Server cluster temperature automatic control management system based on Internet of Things

Abstract

The invention discloses a server cluster temperature automatic control management system based on the Internet of Things, and relates to the technical field of Internet of Things and data center management. Comprising an environmental data acquisition and preprocessing module, a feature extraction and fluctuation analysis module, an intelligent evaluation and mode recognition module and a tolerance adjustment and temperature control response optimization module, the obtained original environment data information is preprocessed, and a data set is constructed; according to the method, environment data are collected in real time, temperature and humidity fluctuation is intelligently evaluated, the tolerance and response time of the temperature control system are automatically adjusted, and it is ensured that response is made only when the environment is abnormal. The accuracy and response efficiency of the temperature control system are improved, unnecessary energy consumption and equipment burden are reduced, excessive operation of cooling equipment is avoided, and the stability of a server cluster is improved.

Description

Technical Field

[0001] This invention relates to the fields of Internet of Things (IoT) and data center management technology, specifically to an IoT-based automatic temperature control and management system for server clusters. Background Technology

[0002] An IoT-based server cluster temperature automatic control and management system utilizes IoT technology to monitor the internal temperature of the server cluster in real time through sensors and smart devices. It then uses automated control methods to adjust the server's cooling system or ambient temperature, thereby maintaining the server cluster within its optimal operating temperature range. Its purpose is to improve server operating efficiency, reduce hardware failures or performance degradation caused by overheating, lower energy consumption, extend server lifespan, and ensure the stability and reliability of the data center. Intelligent temperature control management enables dynamic temperature adjustment, maximizing energy savings and cost control while avoiding risks associated with human error in equipment management.

[0003] Temperature and humidity loggers are used in the automatic temperature control and management of IoT-based server clusters. A temperature and humidity logger is a device used to record ambient temperature and humidity in real time. It is typically equipped with sensors to monitor changes in the surrounding environment's temperature and humidity, storing the data or transmitting it wirelessly to a central system for real-time monitoring and analysis. Its function is to provide accurate temperature and humidity data, helping administrators understand the environmental conditions of the server room where the server cluster is located. In server cluster management, temperature and humidity loggers can detect abnormal fluctuations in ambient temperature, triggering early warning mechanisms in a timely manner to ensure that servers operate in an optimal environment, preventing hardware failures or performance degradation caused by excessively high temperatures or improper humidity. Furthermore, it can provide long-term environmental change trends for data analysis, thereby optimizing temperature control strategies and improving energy efficiency.

[0004] Existing technologies have the following shortcomings: In practical applications, temperature and humidity loggers may overreact to short-term fluctuations in the environment, especially when affected by instantaneous airflow changes or rapid changes in the external environment, such as the opening and closing of server room doors, personnel movement, and the start and stop of air conditioning systems. These instantaneous fluctuations are usually only localized changes in temperature and humidity and do not represent a true problem in the overall environment. However, due to the high sensitivity of temperature and humidity loggers, they may misinterpret these short-term changes as system malfunctions or environmental anomalies, thereby generating unnecessary alarm signals or data fluctuations. This overreaction phenomenon can cause temperature and humidity loggers to frequently trigger alarms and may mislead the temperature control system to make unnecessary adjustments. As a result, the system may initiate excessive cooling or heating processes without intervention, increasing energy consumption and causing cooling equipment to operate excessively, further exacerbating energy waste. In addition, continuous excessive cooling or heating may also overload the cooling equipment, increasing the risk of equipment failure and thus affecting the normal operation of the server cluster and system stability.

[0005] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0006] The purpose of this invention is to provide an IoT-based automatic temperature control and management system for server clusters. By collecting environmental data in real time and intelligently assessing temperature and humidity fluctuations, this method automatically adjusts the tolerance and response time of the temperature control system, ensuring that it only reacts to abnormal environmental conditions. This improves the accuracy and response efficiency of the temperature control system, reduces unnecessary energy consumption and equipment burden, avoids excessive operation of cooling equipment, enhances the stability and reliability of the server cluster, and reduces long-term operating costs, thereby solving the problems mentioned in the background art.

[0007] To achieve the above objectives, the present invention provides the following technical solution: an IoT-based server cluster temperature automatic control and management system, comprising an environmental data acquisition and preprocessing module, a feature extraction and fluctuation analysis module, an intelligent evaluation and pattern recognition module, and a tolerance adjustment and temperature control response optimization module. The environmental data acquisition and preprocessing module collects environmental data from the server cluster's computer room in real time using a temperature and humidity recorder, preprocesses the acquired raw environmental data, and constructs a dataset. The feature extraction and fluctuation analysis module uses feature engineering technology to extract key features from the dataset that reflect short-term fluctuations in the temperature and humidity of the computer room, and performs comprehensive analysis on the extracted features to quantify the magnitude of temperature and humidity fluctuations. The intelligent assessment and pattern recognition module takes the analyzed key indicators as feature vectors and inputs them into a pre-trained machine learning model. The machine learning model then intelligently assesses the temperature and humidity fluctuations in the computer room and identifies the temperature and humidity fluctuation patterns. The tolerance adjustment and temperature control response optimization module automatically increases the tolerance boundary based on the assessment results when the temperature and humidity of the server cluster's data center are fluctuating in the short term. This increases the tolerance for small fluctuations and adaptively extends the temperature control response time to avoid overreacting to short-term fluctuations.

[0008] Preferably, the specific steps for collecting environmental data information of the server cluster's computer room in real time using a temperature and humidity recorder are as follows: First, install temperature and humidity loggers in key locations in the computer room; Then, configure the recorder to communicate with the data acquisition system to ensure that the device can transmit the collected temperature and humidity data to the central monitoring system in real time. Next, set the recorder's sampling frequency and data transmission interval; The temperature and humidity recorder starts working automatically and continuously collects data, which is then transmitted to the data storage system in real time. Finally, the collected data will enter the preprocessing stage to prepare for subsequent analysis.

[0009] Preferably, key features reflecting short-term fluctuations in the temperature and humidity of the computer room are extracted from the dataset using feature engineering techniques. The extracted features include the sensitivity of airflow to changes in the temperature and humidity of the computer room and the coordinated changes in temperature and humidity. The sensitivity of airflow to changes in the temperature and humidity of the computer room and the coordinated changes in temperature and humidity are comprehensively analyzed under the detection window to generate reference values ​​for airflow sensitivity and reference values ​​for coordinated changes in temperature and humidity. The magnitude of temperature and humidity fluctuations is quantified by the reference values ​​for airflow sensitivity and coordinated changes in temperature and humidity.

[0010] Preferably, the specific steps for analyzing the sensitivity of airflow to changes in temperature and humidity in the computer room and generating reference values ​​for airflow sensitivity under the detection window are as follows: Record the air velocity and temperature / humidity data at each moment, and calculate the instantaneous impact of changes in airflow on temperature and humidity. The calculation expression is as follows: ; In the formula, R S It is the airflow response intensity, A t A is the air velocity at time t. t-1 The air velocity at time t-1 is the air velocity at the previous time, T. t and H t T represents the temperature and humidity values ​​at time t, respectively.t-1 and H t-1 These represent the temperature and humidity values ​​at time t-1, i.e., the temperature and humidity values ​​at the previous time. After obtaining the airflow response intensity at each moment, the airflow sensitivity reference value is calculated using the following expression: ; In the formula, AFS is the airflow sensitivity reference value, and N is the total number of detection window times. It is the airflow response intensity at time t.

[0011] Preferably, the specific steps for analyzing the coordinated changes in temperature and humidity within the detection window to generate reference values ​​for the coordinated changes in temperature and humidity are as follows: First, the temperature and humidity data are normalized to eliminate the influence of differences in units. The normalization formula for the temperature and humidity data at each moment is as follows: ; In the formula, T norm These are normalized temperature data, where T is the original temperature data, min(T) is the minimum temperature value among all historical temperature records, and max(T) is the maximum temperature value among all historical temperature records. H norm Here, H is the normalized humidity data, min(H) is the minimum value among all humidity records in the historical data, and max(H) is the maximum value among all humidity records in the historical data. The difference between normalized temperature and humidity is calculated, and the difference rate is used to reflect the synchronous change of the two. The expression for the difference rate is as follows; ; In the formula, D represents the degree of difference between temperature and humidity; By combining the difference rates at multiple time points, the coordinated changes in temperature and humidity are quantified. The reference value for the coordinated changes in temperature and humidity is calculated by weighted integral of the difference rates at multiple time points, as shown in the following expression: ; In the formula, D i w represents the degree of difference between temperature and humidity at the i-th time point. i is the weight of the i-th time point, e is the natural base, M is the total number of time points in the detection window, and THC is the reference value for the coordinated change of temperature and humidity.

[0012] Preferably, the analyzed airflow sensitivity reference value and temperature and humidity co-change reference value are used as feature vectors and input into a pre-trained machine learning model. The machine learning model generates a temperature and humidity fluctuation risk coefficient, and the temperature and humidity fluctuation of the computer room is intelligently evaluated and temperature and humidity fluctuation patterns are identified through the temperature and humidity fluctuation risk coefficient.

[0013] Preferably, a pre-trained machine learning model is used to intelligently assess temperature and humidity fluctuations in the computer room. The temperature and humidity fluctuation risk coefficient generated when identifying the temperature and humidity fluctuation pattern is compared with a pre-set reference threshold for temperature and humidity fluctuation risk coefficient to determine the temperature and humidity fluctuation pattern. The determination steps are as follows: If the temperature and humidity fluctuation risk coefficient is greater than the preset reference threshold for temperature and humidity fluctuation risk coefficient, the current temperature and humidity fluctuation in the computer room will be judged as a short-term fluctuation; if the temperature and humidity fluctuation risk coefficient is less than or equal to the preset reference threshold for temperature and humidity fluctuation risk coefficient, the temperature and humidity in the computer room will be judged as not fluctuating.

[0014] Preferably, when the evaluation results show that the temperature and humidity in the server cluster's data center are experiencing short-term fluctuations, the specific steps for automatically increasing the tolerance boundary based on the evaluation results and adaptively extending the temperature control response time are as follows: The temperature and humidity fluctuation risk coefficient (THFR) is compared with a pre-set reference threshold to determine the nature of the temperature and humidity fluctuations. When the assessment results show that the temperature and humidity in the server cluster's data center are experiencing short-term fluctuations, the tolerance boundary is calculated and increased to increase tolerance for short-term, small fluctuations and avoid overreacting to short-term fluctuations. The calculation expression is as follows: ; In the formula, THFR ref ΔT is the reference threshold for the risk factor of temperature and humidity fluctuations, α is the adjustment coefficient used to control the sensitivity of the tolerance boundary, and ΔT is the temperature and humidity fluctuation risk factor. threshold This is the tolerance boundary adjustment value; Assuming short-term temperature and humidity fluctuations are confirmed, the response time of the temperature control system is adaptively extended to prevent it from reacting too quickly to small fluctuations and causing frequent temperature adjustments. The calculation expression is as follows: ; In the formula, T response It is the extended temperature control response time, and β is the reference coefficient for the extension of response time; The temperature control system will take into account the risk factor (THFR) of temperature and humidity fluctuations and the extended temperature control response time (T). response The actual temperature control adjustment range is determined by the following calculation expression: ; In the formula, ΔT controlγ is the temperature adjustment value of the temperature control system, and γ is the temperature control sensitivity coefficient.

[0015] The technical effects and advantages provided by the present invention in the above technical solution are as follows: This invention ensures that the system only reacts when genuine environmental anomalies occur by real-time acquisition of environmental data, intelligent assessment of temperature and humidity fluctuation patterns, and automatic adjustment of the temperature control system's tolerance boundaries and response time. This significantly improves the accuracy and response efficiency of the temperature control system, while also substantially reducing unnecessary energy consumption and equipment burden, minimizing excessive operation of cooling equipment, and enhancing the stability, reliability, and long-term economic efficiency of the server cluster. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0017] Figure 1 This is a schematic diagram of a module of an IoT-based server cluster temperature automatic control and management system according to the present invention. Detailed Implementation

[0018] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided so that the description of this disclosure will be more complete and fully convey the concept of the exemplary embodiments to those skilled in the art.

[0019] This invention provides, for example Figure 1 The illustrated IoT-based server cluster temperature automatic control and management system includes an environmental data acquisition and preprocessing module, a feature extraction and fluctuation analysis module, an intelligent evaluation and pattern recognition module, and a tolerance adjustment and temperature control response optimization module. The environmental data acquisition and preprocessing module collects environmental data from the server cluster's computer room in real time using a temperature and humidity recorder, preprocesses the acquired raw environmental data, and constructs a dataset. The specific steps for collecting real-time environmental data from the server cluster's data center using a temperature and humidity recorder are as follows: First, install temperature and humidity data loggers in key locations within the computer room, such as near air conditioners, vents, and equipment racks, to ensure comprehensive monitoring of ambient temperature and humidity changes. Next, configure the loggers to communicate with the data acquisition system, ensuring real-time transmission of collected temperature and humidity data to the central monitoring system. Then, set the loggers' sampling frequency and data transmission interval, typically set to a frequency of a few minutes based on actual needs, to ensure data timeliness. The temperature and humidity data loggers then begin automatic operation and continuous data collection, transmitting the data to the data storage system in real-time. During data acquisition, the equipment should have a fault-tolerant mechanism, such as automatically switching to a backup sensor in case of sensor failure, ensuring data integrity and reliability. Finally, the collected data will enter a preprocessing stage to prepare for subsequent analysis.

[0020] Environmental data includes, but is not limited to, information on temperature, humidity, airflow, and wind speed. This data provides a crucial foundation for subsequent analysis and decision-making. Real-time data acquisition helps the system monitor the current environment of the computer room, promptly detecting and responding to any sudden changes in temperature and humidity. Timely and accurate environmental data acquisition is a prerequisite for the smooth operation of the intelligent temperature control system. Through the efficient integration of sensors and data acquisition modules, the system can quickly respond to environmental changes, ensuring that the system will not experience untimely or erroneous temperature control responses due to data delays.

[0021] The goal of preprocessing is to remove noise, outliers, and missing values ​​from data, thereby improving data reliability. This typically includes techniques such as data smoothing, denoising, missing data imputation, standardization, and normalization. By removing irrelevant or outlier data, it ensures high-quality data input to analytical and machine learning models, avoiding misleading analytical results. For example, short-term environmental noise (such as door opening and closing, people walking) or momentary sensor malfunctions may produce erroneous data; preprocessing can effectively eliminate these effects, ensuring that the system bases subsequent processing on accurate and valid data.

[0022] After preprocessing, the data needs to be organized into an organized dataset. This step transforms the raw environmental data into a structured dataset containing temperature, humidity, and related indicators at various time points. This dataset includes not only real-time temperature and humidity data but also historical data, data trends, and rates of change. The establishment of this dataset provides fundamental data support for subsequent feature extraction, machine learning modeling, and system decision-making. By establishing appropriate time windows, aggregation functions, and data structures, the dataset can provide the system with a multi-dimensional environmental perspective.

[0023] The feature extraction and fluctuation analysis module uses feature engineering technology to extract key features from the dataset that reflect short-term fluctuations in the temperature and humidity of the computer room, and performs comprehensive analysis on the extracted features to quantify the magnitude of temperature and humidity fluctuations. Feature engineering techniques are used to extract key features reflecting short-term fluctuations in the temperature and humidity of the computer room from the dataset. The extracted features include the sensitivity of airflow to changes in the temperature and humidity of the computer room and the coordinated changes in temperature and humidity. The sensitivity of airflow to changes in the temperature and humidity of the computer room and the coordinated changes in temperature and humidity are comprehensively analyzed under the detection window to generate reference values ​​for airflow sensitivity and temperature and humidity coordinated changes. The magnitude of temperature and humidity fluctuations is quantified by the reference values ​​for airflow sensitivity and temperature and humidity coordinated changes.

[0024] Sudden changes in airflow sensitivity to temperature and humidity variations in a computer room can indeed indicate short-term fluctuations in these conditions. When airflow changes abruptly, the temperature and humidity within the computer room typically fluctuate rapidly and locally. For example, changes in airflow velocity when the air conditioning system starts, vents suddenly open, or a computer room door opens can trigger rapid fluctuations in localized temperature and humidity. Increased airflow leads to stronger heat dissipation or humidification, resulting in a rapid drop in temperature or an increase in humidity; conversely, decreased airflow may lead to a rise in temperature or a decrease in humidity. These sudden changes in sensitivity are usually short-term phenomena, manifesting as instantaneous changes in temperature and humidity with large amplitude but short duration. If the temperature and humidity fluctuations are caused by rapid changes in airflow, and the temperature and humidity return to their original levels afterward, it indicates that these fluctuations are merely short-term phenomena. By monitoring the immediate response of airflow to temperature and humidity, the system can identify these short-term fluctuations and avoid over-adjusting the temperature control system. Temperature and humidity fluctuations caused by sudden changes in airflow usually do not indicate a system malfunction but rather a short-term adaptive fluctuation in the computer room environment.

[0025] The specific steps for analyzing the sensitivity of airflow to changes in temperature and humidity in the computer room and generating reference values ​​for airflow sensitivity under the detection window are as follows: Record the air velocity and temperature / humidity data at each moment, and calculate the instantaneous impact of changes in airflow on temperature and humidity. The calculation expression is as follows: ; In the formula, R S It is the airflow response intensity, A t A is the air velocity at time t. t-1 The air velocity at time t-1 is the air velocity at the previous time, T. t and H t T represents the temperature and humidity values ​​at time t, respectively. t-1 and H t-1 These represent the temperature and humidity values ​​at time t-1, i.e., the temperature and humidity values ​​at the previous time. This step quantifies the impact of changes in airflow on the temperature and humidity of the computer room by calculating the product of changes in airflow velocity and changes in temperature and humidity. If there is a significant change in airflow velocity, and correspondingly, significant fluctuations in temperature and humidity, it indicates that the airflow has a strong, immediate impact on the temperature and humidity of the computer room, with a relatively high response intensity.

[0026] After obtaining the airflow response intensity at each moment, the airflow sensitivity reference value is calculated. The airflow sensitivity reference value is used to describe the sensitivity of airflow changes to changes in temperature and humidity in the computer room. The calculation expression is as follows: ; In the formula, AFS is the airflow sensitivity reference value, and N is the total number of detection window times. It is the airflow response intensity at time t.

[0027] When the airflow sensitivity reference value is close to 1, it indicates that the temperature and humidity in the data center are highly sensitive to changes in airflow, meaning that the temperature and humidity fluctuate significantly and these changes occur dramatically within a short period of time, usually indicating that the data center is in a state of short-term fluctuation. Conversely, when the airflow sensitivity reference value is close to 0, it indicates that changes in airflow have little impact on temperature and humidity, and the temperature and humidity in the data center are relatively stable. This index can help identify whether the data center environment is affected by airflow, and thus reveal whether there are short-term fluctuations.

[0028] The higher the airflow sensitivity reference value generated after analyzing the sensitivity of airflow to changes in temperature and humidity in a computer room within a detection window, the more likely it is that the room's temperature and humidity are experiencing short-term fluctuations. When airflow changes abruptly, its impact on temperature and humidity changes is captured by this reference value. If changes in airflow cause drastic fluctuations in temperature and humidity within a short period, the airflow sensitivity reference value will increase significantly, indicating that the temperature and humidity are experiencing short-term, localized fluctuations. These fluctuations may be caused by the start / stop of the air conditioning system, adjustments to the air vent positions, or changes in the external environment (such as opening and closing doors and windows). In this case, the abrupt change in airflow directly affects the temperature and humidity changes, hence the higher sensitivity reference value. Conversely, a lower airflow sensitivity reference value indicates that airflow has a smaller impact on temperature and humidity, and the temperature and humidity changes are not significant, suggesting that the computer room's temperature and humidity are stable and have not experienced short-term fluctuations.

[0029] The very similar rate and magnitude of temperature and humidity changes usually indicate short-term fluctuations in the computer room's temperature and humidity. This is because changes in temperature and humidity are often triggered by the same or similar factors, especially under the operation of environmental equipment. When equipment such as air conditioners, dehumidifiers, or humidifiers starts or adjusts, their effects on temperature and humidity are usually synchronous because these devices control the environment by adjusting the heat and moisture content in the air. In this case, temperature and humidity changes often manifest as synchronous short-term fluctuations. For example, when an air conditioning system suddenly starts, the temperature drops rapidly, and the humidity also decreases as the moisture in the air decreases. If the magnitude and rate of change of temperature and humidity are similar, it indicates that they are both affected by the same cause (such as equipment start-up or shutdown or changes in the external environment) within a short period of time, resulting in relatively consistent fluctuations. By monitoring such fluctuations, it can be determined that the computer room is in a short-term fluctuation state, rather than a long-term abnormal change. This pattern of synchronous change usually does not require an immediate overreaction because it is caused by the routine adjustment of environmental equipment and is a controllable, short-term fluctuation.

[0030] The specific steps for analyzing the coordinated changes in temperature and humidity within the detection window to generate reference values ​​for these coordinated changes are as follows: First, the temperature and humidity data are normalized to eliminate the influence of dimensional differences. This step aims to ensure that temperature and humidity are analyzed on the same scale, thereby more accurately capturing the relationship between them. The normalization formula for the temperature and humidity data at each moment is as follows: ; In the formula, T norm This is normalized temperature data, with values ​​limited to between 0 and 1. T is the original temperature data, min(T) is the minimum value among all temperature records in the historical data, and max(T) is the maximum value among all temperature records in the historical data. H norm This is the normalized humidity data, with values ​​limited to between 0 and 1. H is the original humidity data, min(H) is the minimum value among all humidity records in the historical data, and max(H) is the maximum value among all humidity records in the historical data. These normalization formulas compress the numerical range of temperature and humidity to between [0,1], thus avoiding the influence of dimensional differences in the original data on the calculation of co-variance. T and H are the original temperature and humidity data, respectively; min(T) and min(H) are the minimum values ​​in the historical data; and max(T) and max(H) are the maximum values ​​in the historical data. Through this normalization, the scale of the data is standardized, facilitating subsequent co-variance analysis.

[0031] The difference between normalized temperature and humidity is calculated, and the difference rate is used to reflect the synchronous change of the two. The expression for the difference rate is as follows; ; In the formula, D represents the degree of difference between temperature and humidity; If the rates of change of the two are nearly synchronized, the difference will be small. The numerator in the formula is their absolute difference, and the denominator is their normalized sum, ensuring that the difference rate is a standardized ratio. The smaller the ratio, the more synchronized the two are; the larger the ratio, the more asynchronous their changes are.

[0032] By combining the difference rates at multiple time points, the coordinated changes in temperature and humidity are quantified. The reference value for the coordinated changes in temperature and humidity is calculated by weighted integral of the difference rates at multiple time points, as shown in the following expression: ; In the formula, D i w represents the degree of difference between temperature and humidity at the i-th time point. i is the weight of the i-th time point, e is the natural base, M is the total number of time points in the detection window, and THC is the reference value for the coordinated change of temperature and humidity.

[0033] By combining the temperature and humidity difference rates at multiple time points, the degree of coordinated change between temperature and humidity is quantified. This step highlights the impact of short-term synchronous changes, ensuring that short-term temperature and humidity fluctuations contribute more to the final reference value, thus accurately reflecting the short-term fluctuation characteristics of the computer room's temperature and humidity.

[0034] A higher reference value for the coordinated change of temperature and humidity, generated after analyzing the coordinated changes within a detection window, generally indicates that the temperature and humidity in the computer room are in a short-term fluctuating state. This reference value reflects the degree of coordinated change between temperature and humidity. Specifically, when the temperature and humidity in the computer room change at similar rates and amplitudes within the same time frame, it means that they are affected by the same or similar factors, such as the start-up of the air conditioning system, or the adjustment of humidifiers or dehumidifiers. In this case, the reference value for the coordinated change of temperature and humidity will be higher, indicating that the temperature and humidity fluctuations are synchronous and short-term, caused by routine environmental control activities. Conversely, when the amplitude and rate of change of temperature and humidity are relatively independent, and the coordinated change between the two is small, the reference value for the coordinated change will be lower, usually indicating that the temperature and humidity in the computer room have not fluctuated significantly and may be in a stable state.

[0035] The intelligent assessment and pattern recognition module takes the analyzed key indicators as feature vectors and inputs them into a pre-trained machine learning model. The machine learning model then intelligently assesses the temperature and humidity fluctuations in the computer room and identifies the temperature and humidity fluctuation patterns. The analyzed airflow sensitivity reference value and temperature and humidity co-change reference value are used as feature vectors and input into a pre-trained machine learning model. The machine learning model generates a temperature and humidity fluctuation risk coefficient, which is used to intelligently assess the temperature and humidity fluctuations in the computer room and identify temperature and humidity fluctuation patterns.

[0036] A pre-trained machine learning model refers to a model that has been trained and learned based on a large amount of historical data and known feature information before its application. This enables the model to make intelligent predictions and decisions when faced with new and unknown data. In this case, the training process of the machine learning model typically involves extracting relevant features of temperature and humidity changes from historical environmental data, such as the amplitude and period of temperature and humidity fluctuations, air flow sensitivity, and coordinated changes in temperature and humidity. These features represent the dynamic behavior of the data center environment. During the machine learning training process, the system compares these input data with corresponding labels (i.e., known temperature and humidity fluctuation patterns and risk coefficients), allowing the model to automatically learn the inherent patterns in the data. Through training, the model can establish the relationship between input features and temperature and humidity fluctuations, and can effectively predict new data, determining whether the data center is in a normal state or whether there is a risk of short-term temperature and humidity fluctuations.

[0037] The training process typically utilizes a large amount of historical data, including temperature and humidity records, airflow variations, and equipment start-up and shutdown status over different time periods. Labels are usually based on fluctuation patterns identified through expert knowledge, real-world environmental feedback, or equipment malfunction records. Commonly used training algorithms include supervised learning classification models (such as support vector machines and decision trees) or regression models (such as random forests and neural networks). During training, the model adjusts its internal parameters to minimize the difference between predicted and actual results. Once training is complete, the machine learning model can quickly generate temperature and humidity fluctuation risk coefficients from new, real-time collected feature vectors and intelligently assess temperature and humidity fluctuations in the data center. The advantage of this method is its ability to adapt to changes, progressively optimize prediction accuracy, and provide dynamic and accurate assessments under different environmental conditions.

[0038] The machine learning model is not limited here. Any machine learning model that can comprehensively analyze the air flow sensitivity reference value (AFS) and the temperature and humidity co-change reference value (THC) to generate the temperature and humidity fluctuation risk coefficient (THFR) is acceptable. To realize the technical solution of this invention, this invention provides a specific implementation method. The formula for generating the Temperature and Humidity Fluctuation Risk Factor (THFR) is as follows: ; In the formula, π a and π bThese are the preset scaling factors for the Air Flow Sensitivity Reference Value (AFS) and the Temperature and Humidity Co-variance Reference Value (THC), respectively, and π a and π b All are greater than 0.

[0039] Preset proportionality coefficient (π) a and π b In machine learning models, π refers to the weighting parameters used to weight various influencing factors (such as airflow sensitivity and temperature-humidity co-variance) in assessing the risk of temperature and humidity fluctuations. Specifically, these weighting coefficients determine the relative importance of the airflow sensitivity reference value (AFS) and the temperature-humidity co-variance reference value (THC) in the calculation of the temperature and humidity fluctuation risk coefficient (THFR). In this formula, π... a and π b These are fixed values, set based on historical data or expert experience, with the aim of enabling the model to reasonably allocate the impact of each factor on the final evaluation result according to the actual situation.

[0040] These preset scaling factors ensure that the model accurately reflects the impact of different factors on temperature and humidity fluctuations in the data center when calculating the risk of temperature and humidity fluctuations. For example, changes in airflow may have a more direct impact on temperature and humidity fluctuations, while coordinated changes in temperature and humidity may reflect the overall trend of changes in equipment or the environment. By adjusting these preset scaling factors, the flexibility and predictive accuracy of the model can be optimized, making the temperature and humidity risk assessment more consistent with the actual operating conditions of the data center.

[0041] As can be seen from the temperature and humidity fluctuation risk coefficient, the larger the air flow sensitivity reference value generated after analyzing the sensitivity of air flow to changes in temperature and humidity in the computer room under the detection window, and the larger the temperature and humidity co-change reference value generated after analyzing the co-change of temperature and humidity under the detection window, the larger the temperature and humidity fluctuation risk coefficient generated when the pre-trained machine learning model intelligently evaluates and identifies temperature and humidity fluctuation patterns in the computer room. This indicates that the probability of the temperature and humidity fluctuations in the computer room being short-term fluctuations is greater, and vice versa.

[0042] A pre-trained machine learning model will be used to intelligently assess temperature and humidity fluctuations in the data center. The temperature and humidity fluctuation risk coefficient generated when identifying the fluctuation pattern will be compared with a pre-set reference threshold to determine the temperature and humidity fluctuation pattern. The determination steps are as follows: If the temperature and humidity fluctuation risk coefficient is greater than the preset reference threshold for temperature and humidity fluctuation risk coefficient, the current temperature and humidity fluctuation in the computer room will be judged as a short-term fluctuation; if the temperature and humidity fluctuation risk coefficient is less than or equal to the preset reference threshold for temperature and humidity fluctuation risk coefficient, the temperature and humidity in the computer room will be judged as not fluctuating.

[0043] The tolerance adjustment and temperature control response optimization module automatically increases the tolerance boundary based on the assessment results when the temperature and humidity of the server cluster's data center are fluctuating in the short term. This increases the tolerance for small fluctuations and adaptively extends the temperature control response time to avoid overreacting to short-term fluctuations. When assessments indicate short-term fluctuations in temperature and humidity within the server cluster's data center, automatically increasing the tolerance threshold and extending the temperature control response time primarily optimizes the stability and energy efficiency of the temperature and humidity management system. These short-term fluctuations are typically caused by instantaneous changes in temperature and humidity due to environmental factors (such as air conditioning start-up / shutdown, door opening, and changes in equipment operating status), and these fluctuations do not pose a real threat to the server cluster's operation. By automatically increasing the tolerance threshold, the system can tolerate these minor temperature and humidity fluctuations without immediately adjusting the temperature control. This avoids unnecessary energy consumption and equipment load caused by frequent adjustments to the temperature control system due to short-term fluctuations.

[0044] Furthermore, the adaptive extension of the temperature control response time ensures that the temperature control system does not react too quickly to every short-term fluctuation, especially when the fluctuation amplitude is small. This strategy not only effectively prevents the temperature control system from over-adjusting but also reduces wear and tear and energy waste caused by frequent start-stop operations. By dynamically adjusting the response time, the system can maintain the current environmental settings under conditions of small short-term fluctuations, avoiding over-cooling or over-heating, thereby maintaining a stable data center environment while saving energy. This intelligent response mechanism helps improve the overall system reliability, reduce unnecessary operations and maintenance, extend equipment lifespan, and maximize energy management efficiency.

[0045] When the assessment results show that the temperature and humidity in the data center where the server cluster is located are experiencing short-term fluctuations, the specific steps for automatically increasing the tolerance boundary and adaptively extending the temperature control response time based on the assessment results are as follows: The temperature and humidity fluctuation risk coefficient (THFR) is compared with a pre-set reference threshold to determine the nature of the temperature and humidity fluctuations. When the assessment results show that the temperature and humidity in the server cluster's data center are experiencing short-term fluctuations, the tolerance boundary is calculated and increased to increase tolerance for short-term, small fluctuations and avoid overreacting to short-term fluctuations. The calculation expression is as follows: ; In the formula, THFR ref ΔT is the reference threshold for the risk factor of temperature and humidity fluctuations, α is the adjustment coefficient used to control the sensitivity of the tolerance boundary, and ΔT is the temperature and humidity fluctuation risk factor. threshold It is the tolerance boundary adjustment value, indicating the range of tolerance that needs to be increased; In this step, the system compares the generated temperature and humidity fluctuation risk coefficient (THFR) with the reference threshold (THFR). ref By comparison, the tolerance boundary adjustment value ΔT is calculated. threshold If the risk factor for current temperature and humidity fluctuations exceeds the reference threshold, the system will adjust the threshold based on the difference ΔT. threshold The temperature control system automatically adjusts its tolerance limits, thereby increasing its tolerance to short-term, small fluctuations. This adjustment ensures that the system does not make unnecessary adjustments due to short-term fluctuations, avoiding over-response.

[0046] Assuming short-term temperature and humidity fluctuations are confirmed, the response time of the temperature control system is adaptively extended to prevent it from reacting too quickly to small fluctuations and causing frequent temperature adjustments. The calculation expression is as follows: ; In the formula, T response It is the extended temperature control response time, which is dynamically adjusted based on the evaluation results. β is the baseline coefficient for the extended response time, controlling the sensitivity of the response time. The purpose of this step is to extend the temperature control response time so that the system does not adjust the temperature control equipment too quickly during short-term fluctuations, thereby avoiding frequent adjustments caused by small environmental fluctuations and ensuring that the temperature and humidity in the computer room are maintained within a reasonable range.

[0047] The temperature control system will take into account the risk factor (THFR) of temperature and humidity fluctuations and the extended temperature control response time (T). response This determines the actual temperature control adjustment range. This adjustment value characterizes the adjustment range of the temperature control system to the current temperature and humidity fluctuations. The calculation expression is as follows: ; In the formula, ΔT control γ is the temperature adjustment value of the temperature control system, representing the actual temperature and humidity level that the system needs to adjust. γ is the temperature control sensitivity coefficient, which determines the amplitude of the temperature control system's response.

[0048] This step automatically adjusts the operating status of the temperature control equipment based on the assessment of current fluctuation risks and the extended response time. This ensures the system can react flexibly to short-term fluctuations and avoids excessive temperature control operations. The adjustment strategy intelligently optimizes the temperature control process according to the severity of the fluctuation and the extension of the response time, minimizing energy consumption and maintaining stable equipment operation.

[0049] This invention ensures that the system only reacts when genuine environmental anomalies occur by real-time acquisition of environmental data, intelligent assessment of temperature and humidity fluctuation patterns, and automatic adjustment of the temperature control system's tolerance boundaries and response time. This significantly improves the accuracy and response efficiency of the temperature control system, while also substantially reducing unnecessary energy consumption and equipment burden, minimizing excessive operation of cooling equipment, and enhancing the stability, reliability, and long-term economic efficiency of the server cluster.

[0050] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0051] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

[0052] It should be noted that, in this document, the use of relational terms such as "first" and "second" is merely for distinguishing one entity or operation from another, and does not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes the element.

[0053] It should be understood that in the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0054] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0055] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0056] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0057] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0058] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0059] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A server cluster temperature automatic control and management system based on the Internet of Things, characterized in that, It includes modules for environmental data acquisition and preprocessing, feature extraction and fluctuation analysis, intelligent assessment and pattern recognition, and tolerance adjustment and temperature control response optimization. The environmental data acquisition and preprocessing module collects environmental data from the server cluster's computer room in real time using a temperature and humidity recorder, preprocesses the acquired raw environmental data, and constructs a dataset. The feature extraction and fluctuation analysis module uses feature engineering technology to extract key features from the dataset that reflect short-term fluctuations in the temperature and humidity of the computer room, and performs comprehensive analysis on the extracted features to quantify the magnitude of temperature and humidity fluctuations. The intelligent assessment and pattern recognition module takes the analyzed key indicators as feature vectors and inputs them into a pre-trained machine learning model. The machine learning model then intelligently assesses the temperature and humidity fluctuations in the computer room and identifies the temperature and humidity fluctuation patterns. The tolerance adjustment and temperature control response optimization module automatically increases the tolerance boundary based on the assessment results when the temperature and humidity of the server cluster's data center are fluctuating in the short term. This increases the tolerance for small fluctuations and adaptively extends the temperature control response time to avoid overreacting to short-term fluctuations.

2. The IoT-based server cluster temperature automatic control and management system according to claim 1, characterized in that, The specific steps for collecting real-time environmental data from the server cluster's data center using a temperature and humidity recorder are as follows: First, install temperature and humidity loggers in key locations in the computer room; Then, configure the recorder to communicate with the data acquisition system to ensure that the device can transmit the collected temperature and humidity data to the central monitoring system in real time. Next, set the recorder's sampling frequency and data transmission interval; The temperature and humidity recorder starts working automatically and continuously collects data, which is then transmitted to the data storage system in real time. Finally, the collected data will enter the preprocessing stage to prepare for subsequent analysis.

3. The IoT-based server cluster temperature automatic control and management system according to claim 1, characterized in that, Feature engineering techniques are used to extract key features reflecting short-term fluctuations in the temperature and humidity of the computer room from the dataset. The extracted features include the sensitivity of airflow to changes in the temperature and humidity of the computer room and the coordinated changes in temperature and humidity. The sensitivity of airflow to changes in the temperature and humidity of the computer room and the coordinated changes in temperature and humidity are comprehensively analyzed under the detection window to generate reference values ​​for airflow sensitivity and temperature and humidity coordinated changes. The magnitude of temperature and humidity fluctuations is quantified by the reference values ​​for airflow sensitivity and temperature and humidity coordinated changes.

4. The IoT-based server cluster temperature automatic control and management system according to claim 3, characterized in that, The specific steps for analyzing the sensitivity of airflow to changes in temperature and humidity in the computer room and generating reference values ​​for airflow sensitivity under the detection window are as follows: Record the air velocity and temperature / humidity data at each moment, and calculate the instantaneous impact of changes in airflow on temperature and humidity. The calculation expression is as follows: ； In the formula, R S It is the airflow response intensity, A t A is the air velocity at time t. t-1 The air velocity at time t-1 is the air velocity at the previous time, T. t and H t T represents the temperature and humidity values ​​at time t, respectively. t-1 and H t-1 These represent the temperature and humidity values ​​at time t-1, i.e., the temperature and humidity values ​​at the previous time. After obtaining the airflow response intensity at each moment, the airflow sensitivity reference value is calculated using the following expression: ； In the formula, AFS is the airflow sensitivity reference value, and N is the total number of detection window times. It is the airflow response intensity at time t.

5. The IoT-based server cluster temperature automatic control and management system according to claim 3, characterized in that, The specific steps for analyzing the coordinated changes in temperature and humidity within the detection window to generate reference values ​​for these coordinated changes are as follows: First, the temperature and humidity data are normalized to eliminate the influence of differences in units. The normalization formula for the temperature and humidity data at each moment is as follows: ; In the formula, T norm These are normalized temperature data, where T is the original temperature data, min(T) is the minimum temperature value among all historical temperature records, and max(T) is the maximum temperature value among all historical temperature records. H norm Here, H is the normalized humidity data, min(H) is the minimum value among all humidity records in the historical data, and max(H) is the maximum value among all humidity records in the historical data. The difference between normalized temperature and humidity is calculated, and the difference rate is used to reflect the synchronous change of the two. The expression for the difference rate is as follows; ； In the formula, D represents the degree of difference between temperature and humidity; By combining the difference rates at multiple time points, the coordinated changes in temperature and humidity are quantified. The reference value for the coordinated changes in temperature and humidity is calculated by weighted integral of the difference rates at multiple time points, as shown in the following expression: ; In the formula, D i w represents the degree of difference between temperature and humidity at the i-th time point. i is the weight of the i-th time point, e is the natural base, M is the total number of time points in the detection window, and THC is the reference value for the coordinated change of temperature and humidity.

6. The IoT-based server cluster temperature automatic control and management system according to claim 3, characterized in that, The analyzed airflow sensitivity reference value and temperature and humidity co-change reference value are used as feature vectors and input into a pre-trained machine learning model. The machine learning model generates a temperature and humidity fluctuation risk coefficient, which is used to intelligently assess the temperature and humidity fluctuations in the computer room and identify temperature and humidity fluctuation patterns.

7. The IoT-based server cluster temperature automatic control and management system according to claim 6, characterized in that, A pre-trained machine learning model will be used to intelligently assess temperature and humidity fluctuations in the data center. The temperature and humidity fluctuation risk coefficient generated when identifying the fluctuation pattern will be compared with a pre-set reference threshold to determine the temperature and humidity fluctuation pattern. The determination steps are as follows: If the temperature and humidity fluctuation risk coefficient is greater than the preset reference threshold for temperature and humidity fluctuation risk coefficient, the current temperature and humidity fluctuation in the computer room will be judged as a short-term fluctuation; if the temperature and humidity fluctuation risk coefficient is less than or equal to the preset reference threshold for temperature and humidity fluctuation risk coefficient, the temperature and humidity in the computer room will be judged as not fluctuating.

8. The IoT-based server cluster temperature automatic control and management system according to claim 1, characterized in that, When the assessment results show that the temperature and humidity in the data center where the server cluster is located are experiencing short-term fluctuations, the specific steps for automatically increasing the tolerance boundary and adaptively extending the temperature control response time based on the assessment results are as follows: The temperature and humidity fluctuation risk coefficient (THFR) is compared with a pre-set reference threshold to determine the nature of the temperature and humidity fluctuations. When the assessment results show that the temperature and humidity in the server cluster's data center are experiencing short-term fluctuations, the tolerance boundary is calculated and increased to increase tolerance for short-term, small fluctuations and avoid overreacting to short-term fluctuations. The calculation expression is as follows: ; In the formula, THFR ref ΔT is the reference threshold for the risk factor of temperature and humidity fluctuations, α is the adjustment coefficient used to control the sensitivity of the tolerance boundary, and ΔT is the temperature and humidity fluctuation risk factor. threshold This is the tolerance boundary adjustment value; Assuming short-term temperature and humidity fluctuations are confirmed, the response time of the temperature control system is adaptively extended to prevent it from reacting too quickly to small fluctuations and causing frequent temperature adjustments. The calculation expression is as follows: ; In the formula, T response It is the extended temperature control response time, and β is the reference coefficient for the extension of response time; The temperature control system will take into account the risk factor (THFR) of temperature and humidity fluctuations and the extended temperature control response time (T). response The actual temperature control adjustment range is determined by the following calculation expression: ; In the formula, ΔT control γ is the temperature adjustment value of the temperature control system, and γ is the temperature control sensitivity coefficient.

Images