A server liquid cooling rack and a data center cooling system containing the rack

Abstract

The application relates to a server liquid cooling rack and a data center cooling system containing the rack, and belongs to the technical field of data center cooling. The system comprises a liquid cooling rack core heat dissipation unit, a cooling system medium circulation unit, a waste heat recovery and green energy adaptation unit and a system dynamic regulation and fault tolerance unit. The liquid cooling rack core heat dissipation unit is based on the heat generation characteristics of servers to build an integrated liquid cooling circuit, converts the heat generation power into a medium flow parameter and predicts the heat aggregation regulation flow rate. The cooling system medium circulation unit constructs a centralized medium module, connects the circuit and verifies the matching of the medium and the heat dissipation demand. The waste heat recovery and green energy adaptation unit divides the waste heat utilization level, retains a switching interface and drives wind-solar photovoltaic energy storage. The system dynamic regulation and fault tolerance unit constructs an interaction mechanism, adjusts the heat dissipation intensity and processes sensor faults. The heat dissipation efficiency and system adaptability are improved, and the stable operation of high-density computing power of the data center is ensured.

Description

Technical Field

[0001] This invention belongs to the field of data center cooling technology, specifically relating to a server liquid-cooled rack and a data center cooling system containing the rack. Background Technology

[0002] With the development of the digital economy, the demand for computing power in data centers has surged, and server density continues to increase. Equipment heat dissipation has become a key bottleneck restricting its efficient operation. Traditional air-cooling technology is limited by its heat dissipation efficiency, easily leading to localized overheating in high-density server scenarios, and its high fan energy consumption does not meet the requirements of green and low-carbon development. While existing liquid cooling technology can improve heat dissipation efficiency, it still has many limitations: First, it has poor integration and adaptability; liquid cooling loops are mostly fixed designs, making it difficult to flexibly match heterogeneous servers from different manufacturers and with different specifications, requiring extensive customization for installation and debugging. Second, media control lacks intelligent linkage. Centralized media modules are supplied with fixed parameters and are not dynamically linked to server computing load, which can easily lead to insufficient media under high load and wasted media under low load. Third, energy utilization is fragmented, with waste heat mostly being directly discharged, and only a few achieving small-scale heating. It has not formed a synergy with surrounding heating scenarios, and its adaptation with wind, solar and energy storage devices is limited to simple power supply without dynamic switching and priority scheduling mechanisms. Fourth, it has weak fault tolerance and control capabilities, relying on manual monitoring and adjustment. When sensors fail, it is necessary to shut down for troubleshooting. It is impossible to estimate heat dissipation demand through adjacent data, which can easily interrupt the continuity of data center services. Summary of the Invention

[0003] To address the aforementioned problems in the prior art, the present invention provides a server liquid-cooled rack and a data center cooling system containing the rack.

[0004] The objective of this invention can be achieved through the following technical solutions:

[0005] Includes: liquid-cooled rack core heat dissipation unit, cooling system medium circulation unit, waste heat recovery and green energy adaptation unit, system dynamic control and fault tolerance unit;

[0006] The core heat dissipation unit of the liquid-cooled rack is based on the heat distribution characteristics of heterogeneous server devices and builds an integrated liquid cooling circuit. It converts the heat power of different areas of the server into medium flow control parameters of the liquid cooling circuit. Based on the real-time temperature data inside the rack, it predicts the local heat accumulation trend and adjusts the cooling medium flow rate of the corresponding area to form a set of medium control parameters.

[0007] The cooling system media circulation unit constructs a centralized media module, which is connected to the integrated liquid cooling circuit; it acquires the outlet media temperature and pressure data of each liquid cooling rack, verifies the matching relationship between media flow rate and server heat dissipation requirements; it presets media priority allocation rules and generates a media circulation adaptation report.

[0008] The waste heat recovery and green energy adaptation unit classifies waste heat utilization levels based on the medium control parameter set and medium circulation adaptation report; at the same time, it retains the waste heat utilization mode switching interface and constructs a linkage mechanism between the cooling system and wind, solar and energy storage devices; it adjusts the recovery ratio according to the surrounding heat demand and obtains a waste heat recovery and energy consumption balance report.

[0009] The system's dynamic control and fault-tolerant unit constructs an information interaction mechanism to automatically adjust the heat dissipation intensity of each area's liquid-cooled racks based on changes in computing load, and synchronously feeds back the cooling capacity status. When a liquid-cooled circuit sensor malfunctions, the system estimates the heat dissipation demand of the faulty area using data from adjacent sensors and server power consumption data, and simultaneously sends a sensor fault warning to the maintenance terminal. Finally, a system summary report is generated.

[0010] Specifically, the process of building an integrated liquid cooling circuit is as follows: analyze the heat distribution characteristics of each component of the heterogeneous server device, and mark the location and distribution range of the heat-generating components; plan the overall direction of the liquid cooling circuit based on the marking results, with the circuit path covering all heat-generating components and surrounding areas; set the correspondence between the flow channels and each heat-generating component according to the planned path; connect each flow channel sequentially through the main path to form a complete closed circuit; finally, check the circuit and adjust the local flow channel paths until all heat-generating components are covered.

[0011] Specifically, the process of converting the heat generation power of different areas of the server into medium flow rate control parameters for the liquid cooling circuit is as follows: acquiring the real-time heat generation power of each area at a preset fixed period; constructing a power-flow rate correspondence model based on heat dissipation test data under different loads, and calling the power-flow rate correspondence model; inputting the real-time heat generation power into the model, calculating the medium flow rate required for each area's flow channel, and cross-validating the calculated flow rate parameters with historical flow rate parameters under similar loads; organizing the calculation results into structured flow rate control parameters, and marking the applicable load range of the parameters.

[0012] Specifically, the process of predicting local heat accumulation trends based on real-time acquired rack temperature data and adjusting the cooling medium flow rate in the corresponding area is as follows: acquiring real-time temperature data at different locations within the rack and filtering abnormal data in the surrounding areas of heat-generating components; comparing the collected temperature data with a preset safe temperature range and judging based on temperature change trends, issuing an early warning if the temperature continues to rise; determining that the area is a potential heat accumulation area, calculating the required flow rate adjustment range, and adopting a gradual adjustment strategy; sending adjustment commands to the flow control link in the corresponding area, gradually increasing the medium flow rate according to the calculated range; and recording the temperature changes, flow rate parameters, and server load status before and after adjustment.

[0013] As a preferred technical solution of the present invention, the specific process of the core heat dissipation unit of the liquid-cooled rack is as follows: Heat characteristic analysis is performed on each component of the heterogeneous server equipment to distinguish between core high-heat-generating components and low-heat-consumption components, and the location, range, and heat diffusion law of the high-heat-generating area are clarified; based on the analysis results, an integrated liquid-cooling loop layout is planned, with dense flow channels designed in the high-heat-generating area and simplified flow channels used in the low-heat-consumption area to reduce losses. Each flow channel is connected in series through the main path to form a closed loop, and zonal verification is used to ensure that all heat-generating components are covered without dead zones; then, real-time heat power in different areas of the server is continuously collected, and a preset power-flow correspondence model is called to convert the heat power into the medium flow rate control parameters of the corresponding flow channel, which are then organized into structured data and stored in the control unit; finally, temperature data in each area of ​​the rack is collected in real time, with a focus on monitoring temperature changes around high-heat-generating components. If a risk of local heat accumulation is determined, the adjustment range of the medium flow rate is calculated based on the temperature deviation and the current load, and an adjustment command is sent using a gradual strategy, while simultaneously recording the temperature, flow rate, and load data before and after the adjustment.

[0014] Specifically, the process of constructing the centralized media module and connecting it to the integrated liquid cooling circuit is as follows: planning the connection path between the centralized media module and each liquid cooling rack; establishing a main media supply channel, pre-setting segmented control modules at key nodes of the channel to connect the centralized media module to all liquid cooling circuits; pre-setting basic media supply parameters; dynamically adjusting the media supply quantity and temperature according to the real-time needs of each circuit and in conjunction with feedback data; balancing the media distribution of each circuit and performing closed-loop control.

[0015] Specifically, the process for verifying the matching relationship between the medium flow rate and the server's heat dissipation requirements is as follows: acquire the medium temperature and pressure data at the outlet of each liquid-cooled rack, and record the acquisition time point synchronously; retrieve the service load information of each server in the current period, distinguish the current load from the historical average load to analyze the fluctuation; combine the correlation rules between load and heat dissipation requirements to determine the range of medium parameters that should be matched; compare the acquired actual data with the parameter range and calculate the deviation rate; record the deviation value, the corresponding rack information, and the ambient temperature at that time in real time.

[0016] Specifically, the process of the preset media priority allocation rule is as follows: classify the service types of the server, distinguishing between core services, ordinary services and backup services; assign priorities to the corresponding liquid-cooled racks according to the importance of the services, review them regularly and update them dynamically according to changes in services; set the media allocation order corresponding to the priorities, and formulate temporary priority adjustment rules in emergency situations; set media usage limits for each priority.

[0017] Specifically, the process of classifying waste heat utilization levels based on the medium control parameter set and the medium circulation adaptation report is as follows: extract the medium inlet and outlet temperature difference of each loop from the medium control parameter set, extract the medium flow rate from the medium circulation adaptation report, and correct outliers; calculate the total waste heat of the system by combining the temperature difference and flow rate, and use a weighted algorithm; classify the levels according to the numerical range of the total waste heat and set buffer zones; associate each level with typical heat use scenarios according to seasonal factors.

[0018] Specifically, the process of the waste heat utilization mode switching interface is as follows: preset switchable waste heat output paths; establish a mode switching operation mechanism and formulate operation permission management to distinguish permissions; after receiving the switching command, test the receiving capability of the target device and adjust the path direction of waste heat output; monitor the waste heat delivery status during the switching process and track temperature loss; provide feedback on the switching result, along with energy consumption comparison data before and after the switching, and trigger a prompt if an abnormality occurs.

[0019] Specifically, the process of constructing the linkage mechanism between the cooling system and the wind, solar and energy storage device is as follows: establish an information interaction link between the cooling system and the wind, solar and energy storage device to obtain the power supply capacity data of the energy storage device in real time; preset power supply switching conditions, and predict the energy storage power supply capacity in advance based on weather forecast data. When the energy storage power supply capacity meets the needs of the cooling system and has sufficient reserves, switch to energy storage power supply; conduct a power supply stability test before switching, and switch only after confirming that there are no fluctuations. If the energy storage power supply capacity is insufficient, automatically switch back to the conventional power supply; if the power supply is abnormal, preset a switching time threshold.

[0020] Specifically, the process of adjusting the recovery ratio according to the surrounding heat demand is as follows: obtain real-time demand information, including the required heat and time period; analyze the demand information and distinguish between rigid demand and flexible demand; adjust the waste heat recovery ratio according to the analysis results, increasing the ratio when demand increases and decreasing the ratio when demand decreases; simultaneously adjust the intensity of waste heat transmission, monitor the loss in the waste heat transmission process and optimize it in a timely manner; and record the correspondence between demand changes and the recovery ratio.

[0021] Specifically, the information interaction mechanism is used to automatically adjust the heat dissipation intensity of liquid-cooled racks in each region. The specific process is as follows: establish a two-way information channel with the computing power scheduling platform, set up a data verification mechanism, and receive computing power load data from servers in each region; calculate the required heat dissipation intensity for each region's liquid-cooled racks by adjusting the heat dissipation intensity requirement based on the correlation model between load data and heat dissipation requirements and in conjunction with real-time ambient temperature; generate heat dissipation adjustment instructions, simulate the execution effect before generating the instructions, and send them to the corresponding region's liquid-cooled racks; and collect the current operating data of the cooling system.

[0022] The beneficial effects of this invention are as follows:

[0023] (1) By setting up a collaborative mechanism between the core heat dissipation unit of the liquid-cooled rack and the system dynamic control unit, the accuracy of heat dissipation and system adaptability are significantly improved, ensuring stable operation in high-density computing scenarios; an integrated liquid cooling circuit is built based on the heat distribution characteristics of heterogeneous server devices, which can accurately cover heat-generating components in different areas such as CPU and GPU, avoiding the problem of "uneven overall cooling and local overheating" in traditional cooling technology; at the same time, by converting heat power into medium flow rate control parameters, and combining real-time temperature data to predict the heat accumulation trend and adjust the flow rate, the heat dissipation efficiency is greatly improved, and the temperature of core components is kept stable; the system dynamic control unit automatically adjusts the heat dissipation intensity of each area according to the changes in computing load through the information interaction mechanism, increases the medium supply in high-heat areas during peak load, and reduces redundant supply during low load, effectively reducing ineffective energy consumption;

[0024] (2) By setting up a linkage design with waste heat recovery and green energy adaptation unit and cooling system medium circulation unit, the green and low carbon attributes and energy comprehensive utilization rate are enhanced, and the total life cycle cost is reduced; the waste heat utilization level is divided based on the medium control parameter set and circulation adaptation report. The high-level waste heat is supplied to the surrounding heating and industrial heat, the medium level is used for domestic hot water, and the low level meets the needs of the machine room insulation, which greatly improves the waste heat utilization rate and reduces the consumption of traditional energy; at the same time, the unit builds a linkage mechanism with wind and solar energy storage devices, and dynamically switches the power supply according to the energy storage output and the amount of stored energy, which significantly improves the green electricity utilization rate and reduces carbon emissions; the centralized control of the cooling system medium circulation unit avoids the imbalance of medium distribution and further reduces the energy consumption of transmission. Attached Figure Description

[0025] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings.

[0026] Figure 1 This is a system architecture diagram of a server liquid-cooled rack and a data center cooling system containing the rack, according to the present invention.

[0027] Figure 2 This is a state transition diagram of a server liquid-cooled rack and a data center cooling system containing the rack, according to the present invention. Detailed Implementation

[0028] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided.

[0029] Please see Figure 1-2 A server liquid-cooled rack and a data center cooling system containing the rack.

[0030] Includes: liquid-cooled rack core heat dissipation unit, cooling system medium circulation unit, waste heat recovery and green energy adaptation unit, system dynamic control and fault tolerance unit;

[0031] The core heat dissipation unit of the liquid-cooled rack is based on the heat distribution characteristics of heterogeneous server devices and builds an integrated liquid cooling circuit. It converts the heat power of different areas of the server into medium flow control parameters of the liquid cooling circuit. Based on the real-time temperature data inside the rack, it predicts the local heat accumulation trend and adjusts the cooling medium flow rate of the corresponding area to form a set of medium control parameters.

[0032] The cooling system media circulation unit constructs a centralized media module, which is connected to the integrated liquid cooling circuit; it acquires the outlet media temperature and pressure data of each liquid cooling rack, verifies the matching relationship between media flow rate and server heat dissipation requirements; it presets media priority allocation rules and generates a media circulation adaptation report.

[0033] The waste heat recovery and green energy adaptation unit classifies waste heat utilization levels based on the medium control parameter set and medium circulation adaptation report; at the same time, it retains the waste heat utilization mode switching interface and constructs a linkage mechanism between the cooling system and wind, solar and energy storage devices; it adjusts the recovery ratio according to the surrounding heat demand and obtains a waste heat recovery and energy consumption balance report.

[0034] The system's dynamic control and fault-tolerant unit constructs an information interaction mechanism to automatically adjust the heat dissipation intensity of each area's liquid-cooled racks based on changes in computing load, and synchronously feeds back the cooling capacity status. When a liquid-cooled circuit sensor malfunctions, the system estimates the heat dissipation demand of the faulty area using data from adjacent sensors and server power consumption data, and simultaneously sends a sensor fault warning to the maintenance terminal. Finally, a system summary report is generated.

[0035] Specifically, the process of building an integrated liquid cooling circuit is as follows: analyze the heat distribution characteristics of each component of the heterogeneous server device, and mark the location and distribution range of the heat-generating components; plan the overall direction of the liquid cooling circuit based on the marking results, with the circuit path covering all heat-generating components and surrounding areas; set the correspondence between the flow channels and each heat-generating component according to the planned path; connect each flow channel sequentially through the main path to form a complete closed circuit; finally, check the circuit and adjust the local flow channel paths until all heat-generating components are covered.

[0036] In this embodiment, when constructing the integrated liquid cooling loop of the core heat dissipation unit of the liquid-cooled rack, the server is first run continuously for a period of time under typical workloads in daily business scenarios. A thermal imager is used to record the heat distribution of each component in real time, focusing on capturing the peak heat dissipation areas of core components such as the GPU and CPU under high load, as well as the paths through which heat diffuses to the surrounding areas. When planning the flow channels, critical interfaces such as power connectors and network cable slots within the rack are deliberately avoided to prevent the flow channel layout from affecting daily equipment maintenance. After the loop is constructed, the server is run at full load for an extended period. During this time, the rack casing in the flow channel coverage area is manually touched to confirm the absence of localized overheating. Simultaneously, the medium pressure data of each flow channel is monitored through the control unit to ensure balanced pressure distribution and prevent insufficient flow rate due to unreasonable local path design, which would hinder effective heat removal.

[0037] Specifically, the process of converting the heat generation power of different areas of the server into medium flow rate control parameters for the liquid cooling circuit is as follows: acquiring the real-time heat generation power of each area at a preset fixed period; constructing a power-flow rate correspondence model based on heat dissipation test data under different loads, and calling the power-flow rate correspondence model; inputting the real-time heat generation power into the model, calculating the medium flow rate required for each area's flow channel, and cross-validating the calculated flow rate parameters with historical flow rate parameters under similar loads; organizing the calculation results into structured flow rate control parameters, and marking the applicable load range of the parameters.

[0038] Specifically, the process of predicting local heat accumulation trends based on real-time acquired rack temperature data and adjusting the cooling medium flow rate in the corresponding area is as follows: acquiring real-time temperature data at different locations within the rack and filtering abnormal data in the surrounding areas of heat-generating components; comparing the collected temperature data with a preset safe temperature range and judging based on temperature change trends, issuing an early warning if the temperature continues to rise; determining that the area is a potential heat accumulation area, calculating the required flow rate adjustment range, and adopting a gradual adjustment strategy; sending adjustment commands to the flow control link in the corresponding area, gradually increasing the medium flow rate according to the calculated range; and recording the temperature changes, flow rate parameters, and server load status before and after adjustment.

[0039] As a preferred technical solution of the present invention, the specific process of the core heat dissipation unit of the liquid-cooled rack is as follows: Heat characteristic analysis is performed on each component of the heterogeneous server equipment to distinguish between core high-heat-generating components and low-heat-consumption components, and the location, range, and heat diffusion law of the high-heat-generating area are clarified; based on the analysis results, an integrated liquid-cooling loop layout is planned, with dense flow channels designed in the high-heat-generating area and simplified flow channels used in the low-heat-consumption area to reduce losses. Each flow channel is connected in series through the main path to form a closed loop, and zonal verification is used to ensure that all heat-generating components are covered without dead zones; then, real-time heat power in different areas of the server is continuously collected, and a preset power-flow correspondence model is called to convert the heat power into the medium flow rate control parameters of the corresponding flow channel, which are then organized into structured data and stored in the control unit; finally, temperature data in each area of ​​the rack is collected in real time, with a focus on monitoring temperature changes around high-heat-generating components. If a risk of local heat accumulation is determined, the adjustment range of the medium flow rate is calculated based on the temperature deviation and the current load, and an adjustment command is sent using a gradual strategy, while simultaneously recording the temperature, flow rate, and load data before and after the adjustment.

[0040] Specifically, the process of constructing the centralized media module and connecting it to the integrated liquid cooling circuit is as follows: planning the connection path between the centralized media module and each liquid cooling rack; establishing a main media supply channel, pre-setting segmented control modules at key nodes of the channel to connect the centralized media module to all liquid cooling circuits; pre-setting basic media supply parameters; dynamically adjusting the media supply quantity and temperature according to the real-time needs of each circuit and in conjunction with feedback data; balancing the media distribution of each circuit and performing closed-loop control.

[0041] Specifically, the process for verifying the matching relationship between the medium flow rate and the server's heat dissipation requirements is as follows: acquire the medium temperature and pressure data at the outlet of each liquid-cooled rack, and record the acquisition time point synchronously; retrieve the service load information of each server in the current period, distinguish the current load from the historical average load to analyze the fluctuation; combine the correlation rules between load and heat dissipation requirements to determine the range of medium parameters that should be matched; compare the acquired actual data with the parameter range and calculate the deviation rate; record the deviation value, the corresponding rack information, and the ambient temperature at that time in real time.

[0042] Specifically, the process of the preset media priority allocation rule is as follows: classify the service types of the server, distinguishing between core services, ordinary services and backup services; assign priorities to the corresponding liquid-cooled racks according to the importance of the services, review them regularly and update them dynamically according to changes in services; set the media allocation order corresponding to the priorities, and formulate temporary priority adjustment rules in emergency situations; set media usage limits for each priority.

[0043] In this embodiment, when the cooling system's media circulation unit presets media priority allocation rules, it first collaborates with the data center's business departments to identify the types of services carried by all servers. Core services include online transaction systems for users and real-time backup systems for core data; ordinary services include OA systems for employees' daily office work and storage systems for non-critical data; and backup services are temporary servers used for testing new features. After prioritizing the corresponding liquid-cooled racks according to this classification, the unit also synchronizes business changes with the business departments monthly. For example, if a new core service is launched, the priority of the server rack carrying that service is promptly increased. In case of emergencies, such as a core service requiring temporary expansion, the media allocation order of the rack corresponding to that service is temporarily adjusted to prioritize its needs. The original rules are restored after the expansion is completed to prevent heat dissipation problems caused by insufficient media in core services.

[0044] Specifically, the process of classifying waste heat utilization levels based on the medium control parameter set and the medium circulation adaptation report is as follows: extract the medium inlet and outlet temperature difference of each loop from the medium control parameter set, extract the medium flow rate from the medium circulation adaptation report, and correct outliers; calculate the total waste heat of the system by combining the temperature difference and flow rate, and use a weighted algorithm; classify the levels according to the numerical range of the total waste heat and set buffer zones; associate each level with typical heat use scenarios according to seasonal factors.

[0045] Specifically, the process of the waste heat utilization mode switching interface is as follows: preset switchable waste heat output paths; establish a mode switching operation mechanism and formulate operation permission management to distinguish permissions; after receiving the switching command, test the receiving capability of the target device and adjust the path direction of waste heat output; monitor the waste heat delivery status during the switching process and track temperature loss; provide feedback on the switching result, along with energy consumption comparison data before and after the switching, and trigger a prompt if an abnormality occurs.

[0046] Specifically, the process of constructing the linkage mechanism between the cooling system and the wind, solar and energy storage device is as follows: establish an information interaction link between the cooling system and the wind, solar and energy storage device to obtain the power supply capacity data of the energy storage device in real time; preset power supply switching conditions, and predict the energy storage power supply capacity in advance based on weather forecast data. When the energy storage power supply capacity meets the needs of the cooling system and has sufficient reserves, switch to energy storage power supply; conduct a power supply stability test before switching, and switch only after confirming that there are no fluctuations. If the energy storage power supply capacity is insufficient, automatically switch back to the conventional power supply; if the power supply is abnormal, preset a switching time threshold.

[0047] In this embodiment, when the waste heat recovery and green energy adaptation unit constructs the linkage mechanism between the cooling system and the wind, solar, and energy storage devices, it first establishes regular data interaction with the local meteorological department to obtain meteorological information such as wind intensity and sunshine duration for the next few days. If it is predicted that there will be sufficient sunshine or strong winds the next day, the wind, solar, and energy storage devices will be controlled to enter the priority charging mode the day before to maximize the energy storage capacity. The next day, after the output power of the energy storage devices gradually stabilizes, the power supply of the circulating pump in the cooling system is switched to the energy storage power supply first. After observing for a period of time and confirming that the operating parameters of the circulating pump are normal, the power supply of other equipment such as the temperature control components and waste heat transfer pumps is gradually switched over as well. If fluctuations in the energy storage output are detected during the process, such as a decrease in photovoltaic output due to cloud cover, some non-critical equipment will be immediately switched back to municipal power supply to ensure that the core functions of the cooling system are not affected and to maintain stable server heat dissipation.

[0048] Specifically, the process of adjusting the recovery ratio according to the surrounding heat demand is as follows: obtain real-time demand information, including the required heat and time period; analyze the demand information and distinguish between rigid demand and flexible demand; adjust the waste heat recovery ratio according to the analysis results, increasing the ratio when demand increases and decreasing the ratio when demand decreases; simultaneously adjust the intensity of waste heat transmission, monitor the loss in the waste heat transmission process and optimize it in a timely manner; and record the correspondence between demand changes and the recovery ratio.

[0049] In this embodiment, when the system's dynamic control and fault-tolerant unit handles liquid-cooled circuit sensor faults, once it detects that a sensor in a certain area has no data feedback, it immediately retrieves the real-time temperature data of adjacent sensors around that sensor. Simultaneously, it obtains the current load status of the corresponding hardware in that area from the server's management system. Combining this information with preset load-temperature correlation logic, it estimates the approximate temperature of the faulty area. Then, the control unit automatically fine-tunes the medium flow rate in the flow channel of that area to ensure heat dissipation. At the same time, it sends detailed warning information to the maintenance terminal, specifying the location of the faulty sensor, the estimated current temperature, and the suggested processing time. Before maintenance personnel replace the sensor, the system re-estimates the temperature of the faulty area at regular intervals and compares it with adjacent sensor data to ensure that the temperature remains within a safe range. During this process, server operation does not need to be stopped, ensuring normal business operations.

[0050] In this embodiment, when the waste heat recovery and green energy adaptation unit adjusts the recovery ratio according to the surrounding heat demand, it will regularly communicate with the property management of residential communities and agricultural greenhouse management teams around the data center to understand their recent changes in heat demand. For example, in winter, residential communities need to maintain indoor warmth, especially during the morning and evening when heat demand is higher. During these periods, the waste heat recovery ratio will be appropriately increased, and the waste heat will be transported to the community's heat exchange station through well-insulated pipes. At the same time, personnel will be arranged to regularly check the connection of the transport pipes to prevent heat leakage. In spring and autumn, agricultural greenhouses need to provide a stable temperature environment for crop seedling cultivation. At this time, the stability of waste heat recovery will be adjusted to avoid frequent changes in the recovery ratio that cause the greenhouse temperature to fluctuate. Temperature data fed back by the greenhouse will also be collected every day, and the recovery ratio will be fine-tuned according to the actual situation.

[0051] Specifically, the information interaction mechanism is used to automatically adjust the heat dissipation intensity of liquid-cooled racks in each region. The specific process is as follows: establish a two-way information channel with the computing power scheduling platform, set up a data verification mechanism, and receive computing power load data from servers in each region; calculate the required heat dissipation intensity for each region's liquid-cooled racks by adjusting the heat dissipation intensity requirement based on the correlation model between load data and heat dissipation requirements and in conjunction with real-time ambient temperature; generate heat dissipation adjustment instructions, simulate the execution effect before generating the instructions, and send them to the corresponding region's liquid-cooled racks; and collect the current operating data of the cooling system.

[0052] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A server liquid-cooled rack and a data center cooling system containing the rack, characterized in that, include: Liquid-cooled rack core heat dissipation unit, cooling system medium circulation unit, waste heat recovery and green energy adaptation unit, system dynamic control and fault tolerance unit; The core heat dissipation unit of the liquid-cooled rack is based on the heat distribution characteristics of heterogeneous server devices and builds an integrated liquid cooling circuit. It converts the heat power of different areas of the server into medium flow control parameters of the liquid cooling circuit. Based on the real-time temperature data inside the rack, it predicts the local heat accumulation trend and adjusts the cooling medium flow rate of the corresponding area to form a set of medium control parameters. The cooling system media circulation unit constructs a centralized media module, which is connected to the integrated liquid cooling circuit; it acquires the outlet media temperature and pressure data of each liquid cooling rack to verify the matching relationship between media flow rate and server heat dissipation requirements; Preset media priority allocation rules and generate media cycle adaptation reports; The waste heat recovery and green energy adaptation unit classifies waste heat utilization levels based on the medium control parameter set and medium circulation adaptation report; at the same time, it retains the waste heat utilization mode switching interface and constructs a linkage mechanism between the cooling system and wind, solar and energy storage devices; it adjusts the recovery ratio according to the surrounding heat demand and obtains a waste heat recovery and energy consumption balance report. The system's dynamic control and fault-tolerant unit constructs an information interaction mechanism to automatically adjust the heat dissipation intensity of each area's liquid-cooled racks based on changes in computing load, and synchronously feeds back the cooling capacity status. When a liquid-cooled circuit sensor malfunctions, the system estimates the heat dissipation demand of the faulty area using data from adjacent sensors and server power consumption data, and simultaneously sends a sensor fault warning to the maintenance terminal. Finally, a system summary report is generated.

2. The system according to claim 1, characterized in that, The specific process of building an integrated liquid cooling circuit is as follows: analyze the heat distribution characteristics of each component of the heterogeneous server device, and mark the location and distribution range of the heat-generating components; plan the overall direction of the liquid cooling circuit based on the marking results, with the circuit path covering all heat-generating components and surrounding areas; set the correspondence between the flow channels and each heat-generating component according to the planned path; connect each flow channel sequentially through the main path to form a complete closed circuit; finally, check the circuit and adjust the local flow channel paths until all heat-generating components are covered.

3. The system according to claim 1, characterized in that, The specific process of converting the heat generation power of different areas of the server into medium flow rate control parameters for the liquid cooling circuit is as follows: real-time heat generation power of each area is acquired at a preset fixed period; a power-flow correspondence model is constructed based on heat dissipation test data under different loads, and the power-flow correspondence model is called; the real-time heat generation power is input into the model to calculate the medium flow rate required for each area's flow channel, and the calculated flow rate is cross-validated with the flow rate parameters under similar historical loads; the calculation results are organized into structured flow rate control parameters, and the applicable load range of the parameters is marked.

4. The system according to claim 1, characterized in that, The specific process of predicting local heat accumulation trends and adjusting the cooling medium flow rate in the corresponding area based on real-time acquired rack temperature data is as follows: real-time temperature data at different locations within the rack is acquired, and abnormal data in the surrounding areas of heat-generating components is filtered out; the collected temperature data is compared with a preset safe temperature range, and a warning is issued if the temperature continues to rise; the area is determined to be a potential heat accumulation area, the required flow rate adjustment range is calculated, and a gradual adjustment strategy is adopted; an adjustment command is sent to the flow control link in the corresponding area, and the medium flow rate is gradually increased according to the calculated range; the temperature changes, flow rate parameters, and server load status before and after adjustment are recorded.

5. The system according to claim 1, characterized in that, The specific process of constructing the centralized media module and connecting it with the integrated liquid cooling circuit is as follows: planning the connection path between the centralized media module and each liquid cooling rack; establishing a main media supply channel, pre-setting segmented control modules at key nodes of the channel to connect the centralized media module with all liquid cooling circuits; pre-setting basic media supply parameters; dynamically adjusting the media supply quantity and temperature according to the real-time needs of each circuit and in combination with feedback data; balancing the media distribution of each circuit and performing closed-loop control.

6. The system according to claim 1, characterized in that, The specific process for verifying the matching relationship between the medium flow rate and the server's heat dissipation requirements is as follows: acquire the medium temperature and pressure data at the outlet of each liquid cooling rack, and record the acquisition time point simultaneously; retrieve the business load information of each server in the current period, distinguish the current load from the historical average load to analyze the fluctuation; combine the correlation rules between load and heat dissipation requirements to determine the range of medium parameters that should be matched; compare the collected actual data with the parameter range and calculate the deviation rate; record the deviation value, the corresponding rack information, and the ambient temperature at that time in real time.

7. The system according to claim 1, characterized in that, The specific process of the preset media priority allocation rule is as follows: classify the service types of the server, distinguishing between core services, ordinary services and backup services; assign priorities to the corresponding liquid-cooled racks according to the importance of the services, review them regularly and update them dynamically according to changes in services; set the media allocation order corresponding to the priorities, and formulate temporary priority adjustment rules in emergency situations; set media usage limits for each priority.

8. The system according to claim 1, characterized in that, The specific process of classifying waste heat utilization levels based on the medium control parameter set and the medium circulation adaptation report is as follows: extract the medium inlet and outlet temperature difference of each loop from the medium control parameter set, extract the medium flow rate from the medium circulation adaptation report, and correct outliers; calculate the total waste heat of the system by combining the temperature difference and flow rate, and use a weighted algorithm; classify the levels according to the numerical range of the total waste heat and set buffer zones; associate each level with typical heat use scenarios according to seasonal factors.

9. The system according to claim 1, characterized in that, The specific process of the waste heat utilization mode switching interface is as follows: preset switchable waste heat output paths; establish a mode switching operation mechanism and formulate operation permission management to distinguish permissions; after receiving the switching command, test the receiving capability of the target device and adjust the path direction of waste heat output. Monitor the waste heat transfer status during the switching process and track temperature loss; provide feedback on the switching results, along with energy consumption comparison data before and after the switching; and trigger prompts if any abnormalities occur.

10. The system according to claim 1, characterized in that, The specific process of constructing the linkage mechanism between the cooling system and the wind, solar and energy storage device is as follows: establish an information interaction link between the cooling system and the wind, solar and energy storage device to obtain the power supply capacity data of the energy storage device in real time; preset power supply switching conditions, and predict the energy storage power supply capacity in advance based on weather forecast data. When the energy storage power supply capacity meets the needs of the cooling system and has sufficient reserves, switch to energy storage power supply; conduct a power supply stability test before switching, and switch only after confirming that there are no fluctuations. If the energy storage power supply capacity is insufficient, automatically switch back to the conventional power supply; if the power supply is abnormal, preset a switching time threshold.

11. The system according to claim 1, characterized in that, The specific process of adjusting the recovery ratio according to the surrounding heat demand is as follows: obtain real-time demand information, including the required heat and time period; analyze the demand information and distinguish between rigid demand and flexible demand; adjust the waste heat recovery ratio according to the analysis results, increasing the ratio when demand increases and decreasing the ratio when demand decreases; simultaneously adjust the intensity of waste heat transmission, monitor the loss in the waste heat transmission process and optimize it in a timely manner; record the correspondence between demand changes and the recovery ratio.

12. The system according to claim 1, characterized in that, The aforementioned information interaction mechanism is used to automatically adjust the heat dissipation intensity of liquid-cooled racks in each region. The specific process is as follows: establish a two-way information channel with the computing power scheduling platform, set up a data verification mechanism, and receive computing power load data from servers in each region; calculate the required heat dissipation intensity for each region's liquid-cooled racks based on the correlation model between load data and heat dissipation requirements, combined with real-time ambient temperature correction; generate heat dissipation adjustment commands, simulate the execution effect before command generation, and send them to the corresponding region's liquid-cooled racks; and collect the current operating data of the cooling system.

Images