature peak occurring between 14:00 and 16:00 and the temperature trough occurring between 02:00 and 04:00.

[0083] The technical team established a power matching model for the thermal storage medium, setting the turbine efficiency coefficient to 0.88 and the steam system efficiency coefficient to 0.93. During peak grid load periods, the thermal storage tank releases stored heat to heat the feedwater system, generating high-temperature, high-pressure steam at 485℃ and 12.5MPa to drive an 85MW auxiliary turbine for power generation. The thermal storage medium heating power requirement is 85MW, the thermal storage medium cooling power requirement is 0MW, the main turbine load variation is -75MW, and the main steam power variation is -80MW. During off-peak grid load periods, the technical team uses a flue gas diversion system to introduce high-temperature flue gas from the boiler into the thermal storage tank. The flue gas temperature is 650℃, and the flow rate is 45... After heat exchange with the thermal storage medium, the temperature drops to 280℃, while the temperature of the thermal storage medium rises from the initial 320℃ to 560℃. The heating power requirement of the thermal storage medium is 0MW, the cooling power requirement is 95MW, the main turbine load change is +85MW, and the main steam power change is +90MW. The main steam flow rate change parameter is -12 in thermal storage mode. In exothermic mode, it is +15. .

[0084] like Figure 4 As shown, there is a strong correlation between the grid load forecast data and the thermal storage capacity demand, with a correlation coefficient reaching 0.92. The technical team used a power matching relationship model of the thermal storage medium to calculate the heat release rate and heat absorption rate of the thermal storage medium. The average temperature gradient parameter of the thermal storage medium is 8.5 K / m during the thermal storage process and 6.2 K / m during the heat release process. The heat flux density distribution field data shows that the surface heat flux density of the thermal storage medium is 2800 K / m under the thermal storage mode. In exothermic mode, it is 2200. The heat release rate parameter of the thermal storage medium is 18.7 kJ / m³. The heat absorption rate parameter of the heat storage medium is 23.8kJ / m³. The optimal operating temperature range parameters for the thermal storage medium are set as follows: 280℃ to 560℃ for high-temperature molten salt thermal storage medium, 800℃ to 1200℃ for high-temperature ceramic particle thermal storage medium, and 600℃ to 900℃ for high-temperature metal alloy thermal storage medium.

[0085] As shown in Table 3, the operating parameters and performance indicators of different thermal storage media types reflect their respective technical characteristics.

[0086] Table 3 Comparison of Operating Parameters for Different Types of Thermal Storage Media

[0087] Based on the optimal operating temperature range parameters of the thermal storage medium and real-time load demand data of the thermal power plant, the technical team calculated the thermal storage depth and heat release depth parameters of the thermal storage medium. The thermal storage depth parameter reaches 2850 MJ during off-peak load periods, and the heat release depth parameter reaches 2650 MJ during peak load periods. The temperature change range parameter of the thermal storage medium is 85 K during the thermal storage process, and the temperature decrease parameter is 78 K during the heat release process. The thermal capacity parameter of the thermal storage medium is obtained as 4.2 × (the calculated value is missing from the original text). J / K. The thermal response time parameter of the thermal storage medium is 1350 seconds, and the thermal inertia coefficient parameter of the thermal storage medium is 2.8 × .

[0088] like Figure 5 As shown, the power matching relationship of the thermal storage medium maintains a good balance under different operating modes. The thermal storage parameters for peak-shaving operation of the thermal power plant are set as follows: flue gas diversion power 95MW, storage duration 4.5 hours, and storage capacity 2850MJ. The heat release parameters for peak-shaving operation of the thermal power plant are set as follows: auxiliary power generation power 85MW, heat release duration 3.8 hours, and heat release capacity 2650MJ. Based on the output peak-shaving operation parameters, operators adjust equipment parameters such as the number of boiler burners in operation, the opening of the flue gas diversion valve, the opening of the auxiliary turbine regulating valve, and the feedwater pump speed to achieve flexible peak-shaving operation of the thermal power plant. The average flue gas diversion volume during the thermal storage process is 162... The flue gas temperature dropped from 650℃ to 280℃, the heat exchange efficiency reached 87%, and the temperature gradient of the heat storage medium remained at 8.5 K / m under flue gas heating. The steam flow rate of the auxiliary power generation system during the heat release process was 78... The steam temperature is 485℃, the auxiliary steam turbine has a power generation efficiency of 91%, and it can achieve an auxiliary power output of 85MW.

[0089] On a typical operating day, the thermal power plant successfully achieved a peak-shaving depth of 150MW through the thermal storage system, reducing the peak-shaving response time to 18 minutes and extending the peak-shaving duration to 8 hours. As shown in Table 4, the operating data of the thermal storage system under different peak-shaving conditions validated the effectiveness of the technical solution.

[0090] Table 4 Peak Shaving Operation Data of Thermal Storage System

[0091] After the thermal storage system was put into operation, the peak-shaving operation parameters of the thermal power plant exhibited good dynamic response characteristics, with all parameters changing smoothly and responding rapidly. The temperature distribution of the thermal storage medium remained uniform during the peak-shaving process, and the temperature gradient was controlled within a reasonable range, avoiding local overheating or undercooling. The system's operational stability and reliability were significantly improved, and equipment utilization and energy conversion efficiency were markedly enhanced.

[0092] This invention represents a significant technological advancement over traditional peak-shaving methods in thermal power plants. Traditional peak-shaving methods primarily rely on boiler combustion regulation and turbine load adjustment, resulting in slow response times, limited peak-shaving depth, and severe equipment wear. This invention, by introducing a high-temperature thermal energy storage system, achieves coordinated and optimized control of the thermal storage medium and thermal power plant equipment. The thermal storage medium acts as a heat buffer, effectively alleviating the pressure of rapid load changes in the boiler and turbine. The optimization model based on the Transformer-XL architecture possesses powerful time-series modeling capabilities, accurately predicting grid load changes and proactively formulating thermal storage strategies, avoiding the lag inherent in traditional passive regulation modes. The coordinated configuration of multiple types of thermal storage media fully leverages the advantages of different materials: high-temperature molten salt thermal storage provides rapid response, high-temperature ceramic particle thermal storage provides large-capacity long-term thermal storage, and high-temperature metal alloy thermal storage provides moderate response and uniform heat transfer. The synergistic effect of these three materials achieves optimal performance of the thermal storage system across all operating conditions. The temperature field mathematical model established by the Fourier thermal conductivity differential equation accurately describes the heat transfer process inside the thermal storage medium, providing a theoretical basis for precise control of the temperature distribution of the thermal storage medium and avoiding the uncertainties of traditional empirical control methods. The thermal storage medium power matching relationship model establishes a precise mathematical relationship between the thermal storage system and the thermal power plant equipment, realizing precise balance and optimal allocation of energy flow, and improving the overall system's operating efficiency and economy.

[0093] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A peak-shaving method for thermal power plants utilizing high-temperature thermal energy storage, characterized in that, include: Based on the steam temperature and pressure parameters of the boiler system in a thermal power plant, a mathematical model of the temperature field of the thermal storage medium in a high-temperature thermal energy storage system is established. This model uses the Fourier differential equation of thermal conductivity to describe the temperature distribution within the thermal storage medium. The temperature distribution field and heat flux density distribution field data of the thermal storage medium are calculated using its density, specific heat capacity, and thermal conductivity parameters. The thermal storage medium comprises high-temperature molten salt, high-temperature ceramic particles, and high-temperature metal alloys to form a layered thermal storage structure. The density of the high-temperature molten salt thermal storage medium is 1800... Specific heat capacity is 1650 The thermal conductivity is 0.

85. The density of the high-temperature ceramic particle heat storage medium is 3200. Specific heat capacity is 1100 Its thermal conductivity is 2.

1. The density of the high-temperature metal alloy heat storage medium is 7200. Specific heat capacity is 520 The thermal conductivity is 68. ; Based on the expected temperature gradient distribution, sensors are arranged at a density of 16 monitoring points per cubic meter in areas with a large expected temperature gradient of the thermal storage medium, and at a density of 8 monitoring points per cubic meter in areas with a small expected temperature gradient. A temperature sensor array arranged inside the thermal storage medium is used to collect real-time temperature data of the thermal storage medium. A mathematical model of the temperature change rate of the thermal storage medium is established through time series analysis of the temperature data. The mathematical model of the temperature change rate of the thermal storage medium processes the real-time temperature data through the sliding window method and calculates the temperature change rate parameter and temperature gradient parameter of the thermal storage medium at different locations, so as to realize real-time monitoring of the temperature status of the thermal storage medium and provide temperature change information for peak shaving control decision-making. A mathematical model for predicting thermal storage capacity demand is established based on power grid load forecast data and thermal storage medium temperature change rate parameters. Input the target temperature distribution parameters of the thermal storage medium and the thermal storage time window parameters into the thermal energy storage peak shaving optimization model to calculate the heating power demand parameters and cooling power demand parameters of the thermal storage medium. Based on the heating time parameters and cooling time parameters of the thermal storage medium, calculate the main steam flow regulation parameters and turbine load regulation parameters of the thermal power plant; Using the power matching relationship model of the thermal storage medium, the heat release rate parameter and the heat absorption rate parameter of the thermal storage medium are calculated, and the optimal operating temperature range parameter of the thermal storage medium is output. Based on the optimal operating temperature range parameters of the thermal storage medium and the real-time load demand data of the thermal power plant, the thermal storage depth parameters and the thermal release depth parameters of the thermal storage medium are calculated. Operators adjust the corresponding equipment parameters according to the calculation results to achieve peak shaving.

2. The peak-shaving method for thermal power plants utilizing high-temperature thermal energy storage according to claim 1, characterized in that, High-temperature molten salt thermal storage media are classified based on their melting point being below 600℃ and their specific heat capacity being greater than 1500 J / kg·K. High-temperature ceramic particle thermal storage media are classified based on their melting point being above 1000℃ and their density being greater than 3000 kg / kJ. The conditions for obtaining.

3. The peak-shaving method for thermal power plants utilizing high-temperature thermal energy storage according to claim 1, characterized in that, The mathematical model of the temperature field of the thermal storage medium is established using the Fourier thermal conductivity differential equation. It includes numerical calculations of the temperature distribution field data and heat flux density distribution field data of the thermal storage medium. The numerical simulation of the heat transfer process of the thermal storage medium is achieved through grid generation and boundary condition setting.

Images