A prediction cooling load correction method and device based on a pre-trained large language model

Abstract

This application discloses a method and device for predicting and correcting cooling load based on a pre-trained large language model, applied to a high-speed rail transfer hub. The high-speed rail transfer hub structure includes a double-layer coaxial shaft and a cable trench. The method relates to the field of cooling load correction and includes: acquiring sample multi-source data of the high-speed rail transfer hub during a set time period, specifically including: air density, average wind speed, temperature and humidity parameters, specific heat at constant pressure of air within the station, latent heat of vaporization of air within the station, reference volumetric flow rate of fans in the underground high-speed rail transfer hub, and baseline cooling load of the underground high-speed rail transfer hub; calculating the theoretical true cooling load sequence based on the sample multi-source data of the set time period and the equivalent height of the double-layer coaxial shaft; inputting the historical cooling load sequence into the pre-trained large language model to obtain the predicted cooling load sequence, and dynamically correcting it using the theoretical true cooling load sequence to obtain the corrected predicted cooling load sequence. This application can improve the accuracy of the predicted cooling load sequence.

Description

Technical Field

[0001] This application relates to the field of cooling load correction technology, and in particular to a method and device for predicting cooling load correction based on a pre-trained large language model. Background Technology

[0002] With the rapid development of underground rail transit networks, the spatial structure of high-speed rail transfer hubs is becoming increasingly complex. Especially in traction substations and equipment areas, a composite ventilation structure consisting of double-layer coaxial shafts and cable trenches is commonly used to integrate cable laying and smoke extraction ventilation functions within a limited space. During peak train arrival and departure times in the morning and evening, the combined effect of train piston wind and the heat storage effect of the enclosure structure creates a significant vertical temperature and pressure difference. At this time, the air density distribution between the ground and underground spaces changes. If the outside is cold and the inside is hot, and the shaft is high with low resistance along the path, cold air from the ground will sink into the underground space through the shaft, while hot and humid air inside the station will flow back and stagnate in the equipment layer and cable trench area, creating a reverse chimney effect. This effect significantly alters the local thermal and humidity environment of the traction substation, causing dynamic fluctuations in the operating load of the ventilation system, chiller system, and power distribution equipment, affecting overall energy consumption stability and peak load patterns.

[0003] Current load forecasting methods primarily rely on fixed meteorological variables and historical energy consumption data, using traditional models or deep learning algorithms for prediction. These methods often assume that environmental disturbances are statistically stationary, failing to identify sudden load drift problems caused by the geometry of underground complex ventilation systems, air density gradients, and flow direction reversals. When the reverse chimney effect occurs, the actual load curve exhibits abnormal bulges or shifts in peak sections, leading to systematic errors in the prediction model output. Summary of the Invention

[0004] The purpose of this application is to provide a method and device for predicting and correcting cold load based on a pre-trained large language model, which can improve the accuracy of predicting cold load sequences.

[0005] To achieve the above objectives, this application provides the following solution.

[0006] Firstly, this application provides a method for predicting cold load correction based on a pre-trained large language model, applied to a high-speed rail transfer hub. The high-speed rail transfer hub's structure includes a double-layer coaxial shaft and a cable trench. The method for predicting cold load correction based on a pre-trained large language model includes: Acquire sample multi-source data of high-speed rail transfer hubs during a set time period; the sample multi-source data includes: air density, average wind speed, temperature and humidity parameters, specific heat of air at constant pressure in the station, latent heat of vaporization of air in the station, reference volumetric flow rate of fans in underground high-speed rail transfer hubs, and baseline cooling load of underground high-speed rail transfer hubs. Based on multi-source sample data for a set time period and the equivalent height of the double-layer coaxial shaft, the theoretical real cooling load sequence is calculated. Obtain the historical cold load sequence of high-speed rail transfer hubs; The historical cold load sequence is input into a pre-trained large language model to obtain a predicted cold load sequence; the pre-trained large language model is obtained by training a Transformer neural network based on the sample cold load sequence. The predicted cooling load sequence is dynamically corrected based on the theoretical real cooling load sequence to obtain the corrected predicted cooling load sequence.

[0007] Secondly, this application provides a computer device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-described method for predicting cold load correction based on a pre-trained large language model.

[0008] According to the specific embodiments provided in this application, this application has the following technical effects: This application introduces multi-source sample data for a set period of time, based on the reverse chimney effect formed by the structure of high-speed rail transfer hubs, including double-layer coaxial shafts and cable trenches. Specifically, this data includes air density, average wind speed, temperature and humidity parameters, specific heat at constant pressure of air within the station, latent heat of vaporization of air within the station, reference volumetric flow rate of fans in underground high-speed rail transfer hubs, and baseline cooling load of underground high-speed rail transfer hubs. Based on the multi-source sample data for the set period and the equivalent height of the double-layer coaxial shafts, the theoretical true cooling load sequence is calculated. This allows for the accurate early identification of sudden load shifts that the prediction model cannot capture, presented in the form of physical quantities. Finally, the predicted cooling load sequence of the pre-trained large-scale language model is dynamically corrected, resulting in a corrected predicted cooling load sequence. This avoids continuity errors in the pre-trained large-scale language model, enables timely correction of abnormal spikes and shifts in the predicted cooling load sequence, and improves the accuracy of the predicted cooling load sequence. Therefore, it solves the problem of systematic errors in the predicted cooling load sequence output by the pre-trained large-scale language model. Attached Figure Description

[0009] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0010] Figure 1 This is a flowchart illustrating a method for predicting cold load correction based on a pre-trained large language model, as provided in this application.

[0011] Figure 2 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0012] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0013] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0014] In one exemplary embodiment, such as Figure 1 As shown, a method for predicting cold load correction based on a pre-trained large language model is provided and applied to a high-speed rail transfer hub. The structure of the high-speed rail transfer hub includes a double-layer coaxial shaft and a cable trench. The method is executed by computer equipment, specifically by a computer device such as a terminal or server alone, or by a terminal and a server together. In this embodiment, the method is described using a server as an example, including the following steps S1 to S8.

[0015] Step S1: Obtain multi-source sample data of the high-speed rail transfer hub during a set time period; the multi-source sample data includes: air density, average wind speed, temperature and humidity parameters, specific heat of air at constant pressure in the station, latent heat of vaporization of air in the station, reference volumetric flow rate of fans in the underground high-speed rail transfer hub, and baseline cooling load of the underground high-speed rail transfer hub.

[0016] As an feasible approach, air density includes: air density outside the station and air density inside the station; average wind speed includes: average wind speed in the shaft of the double-layer coaxial shaft and average wind speed in the cable trench ventilation channel; temperature and humidity parameters include: air temperature in the cable trench, relative humidity in the cable trench, air temperature in the station hall, and relative humidity in the station hall.

[0017] The spatial location outside the station (corresponding to the area of ​​the air density measurement point outside the station) is the outdoor atmospheric area outside the top opening of the double-layer coaxial shaft exposed above the ground, belonging to the open outdoor space outside the underground hub building envelope. The double-layer coaxial shaft is a ventilation shaft that vertically connects the ground and the underground traction substation and cable trench, with the upper end of the shaft directly connected to the outdoor ground. The air outside the station is the natural atmosphere outside the shaft opening, which is the source of external cold air that forms the buoyancy pressure difference and reverse chimney effect of the shaft. During the morning and evening peak hours, the outdoor temperature is lower than that of the underground hub, and the cold air outside the station sinks through the shaft and flows into the underground cable trench and equipment area, which is the external air source that generates reverse airflow.

[0018] The spatial location within the station (corresponding to the air density measurement area within the station): The entire enclosed underground main body of the underground hub (including the double-layer coaxial vertical shaft ring passage, the interior of the cable trench, the traction substation equipment layer, underground platforms / equipment rooms, and all other underground sealed spaces) is the collective term for all enclosed areas completely buried underground within the building envelope. It includes the double-layer coaxial vertical shaft ring ventilation passage (the air circulation cavity between the inner and outer cylinders). It also includes the cable connection ventilation passage (the narrow underground ventilation trench for laying power supply cables, connected to the bottom of the vertical shaft); the air inside the vertical shaft and cable trench is collectively referred to as "station air," used to calculate fluid parameters such as frictional pressure drop, buoyancy pressure difference, and effective exchange rate along the vertical shaft and cable trench.

[0019] Spatial location of the station hall (corresponding to the area for measuring air temperature and humidity parameters in the station hall): The independent passenger public hall on the first basement floor of the station (a functional area for passenger flow within the station, separated from the cable trench and vertical shaft equipment area); The passenger public waiting / transfer hall of the underground high-speed rail transfer hub (the station hall level on the first basement floor) is an independent functional sub-zone within the "station", physically separated from the cable trench and vertical shaft equipment area, and is a tall public space for passenger activities. Spatial separation: The station hall is a public area for passenger flow, while the cable trench and double-layer coaxial vertical shaft are located in the substation equipment mezzanine and the bottom underground floor, and the two are weakly connected through ventilation gaps and inspection holes; Heat and humidity exchange relationship: Under the reverse chimney effect, a small amount of hot and humid air from the cable trench will flow back to the station hall. Therefore, it is necessary to collect the air temperature and humidity of the cable trench and the station hall separately, and calculate the difference in moisture content and temperature between the two as the basis for calculating the additional load of sensible heat and latent heat.

[0020] Specifically, the process of obtaining the density of the air outside the station is as follows: using the ideal gas approximation method, the defined air constant is called, and the corresponding outside pressure and temperature are obtained. The outside pressure is used as the numerator of the calculation, and the product of the air constant and the outside temperature is used as the denominator. The outside air density is obtained by dividing the outside pressure by the product of the air constant and the outside temperature. The process of obtaining the density of the air inside the station is also as follows: using the ideal gas approximation method, the defined air constant is called, and the corresponding inside pressure and temperature are obtained. The inside pressure is used as the numerator of the calculation, and the product of the air constant and the inside temperature is used as the denominator. The inside air density is obtained by dividing the inside pressure by the product of the air constant and the inside temperature.

[0021] Specifically, the process for obtaining the specific heat at constant pressure and latent heat of vaporization of the air inside the station is as follows: At least three representative measuring points are evenly selected in the functional areas of the station hall and the cable connection ventilation ducts. Temperature data of the air inside the station at each measuring point is collected using temperature sensors with accuracy meeting the requirements of engineering thermal measurement. At the same time, relative humidity data of the air inside the station at each measuring point is collected synchronously using relative humidity sensors of the same accuracy. After removing abnormal fluctuations in the data, the arithmetic mean of the air temperature and relative humidity inside the station is calculated. Based on the thermodynamic properties of moist air, combined with the empirical formula for the change of the specific heat at constant pressure of dry air with temperature, and the change of the specific heat at constant pressure of water vapor with temperature and humidity... The corresponding relationship is used to substitute the obtained station air temperature and relative humidity into the mixed calculation formula of the specific heat of humid air at constant pressure, or to look up the thermodynamic property table of humid air under the corresponding temperature and humidity conditions to obtain the specific heat of air at constant pressure that can characterize the heat capacity characteristics of the station air; the steps to obtain the latent heat of vaporization of the station air are as follows: call the arithmetic mean of the station air temperature that has been calculated above, and according to the empirical formula for the change of the latent heat of vaporization of water vapor with temperature, such as the Trudeau rule, or by looking up the latent heat value of vaporization at the corresponding temperature in the thermodynamic property table of saturated water vapor, obtain the heat parameter required for the phase change of water vapor in the station air under the temperature conditions, that is, the latent heat of vaporization of the station air.

[0022] Specifically, the process for obtaining the baseline cooling load of an underground high-speed rail transfer hub is as follows: First, collect the baseline operating parameters of the underground high-speed rail transfer hub, including thermal parameters such as the thermal conductivity of the building envelope, the thickness of the insulation layer, and the heat transfer area; internal heat source parameters such as the station's personnel density, per capita sensible heat dissipation, per capita latent heat dissipation, equipment operating power, and lighting power under the design benchmark conditions; temperature and humidity parameters such as the station hall air temperature, station hall air relative humidity, outdoor design temperature, and outdoor design relative humidity; and ventilation system design parameters such as the reference volumetric flow rate of the fans. Simultaneously, it is clarified that additional influencing factors such as the reverse chimney effect and additional environmental heat and humidity exchange are excluded from this operating condition. Then, calculate the baseline cooling load of each component. The heat transfer cooling load of the building envelope is calculated by multiplying the thermal conductivity and heat transfer area of ​​the building envelope by the difference between the indoor and outdoor design temperatures. The sensible heat cooling load of personnel is calculated by multiplying the per capita sensible heat dissipation of personnel density in the station by the usable area of ​​the hub. The latent heat cooling load of personnel is calculated by multiplying the per capita latent heat dissipation of personnel density in the station by the usable area of ​​the hub. The cooling load of equipment and lighting is directly calculated using their design operating power (since this part of the energy is basically converted into heat to form a cooling load). The ventilation and infiltration cooling load is calculated by multiplying the reference volumetric flow rate of the fan, the air density in the station, the specific heat of air at constant pressure, and the indoor and outdoor design temperature difference. Finally, the heat transfer cooling load of the building envelope, the sensible heat cooling load of personnel, the latent heat cooling load of personnel, the cooling load of equipment and lighting, and the ventilation and infiltration cooling load are summed. The result is the basic cooling load that the underground high-speed rail transfer hub needs to handle under the design reference condition without additional environmental influences, which is the baseline cooling load.

[0023] Step S2: Calculate the theoretical real cooling load sequence based on sample multi-source data for a set time period and the equivalent height of the double-layer coaxial shaft.

[0024] As one feasible approach, step S2 specifically includes steps S21 to S28: Step S21: Calculate the buoyancy pressure difference based on air density and the equivalent height of the double-layer coaxial shaft.

[0025] As one feasible approach, step S21 specifically includes steps S211 to S212: Step S211: Add the first effective vertical height and the second effective vertical height of the first section of the double-layer coaxial shaft to obtain the equivalent height of the shaft.

[0026] Specifically, the double-layer coaxial shaft is a composite ventilation shaft structure applied to the traction substation area of ​​an underground high-speed rail transfer hub with ultra-deep burial and limited site space. It consists of two cylindrical structures arranged on the same central axis. The inner cylindrical structure is mainly used for laying the power supply cables of the traction substation, while the outer cylindrical structure mainly undertakes the fire smoke exhaust function of the underground hub area. The inner and outer cylindrical structures are spatially connected through locally set maintenance holes. This structure can integrate the functions of cable laying channel and fire smoke exhaust channel in a limited underground space. It is also the core carrier for the reverse chimney effect generated by the combined action of train piston wind and air heat difference during the morning and evening peak hours of the underground hub.

[0027] Step S212: Calculate the buoyancy pressure difference based on the equivalent height of the shaft, the air density outside the station, and the air density inside the station.

[0028] Specifically, the buoyancy pressure difference is obtained by multiplying the gravitational acceleration by the equivalent height of the shaft and then by the difference between the air density outside the station and the air density inside the station.

[0029] The first effective vertical height of the first section of the shaft refers to the effective length of the first section of the double-layer coaxial shaft that actually participates in air circulation and pressure transmission in the vertical direction. The second effective vertical height of the first section of the shaft refers to the effective vertical length of another part of the first section that actually participates in air circulation and pressure transmission. The equivalent height of the shaft refers to the equivalent length that can comprehensively reflect the influence of the shaft on air buoyancy in the vertical dimension after superimposing the two effective vertical heights of the first section of the shaft. The external air density refers to the mass of air per unit volume in the external environment of the underground hub where the double-layer coaxial shaft is located. It is determined by factors such as temperature and pressure in the external environment. The internal air density refers to the mass of air per unit volume in the internal space of the underground hub. It is determined by environmental parameters such as temperature and humidity in the internal space. The buoyancy pressure difference refers to the air pressure difference obtained by multiplying the gravitational acceleration by the equivalent height of the shaft and then by the difference between the external air density and the internal air density under the action of gravity. This difference is the core parameter related to the natural force driving the air flow in the double-layer coaxial shaft.

[0030] Among them, the double-layer coaxial shaft is a composite ventilation shaft structure applied to the traction substation area of ​​an underground high-speed rail transfer hub with ultra-deep burial and limited site space. Its vertical part is not a single continuous height but is composed of segmented effective vertical heights. The buoyancy pressure difference is the core natural force parameter driving the reverse chimney effect within this structure. Its value is directly related to the actual length of the shaft participating in the air buoyancy effect in the vertical dimension. If only the effective vertical height of a single segment of the shaft is used to calculate the buoyancy pressure difference, it will not be able to fully and accurately reflect the contribution of the entire double-layer coaxial shaft in the vertical direction to the buoyancy effect generated by the combination of the air density difference and gravity effect inside and outside the station. At the same time, combined with the specific characteristics of this structure, it can be seen that the occurrence and strength of the reverse chimney effect are determined by the effective height of the double-layer coaxial section. The vertical shaft is determined by structural parameters, and the equivalent height of the shaft is a parameter that comprehensively represents the actual length of the vertical dimension of the entire shaft by superimposing the effective vertical heights of the segments. The buoyancy pressure difference is calculated based on the equivalent height of the shaft, which can accurately obtain the magnitude of the buoyancy force driving the airflow in the double-layer coaxial shaft. This allows for a precise determination of whether the total friction loss of the composite channel is sufficient to offset the buoyancy pressure difference, thereby clarifying the occurrence and intensity of the reverse chimney effect. If the buoyancy pressure difference is not calculated based on the equivalent height of the shaft, the vertical characteristics of the structure and the buoyancy driving effect cannot be accurately correlated, which will lead to the inability to accurately integrate the correlation between the structure and the heat and humidity load. Ultimately, this will result in a systematic underestimation or overestimation of the predicted cooling load, affecting the prior arrangement and on-site scheduling of ventilation and cooling strategies.

[0031] The reverse chimney effect is an abnormal airflow phenomenon observed in the composite ventilation shaft structure of a double-layer coaxial vertical shaft cable trench in the traction substation area of ​​an underground high-speed railway transfer hub. It typically occurs during peak morning and evening hours when the piston-like airflow generated by high-density train arrivals and departures overlaps with the vertical temperature and pressure field formed by the heat storage of the station's enclosure structure. Simultaneously, there is a density difference between the air outside and inside the station due to the external cold and internal heat. Normally, this composite ventilation shaft should exhaust air from bottom to top. However, when the external air density is greater than the internal air density, and the total friction loss of the composite channel is insufficient to offset the buoyancy pressure difference formed by the equivalent height of the vertical shaft due to gravity acceleration and the density difference between the air inside and outside the station, the airflow direction reverses. Specifically, cold air from the ground sinks through the double-layer coaxial vertical shaft and cable trench into the area surrounding the equipment, while hot and humid air from inside the station flows back and stagnates in this area, forming a heat and moisture accumulation zone. This abnormal flow state alters the thermal environment within the cable trench, increasing the cable's temperature rise resistance and thus generating additional It. 2 R loss causes additional bulges or shape shifts in the peak total load curve, and forces auxiliary systems such as fans and chillers to operate under unconventional conditions. Whether it occurs and its intensity are not determined solely by meteorological conditions, but are jointly determined by structural parameters such as the effective height of the cross-section of the double-layer coaxial shaft and the size and roughness of the cable trench. It is one of the key factors causing systematic deviations in peak load forecasting.

[0032] Step S22: Based on the average wind speed, calculate the total loss of the composite channel between the ventilation channel of the double-layer coaxial shaft and the ventilation channel of the cable trench.

[0033] As one feasible approach, step S22 specifically includes steps S221 to S225: Step S221: Calculate the frictional pressure drop along the shaft section based on the shaft friction coefficient, equivalent shaft height, equivalent shaft diameter, air density in the station, and average wind speed of the double-layer coaxial shaft.

[0034] Specifically, the process of obtaining the equivalent diameter of a double-layer coaxial shaft is as follows: Using a laser rangefinder or other distance measuring equipment with the required accuracy, the inner diameter of the outer cylindrical structure (i.e., the inner diameter of the outer cylinder) and the outer diameter of the inner cylindrical structure (i.e., the outer diameter of the inner cylinder) are measured respectively. Based on the calculation method for the flow cross-sectional area of ​​an annular channel, the cross-sectional area of ​​the outer cylinder is calculated using the inner diameter, and the cross-sectional area of ​​the inner cylinder is calculated using the outer diameter. The actual flow cross-sectional area of ​​the shaft is obtained by subtracting the cross-sectional area of ​​the inner cylinder from the outer cross-sectional area. Simultaneously, according to the calculation rules for the wetted perimeter of an annular channel, the circumference of the inner wall of the outer cylinder is added to the circumference of the outer wall of the inner cylinder to obtain the wetted perimeter of the shaft. Then, following the general calculation logic for the equivalent diameter of a pipe, the actual flow cross-sectional area of ​​the shaft is multiplied by four and divided by the wetted perimeter of the shaft to obtain a uniform representation of the cross-section of the double-layer coaxial shaft. The equivalent diameter of the shaft is used to measure the impact of dimensions on airflow resistance. The process for obtaining the average wind speed of a double-layer coaxial shaft is as follows: At least three measuring points are set up at the vertical height of the double-layer coaxial shaft according to the principle of uniform distribution. At least four measuring positions are evenly arranged at each measuring point in the circumferential direction of the shaft. Using wind speed measuring equipment with real-time measurement capabilities, such as hot-wire anemometers or ultrasonic anemometers, the instantaneous wind speed data of each measuring point and each measuring position are collected synchronously under the typical working condition of trains passing through the hub during the morning and evening peak hours. After removing abnormal fluctuation data, the arithmetic mean of all valid instantaneous wind speed data is calculated. Alternatively, the volumetric flow rate of the air in the shaft during the same period is collected by the measuring equipment. The volumetric flow rate is divided by the actual flow cross-sectional area of ​​the shaft calculated above to obtain the average wind speed of the shaft, which reflects the overall airflow intensity in the shaft.

[0035] Based on the shaft friction coefficient, equivalent shaft height, equivalent shaft diameter, air density within the station, and average wind speed of the double-layer coaxial shaft, the frictional pressure drop along the shaft section is calculated. The formula for calculating the frictional pressure drop along the shaft section is as follows: .

[0036] In the formula, It is the frictional pressure drop along the shaft section. It is the vertical shaft friction coefficient. It is the equivalent height of the shaft. It is the equivalent diameter of the shaft. It is the average wind speed in the shaft. It refers to the air density inside the station.

[0037] Specifically, the vertical shaft friction coefficient refers to a quantitative parameter representing the resistance to airflow caused by the friction between the inner wall of a double-layer coaxial vertical shaft and the flowing air. Its value is related to structural characteristics such as the roughness of the inner wall and is a core indicator characterizing the degree to which the inner wall of the vertical shaft obstructs airflow. The equivalent height of the vertical shaft refers to the sum of the effective vertical heights of the segments of the double-layer coaxial vertical shaft, comprehensively reflecting the equivalent length of the vertical shaft that actually participates in airflow and pressure transmission in the vertical dimension. It is a key parameter relating the vertical structure of the shaft to the interaction with airflow. The equivalent diameter of the vertical shaft refers to the equivalent conversion of the actual cross-sectional shape of the double-layer coaxial vertical shaft into... The diameter corresponding to the circular cross-section is used to uniformly characterize the influence of the shaft cross-sectional dimensions on airflow resistance; the air density inside the station refers to the mass per unit volume of air in the internal space of the underground high-speed rail transfer hub, which is determined by environmental parameters such as temperature and humidity inside the station, and is one of the basic physical quantities for calculating pressure changes during airflow; the average wind speed in the shaft refers to the average speed of air flowing in the double-layer coaxial shaft, and is a parameter reflecting the intensity of airflow in the shaft; the frictional pressure drop along the shaft section refers to the pressure loss caused by the friction between the inner wall of the shaft and the air during the airflow in the double-layer coaxial shaft.

[0038] Specifically, the calculation of the frictional pressure drop along the shaft section is derived from the classic fluid friction loss formula, namely the Darcy-Visbach formula. This formula is a general formula characterizing the frictional pressure loss of fluids flowing in pipe-like channels, and is applicable to the calculation of frictional losses in vertical channels such as double-layer coaxial shafts. The shaft friction coefficient corresponds to the frictional resistance coefficient in the Darcy-Visbach formula. It is used to quantify the influence of the roughness of the shaft wall and the airflow state on the frictional resistance and is a core parameter reflecting the intensity of friction between the wall and the air. The equivalent height of the shaft corresponds to the channel length in the Darcy-Visbach formula. It is the vertical spatial range in which friction occurs when air flows in the shaft. The longer the channel length, the higher the cumulative degree of friction between the air and the wall, and the greater the corresponding pressure loss. The equivalent diameter of the shaft corresponds to the channel diameter in the Darcy-Visbach formula. It is the characteristic dimension after the actual cross-section of the double-layer coaxial shaft is equivalent to a circular cross-section, used to unify... The influence of channel cross-sectional constraints on frictional resistance is characterized by a smaller equivalent diameter, stronger wall constraint during airflow, and greater frictional loss per unit length. The air density within the station corresponds to the fluid density in Darcy-Visbach's formula. As the flowing medium, the greater the air density, the greater the pressure loss corresponding to the momentum loss caused by friction during flow. The average wind speed in the shaft corresponds to the fluid velocity in Darcy-Visbach's formula. The greater the air velocity, the higher the relative motion intensity between the air and the shaft wall, and the more significant the energy loss caused by friction. Furthermore, the pressure loss is proportional to the square of the velocity, which is because the change in fluid kinetic energy is positively correlated with the square of the velocity. Combining the ratio of the shaft friction coefficient, the equivalent height of the shaft, and the equivalent diameter of the shaft, and then multiplying it by half of the air kinetic energy term, which is the product of the air density in the station and the square of the average wind speed in the shaft, follows the physical logic of Darcy-Visbach's formula. That is, the frictional pressure drop along the shaft is directly proportional to the friction coefficient and the channel length, inversely proportional to the equivalent diameter, and directly proportional to the kinetic energy of the fluid. The result obtained by combining these factors can accurately quantify the pressure loss caused by the continuous friction of the walls during the air flow in the double-layer coaxial shaft, that is, the frictional pressure drop along the shaft section.

[0039] Step S222: Calculate the frictional pressure drop along the cable trench section based on the cable trench friction coefficient, cable trench length, and cable trench equivalent height of the ventilation channel of the cable trench.

[0040] Specifically, the process for obtaining the friction coefficient of the cable trench in the cable communication ventilation channel is as follows: First, at least five measuring points are evenly selected in different areas of the inner wall of the cable trench using a surface roughness measuring instrument. The arithmetic mean roughness value of each measuring point is measured and its average value is calculated. At the same time, the Reynolds number of air flowing in the cable trench is calculated using the average wind speed, air density, and aerodynamic viscosity of the cable trench obtained in the previous step. Combined with the material type of the inner wall of the cable trench, and based on the empirical formula for the friction coefficient of the corresponding material in fluid mechanics, such as the Nikolaiz formula, the average roughness value obtained from the above measurement and the calculated Reynolds number are substituted to calculate the friction coefficient of the cable trench under the corresponding morning and evening peak ventilation conditions. The process for obtaining the length of the cable trench is as follows: The length is measured along the actual ventilation flow path of the cable trench using a laser rangefinder, starting from the connection between the cable trench and the double-layer coaxial vertical shaft. From the starting point to the end outlet of the cable trench, the path is divided into several straight or slightly inclined segments. The length of each segment is measured and summed to obtain the actual path length of air flow in the cable trench, i.e., the cable trench length. The process of obtaining the equivalent height of the cable trench is as follows: the actual cross-sectional width and clearance height of the cable trench are measured with a steel ruler. The total cross-sectional area is calculated by multiplying the width and clearance height. The projected area of ​​obstacles such as cables and supports laid in the cable trench is measured and summed. The actual flow cross-sectional area of ​​the cable trench is obtained by subtracting the projected area of ​​the obstacle from the total cross-sectional area. At the same time, the length of the wall surface in contact with air in the cross-section of the cable trench is measured, i.e., the wetted perimeter. According to the general calculation rules of the equivalent height of the channel, the actual flow cross-sectional area is multiplied by four and then divided by the wetted perimeter to obtain the equivalent height of the cable trench that can characterize the influence of the cable trench cross-sectional constraints on the air flow resistance.

[0041] Based on the cable trench friction coefficient, cable trench length, and equivalent height of the cable trench, the frictional pressure drop along the cable trench section is calculated. The formula for calculating the frictional pressure drop along the cable trench section is as follows: .

[0042] In the formula, It is the frictional voltage drop along the cable trench section. It is the coefficient of friction of the cable trench. It is the length of the cable trench. It is the equivalent height of the cable trench. It refers to the air density inside the station. It is the average wind speed in the cable trench.

[0043] Specifically, the cable trench friction coefficient refers to a quantitative parameter representing the resistance to airflow caused by friction between the inner wall of the cable trench and the flowing air. Its value is related to the roughness of the inner wall and the airflow state, and it is a core indicator reflecting the intensity of friction between the cable trench wall and the air. The cable trench length refers to the path length traveled by the air flowing within the cable trench, representing the spatial range where friction occurs between the air and the cable trench wall. The longer the length, the higher the cumulative degree of friction, and the greater the corresponding pressure loss. The equivalent height of the cable trench refers to the characteristic height corresponding to the equivalent conversion of the actual cross-sectional shape of the cable trench into a regular cross-section, used to uniformly characterize the cable trench cross-section. The influence of size on airflow resistance: the smaller the equivalent height, the stronger the constraint of the wall on airflow, and the greater the friction loss per unit length; the air density inside the station refers to the mass per unit volume of air in the cable trench area inside the underground high-speed rail transfer hub, which is determined by environmental parameters such as temperature and humidity inside the station, and is one of the basic physical quantities for calculating pressure changes during airflow; the average wind speed in the cable trench refers to the average speed of air flowing in the cable trench, and is a parameter reflecting the intensity of airflow in the cable trench; the frictional pressure drop along the cable trench section refers to the pressure loss caused by the continuous friction between the inner wall of the cable trench and the air during airflow in the cable trench.

[0044] Specifically, the calculation of the frictional pressure drop along the cable trench section is derived from the classic Darcy-Visbach formula in the field of fluid flow. This formula is a general physical formula for quantifying the frictional pressure loss of fluids flowing in pipe-like channels, and is applicable to the calculation of frictional losses in horizontal or slightly inclined channels such as cable trenches. The cable trench friction coefficient corresponds to the frictional resistance coefficient in the Darcy-Visbach formula. It is used to characterize the influence of the roughness of the inner wall of the cable trench and the air flow state on the frictional resistance, and is a core parameter reflecting the intensity of friction between the wall and the air. The cable trench length corresponds to the channel length in the Darcy-Visbach formula. It represents the spatial path range of frictional action when air flows in the cable trench. The longer the channel length, the more significant the cumulative effect of friction between the air and the wall, and the greater the corresponding pressure loss. The equivalent height of the cable trench corresponds to the channel characteristic dimension in the Darcy-Visbach formula. It is the characteristic height after converting the actual cross-sectional shape of the cable trench into a regular cross-section, used to uniformly characterize... The influence of cable trench cross-sectional constraints on airflow resistance is as follows: the smaller the equivalent height, the stronger the constraint effect of the wall on airflow, and the greater the frictional loss per unit length; the air density in the station corresponds to the fluid density in Darcy-Visbach's formula. As the flow medium, the greater the density of air, the greater the pressure loss corresponding to the momentum loss caused by friction during flow; the average wind speed in the cable trench corresponds to the fluid velocity in Darcy-Visbach's formula. The greater the air velocity, the higher the relative motion intensity between the air and the cable trench wall, and the more obvious the energy loss caused by friction. Moreover, the pressure loss is proportional to the square of the velocity, because the change in the kinetic energy of the fluid is positively correlated with the square of the velocity. Combining the ratio of the cable trench friction coefficient, the cable trench length, and the cable trench equivalent height, and then multiplying it by half of the air kinetic energy term, which is the product of the air density in the station and the square of the average wind speed in the cable trench, follows the physical logic of Darcy-Visbach's formula. That is, the frictional pressure drop along the cable trench is directly proportional to the friction coefficient and the length of the channel, inversely proportional to the equivalent height, and directly proportional to the kinetic energy of the fluid. The result obtained by combining these factors can accurately quantify the pressure loss caused by the continuous friction of the wall during the flow of air in the cable trench, that is, the frictional pressure drop along the cable trench section.

[0045] Step S223: Add the frictional pressure drop along the shaft section and the frictional pressure drop along the cable trench section to obtain the total loss of the composite channel.

[0046] Step S23: Calculate the reverse strength index based on the buoyancy pressure difference and the total loss of the composite channel.

[0047] As one feasible approach, step S23 specifically includes steps S231 to S232: Step S231: Calculate the difference between the buoyancy pressure difference and the total loss of the composite channel to obtain the buoyancy-driven net pressure difference.

[0048] Specifically, the buoyancy-driven net pressure difference refers to the difference between the buoyancy pressure difference used to drive the air flow in the composite channel and the total loss of the composite channel. It reflects the effective pressure difference that can actually drive the air flow in the composite channel. The reverse strength index refers to the ratio of the absolute value of the buoyancy-driven net pressure difference to the buoyancy pressure difference. It is a key indicator for quantifying the intensity of the reverse chimney effect in the composite ventilation shaft structure of the double-layer coaxial vertical cable trench. Through this index, the degree of reverse air flow in the composite channel can be clearly defined, thereby providing a quantitative basis for judging the heat and moisture accumulation state in the station and the additional fluctuation of peak load.

[0049] Step S232: Calculate the reverse strength index based on the net pressure difference driven by buoyancy and the pressure difference of buoyancy.

[0050] Specifically, the reverse strength index is obtained by dividing the net pressure difference driven by buoyancy by the absolute value of the buoyancy pressure difference.

[0051] Specifically, the above calculation process is derived from the dynamic and resistance balance relationship of fluid flow and the quantitative logic of reverse flow. First, the total loss of the composite channel is obtained by adding the frictional pressure drop along the shaft section and the frictional pressure drop along the cable trench section. This is because when air flows in the composite ventilation channel composed of the shaft and the cable trench, it must overcome the frictional resistance of the inner walls of the shaft and the cable trench in sequence. The total resistance loss is the superposition of the frictional losses of the two sections, which conforms to the principle of superposition of resistance losses in fluid flow. Then, the buoyancy-driven net pressure difference is calculated by the difference between the buoyancy pressure difference and the total loss of the composite channel. This is because the buoyancy pressure difference is the natural driving force for air to flow in the composite channel, and the total loss of the composite channel is the resistance that hinders the air flow. The difference between the two is... The value represents the effective pressure difference that can actually drive airflow. Its sign and magnitude directly reflect the actual direction and strength of airflow. Finally, the reverse intensity index is obtained by dividing the net pressure difference driven by buoyancy by the absolute value of the buoyancy pressure difference. This is because the direction of the buoyancy pressure difference determines the theoretical direction of airflow, while the sign of the net pressure difference driven by buoyancy corresponds to whether the actual flow direction is consistent with the theoretical direction. Dividing by the absolute value of the buoyancy pressure difference can normalize the difference into a relative index. Its value can quantify the degree to which the actual flow direction deviates from the theoretical direction, that is, the intensity of the reverse chimney effect. Therefore, the result obtained through this series of calculations can accurately characterize the intensity of reverse airflow in the composite ventilation channel, which is the reverse intensity index.

[0052] Step S24: Based on the average wind speed and reverse intensity index, calculate the effective exchange volume between the double-layer coaxial vertical shaft ventilation channel and the cable communication ventilation channel.

[0053] As one feasible approach, step S24 specifically includes steps S241 to S248: Step S241: Calculate the equivalent cross-sectional area of ​​the shaft based on the inner coaxial shaft radius of the inner shaft and the outer coaxial shaft radius of the outer shaft.

[0054] Specifically, firstly, using a laser rangefinder with accuracy meeting engineering surveying requirements, at least four measuring points are uniformly selected along the circumference of the same horizontal cross-section of the inner shaft. The distance from each measuring point to the central axis of the inner shaft is measured, i.e., the radius of the inner coaxial shaft corresponding to each measuring point. The arithmetic mean of the inner coaxial shaft radii at all measuring points is calculated as the final inner coaxial shaft radius. Simultaneously, at least four measuring points are uniformly selected along the circumference of the same horizontal cross-section of the outer shaft. The distance from each measuring point to the central axis of the outer shaft is measured, i.e., the radius of the outer coaxial shaft corresponding to each measuring point. The arithmetic mean of the inner coaxial shaft radii at all measuring points is calculated. The arithmetic mean of the radii of all measuring points on the outer coaxial shaft is taken as the final outer coaxial shaft radius. Then, according to the general formula for calculating the area of ​​a circle, the area of ​​the outer circle corresponding to the outer coaxial shaft radius and the area of ​​the inner circle corresponding to the inner coaxial shaft radius are calculated respectively. The area of ​​the outer circle is obtained by multiplying pi by the square of the final outer coaxial shaft radius, and the area of ​​the inner circle is obtained by multiplying pi by the square of the final inner coaxial shaft radius. Finally, the calculated outer circle area is subtracted from the inner circle area, and the difference is the effective cross-sectional area of ​​the airflow in the double-layer coaxial shaft, which is the equivalent cross-sectional area of ​​the shaft.

[0055] Step S242: Calculate the volumetric flow rate of the shaft based on the equivalent cross-sectional area and the average wind speed of the shaft.

[0056] Specifically, the equivalent cross-sectional area of ​​the shaft and the average wind speed of the shaft are multiplied to obtain the volumetric flow rate of the shaft.

[0057] Specifically, the inner coaxial shaft radius refers to the radial dimension of the inner shaft cylinder in a double-layer coaxial shaft, which is a fundamental parameter characterizing the spatial scale of the inner shaft cylinder; the outer coaxial shaft radius refers to the radial dimension of the outer shaft cylinder in a double-layer coaxial shaft, which, together with the inner coaxial shaft radius, constitutes the annular cross-sectional boundary of the double-layer coaxial shaft; the shaft equivalent cross-sectional area refers to the area of ​​the annular region of the double-layer coaxial shaft calculated based on the inner and outer coaxial shaft radii, which is the effective spatial cross-sectional size for air circulation within the double-layer coaxial shaft, directly determining the scale of the air circulation channel within the shaft; the shaft average wind speed refers to the average speed of air flowing within the double-layer coaxial shaft, reflecting the airflow intensity within the shaft; and the shaft volumetric flow rate refers to the physical quantity obtained by multiplying the shaft equivalent cross-sectional area by the shaft average wind speed, representing the volume of air passing through the double-layer coaxial shaft per unit time, which is a key indicator for quantifying the shaft's ventilation capacity.

[0058] Step S243: Calculate the cross-sectional area of ​​the cable trench based on the equivalent width and equivalent height of the cable trench ventilation channel.

[0059] Specifically, the cross-sectional area of ​​the cable trench is obtained by multiplying the equivalent width of the cable trench by the equivalent height of the cable trench.

[0060] Specifically, the process of obtaining the equivalent width of the cable trench is as follows: First, along the length of the cable trench ventilation channel, at least five representative measuring points are selected according to the principle of uniform distribution. Each measuring point is located in the middle area of ​​the cable trench cross-section and avoids obvious structural protrusions or areas with dense obstacles. Using a laser rangefinder or steel ruler with accuracy conforming to engineering measurement standards, the actual distance between the two ends of the inner wall of the cable trench at each measuring point is measured, which is the actual lateral dimension of the cable trench at that measuring point. Then, for each measuring point, the projected width of all obstacles such as cable bundles, supports, and fixing clamps laid in the cable trench in the lateral direction is measured one by one. The lateral projected widths of all obstacles at the same measuring point are summed to obtain the total lateral projected width of obstacles at that measuring point. Then, by subtracting the total lateral projected width of obstacles at the corresponding measuring point from the actual lateral dimension of the cable trench at each measuring point, the effective lateral flow dimension at each measuring point that allows air to flow freely is calculated. Finally, the effective lateral flow dimensions of all measuring points are arithmetically averaged to obtain the equivalent width of the cable trench that can comprehensively characterize the scale of the effective lateral flow space of the cable trench ventilation channel. This width can uniformly reflect the constraint effect of the lateral dimension of the cable trench on airflow.

[0061] Step S244: Calculate the volumetric flow rate of the cable trench based on the cross-sectional area of ​​the cable trench and the average wind speed of the cable trench.

[0062] Specifically, the volumetric flow rate of the cable trench is obtained by multiplying the cross-sectional area of ​​the cable trench and the average wind speed in the cable trench.

[0063] Step S245: Calculate the total exchange volumetric flow rate based on the vertical shaft volumetric flow rate and the cable trench volumetric flow rate.

[0064] Specifically, the volumetric flow rate of the vertical shaft and the volumetric flow rate of the cable trench are added together to obtain the total exchange volumetric flow rate.

[0065] Specifically, the equivalent width of a cable trench refers to the lateral characteristic dimension after converting the actual cross-sectional dimensions of the cable trench into a regular flow section. It is a fundamental parameter characterizing the effective lateral flow space of the cable trench. The equivalent height of a cable trench refers to the longitudinal characteristic dimension after converting the actual cross-sectional dimensions of the cable trench into a regular flow section. Together with the equivalent width, it defines the effective flow section range of the cable trench. The cross-sectional area of ​​a cable trench is a physical quantity obtained by multiplying the equivalent width and equivalent height. It represents the effective cross-sectional size of the airflow space within the cable trench and directly determines the scale of the airflow channel. The volumetric flow rate of a cable trench is a physical quantity obtained by multiplying the cross-sectional area and the average wind speed. It represents the volume of air passing through the cable trench per unit time and is a key indicator for quantifying the ventilation capacity of the cable trench. The total exchange volumetric flow rate is a physical quantity obtained by adding the volumetric flow rate of the shaft and the volumetric flow rate of the cable trench. It represents the total air exchange volume per unit time of the composite ventilation channel composed of the double-layer coaxial shaft and the cable trench, and can comprehensively reflect the overall ventilation exchange capacity of the composite channel.

[0066] Step S246: Calculate the return flow distribution coefficient based on the cable trench volumetric flow rate and the total exchange volumetric flow rate.

[0067] Specifically, the ratio of the volumetric flow rate of the cable trench to the total exchange volumetric flow rate is calculated to obtain the return flow distribution coefficient.

[0068] Step S247: Perform a minimum value operation on the reverse strength index and 0 to obtain the effective value of the reverse strength.

[0069] Step S248: Calculate the effective exchange volume based on the effective value of the reverse intensity, the total exchange volume flow rate, and the return distribution coefficient.

[0070] Specifically, the effective exchange rate is obtained by multiplying the effective value of the reverse intensity by the total exchange volume flow rate and the reflux distribution coefficient.

[0071] Specifically, the return flow distribution coefficient is a parameter calculated by comparing the volumetric flow rate of the cable trench with the total exchange volumetric flow rate. It represents the proportion of air volume that flows through the cable trench area and forms a return flow in the total air exchange volume of the composite ventilation channel composed of the double-layer coaxial shaft and the cable trench. It is the core indicator for quantifying the proportion of air return flow in the cable trench. The effective value of reverse strength is the result of taking the minimum value of the reverse strength index and 0. When the reverse strength index is negative, the index itself is taken; when the reverse strength index is positive, 0 is taken. This is used to screen out the effective strength of actual reverse flow and exclude invalid values ​​under the forward airflow state. The total exchange volumetric flow rate is the total air exchange volume of the composite ventilation channel per unit time, obtained by adding the volumetric flow rate of the shaft and the volumetric flow rate of the cable trench. It is a parameter that represents the overall ventilation exchange capacity of the composite channel. The effective exchange quantity is a physical quantity obtained by multiplying the effective value of reverse strength by the total exchange volumetric flow rate and the return flow distribution coefficient. It represents the actual volume of air participating in heat and moisture transfer in the composite ventilation channel and can accurately quantify the effective exchange scale between the hot and humid air in the station and the external air under the reverse chimney effect.

[0072] Specifically, the above calculation process is derived from the quantitative logic of airflow distribution and reverse exchange within the composite ventilation channel. First, the equivalent width and equivalent height of the cable trench are multiplied to obtain the cross-sectional area of ​​the cable trench. This cross-sectional area determines the airflow cross-sectional scale within the cable trench through the effective transverse and longitudinal dimensions. Then, this cross-sectional area is multiplied by the average wind speed of the cable trench to obtain the cable trench volumetric flow rate. This conforms to the physical law that fluid volumetric flow rate is the product of the flow cross-section and the velocity, and can characterize the volume of air flowing through the cable trench per unit time. The volumetric flow rate of the vertical shaft is added to the volumetric flow rate of the cable trench to obtain the total exchange volumetric flow rate, because both together constitute the airflow path of the composite ventilation channel, and the total flow rate represents the total air exchange scale of the channel per unit time. Finally, the ratio of the cable trench volumetric flow rate to the total exchange volumetric flow rate is calculated to obtain the return flow distribution coefficient. The reason for this is that the backflow phenomenon mainly occurs in the cable trench area. This ratio can quantify the proportion of air flowing through the cable trench and participating in the backflow in the total exchange volume. The effective value of the reverse intensity is obtained by taking the minimum value of the reverse intensity index and 0. This is because when the reverse intensity index is negative, it means that the actual air flow direction is opposite to the theoretical direction, that is, there is reverse flow. Taking the minimum value can filter out the effective intensity that only reflects the reverse flow and exclude invalid values ​​under the forward flow state. Finally, the effective value of the reverse intensity is multiplied by the total exchange volume flow rate and the backflow distribution coefficient. By combining the effective intensity of the reverse flow with the total exchange scale and the backflow proportion, the actual air volume participating in the reverse heat and moisture exchange in the composite ventilation channel is accurately quantified. This result is the effective exchange volume, which can accurately reflect the effective exchange scale between the station and the outside air under the action of the reverse chimney effect.

[0073] Step S25: Calculate the moisture content difference and sensible heat addition based on temperature and humidity parameters and air specific heat at constant pressure.

[0074] As one feasible approach, step S25 specifically includes steps S251 to S254: Step S251: Calculate the humidity of the air in the station hall based on the air temperature, relative humidity, and density in the station hall.

[0075] Specifically, the steps for obtaining the moisture content of the air in the cable trench are as follows: At least four measuring points are evenly selected along the length of the cable trench ventilation channel, avoiding areas with dense obstacles. Temperature data at each measuring point is collected using a temperature sensor with accuracy meeting engineering requirements. After removing abnormal data, the arithmetic mean is calculated to obtain the cable trench air temperature. Simultaneously, relative humidity data is collected at the same measuring point using a relative humidity sensor of the same accuracy. After removing abnormal data, the arithmetic mean is calculated to obtain the cable trench air relative humidity. The previously obtained air density within the station is used. Based on the general calculation method for humid air content—that is, by looking up the saturated water vapor partial pressure at the corresponding temperature using the cable trench air temperature—the actual water vapor partial pressure is calculated by combining it with the cable trench air relative humidity. Then, the water vapor is substituted into the humid air state equation, using the station air density as the basic parameter, to calculate the mass of water vapor contained in a unit mass of cable trench air. The steps for obtaining the humidity content of the air in the cable trench and the humidity content of the air in the station hall are as follows: At least six measuring points are evenly selected in different functional areas of the station hall. The air temperature data at each measuring point is collected using a temperature sensor of the same precision as mentioned above. After removing abnormal data, the arithmetic mean is calculated to obtain the air temperature in the station hall. At the same time, the relative humidity data of the air is collected at the same measuring point using a relative humidity sensor of the same precision. After removing abnormal data, the arithmetic mean is calculated to obtain the relative humidity in the station hall. The air density in the station obtained above is used, and the same calculation method as for the humidity content of the air in the cable trench is adopted. That is, the saturated water vapor partial pressure at the corresponding temperature is looked up by the air temperature in the station hall. The actual water vapor partial pressure is calculated by combining the relative humidity in the station hall. Then, it is substituted into the wet air state equation. The mass of water vapor contained in a unit mass of air in the station hall is calculated based on the air density in the station hall as the basic parameter to obtain the humidity content of the air in the station hall.

[0076] Step S252: Calculate the difference between the moisture content of the air in the cable trench and the moisture content of the air in the station hall to obtain the moisture content difference.

[0077] Specifically, the calculation of humid air moisture content refers to the method of calculating the mass of water vapor contained in a unit mass of air by combining parameters such as air temperature, relative humidity, and density. It is a general calculation method for quantifying air humidity. The moisture content of cable trench air refers to the mass of water vapor contained in a unit mass of cable trench air, and is a core parameter characterizing the humidity state of cable trench air. Station hall air temperature refers to the degree of hotness or coldness of the air within the station hall space, and is one of the key parameters affecting the air's water vapor carrying capacity. Station hall air relative humidity refers to the actual amount of water vapor contained in the station hall air and the maximum amount of water vapor that air can hold at the same temperature. The ratio of water vapor content is a parameter reflecting the degree of humidity saturation in the station hall air; the air density in the station refers to the mass of air per unit volume in the station hall area, and is one of the basic physical quantities in the calculation of the moisture content of humid air; the moisture content of the station hall air refers to the mass of water vapor contained in a unit mass of station hall air, and is a core parameter characterizing the humidity state of the station hall air; the moisture content difference refers to the result obtained by calculating the difference between the moisture content of the cable trench air and the moisture content of the station hall air, which reflects the degree of difference in water vapor content between the cable trench air and the station hall air, and is a key indicator for quantifying the moisture exchange potential between the two.

[0078] Step S253: Calculate the difference between the air temperature in the cable trench and the air density in the station to obtain the temperature difference.

[0079] Step S254: Calculate the sensible heat addition based on the specific heat of air at constant pressure, the air density in the station, the effective exchange rate, and the temperature difference.

[0080] Specifically, the sensible heat addition is obtained by multiplying the specific heat of air at constant pressure by the air density, effective exchange rate, and temperature difference within the station.

[0081] Specifically, the isobaric specific heat of the air inside the station refers to the amount of heat absorbed by a unit mass of air when its temperature increases by a unit value under constant pressure; it is a core parameter characterizing the heat capacity of the air inside the station. The latent heat of vaporization of the air inside the station refers to the amount of heat absorbed by a unit mass of water vapor when it changes from liquid to gas under corresponding environmental conditions; it is a parameter characterizing the phase change heat characteristics of the air inside the station. The temperature difference is the result of calculating the difference between the temperature of the air in the cable trench and the air inside the station; it characterizes the difference in the degree of hotness between the air in the cable trench and the air inside the station. The sensible heat addition is a physical quantity obtained by multiplying the isobaric specific heat of the air by the effective exchange rate of the air density inside the station and the temperature difference; it represents the additional heat transferred through the effective exchange rate due to the temperature difference between the air in the cable trench and the air inside the station, and is the heat increment in the heat load inside the station that is only related to temperature changes.

[0082] Specifically, the sensible heat gain corresponds to the increase in heat transferred by air due to temperature changes alone. Its calculation follows the physical laws of sensible heat transfer under constant pressure conditions, meaning the sensible heat is proportional to the mass of air participating in the heat exchange, the air's specific heat at constant pressure, and the temperature difference. First, the specific heat at constant pressure of the air inside the station is obtained because the station environment is under stable pressure; sensible heat calculations require specific heat parameters under constant pressure to accurately characterize the air's heat capacity. The latent heat of vaporization is obtained as a fundamental thermal parameter of this heat exchange system. Although sensible heat transfer does not involve phase change, it needs to be retained as a system thermal characteristic parameter. The temperature difference between the cable trench air temperature and the station air temperature is calculated; this difference is the temperature gradient driving sensible heat transfer and is the core driving force for sensible heat generation. Multiply the specific heat of air at constant pressure by the air density, effective heat exchange, and temperature difference within the station. The product of the air density and the effective heat exchange corresponds to the mass of air participating in heat exchange per unit time. Combining the specific heat of air at constant pressure and the temperature difference, this perfectly matches the general formula for sensible heat transfer under constant pressure conditions: sensible heat equals the mass of air multiplied by the specific heat at constant pressure and then multiplied by the temperature difference. The result calculated in this way is the additional heat transferred per unit time due to the temperature difference between the cable trench and the air within the station through the effective heat exchange, i.e., the additional sensible heat.

[0083] Step S26: Calculate the equivalent environmental additional power based on latent heat of vaporization, air density, reference volumetric flow rate, effective exchange capacity, moisture content difference, and sensible heat addition.

[0084] As one feasible approach, step S26 specifically includes steps S261 to S265: Step S261: Calculate the additional latent heat based on the latent heat of vaporization, air density in the station, effective exchange rate, and moisture content difference.

[0085] Specifically, the latent heat of vaporization is multiplied by the air density, effective exchange rate, and moisture content difference within the station to obtain the additional latent heat.

[0086] Step S262: Calculate the wind turbine power increment based on the effective exchange volume and the reference volumetric flow rate.

[0087] Specifically, the increase in the electric power of the wind turbine is calculated based on the effective exchange capacity and the reference volumetric flow rate of the wind turbine. The formula for calculating the increase in the electric power of the wind turbine is as follows: .

[0088] In the formula, It is the increase in the electrical power of the wind turbine. It is a wind turbine characteristic factor. It is the effective exchange quantity. This is the reference volumetric flow rate of the fan.

[0089] Specifically, latent heat addition refers to the additional heat generated due to the phase change process of air. It is the heat increment related to the phase change of water vapor in the heat load of the station, and it represents the additional heat transfer caused by changes in air humidity. The reference volumetric flow rate of the fan refers to the volume of air that the fan can transport per unit time under design conditions. It is the benchmark flow scale in the fan performance parameters and is used to measure the ventilation capacity of the fan under standard conditions. The fan power increment refers to the additional power consumed by the fan due to the change in ventilation volume compared with the reference conditions. It is a parameter that represents the degree of additional load on the fan. The fan characteristic factor refers to the characteristic coefficient determined by the fan's structural form, performance parameters, and operating conditions. Its value reflects the correlation between the fan power and the ventilation volumetric flow rate. It is the core parameter for quantifying the relationship between fan power and flow rate. The effective exchange volume refers to the volume of air actually involved in heat and moisture transfer in the composite ventilation channel, which is calculated by the reverse intensity effective value total exchange volumetric flow rate return distribution coefficient. It is a parameter that represents the effective exchange scale between the station and the outside air.

[0090] Specifically, the calculation of the fan's electrical power increment is derived from the fan similarity law, a core physical law describing the performance correlation of similar fans under different operating conditions. In similar operating conditions, the fan's electrical power is proportional to the cube of the ventilation volumetric flow rate. This is because the fan's shaft power is directly related to the ventilation pressure and flow rate, and the ventilation pressure is proportional to the square of the flow rate. The combination of these two factors results in a cubic relationship between power and flow rate. The fan characteristic factor integrates the fan's structural design parameters and power characteristics under reference operating conditions, serving as a coefficient relating the power changes between the reference and actual operating conditions. The effective exchange capacity represents the additional air volume the fan needs to deliver under actual operating conditions. The fan's reference volumetric flow rate is the baseline flow scale under the fan's design operating conditions. This represents the degree of change in actual flow rate relative to the reference flow rate. Taking the cube of this ratio aligns with the cubic relationship between power and flow rate in the wind turbine similarity law. Multiplying this by the wind turbine characteristic factor quantifies the additional electrical power consumed by the wind turbine under actual operating conditions relative to the reference operating conditions; this result is the wind turbine's electrical power increment.

[0091] Step S263: Based on the preset sensible heat conversion factor and latent heat conversion factor, perform a weighted calculation of the latent heat additional amount, sensible heat additional amount and moisture content difference to obtain the total cooling load.

[0092] Specifically, the process first calls upon the pre-set sensible heat conversion factor and latent heat conversion factor. The sensible heat conversion factor is determined based on the building thermal characteristics and ventilation system operation conditions of the underground high-speed rail transfer hub, according to industry heat load calculation standards and practical engineering experience. The latent heat conversion factor is determined based on the station's air humidity control requirements and the actual impact of moisture exchange, also according to relevant standards and experience. Next, the previously calculated sensible heat addition, along with the previously obtained moisture content difference and the latent heat of vaporization of the station's air, are used. The moisture content difference is multiplied by the latent heat of vaporization of the station's air to obtain the final result. The latent heat difference corresponding to the moisture content difference is used to convert the humidity difference into a corresponding heat load. Then, according to the weighted calculation rules, the sensible heat difference is multiplied by the sensible heat conversion factor to obtain the weighted cooling load contribution corresponding to the sensible heat part. At the same time, the latent heat difference corresponding to the moisture content difference is multiplied by the latent heat conversion factor to obtain the weighted cooling load contribution corresponding to the latent heat part. Finally, the weighted cooling load contribution of the sensible heat part and the weighted cooling load contribution of the latent heat part are added together to obtain the total heat load that needs to be processed by the refrigeration unit in the underground high-speed rail transfer hub, which is the total cooling load.

[0093] Specifically, the basic parameters of the underground high-speed rail transfer hub were first collected, including thermal parameters such as the thermal conductivity of the building envelope and the thickness of the insulation layer, the thermal and humidity characteristics of the functional zones within the station, the design air exchange rate of the ventilation system, and climatic parameters such as the average temperature and humidity in the local summer, serving as the basis for coefficient generation. Then, regarding the sensible heat conversion coefficient, the conversion criteria for sensible heat load in underground transportation hubs were referenced in industry standards such as the Energy Conservation Design Standard for Public Buildings. Combined with actual operating data on personnel density and equipment heating within the building envelope and the station, the proportion of sensible heat additional load requiring chiller treatment was analyzed, and the range of the sensible heat conversion coefficient was initially determined. Simultaneously, regarding the latent heat conversion coefficient, the design specifications for heating, ventilation, and air conditioning in civil buildings for underground spaces were referenced. The calculation of wet load is based on the actual monitoring data of moisture dissipation from personnel and equipment within the station, as well as the dehumidification capacity of the ventilation system. The proportion of latent heat requiring chiller processing is analyzed, and a preliminary range of latent heat conversion coefficients is determined. Next, operational projects with similar building scale and functional layout to this hub are selected, and their actual sensible and latent heat conversion coefficient data are collected. These are then adjusted to narrow the range based on the operating conditions of this hub. Finally, the basic parameters and preliminary coefficient range of this hub are input into building energy consumption simulation software to perform multi-condition simulation calculations. The matching degree between the total cooling load under different coefficients and the actual chiller operating load is compared, ultimately determining the specific values ​​of the sensible and latent heat conversion coefficients that best suit the actual needs of this hub, thus completing the pre-setting process.

[0094] Step S264: Calculate the increase in the power of the refrigeration unit based on the total cooling load and the refrigeration efficiency of the refrigeration unit in the underground high-speed rail transfer hub.

[0095] Specifically, the total cooling load is divided by the cooling efficiency of the refrigeration unit in the underground high-speed rail transfer hub to obtain the increment of the refrigeration unit's electrical power.

[0096] Step S265: Calculate the equivalent environmental additional power based on the wind turbine power increment and the chiller power increment.

[0097] Specifically, the increase in the power output of the wind turbine is added to the increase in the power output of the chiller to obtain the equivalent environmental additional power.

[0098] Specifically, the sensible heat conversion factor refers to a preset coefficient used to weigh the contribution of sensible heat addition to the cooling load, and its value reflects the proportional relationship of sensible heat being converted into cooling load; the latent heat conversion factor refers to a preset coefficient used to weigh the contribution of latent heat corresponding to the moisture content difference to the cooling load, and its value reflects the proportional relationship of latent heat caused by humidity difference being converted into cooling load; sensible heat addition refers to the increase in heat related only to temperature change transferred through effective exchange due to the temperature difference between the cable trench and the station air; moisture content difference refers to the degree of difference in water vapor content between the air in the cable trench and the station hall air, which is a key indicator for quantifying the moisture exchange potential between the two; the total cooling load refers to the ratio of sensible heat addition to moisture content difference based on the sensible heat conversion factor and the latent heat conversion factor. The weighted calculation of the latent heat component represents the total heat load that needs to be handled by chillers within the underground high-speed rail transfer hub; the chiller's refrigeration efficiency is the ratio between the chiller's output cooling capacity and its consumed electrical power, a core parameter reflecting the chiller's energy conversion capability; the chiller's electrical power increment is the result of dividing the total cooling load by the chiller's refrigeration efficiency, representing the additional electrical power consumed by the chiller due to the increase in cooling load; the fan's electrical power increment refers to the additional electrical power consumed by the fan under actual operating conditions relative to the reference operating conditions, a parameter representing the degree of additional load on the fan; the equivalent environmental additional power is the result of adding the fan's electrical power increment to the chiller's electrical power increment, representing the total additional power caused by changes in environmental heat and humidity exchange and ventilation load.

[0099] Specifically, the above calculation process is derived from the logical deduction of the additional impact of environmental heat and humidity exchange on system energy consumption. First, the latent heat is calculated because the difference in moisture content represents the difference in water vapor content between the cable trench and the station air. The latent heat of vaporization is the heat parameter of water vapor phase change. Combining the station air density and effective exchange rate, their product corresponds to the latent heat transferred per unit time due to the humidity difference. This is the additional heat load caused by environmental moisture exchange, which conforms to the physical laws of phase change heat transfer. Next, the increase in fan power is calculated based on the fan similarity law. Fan power is proportional to the cube of the ventilation volumetric flow rate. The ratio of the effective exchange rate to the reference volumetric flow rate reflects the change in actual flow rate relative to the design conditions. Multiplying this by the fan characteristic factor quantifies the additional power consumed by the fan due to changes in environmental load. The first is the increased energy consumption of the ventilation system due to environmental influences. The second is the total cooling load, which is obtained by weighting the sensible and latent heat conversion factors. This takes into account that the additional sensible and latent heat is not entirely borne by the chiller. After conversion based on the system's heat and humidity sharing characteristics, the total load that the chiller needs to handle is obtained. This is then divided by the refrigeration efficiency to obtain the chiller's power increment. This is the additional power consumed by the refrigeration system due to the increased environmental heat and humidity load, which is consistent with the energy conversion relationship that the ratio of the chiller's cooling capacity to its power is the refrigeration efficiency. Finally, the power increments of the fan and the chiller are added together because both are additional power consumption caused by environmental factors such as environmental heat and humidity exchange and the reverse chimney effect. Their sum can comprehensively represent the total additional power consumed by the entire underground high-speed rail transfer hub system due to environmental influences, i.e., the equivalent environmental additional power.

[0100] Step S27: Calculate the cooling load adjustment factor based on the baseline cooling load and the equivalent environmental additional power.

[0101] Specifically, based on the baseline cooling load and equivalent environmental additional power, the cooling load adjustment factor is calculated. The formula for calculating the cooling load adjustment factor is as follows: .

[0102] In the formula, It is the equivalent environmental added power. It is the cooling load gain coefficient. This is the baseline cooling load. It is the cooling load adjustment factor.

[0103] Specifically, in the formula for calculating the cooling load adjustment factor, 1 represents the baseline proportion corresponding to the baseline cooling load. Since the baseline cooling load is the basic cooling load that the hub needs to handle under the design baseline condition without additional environmental impact, no additional adjustment is needed. Therefore, it is used as the baseline for cooling load adjustment, which aligns with the logic of no additional correction under the baseline condition in load calculation. Equivalent environmental additional power is the additional power consumption of the system caused by factors such as environmental heat and humidity exchange and the reverse chimney effect. It is the core quantitative indicator of environmental additional impact. The baseline cooling load is the basic cooling load under the baseline condition. The ratio of the two is used to measure the proportion of equivalent environmental additional power relative to the baseline cooling load, converting the absolute value of the additional power into a proportion relative to the base load, avoiding interference from the difference in absolute values ​​between hubs of different scales. The cooling load gain coefficient is a parameter determined by combining the building thermal characteristics and system response efficiency of the hub. It is used to quantify the degree of conversion of equivalent environmental additional power into cooling load increment. Since not all equivalent environmental additional power is directly converted into cooling load, some is dissipated through heat dissipation of the building envelope, ventilation, and other means. This coefficient is needed to weigh the actual conversion correlation.

[0104] Step S28: Calculate the theoretical actual cooling load based on the baseline cooling load and the cooling load adjustment factor.

[0105] Specifically, the actual cooling load is obtained by multiplying the baseline cooling load and the cooling load adjustment factor.

[0106] Specifically, the baseline cooling load of an underground high-speed rail transfer hub refers to the cooling load that the hub needs to handle under the design benchmark operating condition without additional environmental impact. It is the benchmark value for cooling load calculation and represents the basic cooling demand of the hub under the benchmark operating condition. The equivalent environmental additional power refers to the total additional power consumed by the fans and chillers in the hub due to environmental factors such as environmental heat and humidity exchange and the reverse chimney effect. It is the additional power consumption of the system caused by environmental impact. The cooling load gain coefficient refers to a coefficient preset based on the building thermal characteristics and system response characteristics of the hub. It is used to weigh the degree of influence of the equivalent environmental additional power on the cooling load and reflects the degree of correlation between the additional power and the increase in cooling load. The cooling load adjustment factor refers to the parameter calculated based on the baseline cooling load and the equivalent environmental additional power. It represents the adjustment ratio of the cooling load relative to the baseline cooling load under the influence of environmental additional factors. The actual cooling load refers to the result obtained by multiplying the baseline cooling load and the cooling load adjustment factor. It represents the cooling load that needs to be handled by the chillers during the actual operation of the hub and is the actual basis for chiller operation and control.

[0107] Specifically, the steps for generating the cooling load gain coefficient are as follows: First, collect the core influencing parameters of the underground high-speed rail transfer hub, including thermal characteristic parameters such as the thermal conductivity of the building envelope, the thickness of the insulation layer, and the heat transfer area; system response characteristic parameters such as the refrigeration efficiency of the chiller and the ventilation efficiency of the fan; and measured data such as the baseline cooling load, equivalent environmental additional power, and actual cooling load from the hub's historical operation. Simultaneously, collect relevant data and operating parameters of the cooling load gain coefficient from existing underground transportation hubs with similar building scale, functional layout, and system configuration as this hub, as the basis for coefficient generation. Then, referencing the principles for determining the load conversion coefficient in industry standards such as the Energy-Saving Design Standard for Public Buildings and the Design Code for Heating, Ventilation, and Air Conditioning of Civil Buildings, and combining the thermal characteristics and system response characteristics of this hub, analyze the key influencing factors for the conversion of equivalent environmental additional power into cooling load increments, including the thermal insulation capacity of the building envelope, the heat and moisture exchange efficiency of the ventilation system, and the load regulation characteristics of the chiller, to initially determine the range of values ​​for the cooling load gain coefficient. Next, combine the collected basic parameter influencing factors and preliminary values... The system inputs data into building energy consumption simulation software, sets multiple typical operating conditions such as different outdoor temperatures and humidity levels, indoor personnel density, and equipment operating loads, and performs matching simulation calculations between the cooling load adjustment factor and the actual cooling load to obtain candidate values ​​for the cooling load gain coefficient under different operating conditions. Then, combined with measured data from the actual operation of the hub, each candidate value is substituted into the cooling load adjustment factor calculation formula to calculate the corresponding theoretical actual cooling load. The deviation is compared with the measured actual cooling load, and candidate values ​​with deviations exceeding the engineering allowable range are eliminated to narrow the coefficient value range. Finally, the remaining candidate values ​​are weighted and averaged, with the weighting weights determined based on the frequency of occurrence of the corresponding operating conditions in the actual operation of the hub, to obtain the initial value of the cooling load gain coefficient. This initial value is then verified by measured data over a continuous operating cycle. If the average deviation between the theoretically calculated actual cooling load and the measured value is controlled within the engineering allowable range, the initial value is determined as the final cooling load gain coefficient. If the deviation exceeds the range, the above simulation comparison and correction process is repeated until the coefficient can accurately quantify the degree of conversion correlation between the equivalent environmental additional power and the cooling load increment.

[0108] Specifically, underground high-speed rail transfer hubs rely on double-layer coaxial shafts and cable trenches to form a composite ventilation shaft. The combination of peak morning and evening passenger flows and high-density train arrivals and departures causes a short-term concentration of heat and humidity loads. Simultaneously, the temperature difference between the surface and underground creates a vertical thermo-pressure field, triggering a reverse chimney effect. This causes external cold air to sink through the shafts and cable trenches, stagnating and flowing back around the equipment, leading to heat and humidity accumulation. Furthermore, auxiliary systems such as fans and chillers operate under unconventional conditions, resulting in additional spikes or shifts in the total load during peak periods. Conventional load forecasting primarily relies on passenger flow maps and external weather forecasts, failing to incorporate the additional heat and humidity load impact of the reverse chimney effect. This leads to discrepancies in the total load during peak periods, resulting in systematic underestimation or overestimation of peak loads. The purpose of calculating the actual cooling load at this time is to take into account the equivalent environmental additional impact caused by environmental factors such as the reverse chimney effect. Based on the baseline cooling load, the load increment brought about by environmental additional factors is combined to accurately represent the cooling load demand that the hub needs to handle in actual operation. This avoids the a priori arrangement of ventilation and cooling strategies and improper on-site scheduling due to load prediction deviations. At the same time, it ensures that the fan and chiller system operates under conditions that are adapted to the actual load, thereby improving the stability and energy efficiency of the system operation.

[0109] Step S29: Continue until the multi-source data of samples at all times within the set time period are calculated to obtain the theoretical true cooling load sequence.

[0110] Step S3: Obtain the historical cold load sequence of high-speed rail transfer hubs.

[0111] Step S4: Input the historical cold load sequence into the pre-trained large language model to obtain the predicted cold load sequence; the pre-trained large language model is obtained by training the Transformer neural network based on the sample cold load sequence.

[0112] Specifically, the pre-trained large language model is a model that has been pre-trained on a large-scale related data to train the Transformer neural network and has the ability to predict cold load. Here, it is used to analyze and predict the load status of the hub.

[0113] Specifically, the predicted cooling load sequence can be the predicted cooling load sequence for morning and evening peak periods. The pre-trained large language model outputs the predicted cooling load value for the peak periods of concentrated passenger flow and high-density train arrivals and departures at the hub in the morning and evening. This is the preliminary prediction result obtained by the model based on historical data and preset parameters. Subsequent dynamic correction refers to the operation of adjusting and optimizing the predicted cooling load of the morning and evening peak periods output by the pre-trained large language model in real time based on the theoretical real cooling load. The aim is to correct the prediction deviation caused by the model not fully incorporating actual operating conditions such as the reverse chimney effect, so that the predicted cooling load of the morning and evening peak periods is more in line with the actual load status of the hub.

[0114] Step S5: Dynamically correct the predicted cooling load sequence based on the theoretical real cooling load sequence to obtain the corrected predicted cooling load sequence.

[0115] As one feasible approach, step S5 specifically includes steps S51 to S57: Step S51: Based on the theoretical real cooling load sequence and the predicted cooling load sequence, calculate the load deviation value sequence and the relative deviation rate sequence.

[0116] Step S52: Calculate the average deviation value and the average relative deviation rate based on the load deviation value sequence and the relative deviation rate sequence.

[0117] Step S53: Based on the average deviation value, average relative deviation rate, and multi-source sample data for a set time period, and combined with the building thermal characteristics, determine the dynamic correction coefficient sequence.

[0118] Step S54: Calculate the correction amount sequence based on the dynamic correction coefficient sequence.

[0119] Step S55: Overlay the predicted cooling load sequence and the corresponding correction sequence to obtain the preliminarily corrected predicted cooling load sequence.

[0120] Step S56: Calculate the corrected average deviation based on the theoretical true cooling load sequence and the preliminarily corrected predicted cooling load sequence.

[0121] Step S57: Determine whether the corrected average deviation is less than or equal to the set deviation; if yes, then determine the preliminarily corrected predicted cold load sequence as the corrected cold load sequence; if no, then re-execute the step of obtaining sample multi-source data of high-speed rail transfer hubs to calculate the theoretical real cold load, obtain the updated theoretical real cold load, and dynamically correct the predicted cold load based on the updated theoretical real cold load.

[0122] Specifically, the time period is first determined as the morning and evening peak hours of the underground high-speed rail transfer hub, based on passenger flow patterns and train arrival and departure schedules. For example, the morning peak is from 7:00 to 9:00 and the evening peak is from 17:00 to 19:00. The prediction intervals within the peak hours are divided into 5-minute intervals to ensure accurate matching in the time dimension. Then, the theoretical actual cold load corresponding to each prediction interval calculated above is used, and the predicted cold load of each prediction interval for the corresponding peak hours is extracted from the output of the pre-trained large language model to ensure that the data is completely aligned in terms of time granularity and load dimension. Next, the load of each prediction interval is calculated. The deviation value (theoretical actual cooling load minus predicted cooling load) and the relative deviation rate (the ratio of load deviation value to theoretical actual cooling load) are used to calculate the weighted average deviation value and average relative deviation rate after removing extreme deviation values ​​caused by data acquisition anomalies. This is then combined with historical deviation data for the corresponding intervals during the same peak period over the past 30 days, with the principle that the more recent the data, the greater the weight. This weighted average deviation value and average relative deviation rate serve as the core correction basis. Based on the average deviation value and average relative deviation rate, and in conjunction with the hub's real-time environmental parameters (such as outdoor temperature and humidity, and the intensity of the reverse chimney effect) (i.e., multi-source data of the sample) and the building's thermal characteristics, the load for each prediction interval is determined. The dynamic correction coefficient (i.e., the dynamic correction coefficient sequence) has a higher weight due to the larger the deviation and the more significant the real-time environmental interference. This coefficient is used to calculate the targeted correction amount for each interval (the product of the correction coefficient and the average deviation value). Then, the prediction results for each interval (i.e., the predicted cooling load sequence) output by the pre-trained large language model are superimposed with the corresponding correction amount (i.e., the corresponding correction amount sequence) to obtain the initially corrected predicted cooling load sequence. Next, the correction effect is verified by calculating the average correction deviation between the initially corrected predicted cooling load sequence and the theoretical actual cooling load sequence. If the average correction deviation is controlled within the engineering allowable range (e.g., ±5%), it is determined as the final correction result, and the corrected cooling load sequence is obtained. If it exceeds the range, parameters such as the real-time equivalent environmental additional power and the cooling load gain coefficient are reintroduced to perform a secondary calibration of the correction coefficient. The correction amount is adjusted, and the correction result is recalculated until the average correction deviation meets the requirements. Finally, the deviation data, correction coefficients, final results, and corresponding real-time operating parameters from this correction process are stored in the model database to update the prediction parameters of the pre-trained large language model, improving the initial accuracy and correction efficiency of subsequent peak load predictions.

[0123] The beneficial effects of the cold load prediction correction method based on pre-trained large language models proposed in this application are mainly reflected in the following aspects: 1. This application introduces a mechanism to quantify the additional environmental impact caused by the composite ventilation structure of a double-layer coaxial vertical shaft cable trench in the complete load prediction link. First, the buoyancy pressure difference of the vertical shaft, the total loss of the composite channel and the reverse strength index are calculated. Then, the effective exchange volume and the equivalent environmental additional power are derived from these. Finally, the predicted cold load sequence of the pre-trained large language model is dynamically corrected, so that the pre-trained large language model can automatically sense the additional energy consumption deviation caused by the change of air flow direction during the morning and evening peak periods of the underground hub. This avoids the continuity error of the predicted cold load sequence output by the traditional model, improves the accuracy of the predicted cold load sequence, and realizes timely correction of abnormal bulges and deviations in the predicted cold load sequence.

[0124] 2. This application quantifies the frictional loss along the path, air density gradient, return flow distribution ratio, and effective exchange volume of the composite ventilation path composed of double-layer coaxial shafts and cable trenches. This allows the thermal and humid disturbances caused by the reverse chimney effect to enter the load prediction stage in a calculable form for the first time. It unifies the temperature difference, water vapor content difference, sensible heat addition, latent heat addition, and fan incremental power into equivalent environmental additional power. This enables sudden load drifts that traditional prediction models cannot capture to be identified in advance in the form of physical quantities, thereby improving the accuracy of predicting the cold load sequence of underground high-speed rail transfer hubs during the morning and evening peak transportation periods.

[0125] In one exemplary embodiment, a computer device is provided, which may be a server or a terminal, and its internal structure diagram may be as follows. Figure 2 As shown, this computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs stored in the non-volatile storage media. The database stores multi-source sample data. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communicating with external terminals via a network connection. When the computer program is executed by the processor, it implements a predictive cold load correction method based on a pre-trained large language model.

[0126] Those skilled in the art will understand that Figure 2The structures shown are merely block diagrams of some structures related to the present application and do not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than shown in the figures, or combine certain components, or have different component arrangements. In an exemplary embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0127] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0128] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for predicting cold load correction based on a pre-trained large language model, applied to a high-speed rail transfer hub, wherein the structure of the high-speed rail transfer hub includes a double-layer coaxial shaft and a cable trench; characterized in that, The method includes: Acquire sample multi-source data of high-speed rail transfer hubs during a set time period; the sample multi-source data includes: air density, average wind speed, temperature and humidity parameters, specific heat of air at constant pressure in the station, latent heat of vaporization of air in the station, reference volumetric flow rate of fans in underground high-speed rail transfer hubs, and baseline cooling load of underground high-speed rail transfer hubs. Based on multi-source sample data for a set time period and the equivalent height of the double-layer coaxial shaft, the theoretical real cooling load sequence is calculated. Obtain the historical cold load sequence of high-speed rail transfer hubs; The historical cold load sequence is input into a pre-trained large language model to obtain the predicted cold load sequence; the pre-trained large language model is obtained by training a Transformer neural network based on the sample cold load sequence. The predicted cooling load sequence is dynamically corrected based on the theoretical real cooling load sequence to obtain the corrected predicted cooling load sequence.

2. The method for predicting cold load correction based on a pre-trained large language model according to claim 1, characterized in that, Based on multi-source sample data for a given time period and the equivalent height of the double-layer coaxial shaft, the theoretical real cooling load sequence is calculated, specifically including: Calculate the buoyancy pressure difference based on the air density and the equivalent height of the double-layer coaxial shaft; Based on the average wind speed, calculate the total loss of the composite channel between the ventilation channel of the double-layer coaxial shaft and the ventilation channel of the cable trench; Calculate the reverse strength index based on the buoyancy pressure difference and the total loss of the composite channel; Based on the average wind speed and the reverse strength index, the effective exchange volume between the double-layer coaxial vertical shaft ventilation channel and the cable communication ventilation channel is calculated. Based on the temperature and humidity parameters and the specific heat of air at constant pressure, calculate the moisture content difference and the sensible heat addition. The equivalent environmental additional power is calculated based on the latent heat of vaporization, the air density, the reference volumetric flow rate, the effective exchange capacity, the moisture content difference, and the sensible heat addition. Calculate the cooling load adjustment factor based on the baseline cooling load and the equivalent environmental additional power. Calculate the theoretical actual cooling load based on the baseline cooling load and the cooling load adjustment factor; The calculation continues until all multi-source data from samples at all times within the set time period are used to obtain the theoretical true cooling load sequence.

3. The method for correcting predicted cold load based on a pre-trained large language model according to claim 2, characterized in that, The air density includes: the air density outside the station and the air density inside the station; Based on the air density and the equivalent height of the double-layer coaxial shaft, the buoyancy pressure difference is calculated, specifically including: The first effective vertical height and the second effective vertical height of the first section of the double-layer coaxial shaft are added together to obtain the equivalent height of the shaft. The buoyancy pressure difference is calculated based on the equivalent height of the shaft, the air density outside the station, and the air density inside the station.

4. The method for correcting predicted cold load based on a pre-trained large language model according to claim 2, characterized in that, The average wind speed includes: the average wind speed of the double-layer coaxial shaft and the average wind speed of the cable trench ventilation channel. Based on the average wind speed, the total loss of the combined channel between the ventilation channel of the double-layer coaxial shaft and the ventilation channel of the cable trench is calculated, specifically including: Based on the vertical shaft friction coefficient, equivalent vertical shaft height, equivalent vertical shaft diameter, air density in the station and average wind speed of the vertical shaft, calculate the frictional pressure drop along the vertical shaft section; Based on the cable trench's ventilation channel, cable trench friction coefficient, cable trench length, and equivalent height, calculate the frictional pressure drop along the cable trench section. The frictional pressure drop along the shaft section and the frictional pressure drop along the cable trench section are added together to obtain the total composite channel loss between the ventilation channel of the double-layer coaxial shaft and the ventilation channel of the cable trench.

5. The method for correcting predicted cold load based on a pre-trained large language model according to claim 2, characterized in that, Based on the buoyancy pressure difference and the total loss of the composite channel, the reverse strength index is calculated, specifically including: The difference between the buoyancy pressure difference and the total loss of the composite channel is calculated to obtain the buoyancy-driven net pressure difference; The reverse strength index is calculated based on the net pressure difference driven by buoyancy and the buoyancy pressure difference.

6. The method for correcting predicted cold load based on a pre-trained large language model according to claim 4, characterized in that, Based on the average wind speed and the reverse intensity index, the effective exchange rate between the double-layer coaxial vertical shaft ventilation channel and the cable-connected ventilation channel is calculated, specifically including: Calculate the equivalent cross-sectional area of ​​the shaft based on the inner coaxial shaft radius of the inner shaft cylinder and the outer coaxial shaft radius of the outer shaft cylinder in the double-layer coaxial shaft. Calculate the volumetric flow rate of the shaft based on the equivalent cross-sectional area of ​​the shaft and the average wind speed of the shaft; Calculate the cross-sectional area of ​​the cable trench based on the equivalent width and equivalent height of the cable trench ventilation channel; Calculate the volumetric flow rate of the cable trench based on the cross-sectional area of ​​the cable trench and the average wind speed of the cable trench; Calculate the total exchange volumetric flow rate based on the vertical shaft volumetric flow rate and the cable trench volumetric flow rate; Calculate the return flow distribution coefficient based on the cable trench volumetric flow rate and the total exchange volumetric flow rate; The effective value of the reverse strength is obtained by taking the minimum value of the reverse strength index and 0. The effective exchange volume is calculated based on the effective value of the reverse intensity, the total exchange volume flow rate, and the return flow distribution coefficient.

7. The method for correcting predicted cold load based on a pre-trained large language model according to claim 2, characterized in that, The temperature and humidity parameters include: cable trench air temperature, cable trench air relative humidity, station hall air temperature, and station hall air relative humidity; Based on the temperature and humidity parameters and the specific heat of air at constant pressure, the moisture content difference and sensible heat addition are calculated, specifically including: The humidity content of the air in the station hall is calculated based on the air temperature, relative humidity, and density in the station hall. The difference between the moisture content of the air in the cable trench and the moisture content of the air in the station hall is calculated to obtain the moisture content difference; The temperature difference is obtained by calculating the difference between the air temperature in the cable trench and the air density in the station. The sensible heat addition is calculated based on the specific heat of air at constant pressure, the air density in the station, the effective exchange rate, and the temperature difference.

8. The method for predicting cold load correction based on a pre-trained large language model according to claim 2, characterized in that, Based on the latent heat of vaporization, air density, reference volumetric flow rate, effective exchange capacity, moisture content difference, and sensible heat addition, the equivalent environmental additional power is calculated, specifically including: The latent heat addition is calculated based on the latent heat of vaporization, the air density in the station, the effective exchange rate, and the moisture content difference. Calculate the wind turbine power increment based on the effective exchange volume and the reference volume flow rate; The total cooling load is obtained by weighting the latent heat additional amount, the sensible heat additional amount, and the moisture content difference based on the preset sensible heat conversion factor and latent heat conversion factor. Based on the total cooling load and the refrigeration efficiency of the refrigeration unit in the underground high-speed rail transfer hub, calculate the increment of the refrigeration unit's electrical power. The equivalent environmental additional power is calculated based on the incremental power of the wind turbine and the incremental power of the chiller.

9. The method for predicting cold load correction based on a pre-trained large language model according to claim 2, characterized in that, Based on the theoretical actual cooling load sequence, the predicted cooling load sequence is dynamically corrected to obtain the corrected predicted cooling load sequence, specifically including: Based on the theoretical real cooling load sequence and the predicted cooling load sequence, calculate the load deviation value sequence and the relative deviation rate sequence; Based on the load deviation value sequence and the relative deviation rate sequence, calculate the average deviation value and the average relative deviation rate; Based on the average deviation value, the average relative deviation rate, and multi-source sample data for a set period, combined with the building thermal characteristics, a dynamic correction coefficient sequence is determined. Based on the dynamic correction coefficient sequence, calculate the correction amount sequence; The predicted cooling load sequence and the corresponding correction sequence are superimposed to obtain the preliminarily corrected predicted cooling load sequence. Based on the theoretical actual cooling load sequence and the preliminary corrected predicted cooling load sequence, the corrected average deviation is calculated; Determine whether the corrected average deviation is less than or equal to the set deviation; If so, the preliminary revised predicted cooling load sequence will be determined as the revised cooling load sequence; If not, the step of obtaining sample multi-source data of high-speed rail transfer hubs to calculate the theoretical real cold load is repeated to obtain the updated theoretical real cold load, and the predicted cold load is dynamically corrected based on the updated theoretical real cold load.

10. A computer device, comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the computer program to implement the predictive cold load correction method based on a pre-trained large language model according to any one of claims 1-9.

Images