System and method for regulating and controlling operation state of data center server based on thermal imaging

Abstract

The invention discloses a data center server operation state regulation and control system and method based on thermal imaging. The system comprises a thermal imaging monitoring module used for collecting infrared thermal imaging data of a server and a machine room environment, a sliding rail fan system used for conducting directional heat dissipation on a designated area, a data processing and analyzing module used for recognizing the running state of the server and positioning a thermal abnormal area, and an intelligent control module. The data processing and analyzing module is in communication connection with the thermal imaging monitoring module and is used for receiving the infrared thermal imaging data and predicting the temperature change trend of the thermal abnormal area based on historical data; the intelligent control module is in communication connection with the data processing and analysis module, the sliding rail fan system and the machine room cooling system and used for generating a control instruction according to the analysis result of the data processing and analysis module and driving the sliding rail fan system and the machine room cooling system to execute adaptive regulation and control strategies. And the high-efficiency and low-carbon development requirements of the data center are met.

Description

Technical Field

[0001] This invention belongs to the field of data center thermal management technology, and more specifically, relates to a data center server operation status control system and method based on thermal imaging. Background Technology

[0002] During data center operation, server clusters continuously operate under high load, generating a large amount of heat. If this heat cannot be dissipated in time, it will cause abnormally high server temperatures, affecting operational stability and even leading to serious problems such as hardware failure and data loss. Therefore, temperature control is a core and critical aspect of data center operation and maintenance. With the rapid development of the digital economy, data centers are constantly expanding in scale, server density is continuously increasing, and heat load is significantly increasing, placing higher demands on the accuracy, timeliness, and efficiency of temperature control.

[0003] Currently, data centers generally employ traditional fixed cooling methods, relying on fixed fans and air conditioning systems distributed throughout the server room for overall cooling. This approach has significant limitations. Traditional monitoring methods often use point sensors, which can only acquire temperature data at localized locations, making it difficult to achieve full coverage monitoring of the server room's thermal field. This lack of comprehensive understanding of temperature distribution makes it difficult to detect scattered or sudden localized thermal anomalies in a timely manner. Furthermore, the airflow direction and intensity of fixed cooling equipment cannot be flexibly adjusted, only providing uniform global cooling. This makes it difficult to specifically address localized thermal anomalies, resulting in wasted cooling resources and failing to guarantee effective cooling.

[0004] Furthermore, the external environment and internal load of data centers are constantly changing. Factors such as server load fluctuations and changes in ambient temperature can cause real-time changes in the thermal field of the data center. Traditional control methods lack effective dynamic response mechanisms, and their fixed and rigid control strategies cannot be adjusted in a timely manner according to changes in the thermal field, easily leading to control lag or over-control. These problems not only affect the safe and stable operation of servers but also cause excessive energy consumption, increasing the operation and maintenance costs of data centers, which is inconsistent with the trend of green and low-carbon development. Therefore, solving the practical problems of incomplete temperature monitoring, insufficient heat dissipation accuracy, and delayed control response in data centers, and improving the level of intelligent temperature control, is of great significance for ensuring the stable operation of data centers, reducing energy consumption, and saving operation and maintenance costs. This has become a technological direction that urgently needs to be broken through in the field of data center thermal management. Summary of the Invention

[0005] This invention aims to solve the problems of incomplete temperature monitoring, insufficient heat dissipation accuracy, and lagging control response in current data centers. By using full-coverage thermal imaging monitoring, thermal anomaly location and trend prediction, directional heat dissipation and collaborative control, and a closed-loop feedback mechanism, it achieves accurate, timely, and dynamic temperature control, ensuring the safe and stable operation of servers, improving heat dissipation efficiency, reducing energy consumption and operation and maintenance costs, and meeting the needs of efficient and low-carbon development of data centers.

[0006] To address the aforementioned deficiencies or improvement needs of existing technologies, as a first aspect of this invention, the present invention provides a data center server operation status control system based on thermal imaging, comprising: It includes a thermal imaging monitoring module for collecting infrared thermal imaging data of servers and computer room environment and realizing full coverage monitoring of computer room thermal field, a sliding rail fan system for directional heat dissipation of designated areas, a data processing and analysis module for identifying server operating status and locating thermal anomaly areas, and an intelligent control module for maintaining the server operating within a safe temperature range. The data processing and analysis module is communicatively connected to the thermal imaging monitoring module, and is used to receive infrared thermal imaging data and predict the temperature change trend of thermal anomaly areas based on historical data. The intelligent control module is communicatively connected to the data processing and analysis module, the sliding fan system, and the computer room cooling system, and is used to generate control commands based on the analysis results of the data processing and analysis module, driving the sliding fan system and the computer room cooling system to execute the appropriate control strategies.

[0007] Furthermore, the slide rail fan system is deployed above the hot aisle and includes a guide rail, a slider that moves on the guide rail, a servo motor whose top is fixedly connected to the bottom of the slider and is used to adjust the fan's airflow direction and can control at least two degrees of freedom, and a fan installed at the bottom of the servo motor.

[0008] Furthermore, in the aforementioned sliding rail fan system, all fans are initially positioned at one end of the sliding rail; when a hot spot is detected, the system controls the fans to move to the area corresponding to the hot spot, thus enabling the fans to... Axis coordinates With hot topics Axis coordinates To align the fan, the system program determines the fan's rotation direction using the following mathematical logic: when At that time, the decision result is "rotate in the negative direction of the Y-axis"; when At that time, the decision result is "rotate in the positive direction of the Y-axis"; in, The Y-axis coordinate of the hot spot Let Y be the fan's Y-axis coordinate. .

[0009] Furthermore, the specific steps for adjusting the fan pitch angle via a servo motor in the aforementioned sliding rail fan system are as follows: Get the fan's installation height 3D coordinates of local hotspots and the fixed horizontal distance between the fan and the hotspot The target pitch angle is calculated using the following mathematical formula. Based on the calculation result, the actuator is driven to complete the angle adjustment: in, This is the arctangent function, used to calculate the vertical height difference between the fan and the hotspot. and fixed horizontal distance Find the pitch angle required for the fan to deliver air in a directional manner.

[0010] Furthermore, the temperature trend prediction function of the data processing and analysis module is based on time series analysis methods, which model historical thermal imaging data and server load data to predict temperature changes in various regions within a specific future time period. The specific process is as follows: Select historical temperature series of each area of ​​the computer room within a preset historical period. ,in For area code, Number the thermal imaging monitoring points. For a historic moment, Indicates time Lower region Central monitoring point Temperature value; Synchronously select the server load sequence at the corresponding time. ,in For server number, Indicates time Download server The load value is obtained by normalizing real-time monitoring data of CPU utilization, memory usage, and total power consumption, and directly corresponds to the instantaneous heat generation of the server; at the same time, the data center environmental parameter sequence of the same period in history is extracted. Including ambient air density and ambient temperature It is adapted to the characteristics of the plateau environment; temperature sequence Load sequence and environmental parameter sequences Align by timestamp to build a multi-dimensional related dataset A prediction model was built based on time series analysis methods, and a heat transfer time series correction term was introduced to correlate the dataset. The data within a continuous time window of a preset length is used as the model input vector: ,in The time window length is determined based on the thermal inertia characteristics of the high-altitude computer room to ensure the complete lag cycle of heat transfer coverage. Based on the preset prediction duration after this time window Temperature values ​​at monitoring points in various areas As model output vector By analyzing related datasets Through traversal learning, integrating time series autocorrelation analysis and the physical laws of heat transfer, an input vector is established. With output vector The coupling mapping relationship between them can be expressed mathematically as follows: in The time series mapping function is obtained through data-driven training. For the current time zone monitoring points rate of temperature change The air density at standard sea level, Used to correct the effect of low air density at high altitudes on heat transfer efficiency; This coupling mapping relationship is used to correlate the input data in real time. Perform the calculation, where Given the current moment, obtain the future time period. Predicted temperature values ​​at monitoring points in various areas By calculating the deviation between the predicted temperature and the safety threshold ,in The temperature threshold for safe server operation, combined with the rate of temperature change. This allows for the prediction of the timing and intensity of thermal anomaly risks, providing core data support for the intelligent control module to formulate predictive intervention strategies.

[0011] Furthermore, the control logic executed by the intelligent control module includes: When the temperature of the identified local hot spot exceeds the first real-time threshold, the real-time intervention mode is triggered, controlling the sliding fan system to move above the hot spot and deliver the highest intensity directional airflow, while simultaneously turning on nearby devices for heat dissipation. When the predicted local hotspot temperature will exceed the second warning threshold in the future, the predictive intervention mode is triggered, and the sliding fan system is controlled to move to the predicted position in advance and provide preventative medium-intensity airflow.

[0012] Furthermore, the regulation strategy of the intelligent control module includes: Basic adaptive heat dissipation strategy: Based on the overall heat load and average temperature of the computer room, dynamically adjust the operating parameters of the basic heat dissipation equipment and compensate for the low air density at high altitudes. Local hotspot dynamic elimination strategy: When a local hotspot is identified or a hotspot is predicted to be generated, the sliding rail fan system is controlled to move directly above the hotspot area to blow air in a directional manner, and the fixed cooling equipment near the hotspot is turned on at the same time. Energy-saving operation strategy: When nighttime is detected and the external environment is below the set temperature threshold, the power and airflow of the basic heat dissipation equipment are reduced, and natural cooling sources are used preferentially.

[0013] Furthermore, in the energy-saving operation strategy, while reducing the power and airflow of basic heat dissipation equipment, the target set value of the overall temperature of the computer room and the alarm threshold of local hot spots are dynamically increased.

[0014] Furthermore, the compensation for low air density at high altitudes in the basic adaptive heat dissipation strategy is achieved by multiplying the basic airflow command with a high altitude compensation coefficient greater than 1.

[0015] As a second aspect of the present invention, a method for controlling the operating status of a data center server based on thermal imaging is also provided, applied to a data center server operating status control system based on thermal imaging as described in any one of the claims, comprising the following steps: S1. The thermal imaging monitoring module continuously collects infrared thermal imaging data of the server and computer room environment to achieve full coverage monitoring of the thermal field of the computer room, and transmits the collected infrared thermal imaging data to the data processing and analysis module in real time. S2. After receiving infrared thermal imaging data, the data processing and analysis module analyzes and processes it, identifies the server's operating status and locates the thermal anomaly area. At the same time, it calls historical thermal imaging data and server load data, predicts the temperature change trend of the thermal anomaly area based on a preset prediction model, and sends the identification results, location information and temperature prediction results to the intelligent control module simultaneously. S3. The intelligent control module receives the output information from the data processing and analysis module, combines it with the real-time operating parameters of the computer room and external environmental data, and generates appropriate control commands. S4. The intelligent control module sends control commands to the sliding fan system and the computer room cooling system respectively, driving the sliding fan system to move to the designated position, adjusting the air supply direction and air supply intensity to achieve directional heat dissipation, and at the same time regulating the operating parameters of the computer room cooling system. Through the coordinated execution of the control strategy by the two, the server is maintained to operate within a safe temperature range. S5. The thermal imaging monitoring module continuously feeds back the thermal field data of the computer room after regulation, the data processing and analysis module tracks the temperature changes in the thermal anomaly area in real time, and the intelligent control module dynamically adjusts the control commands according to the feedback results to form a closed-loop regulation.

[0016] In summary, compared with the prior art, the above-described technical solutions conceived by this invention can achieve the following beneficial effects: 1. The data center server operation status control system based on thermal imaging of the present invention collects infrared thermal imaging data of the server and computer room environment through a thermal imaging monitoring module, realizing full-coverage monitoring of the computer room thermal field. After receiving the data, the data processing and analysis module identifies the server operation status, locates thermal anomaly areas, and predicts the temperature change trend of the thermal anomaly areas based on historical data. This technical feature allows the system to comprehensively and in real time grasp the temperature distribution and change patterns of the computer room, breaking through the limitations of traditional point sensors that have limited monitoring range and difficulty in capturing the global thermal field. It accurately locates scattered or sudden thermal anomaly areas and predicts temperature change trends in advance, providing timely and accurate data support for subsequent control and avoiding the problem of untimely heat dissipation due to incomplete or delayed monitoring.

[0017] 2. The thermal imaging-based data center server operation status control system of the present invention establishes a communication connection between the intelligent control module, the data processing and analysis module, the sliding fan system, and the computer room heat dissipation system. Based on the analysis results, it generates control commands to drive the sliding fan system to a designated position, adjust the airflow direction and intensity to achieve directional heat dissipation, and simultaneously regulate the operating parameters of the computer room heat dissipation system. This technical feature achieves precise allocation and coordinated operation of heat dissipation resources. The directional heat dissipation capability of the sliding fan system can specifically solve local thermal anomalies, while the global control of the computer room heat dissipation system ensures overall temperature stability. The synergy between the two avoids resource waste or uneven heat dissipation in traditional fixed heat dissipation modes, improving heat dissipation efficiency while ensuring the safe operation of the server.

[0018] 3. The thermal imaging-based data center server operation status control system of this invention continuously feeds back the controlled thermal field data of the computer room through a thermal imaging monitoring module, while a data processing and analysis module tracks temperature changes in abnormal thermal areas in real time. An intelligent control module dynamically adjusts control commands based on the feedback results, forming a closed-loop control mechanism. This technical feature allows the system to flexibly adjust its control strategy according to real-time changes in the computer room temperature, responding promptly to temperature fluctuations and avoiding over- or under-control. Whether the temperature fluctuations are caused by changes in server load or by changes in the external environment, the system can quickly adapt through closed-loop feedback, always maintaining the server within a safe temperature range and ensuring the stability and reliability of the data center operation. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of a data center server operation status control system based on thermal imaging, according to an embodiment of the present invention. Figure 2 The initial time of the device in this embodiment of the invention A diagram showing the location at any given time; Figure 3 The device in this embodiment of the invention identifies hotspots. A diagram showing the location at any given time; Figure 4 The device in this embodiment of the invention identifies hotspots. A diagram showing the location at any given time; Figure 5 The device in this embodiment of the invention identifies hotspots. A diagram showing the location at any given time; Figure 6 This is a flowchart illustrating the anti-static process of the slide rail fan according to the present invention. Figure 7 This is a schematic diagram of local hotspot adjustment according to an embodiment of the present invention; Figure 8 This is a flowchart illustrating the temperature prediction process according to an embodiment of the present invention. Figure 9 This is a flowchart of the control method of the system according to an embodiment of the present invention; Figure 10 This is a schematic flowchart illustrating the control method of the system according to an embodiment of the present invention; In all the accompanying drawings, the same reference numerals denote the same technical features, specifically: 1-first cold aisle; 3-second cold aisle; 5-third cold aisle; 2-first hot aisle; 4-second hot aisle; 6-cabinet; 7-slide rail; 8-slider; 9-air conditioner; 10-fan; 11-servo motor. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0021] Example 1 Please refer to Figure 1 This embodiment 1 provides a data center server operation status control system based on thermal imaging, including: It includes a thermal imaging monitoring module for collecting infrared thermal imaging data of servers and computer room environment and realizing full coverage monitoring of computer room thermal field, a sliding rail fan system for directional heat dissipation of designated areas, a data processing and analysis module for identifying server operating status and locating thermal anomaly areas, and an intelligent control module for maintaining the server operating within a safe temperature range. The data processing and analysis module is communicatively connected to the thermal imaging monitoring module, and is used to receive infrared thermal imaging data and predict the temperature change trend of thermal anomaly areas based on historical data. The intelligent control module is communicatively connected to the data processing and analysis module, the sliding fan system, and the computer room cooling system, and is used to generate control commands based on the analysis results of the data processing and analysis module, driving the sliding fan system and the computer room cooling system to execute the appropriate control strategies.

[0022] Please refer to Figure 2 , Figure 3 , Figure 4 , Figure 5 , Figure 6 , Figure 7 , Figure 8 This embodiment 1 will further elaborate on the above system modules.

[0023] (1) Thermal imaging monitoring module In a data center where traditional temperature monitoring relies on point sensors and struggles to achieve full-area temperature sensing, thermal imaging monitoring modules play a crucial role in comprehensively capturing the thermal distribution of the server room. This module utilizes an industrial-grade infrared thermal imager suitable for ceiling mounting, with a resolution of at least 384×288 pixels. It is equipped with a wide-angle lens to expand the field of view, achieving a thermal sensitivity of ≤50mk and a temperature measurement range from -20℃ to 100℃, enabling it to capture subtle temperature changes during server operation.

[0024] The thermal imagers are strategically positioned, not only above the first hot aisle (2) and the second hot aisle (4), but also along the four side walls of the server room. This allows for comprehensive monitoring of heat accumulation in both hot aisles 2 and 4, as well as changes in the thermal state of cold aisles 1, 3, and 5. Overlapping field-of-view design ensures complete coverage of the server room without blind spots. The thermal imagers support ONVIF or GenICam protocols, outputting not only video streams but also raw temperature matrix data and temperature statistics for specific areas via API or SDK, providing a comprehensive temperature data source for subsequent data processing. They also support PoE power supply, allowing for simultaneous power and data transmission with a single network cable, simplifying deployment and cabling.

[0025] In addition, the thermal imager features a pan-tilt-zoom (PTZ) function, allowing for remote control of rotation and tilt, and flexible adjustment of the viewing angle within a small range to adapt to subtle changes in the layout of server racks 6 within the data center. This module operates continuously at a second-level scanning frequency, acquiring real-time infrared thermal imaging data of the server rack 6 surface and the data center environment, and continuously obtaining temperature distribution information across the entire area. This breaks through the limitations of traditional point-based monitoring, providing comprehensive, continuous, and accurate basic data support for the data processing and analysis module to identify server operating status and locate areas of thermal anomalies.

[0026] (2) Slide rail fan system Please refer to Figure 2 , Figure 3 , Figure 4 , Figure 5 Given that traditional cooling systems struggle to deliver precise airflow to localized thermal anomalies, and that dry air in high-altitude areas can generate static electricity that may damage equipment, the sliding rail fan system, as the core component for directional cooling, is deployed above the first and second hot aisles 2 and 4 in the computer room. It simultaneously completes the physical spatial loop connection of all hot aisles, specifically designed to cool specific areas with thermal anomalies.

[0027] The core structure of the system includes a guide rail 7, a slider 8, a servo motor 11, and a fan 10. The slider 8 can move flexibly along the guide rail. The top of the servo motor 11 is fixedly connected to the bottom of the slider 8 to adjust the airflow direction of the fan 10. The fan 10 is fixed to the bottom of the servo motor 11 to ensure structural stability during operation.

[0028] Reference Figure 6 Considering the special environment of the plateau region, the system incorporates a targeted anti-static design. The guide rail 7 uses a conductive metal profile with multiple dedicated grounding terminals along its length, connected to the equipotential grounding network of the equipment room via a low-impedance conductor to construct a basic grounding path. The slider 8 integrates an elastic metal brush, which maintains close contact with the conductive surface of the guide rail 7 throughout its movement, forming a continuous dynamic grounding path that discharges static charge generated by friction to the ground in real time. Simultaneously, all metal components mounted on the slider 8, such as the fan 10 and servo motor 11, are electrically interconnected via wires and ultimately connected to the dynamic discharge brush, ensuring the entire moving component is at equipotential and effectively preventing internal discharge from damaging precision electronic components.

[0029] Please refer to Figure 2 , Figure 3 , Figure 4 , Figure 5 The system's positioning logic is highly accurate. All fans 10 are initially positioned at one end of the slide rail 7. When the data processing and analysis module identifies a thermal anomaly area, the system controls the fans to move to the corresponding hotspot area, ensuring the fans... Axis coordinates With hot topics Axis coordinates To align the fan, the system program determines the fan's rotation direction using the following mathematical logic: when At that time, the decision result is "rotate in the negative direction of the Y-axis"; when At that time, the decision result is "rotate in the positive direction of the Y-axis"; in, The Y-axis coordinate of the hot spot Let Y be the fan's Y-axis coordinate. .

[0030] Regarding pitch angle adjustment, the system will first obtain the fan's installation height. 3D coordinates of local hotspots and the fixed horizontal distance between the fan and the hotspot The target pitch angle is calculated using the following mathematical formula. : in, This is the arctangent function, used to calculate the vertical height difference between the fan and the hotspot. and fixed horizontal distance Find the pitch angle required for the fan to deliver air in a directional manner.

[0031] Specifically, such as Figure 2 As shown, in At any given moment, the thermal imaging monitoring module detects a hotspot and its two-dimensional coordinates. The data processing and analysis module calculates the three-dimensional coordinates of the hotspot, obtaining... Then it is transmitted to the slide rail fan system, which starts slider 8 to slide at a constant speed to the hot spot position, thus achieving... ,like Figure 3 As shown, in Arrive at the hotspot location at each time point, compare the size of Y, and obtain... The command is transmitted to servo motor 11, such as Figure 4 As shown, in At that moment, fan 10 rotates 90 degrees in the negative Y-axis direction, simultaneously according to the formula: To obtain the required rotation angle of fan 10 ,in , The arctangent function is used to calculate the vertical height difference between fan 10 and the hotspot. and fixed horizontal distance Determine the pitch angle required for directional airflow from fan 10. Transmit the required rotation angle to servo motor 11, such as... Figure 5 As shown, in At that moment, drive fan 10 to rotate downwards. The fan 10 operates at maximum airflow to cool the hot spot. If a hot spot is not predicted, the required compensation airflow is calculated using the data processing and analysis module.

[0032] (3) Data processing and analysis module In traditional data centers, the identification of thermal anomaly areas relies on manual inspections or single-point sensors, which suffers from problems such as response lag and inaccurate positioning. Furthermore, the low air density in high-altitude environments leads to unique heat transfer patterns, making it difficult for traditional data analysis to predict temperature changes. The data processing and analysis module is designed to solve these problems. As the core analysis unit of the system, it maintains real-time communication with the thermal imaging monitoring module, continuously receiving infrared thermal imaging data transmitted by the latter, and undertakes the key tasks of thermal anomaly location and temperature trend prediction.

[0033] In terms of identifying thermal anomaly areas, after receiving data, the module analyzes the thermal imaging data using image processing algorithms, comparing the temperature of each area with a preset safe operating temperature threshold for the server, and identifying areas where the temperature exceeds the threshold in real time. To achieve localization, the module has a built-in spatial coordinate mapping model. This model pre-stores the correspondence between the thermal imager's monitoring viewpoint and the physical space of the server room, converting the high-temperature areas identified in the image into three-dimensional physical coordinates within the server room, i.e., "hot spot" coordinates. This clearly identifies the specific location of the hot spots, providing precise location information for subsequent targeted heat dissipation, breaking the limitation of traditional monitoring which can only roughly determine high-temperature areas.

[0034] In terms of temperature trend prediction, the module is fully adapted to the characteristics of the plateau environment and constructs an analytical model that integrates multi-dimensional data. The specific process is as follows: Select historical temperature series of each area of ​​the computer room within a preset historical period. ,in For area code, Number the thermal imaging monitoring points. For a historic moment, Indicates time Lower region Central monitoring point Temperature value; Synchronously select the server load sequence at the corresponding time. ,in For server number, Indicates time Download server The load value is obtained by normalizing real-time monitoring data of CPU utilization, memory usage, and total power consumption, and directly corresponds to the instantaneous heat generation of the server; at the same time, the data center environmental parameter sequence of the same period in history is extracted. Including ambient air density and ambient temperature It is adapted to the characteristics of the plateau environment; temperature sequence Load sequence and environmental parameter sequences Align by timestamp to build a multi-dimensional related dataset A prediction model was built based on time series analysis methods, and a heat transfer time series correction term was introduced to correlate the dataset. The data within a continuous time window of a preset length is used as the model input vector: ,in The time window length is determined based on the thermal inertia characteristics of the high-altitude computer room to ensure the complete lag cycle of heat transfer coverage. Based on the preset prediction duration after this time window Temperature values ​​at monitoring points in various areas As model output vector By analyzing related datasets Through traversal learning, integrating time series autocorrelation analysis and the physical laws of heat transfer, an input vector is established. With output vector The coupling mapping relationship between them can be expressed mathematically as follows: in The time series mapping function is obtained through data-driven training. For the current time zone monitoring points rate of temperature change The air density at standard sea level, Used to correct the effect of low air density at high altitudes on heat transfer efficiency; This coupling mapping relationship is used to correlate the input data in real time. Perform the calculation, where Given the current moment, obtain the future time period. Predicted temperature values ​​at monitoring points in various areas By calculating the deviation between the predicted temperature and the safety threshold ,in The temperature threshold for safe server operation, combined with the rate of temperature change. To predict the timing and intensity of thermal anomaly risks.

[0035] Finally, the module compiles the three-dimensional coordinates of the thermal anomaly area, temperature prediction results, and risk assessment information, and sends them synchronously to the intelligent control module to provide data support for formulating predictive intervention strategies, ensuring that control measures can be deployed and implemented in advance.

[0036] (4) Intelligent control module In the context of traditional data center temperature control relying on fixed parameters and being difficult to adapt to the characteristics of high-altitude environments and dynamic heat load changes, the intelligent control module, as the core decision-making and execution unit of the entire control system, maintains real-time communication with the data processing and analysis module, the sliding fan system, and the computer room cooling system. By executing three coordinated control strategies, it ensures that the server always operates within a safe temperature range.

[0037] During the execution of the basic adaptive cooling strategy, the module first acquires the total power consumption data of the computer room through the intelligent PDU, inputs this data into the feedforward controller, and quickly calculates the basic fan speed command. Simultaneously, the module receives the average temperature data of the computer room transmitted by the data processing and analysis module, compares it with the preset target temperature range, obtains the temperature deviation value, and then inputs this deviation value into the PID controller to calculate the airflow command for precise correction. The module adds the basic fan speed command and the precisely corrected airflow command to obtain the uncompensated basic cooling equipment control command. Considering the problem of reduced cooling efficiency due to low air density in high-altitude areas, the module introduces a high-altitude compensation coefficient, multiplying the basic control command by this coefficient to obtain the final execution command.

[0038] The calculation and calibration of the plateau compensation coefficient requires a combination of physical formula derivation and experimental verification. First, key parameters must be obtained, including the altitude of the data center location. Based on the standard atmospheric model, through altitude The local atmospheric pressure was calculated. and local air density At the same time, the air density under standard sea level conditions was selected as the reference air density. Based on the correlation between air density and heat dissipation efficiency, the theoretical compensation coefficient is calculated by the ratio of the reference air density to the local air density, and its formula is expressed as follows: Its value is greater than 1.

[0039] Using the calculated theoretical compensation coefficient as the initial value, multiple rounds of iterative experimental tests were conducted on the cooling system of a specific data center. By simulating heat dissipation scenarios under different load conditions, the deviation between the actual heat dissipation effect and the target heat dissipation effect was compared, and the coefficient value was gradually adjusted. After multiple experimental calibrations, the optimal compensation coefficient that is suitable for the cooling system of this data center was finally determined. This ensures that the coefficient can effectively compensate for heat loss caused by the low air density at high altitudes.

[0040] During the execution of the local hotspot dynamic elimination strategy, the module takes tiered intervention measures based on the hotspot location information and temperature trend prediction results output by the data processing and analysis module. When the temperature of the identified local hotspot exceeds the first real-time threshold, the module triggers the real-time intervention mode, immediately controlling the sliding fan system to move fan 10 above the hotspot, simultaneously activating the highest intensity directional airflow, and simultaneously activating the fixed heat dissipation equipment near the hotspot to quickly eliminate the existing thermal anomaly. When it is predicted that the temperature of the local hotspot will exceed the second warning threshold in the future, the module triggers the predictive intervention mode, proactively controlling fan 10 in the sliding fan system to move to the predicted hotspot location, providing preventative medium-intensity airflow to avoid the occurrence of thermal anomalies at the source.

[0041] During the execution of the energy-saving operation strategy, the module accesses external meteorological data to monitor changes in ambient temperature in real time. When nighttime is detected and the ambient temperature is below the set threshold, the module automatically triggers the strategy, prioritizing the activation of the fresh air system to utilize natural cooling sources for cooling, while simultaneously reducing the speed of the air conditioning fans to decrease energy consumption for mechanical refrigeration. To further enhance energy efficiency, the module dynamically adjusts the target setpoint for the overall server room temperature and the alarm threshold for local hotspots, maximizing the use of natural cooling sources and reducing the overall energy consumption of the data center's operation and maintenance while ensuring the safe operation of the servers.

[0042] Example 2 Please refer to Figure 9 This embodiment 2 provides a method for controlling the operating status of a data center server based on thermal imaging, applied to any of the data center server operating status control systems based on thermal imaging described in any one of the embodiments, and includes the following steps: S1. The thermal imaging monitoring module continuously collects infrared thermal imaging data of the server and computer room environment to achieve full coverage monitoring of the thermal field of the computer room, and transmits the collected infrared thermal imaging data to the data processing and analysis module in real time. S2. After receiving infrared thermal imaging data, the data processing and analysis module analyzes and processes it, identifies the server's operating status and locates the thermal anomaly area. At the same time, it calls historical thermal imaging data and server load data, predicts the temperature change trend of the thermal anomaly area based on a preset prediction model, and sends the identification results, location information and temperature prediction results to the intelligent control module simultaneously. S3. The intelligent control module receives the output information from the data processing and analysis module, combines it with the real-time operating parameters of the computer room and external environmental data, and generates appropriate control commands. S4. The intelligent control module sends control commands to the sliding fan system and the computer room cooling system respectively, driving the sliding fan system to move to the designated position, adjusting the air supply direction and air supply intensity to achieve directional heat dissipation, and at the same time regulating the operating parameters of the computer room cooling system. Through the coordinated execution of the control strategy by the two, the server is maintained to operate within a safe temperature range. S5. The thermal imaging monitoring module continuously feeds back the thermal field data of the computer room after regulation, the data processing and analysis module tracks the temperature changes in the thermal anomaly area in real time, and the intelligent control module dynamically adjusts the control commands according to the feedback results to form a closed-loop regulation.

[0043] Please refer to Figure 10 Meanwhile, given the lag in response and uneven distribution of hot and cold resources in traditional data center cooling systems, this embodiment 2 also achieves efficient handling of thermal anomalies through a hierarchical response mechanism and collaborative strategy. The system first acquires thermal imaging, environmental sensor, and PDU data through a data acquisition phase. After being analyzed by the data processing and analysis module, it uses thermal imaging data to compare the highest temperature with a real-time critical threshold (set as 85% of the equipment's safe temperature) to identify real-time hotspots. Simultaneously, it uses an LSTM neural network prediction model to input the historical temperature sequence of hotspots and the associated server load sequence into a pre-trained network to obtain future temperature predictions, which are then compared with warning thresholds below the real-time threshold to determine potential hotspots.

[0044] For different types of thermal anomalies, this embodiment 2 adopts a layered response: When a real-time hotspot is identified, the intelligent control module calculates the optimal path within 10 seconds, drives the ceiling-mounted sliding fan system to move above the hotspot, and adjusts the pitch angle of fan 10 to achieve directional airflow. Simultaneously, fixed cooling equipment near the hotspot is activated to form coordinated heat dissipation. For potential hotspots warned by the LSTM model, this embodiment 2 schedules fan 10 to move to the predicted position several minutes in advance, providing preventative intervention with moderate-intensity airflow. In terms of cooling strategy, this embodiment 2 follows the principle of "precision first, global assistance," prioritizing the use of sliding fans for localized cooling. If frequent hotspot occurrences or a continuous increase in overall heat load are detected, the computer room air conditioning system is simultaneously adjusted to increase overall airflow to ensure global heat dissipation.

[0045] This embodiment 2 also incorporates dynamic switching logic for energy-saving operation modes. The control module acquires real-time external ambient temperature and time information. When it is during a preset nighttime period, the external ambient temperature is more than 5°C lower than the computer room return air temperature, and the air humidity meets the standard, the system switches to energy-saving mode, turns on the fresh air system to introduce outdoor cold air, and simultaneously increases the base temperature setpoint and hot spot threshold, and reduces the fan speed to maximize the use of natural cooling sources. If daytime arrives and the external environment no longer meets the natural cooling conditions, or the computer room heat load exceeds the safety threshold, the energy-saving mode is exited and conventional cooling is resumed. The entire process continuously monitors the hot spot temperature through closed-loop control. After it stabilizes and drops below the safety threshold and remains below it for the set duration, the fan 10 is controlled to reduce its speed and return to the standby position, shutting down the associated fixed cooling equipment, thus completing the entire autonomous process from perception and decision-making to execution.

[0046] This embodiment has significant application potential in high-density data center scenarios, especially suitable for the operation and maintenance needs of data centers in complex environments such as high altitudes and diverse climates. Its layered response and precise heat dissipation mechanism can not only solve the problem of difficult handling of local thermal anomalies under traditional heat dissipation modes, but also reduce energy consumption costs through energy-saving operation modes. It is in line with the current development trend of "high efficiency and low carbon" in data centers, and can help large data centers achieve dual optimization of operation and maintenance efficiency and energy consumption control.

[0047] Meanwhile, the system's autonomous processes and adaptability design can also be extended to scenarios such as edge data centers and distributed computing nodes. Its combination of thermal imaging monitoring and LSTM prediction can meet the needs of distributed computing facilities for lightweight and intelligent temperature control, providing reliable thermal management support for the stable operation of computing networks, and has broad application potential in the context of the expanding digital economy.

[0048] At the same time, those skilled in the art will readily understand that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A data center server operation status control system based on thermal imaging, characterized in that, include: It includes a thermal imaging monitoring module for collecting infrared thermal imaging data of servers and computer room environment and realizing full coverage monitoring of computer room thermal field, a sliding rail fan system for directional heat dissipation of designated areas, a data processing and analysis module for identifying server operating status and locating thermal anomaly areas, and an intelligent control module for maintaining the server operating within a safe temperature range. The data processing and analysis module is communicatively connected to the thermal imaging monitoring module, and is used to receive infrared thermal imaging data and predict the temperature change trend of thermal anomaly areas based on historical data. The intelligent control module is communicatively connected to the data processing and analysis module, the sliding fan system, and the computer room cooling system, and is used to generate control commands based on the analysis results of the data processing and analysis module, driving the sliding fan system and the computer room cooling system to execute the appropriate control strategies.

2. The data center server operation status control system based on thermal imaging according to claim 1, characterized in that, The slide rail fan system is deployed above the hot aisle and includes a guide rail, a slider that moves on the guide rail, a servo motor whose top is fixedly connected to the bottom of the slider and is used to adjust the fan's airflow direction and can control at least two degrees of freedom, and a fan installed at the bottom of the servo motor.

3. The data center server operation status control system based on thermal imaging according to claim 2, characterized in that, In the slide rail fan system, all fans are initially positioned at one end of the slide rail; when a hotspot is detected, the system controls the fans to move to the area corresponding to the hotspot, thus enabling the fans to... Axis coordinates With hot topics Axis coordinates To align the fan, the system program determines the fan's rotation direction using the following mathematical logic: when At that time, the decision result is "rotate in the negative direction of the Y-axis"; when At that time, the decision result is "rotate in the positive direction of the Y-axis"; in, The Y-axis coordinate of the hot spot Let Y be the fan's Y-axis coordinate. .

4. The data center server operation status control system based on thermal imaging according to claim 2, characterized in that, The specific steps for adjusting the fan pitch angle via a servo motor in the sliding rail fan system are as follows: Get the fan's installation height 3D coordinates of local hotspots and the fixed horizontal distance between the fan and the hotspot The target pitch angle is calculated using the following mathematical formula. Based on the calculation result, the actuator is driven to complete the angle adjustment: in, This is the arctangent function, used to calculate the vertical height difference between the fan and the hotspot. and fixed horizontal distance Find the pitch angle required for the fan to deliver air in a directional manner.

5. The data center server operation status control system based on thermal imaging according to claim 1, characterized in that, The temperature trend prediction function of the data processing and analysis module is based on time series analysis methods. It models historical thermal imaging data and server load data to predict temperature changes in various regions within a specific future time period. The specific process is as follows: Select historical temperature series of each area of ​​the computer room within a preset historical period. ,in For area code, Number the thermal imaging monitoring points. For a historic moment, Indicates time Lower region Central monitoring point Temperature value; Synchronously select the server load sequence at the corresponding time. ,in For server number, Indicates time Download server The load value is obtained by normalizing real-time monitoring data of CPU utilization, memory usage, and total power consumption, and directly corresponds to the instantaneous heat generation of the server; at the same time, the historical data center environmental parameter sequence is extracted. Including ambient air density and ambient temperature It is adapted to the characteristics of the plateau environment; temperature sequence Load sequence and environmental parameter sequences Align by timestamp to build a multi-dimensional related dataset A prediction model was built based on time series analysis methods, and a heat transfer time series correction term was introduced to correlate the dataset. The data within a continuous time window of a preset length is used as the model input vector: ,in The time window length is determined based on the thermal inertia characteristics of the high-altitude computer room to ensure the complete lag cycle of heat transfer coverage. Based on the preset prediction duration after this time window Temperature values ​​at monitoring points in various areas As model output vector By analyzing related datasets Through traversal learning, integrating time series autocorrelation analysis and the physical laws of heat transfer, an input vector is established. With output vector The coupling mapping relationship between them can be expressed mathematically as follows: in The time series mapping function is obtained through data-driven training. For the current time zone monitoring points rate of temperature change The air density at standard sea level, Used to correct the effect of low air density at high altitudes on heat transfer efficiency; This coupling mapping relationship is used to correlate the input data in real time. Perform the calculation, where Given the current moment, obtain the future time period. Predicted temperature values ​​at monitoring points in various areas By calculating the deviation between the predicted temperature and the safety threshold ,in The temperature threshold for safe server operation, combined with the rate of temperature change. This allows for the prediction of the timing and intensity of thermal anomaly risks, providing core data support for the intelligent control module to formulate predictive intervention strategies.

6. The data center server operation status control system based on thermal imaging according to claim 1, characterized in that, The control logic executed by the intelligent control module includes: When the temperature of the identified local hot spot exceeds the first real-time threshold, the real-time intervention mode is triggered, controlling the sliding fan system to move above the hot spot and deliver the highest intensity directional airflow, while simultaneously turning on nearby devices for heat dissipation. When the predicted local hotspot temperature will exceed the second warning threshold in the future, the predictive intervention mode is triggered, and the sliding fan system is controlled to move to the predicted position in advance and provide preventative medium-intensity airflow.

7. The data center server operation status control system based on thermal imaging according to claim 1, characterized in that, The regulation strategy of the intelligent control module includes: Basic adaptive cooling strategy: Dynamically adjust the operating parameters of basic cooling equipment based on the overall heat load and average temperature of the computer room, and compensate for the low air density at high altitudes. Local hotspot dynamic elimination strategy: When a local hotspot is identified or a hotspot is predicted to be generated, the sliding rail fan system is controlled to move directly above the hotspot area to blow air in a directional manner, and the fixed cooling equipment near the hotspot is turned on at the same time. Energy-saving operation strategy: When nighttime is detected and the external environment is below the set temperature threshold, the power and airflow of the basic heat dissipation equipment are reduced, and natural cooling sources are used preferentially.

8. The data center server operation status control system based on thermal imaging according to claim 7, characterized in that, In the energy-saving operation strategy, while reducing the power and air volume of basic heat dissipation equipment, the target set value of the overall temperature of the computer room and the alarm threshold of local hot spots are dynamically increased.

9. A data center server operation status control system based on thermal imaging according to claim 8, characterized in that, The basic adaptive cooling strategy compensates for the low air density at high altitudes by multiplying the basic airflow command for computation by a high altitude compensation coefficient greater than 1.

10. A method for controlling the operating status of a data center server based on thermal imaging, applied to a data center server operating status control system based on thermal imaging as described in any one of claims 1-9, characterized in that, Includes the following steps: S1. The thermal imaging monitoring module continuously collects infrared thermal imaging data of the server and computer room environment to achieve full coverage monitoring of the thermal field of the computer room, and transmits the collected infrared thermal imaging data to the data processing and analysis module in real time. S2. After receiving infrared thermal imaging data, the data processing and analysis module analyzes and processes it, identifies the server's operating status and locates the thermal anomaly area. At the same time, it calls historical thermal imaging data and server load data, predicts the temperature change trend of the thermal anomaly area based on a preset prediction model, and sends the identification results, location information and temperature prediction results to the intelligent control module simultaneously. S3. The intelligent control module receives the output information from the data processing and analysis module, combines it with the real-time operating parameters of the computer room and external environmental data, and generates appropriate control commands. S4. The intelligent control module sends control commands to the sliding fan system and the computer room cooling system respectively, driving the sliding fan system to move to the designated position, adjusting the air supply direction and air supply intensity to achieve directional heat dissipation, and at the same time regulating the operating parameters of the computer room cooling system. Through the coordinated execution of the control strategy by the two, the server is maintained to operate within a safe temperature range. S5. The thermal imaging monitoring module continuously feeds back the thermal field data of the computer room after regulation, the data processing and analysis module tracks the temperature changes in the thermal anomaly area in real time, and the intelligent control module dynamically adjusts the control commands according to the feedback results to form a closed-loop regulation.

Images