Underground space supercomputing cluster cooling and heat storage exchange system

Abstract

The invention relates to the technical field of super-computing heat exchange, and particularly discloses an underground space super-computing cluster cooling and heat storage exchange system which comprises a super-computing heat source monitoring module, an immersed liquid cooling heat absorption module, a supercritical CO2 heat energy increasing module, a dual-mode heat distribution processing module and a nanometer phase change heat storage module. An underground geothermal well coupling and regulating module and a regional heat supply and auxiliary power generation module; on the basis of the supercomputing heat source monitoring module, the operation temperature and heat load information of the supercomputing cluster is collected in real time, and the data is used for controlling the liquid cooling system, the supercritical CO2 heat exchange module and the dynamic heat distribution system. The closed-loop control mechanism can enable the system to automatically adjust cooling and heat storage strategies according to overcalculation load changes, so that the self-adaptability and response speed of heat management are improved, traditional air cooling or cold plate water cooling is replaced with immersed liquid cooling, a server makes direct contact with a cooling medium, the heat exchange efficiency is enhanced, and the local hot spot risk is reduced.

Description

Technical Field

[0001] The present invention belongs to the technical field of supercomputer heat exchange, and particularly relates to an underground space supercomputer cluster cooling and heat storage exchange system. Background Art

[0002] With the rapid development of information technology, supercomputers have become indispensable important tools in key fields such as scientific research, military, and economy. However, with the continuous improvement of computing power, the heat generated by supercomputer clusters during operation has increased sharply, posing unprecedented challenges to heat dissipation and thermal management. Traditional heat dissipation methods, such as air cooling and cold plate water cooling, are difficult to meet the heat dissipation requirements of high-density supercomputer clusters. The air cooling method has problems such as high noise, low heat dissipation efficiency, and difficulty in eliminating local hot spots; while the cold plate water cooling, although the heat dissipation efficiency has been improved, still has problems such as coolant leakage and complex maintenance.

[0003] In terms of thermal energy management, existing technologies often regard the heat generated by supercomputers as waste that needs to be quickly removed, lacking means for effective utilization. This not only leads to energy waste but also increases the environmental burden. With the intensification of the global energy crisis and the improvement of environmental awareness, how to recover and utilize the thermal energy of supercomputers has become an urgent problem to be solved. Most of the existing heat dissipation and thermal management systems for supercomputer clusters adopt static control strategies, that is, on-off control according to preset thresholds. This control method lacks flexibility and cannot adaptively adjust according to the real-time changes of supercomputer loads, resulting in overheating during low loads and insufficient heat dissipation during high loads. This not only affects the performance and stability of supercomputers but also increases energy consumption and operation and maintenance costs.

[0004] In response to this, the present application proposes an underground space supercomputer cluster cooling and heat storage exchange system to solve the above problems. Summary of the Invention

[0005] The purpose of the present invention is to provide an underground space supercomputer cluster cooling and heat storage exchange system to solve the problems raised in the above background art.

[0006] To achieve the above purpose, the present invention provides the following technical solutions: An underground space supercomputer cluster cooling and heat storage exchange system, comprising: A supercomputer heat source monitoring module for real-time monitoring of the temperature and operating status of supercomputer chips to obtain initial thermal energy data; An immersion liquid cooling heat absorption module for directly immersing supercomputer chips in a liquid cooling medium and calculating the absorbed thermal energy of the chips through a dynamic heat transfer model to obtain preheated liquid; A supercritical CO2 thermal energy enhancement module is used to enhance the thermal energy of the preheated liquid by using the supercritical CO2 cycle technology and combining the initial thermal energy data to obtain a high-temperature working fluid; the initial thermal energy data is further input as a reference parameter into the adaptive control algorithm of the supercritical CO2 module to optimize the CO2 flow rate, pressure, and PID parameters to ensure that the preheated liquid is rapidly heated to the optimal state. A dual-mode thermal shunt processing module is used to intelligently shunt the high-temperature working fluid to obtain a shunted thermal energy flow. A nano-phase change heat storage module is used to optimize the storage of the shunted thermal energy flow through a nano-modified phase change material to obtain a stored thermal energy flow. An underground geothermal well coupling and regulation module is used to directly couple the stored thermal energy flow with the underground heat reservoir through a control valve and pipeline system to obtain a regulated thermal energy flow. A district heating and auxiliary power generation module is used to further utilize the regulated thermal energy flow, generate waste heat power through an ORC system, and at the same time transport hot water at 50 - 60 °C to surrounding buildings for district heating to obtain the comprehensive energy utilization efficiency of thermal energy.

[0007] Preferably, the enhancing of the thermal energy of the preheated liquid by using the supercritical CO2 cycle technology and combining the initial thermal energy data to obtain a high-temperature working fluid includes: Based on an adaptive thermal energy enhancement algorithm, combining the non-linear change of the specific heat capacity of supercritical CO2 at different temperatures and pressures, an adaptive PID control algorithm is used to adjust the CO2 cycle parameters in real time according to the temperature and flow rate of the preheated liquid. Where the formula of the adaptive thermal energy enhancement algorithm is: ; Where Qabs is the absorbed thermal energy, unit J; m is the CO2 flow rate, unit kg / s; is the temperature-related specific heat capacity, unit J / (kg·K); ΔT1 is the temperature difference, ΔT = Tout - Tin, unit K; The formula of the adaptive PID control algorithm is: ; Where the error , Ttarget is the target temperature (e.g., 80 °C); Tout(t) is the real-time monitored outlet temperature of the cooling medium; PID parameter range: Kp = 0.2 - 0.5, Ki = 0.05 - 0.1, Kd = 0.01 - 0.05.

[0008] Such as the initial flow rate: about 0.5 kg / s; Specific heat capacity: about 1000 J / (kg·K) under supercritical conditions; Through this algorithm, the flow rate and temperature rise rate of supercritical CO2 can be quickly adjusted when the load changes, realizing rapid heat energy increase, ensuring that the subsequent power generation and heat storage modules all work efficiently at the designed temperature, while reducing the system response time and improving the overall energy efficiency.

[0009] Preferably, the intelligent shunt processing includes two branch processes, one of which is used for continuous enhanced cooling and the other is used for driving turbine power generation; The intelligent shunt processing adopts a dynamic shunt algorithm, which is used to dynamically adjust the heat energy shunt ratio according to the supercomputer cluster load and real-time temperature after obtaining high-temperature working fluid, and reasonably distribute the heat energy to the cooling and power generation branches; The dynamic shunt algorithm includes a shunt ratio calculation method and an energy shunt calculation method; Among them, the formula of the shunt ratio calculation method is: ; Tset is the safety cooling set temperature (such as 60°C); Tfluid is the current fluid temperature; Tmax is the maximum allowable temperature of the fluid (such as 90°C); In the algorithm, the range of β needs to be limited (such as 0 to 1), and it will be automatically adjusted when the calculated value exceeds; The formula of the energy shunt calculation method is: ; Among them, Qcooling is the energy for continuous enhanced cooling; Qpower is the energy for driving turbine power generation; Qtotal is the total heat energy or total heat storage capacity of the high-temperature fluid; For example, when Tfluid = 70°C, Tset = 60°C, Tmax = 90°C, it is calculated that theoretically β = (60 - 70) / (90 - 70) = -0.5. In practice, the lower limit of 0 will be set, which means that the cooling flow rate needs to be increased at this time; This algorithm enables the system to dynamically adjust the energy distribution between cooling and power generation according to real-time load and temperature data, realizing priority cooling under high load and maximizing heat energy recovery under low load, overall improving energy utilization rate and ensuring the safe operation of equipment.

[0010] Preferably, optimizing the stored heat energy of the split heat energy through a nano-modified phase change material to obtain a stored heat energy flow includes: Introducing the heat energy into the phase change material through a heat exchange system, enabling it to absorb heat and undergo a phase change, storing a large amount of latent heat to obtain latent heat of phase change; After the phase change process is completed, the phase change material continues to absorb heat, and the temperature rises without a change in phase state to obtain sensible heat storage; Optimizing the latent heat of phase change and the sensible heat storage to obtain the stored heat energy flow; Among them, the calculation formula for the latent heat of phase change is: ; Where is the mass of the PCM, with the unit kg; Leff = L0×η is the effective latent heat of phase change, with the unit J / kg; L0 is the basic latent heat (such as 250 kJ / kg); η is the nano-enhancement coefficient, with a value of 1.5 - 2.0; The calculation formula for the sensible heat storage is: ; Where is the specific heat capacity of the material; ΔT2 is the temperature change range; According to the latent heat of phase change and the sensible heat storage, the total heat storage capacity Qtotal is:

[0011] If the basic latent heat L0 = 250 kJ / kg, the effective latent heat after nano-enhancement is increased to about 375 kJ / kg (η = 1.5); This optimization formula improves the heat storage density by about 50% compared with traditional materials, not only enhancing the energy storage capacity but also improving the system cycle stability, reducing the impact of temperature fluctuations on the overall system performance, and providing a more stable heat source for subsequent underground heat storage and district heating.

[0012] Preferably, the dynamic heat transfer model uses a dynamic heat transfer model based on Newton's cooling law to describe the temperature drop curve of the supercomputer chip during the liquid cooling heat absorption process. This model provides a theoretical basis for optimizing the immersion liquid cooling structure and guides the fluid parameters and chip surface design. The formula is: ; Among them, h is the convective heat transfer coefficient between the liquid and the chip, with the unit W / (m²·K); A is the contact area between the chip and the cooling medium, with the unit m²; m is the mass of the chip, with the unit of kg; cp is the specific heat capacity of the chip, with the unit of J / (kg·K); T is the chip temperature, and Tfluid is the temperature of the liquid cooling medium; This model helps designers accurately predict the chip temperature decay process, and then optimize the immersion liquid cooling solution to ensure that the high-density supercomputer cluster can still dissipate heat quickly under high-load operation, guaranteeing the stability and reliability of the equipment.

[0013] Preferably, the supercomputer heat source monitoring module includes a temperature sensor and a data acquisition unit. The data acquisition unit is used to continuously collect the temperature and heat load information of the supercomputer chip monitored by the temperature sensor, so as to monitor in real time and timely feedback the initial heat energy data.

[0014] Preferably, the immersion liquid cooling heat absorption module adopts a fully immersed structure, immersing the supercomputer server directly in the liquid cooling medium, so as to achieve rapid and uniform transfer of the chip heat and obtain the preheated liquid.

[0015] Preferably, the nano-phase change heat storage module's nano-modified phase change material includes at least one of graphene-enhanced molten salt and carbon nanotubes.

[0016] Preferably, the district heating and auxiliary power generation module includes an organic Rankine cycle (ORC) power generation device and a district heating pipe network unit; wherein the organic Rankine cycle power generation device is used to drive a turbine to generate waste heat power using the diverted heat energy; The district heating pipe network unit is used to transport the 50 - 60°C hot water released from the underground heat storage to the surrounding buildings to obtain the comprehensive heat energy utilization efficiency.

[0017] Compared with the prior art, the beneficial effects of the present invention are: (1) Based on the supercomputer heat source monitoring module, the present invention collects the operating temperature and heat load information of the supercomputer cluster in real time, and uses this data to control the liquid cooling system, supercritical CO2 heat exchange module, and dynamic heat diversion system. This closed-loop control mechanism enables the system to automatically adjust the cooling and heat storage strategies according to the supercomputer load changes, thereby improving the self-adaptability and response speed of thermal management and ensuring the efficient and safe operation of the supercomputer cluster under different load conditions.

[0018] (2) The present invention uses immersion liquid cooling to replace traditional air cooling or cold plate water cooling, enabling the server to be in direct contact with the cooling medium, enhancing the heat exchange efficiency, and reducing the risk of local hot spots. At the same time, combined with the supercritical CO2 heat energy enhancement technology, it quickly absorbs, transfers, and enhances the heat, ensuring that the cooling system has higher heat conduction efficiency and lower energy loss, and still maintaining stable temperature control during high-load operation.

[0019] (3) Through dual-mode heat diversion processing, the present invention rationally distributes excess heat energy to the heat storage or waste heat power generation system on the premise of ensuring the heat dissipation requirements of the supercomputer cluster. By using nano-modified phase change material PCM, higher-density energy storage can be achieved in a limited space. Compared with the traditional sensible heat storage method, the phase change heat storage technology can store and release a large amount of heat energy under relatively small temperature changes, and combines with the underground heat storage system for deep storage, improving the overall stability and long-term availability of energy storage, and providing support for district heating or future energy peak shaving. Description of the Drawings

[0020] Figure 1 It is a block diagram of the underground space supercomputer cluster cooling and heat storage exchange system of the present invention. Detailed Embodiments

[0021] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0022] Embodiment 1:

[0023] Please refer to Figure 1 As shown, the underground space supercomputer cluster cooling and heat storage exchange system includes: A supercomputer heat source monitoring module for real-time monitoring of the temperature and operating status of the supercomputer chip to obtain initial heat energy data. Among them, a high-precision temperature sensor and a data acquisition unit are used to monitor the heat generated during the operation of the supercomputer chip in real time, obtain the initial heat energy data, and the data is collected by an embedded sensor (such as an RTD, an infrared thermometer, etc.) and processed in real time by a SCADA or an edge controller to form an accurate temperature and heat load curve. The obtained real-time and accurate initial heat energy data can provide a reliable basis for the subsequent modules, ensure the precise response of cooling, heat energy improvement and diversion control, and thus improve the overall energy efficiency and safety of the system. An immersion liquid cooling heat absorption module for directly immersing the supercomputer chip in a liquid cooling medium and calculating the heat energy absorbed by the chip through a dynamic heat transfer model to obtain preheated liquid. The supercomputer chip is directly immersed in a special liquid cooling medium to achieve direct contact between the chip surface and the cooling medium, so as to quickly absorb and dissipate heat and obtain preheated liquid. By designing a special immersion tank and flow channel, it is ensured that the cooling medium evenly covers the chip surface, and at the same time, natural or forced convection is used to improve the heat transfer efficiency, which can greatly improve the heat dissipation efficiency and uniformity, reduce the risk of local hot spots, and realize the stable operation of high-density supercomputer equipment under high load, and provide a preliminary heat carrier for subsequent heat energy improvement. A supercritical CO2 thermal energy boosting module is used to boost the thermal energy of the preheated liquid by utilizing the supercritical CO2 cycle technology and combining with the initial thermal energy data to obtain a high-temperature working fluid. The initial thermal energy data is further input into the adaptive control algorithm of the supercritical CO2 module as a reference parameter to optimize the CO2 flow rate, pressure, and PID parameters, ensuring that the preheated liquid is rapidly heated up in an optimal state. By utilizing the high heat transfer and diffusivity of supercritical CO2, the low-temperature thermal energy in the preheated liquid is rapidly boosted to a high-temperature working fluid. An adaptive PID control algorithm is adopted to adjust the supercritical CO2 flow rate and pressure according to the real-time temperature error, and rapid temperature increase is achieved through a heat exchanger, thus rapidly concentrating the thermal energy to ensure that the system can reach the preset temperature requirement within a short time, guaranteeing that the subsequent energy shunting and utilization stages operate under ideal conditions, and improving the energy efficiency and response speed of the entire system. A dual-mode heat shunting processing module is used to perform intelligent shunting processing on the high-temperature working fluid to obtain a shunted heat energy flow. The intelligent shunting processing of the obtained high-temperature working fluid is carried out. On the one hand, it continues to meet the cooling requirements, and on the other hand, part of the thermal energy is used to drive a turbine generator or injected into a heat storage system to obtain a shunted heat energy flow. A dynamic shunting algorithm is used to calculate the shunting ratio according to the real-time temperature, load, and energy demand. The fluid distribution is precisely adjusted through a controllable valve. Furthermore, it can preferentially meet the cooling demand under high load to ensure the safety of the equipment. When the load is low or fluctuating, efficient recovery and reuse of waste heat are realized, the overall energy utilization rate of the system is improved, and the optimal balance of energy management is achieved. A nano-phase change heat storage module is used to optimize the heat storage of the shunted heat energy flow through a nano-modified phase change material to obtain highly efficient and stable stored thermal energy. A nano-modified phase change material (PCM) is adopted. The nano-modified phase change material in the nano-phase change heat storage module includes at least one of graphene-enhanced molten salt and carbon nanotubes. Phase change heat storage is achieved by absorbing thermal energy, and at the same time, sensible heat storage is combined to obtain highly efficient and stable stored thermal energy. A large amount of energy is stored using the latent heat of phase change, and at the same time, temperature control is achieved through the sensible heat change of the material. The introduction of a new nano-enhancement coefficient significantly increases the heat storage density. Therefore, the heat storage density can be increased (up to 150% or more of traditional materials), the demand for heat storage volume is reduced, and it can be used stably in cycles, reducing the risk of temperature fluctuations, providing sufficient and stable thermal energy support for subsequent energy utilization. Underground geothermal well coupling and regulation module, which is used to directly couple the stored heat energy flow with the underground heat reservoir through the control valve and pipeline system, realize the bidirectional transmission and regulation of heat energy, and obtain the regulated heat energy flow; directly transmit the heat energy flow stored in the heat storage module to the underground heat reservoir through physical coupling, and realize bidirectional heat energy transmission and regulation through pipelines and control valves to obtain the regulated heat energy flow; by constructing a directly coupled geothermal well system, monitor the underground temperature and pressure in real time, and dynamically adjust the injection and recovery flow rates through the regulation device; connect with the underground heat storage layer through the vertical geothermal well pipeline system, adopt automatic control valves and flow regulation devices, integrate temperature and pressure sensors and remote monitoring systems, and realize online data feedback and real-time regulation; thus realizing the bidirectional and efficient transmission of heat energy between the surface and the underground, balancing energy supply and demand; using the underground geothermal reservoir as a heat energy buffer to improve the overall system operation stability and provide a reliable heat source for district heating; District heating and auxiliary power generation module, which is used to further utilize the regulated heat energy flow, realize waste heat power generation through the ORC system, and at the same time transport hot water at 50 - 60 °C to surrounding buildings for district heating, so as to obtain the comprehensive heat energy utilization efficiency; realize the final utilization of the regulated heat energy flow through the organic Rankine cycle ORC device and the district heating pipe network to obtain the actual energy efficiency output; use the ORC system to convert waste heat into electric energy, and at the same time transport hot water at 50 - 60 °C to surrounding buildings for heating and domestic hot water supply; both realize waste heat power generation and meet the district heating demand, improving the overall energy utilization rate; reducing the cooling energy consumption of the supercomputer cluster, and at the same time reducing carbon emissions through energy recovery, with obvious energy conservation and environmental protection advantages.

[0024] As can be seen from the above, based on the supercomputer heat source monitoring module, the present invention collects the operating temperature and heat load information of the supercomputer cluster in real time, and uses this data to control the liquid cooling system, the supercritical CO2 heat exchange module and the dynamic heat diversion system. This closed-loop control mechanism enables the system to automatically adjust the cooling and heat storage strategies according to the changes in the supercomputer load, thereby improving the self-adaptability and response speed of thermal management, and ensuring that the supercomputer cluster can operate efficiently and safely under different load conditions; Adopt immersion liquid cooling to replace traditional air cooling or cold plate liquid cooling, so that the server is in direct contact with the cooling medium, enhancing the heat exchange efficiency and reducing the risk of local hot spots. At the same time, combined with the supercritical CO2 heat energy enhancement technology, quickly absorb, transmit and enhance heat, ensuring that the cooling system has higher heat conduction efficiency and lower energy loss, and still maintaining stable temperature control during high-load operation.

[0025] The present invention uses dual-mode heat diversion processing to reasonably distribute excess thermal energy to the heat storage or waste heat power generation system while ensuring the heat dissipation needs of the supercomputing cluster. It uses nano-modified phase change material PCM to achieve higher-density energy storage in a limited space. Compared with traditional sensible heat storage methods, phase change heat storage technology can store and release a large amount of thermal energy with little temperature change, and combine with underground heat storage systems for deep storage, thereby improving the stability and long-term availability of the overall energy storage, and providing support for regional heating or future energy peak regulation.

[0026] In the waste heat power generation mode, if the ORC system diverts about 4 kW of heat energy to the power generation branch, it can achieve about 0.6 kW of electrical energy output at a conversion efficiency of 15%; At the same time, underground heat energy is transported to surrounding buildings through a pipeline network to meet heating needs.

[0027] From the above, we can see that through the combined effect of immersion liquid cooling and supercritical CO2 module, the chip temperature quickly dropped from 85°C to less than 70°C; The overall heat recovery rate of the system has been improved, and auxiliary power generation and district heating have increased the overall energy efficiency by more than 25% during the experimental phase; Utilizing adaptive PID and dynamic flow-dividing algorithms, the system can respond quickly and ensure safe operation under high load, ensuring that all modules work in coordination.

[0028] Embodiment 2: Real-time rapid cooling and energy recovery under high load conditions: 1. Working condition description: In a supercomputing data center, the supercomputing cluster generates a lot of heat when it is under high load (close to 100% load). The chip temperature monitoring shows that the actual temperature is about 85°C, while the safe operating temperature is required to be controlled within 70°C. In order to achieve rapid cooling and synchronous energy recovery, the system uses the following modules to work together.

[0029] 2. Data collection and parameter measurement: Temperature monitoring: In the supercomputing heat source monitoring module, a high-precision RTD sensor (accuracy ±0.2°C) is used to record the chip surface temperature in real time with a sampling period of 1 second.

[0030] Flow and pressure monitoring: Use flow meters (accuracy ±1%) and pressure sensors (±0.05 MPa) to monitor the flow and pressure of liquid cooling medium and supercritical CO2 to ensure accurate data feedback to the control system.

[0031] SCADA system: All collected data is processed in real time through a centralized control system, triggering subsequent control algorithms to adjust the parameters of each module.

[0032] 3. Module calculation: (1) Immersion Liquid Cooling Heat Absorption Module: Cooling medium flow rate: 0.6 kg / s Specific heat capacity of the cooling medium: approximately 4200 J / (kg·K) Assume the temperature difference between the chip and the cooling medium: ΔT = 50 K Calculation process: = 0.6×4200×50 ≈ 126 kW That is, a single module can absorb approximately 126 kW of thermal energy under high load to ensure a rapid drop in chip temperature.

[0033] (2) Supercritical CO2 Thermal Energy Enhancement Module: CO2 flow rate: 0.5 kg / s Specific heat capacity of supercritical CO2: approximately 1000 J / (kg·K) Inlet temperature: approximately 40°C; Target temperature: 80°C, ΔT = 40 K PID control parameters: Kp = 0.3, Ki = 0.07, Kd = 0.02 Thermal energy enhancement amount: Qboost = 0.5×1000×40 = 20 kW Through the PID formula: u(t) = 0.3 e(t) + 0.07∫e(t) dt + 0.02de(t) / dt The system dynamically adjusts the CO2 circulation parameters according to the actual temperature error: e(t) = Ttarget−Tout(t) (target 80°C) so that the target temperature can be stably reached in about 10 seconds.

[0034] (3) Dual-Mode Thermal Shunt Processing Module: Algorithm idea: At high load, to ensure equipment cooling, the high-temperature fluid is preferentially used for cooling; at the same time, a part of the energy is reserved for auxiliary power generation.

[0035] Shunt: Assume the total thermal energy of the high-temperature fluid is Qtotal = 20 kW, and the system sets the shunt ratio according to the real-time load as: Cooling branch ratio: 80% Power generation / thermal storage branch ratio: 20% Qcooling = 0.8×20 kW = 16 kW, Qpower = 0.2×20 kW = 4 kW This dynamic shunt ensures that the cooling demand is preferentially met at high load, and at the same time, waste heat utilization is achieved within a safe range.

[0036] (4) Nano-Phase Change Thermal Storage Module: Using graphene to enhance molten salt, basic latent heat L0 = 250 kJ / kg, nano-enhancement coefficient η = 1.5 Total PCM mass: 1000 kg Sensible heat storage: specific heat capacity cp = 2000 J / (kg·K), temperature range ΔT = 20 K Calculation process: Latent heat storage: Ql = 1000 × (250 kJ / kg × 1.5) = 1000 × 375 kJ / kg = 375 MJ Sensible heat storage: Qs = 1000 × 2000 × 20 = 40 MJ Total heat storage capacity: Qtotal=375+40=415 MJ This design enables the system to effectively buffer instantaneous thermal energy fluctuations during high loads and provide sufficient heat source for subsequent district heating and auxiliary power generation.

[0037] (5) Underground geothermal well coupling and regional utilization module: Design parameters and monitoring: Underground injection temperature target: 55°C Injection pressure: about 2 MPa Control method: The pipeline flow and temperature are adjusted in real time by controlling the valve, and the underground heat storage temperature is monitored by using distributed temperature sensors (accuracy ±0.2℃).

[0038] Energy utilization: In the waste heat power generation mode, if the ORC system diverts about 4 kW of heat energy to the power generation branch, it can achieve about 0.6 kW of electrical energy output at a conversion efficiency of 15%; At the same time, underground heat energy is transported to surrounding buildings through a pipeline network to meet heating needs.

[0039] From the above, we can see that through the combined effect of immersion liquid cooling and supercritical CO2 module, the chip temperature quickly dropped from 85°C to less than 70°C; The overall heat recovery rate of the system has been improved, and auxiliary power generation and district heating have increased the overall energy efficiency by more than 25% during the experimental phase; Utilizing adaptive PID and dynamic flow-dividing algorithms, the system can respond quickly and ensure safe operation under high load, ensuring that all modules work in coordination.

[0040] Embodiment three:

[0041] Thermal energy storage and delayed utilization under low load / peak conditions: 1. Working condition description: When the supercomputing cluster is at about 50% load, the actual temperature of the chip is about 75 °C, and the cooling requirement is lower compared to high load. At this time, the system focuses on efficiently storing the excess heat energy for use during subsequent load peaks or for district heating to achieve energy balance scheduling.

[0042] 2. Data acquisition and parameter measurement: Temperature and flow rate acquisition: Similarly, high-precision RTD sensors, flow meters, and pressure sensors are used to record data through the SCADA system at a sampling frequency of 1 second / time; Load monitoring: The load rate of the supercomputing cluster is monitored in real time, and a high-proportion heat storage strategy is triggered when it is below 60%.

[0043] 3. Module calculation: (1)Immersion liquid cooling heat absorption module: Cooling medium flow rate: 0.4 kg / s Specific heat capacity: 4200 J / (kg·K) Assumed temperature difference between the chip and the medium: ΔT = 40 K Calculation process: Qabs = 0.4×4200×40 ≈ 67.2 kW. At lower loads, less heat energy is absorbed, but it is sufficient to maintain the chip temperature within a safe range.

[0044] (2)Supercritical CO2 heat energy boosting module: CO2 flow rate: 0.4 kg / s Inlet temperature: 40 °C, target temperature: 80 °C (ΔT = 40 K) PID parameters are moderately reduced: Kp = 0.25, Ki = 0.05, Kd = 0.01 Calculated: Qboost = 0.4×1000×40 = 16 kW Similarly, through PID control, it is ensured to reach a stable high-temperature working state at a lower flow rate.

[0045] (3)Dual-mode heat shunt processing module: Shunt strategy adjustment: In the case of low load, the system focuses on heat storage and delayed utilization. The set shunt ratio is: Cooling branch: 30% Heat storage / generation branch: 70% Then the heat energy distribution is: Qcooling = 0.3×16 = 4.8 kW, Qstorage = 0.7×16 = 11.2 kW This shunt scheme not only meets the basic cooling requirements of the chip but also channels most of the heat energy into the heat storage system to reserve energy for subsequent utilization.

[0046] (4) Nano-phase change heat storage module: Design parameters: The same nano-modified material is used, but the PCM storage is appropriately increased, and the total mass is taken as 1500 kg Basic latent heat L0 = 250 kJ / kg, enhancement coefficient η = 1.5 Sensible heat storage parameters: cp = 2000 J / (kg·K), temperature difference ΔT = 20 K Calculation process: Latent heat storage: Ql = 1500×250 kJ / kg×1.5 = 1500×375 kJ / kg = 562.5 MJ Sensible heat storage: Qs = 1500×2000×20 = 60 MJ Total heat storage capacity: Qtotal = 562.5 + 60 = 622.5 MJ The greatly improved heat storage capacity can store excess heat energy during low load periods, alleviating the energy demand during subsequent load peaks.

[0047] (5) Underground geothermal well coupling and regional utilization module: Regulation measures: The target underground injection temperature is also set at 55°C, and the injection pressure is maintained at about 2 MPa Adjust the valve opening in real time to ensure uniform temperature in the underground heat storage layer and avoid local overheating or insufficiency District heating and auxiliary power generation: In the heat storage system, 11.2 kW of heat energy diverted to the heat storage branch can be converted into about 1.68 kW of electric energy with an efficiency of about 15% through the ORC system under low load; At the same time, the stored heat energy is released through the pipe network during peak demand periods to supply heating or hot water supply for surrounding buildings, forming energy peak shaving in terms of time.

[0048] As can be seen from the above, during the low load stage, the system realizes large-scale heat energy storage, with a cumulative heat storage capacity of 622.5 MJ, providing a stable heat source during peak load periods.

[0049] The dynamic shunt algorithm ensures that while meeting the cooling demand, 70% of the waste heat is efficiently transferred to the heat storage module, achieving energy balance scheduling between low load valleys and high load peaks.

[0050] Through underground geothermal well coupling, the stored heat energy can be used for district heating during peak demand periods, reducing overall energy consumption and improving energy utilization efficiency.

[0051] Although embodiments of the present invention have been shown and described, it will be understood by those of ordinary skill in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention, and the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An underground space supercomputer cluster cooling and heat storage exchange system, characterized in that Comprising: A supercomputer heat source monitoring module for real-time monitoring of the temperature and operating status of a supercomputer chip to obtain initial heat energy data; An immersion liquid cooling heat absorption module for directly immersing the supercomputer chip in a liquid cooling medium and calculating the heat energy absorbed by the chip through a dynamic heat transfer model to obtain preheated liquid; A supercritical CO2 heat energy enhancement module for using supercritical CO2 cycle technology and combining the initial heat energy data to perform heat energy enhancement treatment on the preheated liquid to obtain a high-temperature working fluid; A dual-mode heat shunt processing module for intelligently shunting the high-temperature working fluid to obtain a shunted heat energy flow; A nano-phase change heat storage module for optimizing heat storage of the shunted heat energy flow through a nano-modified phase change material to obtain a stored heat energy flow; An underground geothermal well coupling and regulation module for directly coupling the stored heat energy flow with an underground heat reservoir through a control valve and a pipeline system to obtain a regulated heat energy flow; A district heating and auxiliary power generation module for further utilizing the regulated heat energy flow, realizing waste heat power generation through an ORC system, and simultaneously delivering hot water at 50 - 60 °C to surrounding buildings for district heating to obtain the comprehensive heat energy utilization efficiency.

2. The underground space supercomputer cluster cooling and heat storage exchange system according to claim 1, characterized in that, The step of using supercritical CO2 cycle technology and combining the initial heat energy data to perform heat energy enhancement treatment on the preheated liquid to obtain a high-temperature working fluid includes: Based on an adaptive heat energy enhancement algorithm, combining the non-linear change of the specific heat capacity of supercritical CO2 at different temperatures and pressures, and adopting an adaptive PID control algorithm to adjust the CO2 cycle parameters in real time according to the temperature and flow rate of the preheated liquid; Wherein the formula of the adaptive heat energy enhancement algorithm is: ； Where Qabs is the absorbed heat energy, in units of J; m is the CO2 flow rate, in units of kg / s; is the specific heat capacity related to temperature, with the unit J / (kg·K); ΔT1 is the temperature difference, in units of K; The formula of the adaptive PID control algorithm is: ； Among them, the error , Ttarget is the target temperature; Tout(t) is the real-time monitored outlet temperature of the cooling medium; PID parameter range: Kp = 0.2 - 0.5, Ki = 0.05 - 0.1, Kd = 0.01 - 0.

05.

3. The underground space supercomputer cluster cooling and heat storage exchange system according to claim 1, characterized in that, The intelligent shunt processing includes two branch processes, one of which is for continuous enhanced cooling and the other is for driving turbine power generation; The intelligent shunt processing adopts a dynamic shunt algorithm for dynamically adjusting the heat energy shunt ratio according to the load and real-time temperature of the supercomputer cluster after obtaining the high-temperature working fluid, and reasonably distributing the heat energy to the cooling and power generation branches; The dynamic shunt algorithm includes a shunt ratio calculation method and an energy shunt calculation method; Wherein the formula of the shunt ratio calculation method is: ； Tset is the safe cooling set temperature; Tfluid is the current fluid temperature; Tmax is the maximum allowable temperature of the fluid; In the algorithm, a range limit needs to be set for β, and it will be automatically adjusted when the calculated value exceeds; The formula of the energy shunt calculation method is: ； Where Qcooling is the energy for continuous enhanced cooling; Qpower is the energy for driving turbine power generation; Qtotal is the total heat energy or total heat storage capacity of the high-temperature fluid.

4. The underground space supercomputer cluster cooling and heat storage exchange system according to claim 1, wherein, Optimizing the heat storage of the divided heat energy through the nano-modified phase change material to obtain the stored heat energy flow, including: Importing the heat energy into the phase change material through a heat exchange system, enabling it to absorb heat and undergo a phase change, storing a large amount of latent heat, and obtaining the latent heat of phase change; After the phase change process is completed, the phase change material continues to absorb heat, and the temperature rises without a change in phase state, obtaining sensible heat storage; Optimizing the latent heat of phase change and the sensible heat storage to obtain the stored heat energy flow; Among them, the calculation formula for the latent heat of phase change is: ； Among them is the PCM quality, unit kg; Leff = L0 × η is the effective latent heat of phase change, with the unit J / kg; L0 is the basic latent heat; η is the nano-enhancement coefficient; The calculation formula for the sensible heat storage is: ； Among them is the specific heat capacity of the material; ΔT2 is the temperature change range; According to the latent heat of phase change and the sensible heat storage, the total heat storage capacity Qtotal is: 。 5. The underground space supercomputer cluster cooling and heat storage exchange system according to claim 1, characterized in that The formula for the dynamic heat transfer model is: ； Among them, h is the convective heat transfer coefficient between the liquid and the chip, with the unit W / (m²·K); A is the contact area between the chip and the cooling medium, with the unit m²; m is the mass of the chip, with the unit kg; cp is the specific heat capacity of the chip, with the unit J / (kg·K); T is the chip temperature, and Tfluid is the temperature of the liquid cooling medium.

6. The underground space supercomputer cluster cooling and heat storage exchange system according to claim 1, characterized in that The supercomputer heat source monitoring module includes a temperature sensor and a data acquisition unit. The temperature sensor is used to monitor the temperature and heat load information of the supercomputer chip, and the data acquisition unit is used to continuously collect the temperature and the heat load information.

7. The underground space supercomputer cluster cooling and heat storage exchange system according to claim 1, characterized in that, The immersion liquid cooling heat absorption module adopts a full immersion structure, enabling the supercomputer server to be directly immersed in the liquid cooling medium to obtain the preheated liquid.

8. The underground space supercomputer cluster cooling and heat storage exchange system according to claim 1, characterized in that, The nano-modified phase change material in the nano-phase change heat storage module includes at least one of graphene-enhanced molten salt and carbon nanotubes.

9. The underground space supercomputer cluster cooling and heat storage exchange system according to claim 1, characterized in that, The district heating and auxiliary power generation module includes an organic Rankine cycle power generation device and a district heating pipe network unit; Among them, the organic Rankine cycle power generation device is used to drive a turbine to generate waste heat power using the divided heat energy; The district heating pipe network unit is used to transport the 50-60°C hot water released from the underground heat storage to the surrounding buildings to obtain the comprehensive heat energy utilization efficiency.

Images