Machine room temperature control method and system based on data driving and AI analysis

Abstract

The invention provides a machine room temperature control method and system based on data driving and AI analysis, electronic equipment and a storage medium, and aims to solve the problem that machine room air conditioner management is difficult to meet efficient and intelligent control requirements. The historical data comprises machine room environment data, air conditioner operation data and external environment data; arranging and analyzing the collected historical data, predicting the temperature change trend in the machine room through a preset algorithm model, and generating a temperature regulation and control scheme corresponding to the machine room; acquiring real-time temperature data of the machine room; generating a temperature regulation and control instruction according to the real-time temperature data of the machine room and the temperature regulation and control scheme corresponding to the machine room; and working parameters of the air conditioner are automatically adjusted according to the temperature regulation and control instruction to ensure that the temperature in the machine room is maintained within a preset range. The temperature stability of the machine room is ensured, energy is saved, consumption is reduced, and the efficiency and reliability of machine room management are improved.

Description

Technical Field

[0001] The present disclosure relates to the field of energy-saving control technology, and in particular to a computer room temperature control method based on data drive and AI analysis, a computer room temperature control system based on data drive and AI analysis, an electronic device, and a computer-readable storage medium. Background Art

[0002] In modern computer room environments, temperature and humidity control are crucial for stable equipment operation and data security. Overheating, overcooling, or excessively high or low humidity can adversely affect equipment and, in turn, disrupt overall business operations. High temperatures can cause electronic equipment to overheat, resulting in performance degradation and even failure. Low temperatures can affect device battery life and waste energy.

[0003] Generally, an air conditioning system (dedicated or general-purpose) can effectively control the temperature and humidity in a computer room, ensuring that the equipment operates in a suitable environment. Professional temperature monitoring equipment can monitor the temperature in the computer room in real time. However, current air conditioning systems are primarily based on static settings and manual on-site management, which has the following major drawbacks:

[0004] 1. Static settings and manual management: At present, the air conditioning system in the computer room mainly relies on static settings and regular manual inspections. This method means that after the initial setting of various parameters (such as air supply temperature, power on / off operation, working mode, etc.), the parameters remain almost unchanged for a long time until the next manual inspection or adjustment; this management method is inefficient and highly dependent on manual operation, with a high possibility of manual error. Moreover, the set parameters cannot be well matched with the computer room itself, and the temperature control settings lack a scientific basis. It is impossible to develop temperature control models with different strategies based on the different individual differences and seasonal changes in the computer room; it may lead to high energy consumption, and cannot ensure the stability of the computer room temperature, and is not suitable for the dynamic changes in the modern computer room environment.

[0005] 2. Lack of dynamic response capabilities: While existing air conditioning systems can monitor temperature using temperature sensors, their regulation is still based on simple on / off switching based on preset thresholds, lacking refined dynamic adjustment capabilities. Faced with real-time temperature fluctuations within the computer room, the system struggles to quickly and accurately adjust, easily leading to equipment overheating or overcooling, impacting operational stability and lifespan.

[0006] 3. Reliance on manual inspections: Existing technologies rely heavily on manual inspections, which are labor-intensive and require varying degrees of expertise. By the time problems are discovered, they have already caused significant damage to the equipment, making timely prevention and resolution impossible. This approach lacks proactiveness, making it difficult to quickly respond and adjust temperature control settings, impacting the safe operation of equipment in the computer room.

[0007] It can be seen that the current management methods are difficult to meet the needs of modern computer rooms for efficient and intelligent environmental control. Summary of the Invention

[0008] In order to at least solve the problem that the existing computer room air conditioning management methods are unable to meet the needs of modern computer rooms for efficient and intelligent environmental control. The present disclosure provides a computer room temperature control method based on data-driven and AI analysis, a computer room temperature control system based on data-driven and AI analysis, an electronic device, and a computer-readable storage medium. Through data-driven and AI optimization, intelligent adjustment of the computer room temperature environment is achieved, and a one-stop, one-scenario refrigeration equipment control strategy is generated for the computer room and automatically issued for execution. This ensures the stability of the computer room temperature while achieving energy conservation and consumption reduction, thereby improving the efficiency and reliability of computer room environmental management.

[0009] In a first aspect, the present disclosure provides a method for controlling temperature in a computer room based on data-driven and AI analysis, the method comprising:

[0010] Collect historical data corresponding to the computer room within a preset time period, the historical data including computer room environment data, air conditioning operation data and external environment data;

[0011] Organize and analyze the collected historical data, predict the temperature change trend in the computer room through the preset algorithm model, and generate the corresponding temperature control plan for the computer room;

[0012] Collect real-time temperature data of the computer room;

[0013] Generate temperature control instructions based on the real-time temperature data of the computer room and the corresponding temperature control plan of the computer room;

[0014] Automatically adjust the air conditioning operating parameters according to temperature control instructions to ensure that the temperature in the computer room is maintained within the preset range.

[0015] Furthermore, the method further comprises:

[0016] Based on historical overheating and overcooling data and outdoor weather, and using a preset algorithm model, possible overheating and overcooling failures are predicted, warning signals are issued in advance, and guidance is provided on setting temperature control plans.

[0017] Further,

[0018] The computer room environmental data includes: indoor temperature, humidity, return air temperature, and supply air temperature;

[0019] The air conditioning operation data includes: air conditioning operation mode, power on / off status, air supply temperature, and air supply speed;

[0020] The external environment data includes: outdoor temperature, humidity, and weather conditions.

[0021] Furthermore, the prediction of the temperature change trend in the computer room and the generation of a temperature control plan corresponding to the computer room include:

[0022] If the room temperature or the air conditioner return air temperature exceeds the threshold value a°C, the air conditioner supply air temperature will be automatically adjusted to the set value X°C. If this adjustment process is triggered N times continuously and the room temperature continues to be higher than the threshold value a°C, control will be stopped and an air conditioner performance degradation notice will be issued.

[0023] If the room temperature or the air conditioner return air temperature is within the threshold range [b°C, a°C], the air conditioner supply air temperature is adjusted downwards to (x1°C, x2°C, x3°C, ...). If the room temperature remains within the threshold range after M consecutive adjustments, the air conditioner supply air temperature is set to the absolute threshold value Y°C, control is stopped, and an air conditioner performance degradation notice is issued.

[0024] Among them, a>b>x1>x2>x3≥X, x3>Y, a, b, X, x1, x2, x3, Y, N, M are confirmed based on the historical data of the computer room and the predicted results of temperature change trend.

[0025] Furthermore, the temperature control scheme also includes:

[0026] When the ambient temperature is higher than the threshold value c℃, the system will turn on all air conditioners;

[0027] When the ambient temperature is not higher than the threshold value c°C, the energy-saving mode is activated. The energy-saving mode includes:

[0028] When the ambient temperature is within the threshold value [d℃, c℃] and the difference δ between the current temperature value and the last collected value is greater than a predetermined value, the system will determine the current number of air conditioners. If the number of air conditioners is less than the threshold value T, K air conditioners will be turned off. If the number of air conditioners is greater than the threshold value W, L air conditioners will be turned off.

[0029] When the ambient temperature is lower than the threshold value d℃, all air conditioners are turned off;

[0030] Among them, b>c>d, L≥K, c, d, K, W, L are confirmed based on the historical data of the computer room and the prediction results of the temperature change trend.

[0031] Furthermore, the temperature control scheme also includes:

[0032] When the number of times the air conditioner is turned on and off exceeds the threshold value S within the preset time range, the energy-saving mode is turned off.

[0033] Furthermore, the method further comprises:

[0034] Collect performance indicators related to the operation of air-conditioning equipment, conduct real-time monitoring and analysis of the collected performance indicators, and combine the issuance of control instructions and status changes to determine whether the air-conditioning equipment has executed the instructions as expected.

[0035] In a second aspect, the present disclosure provides a computer room temperature control system based on data-driven and AI analysis, the system comprising:

[0036] A collection module configured to collect historical data corresponding to the computer room within a preset time period, wherein the historical data includes computer room environment data, air conditioning operation data, and external environment data;

[0037] The control module is configured to organize and analyze the collected historical data, predict the temperature change trend in the computer room through a preset algorithm model, and generate a corresponding temperature control plan for the computer room;

[0038] The acquisition module is further configured to collect real-time temperature data of the computer room;

[0039] A generation module configured to generate a temperature control instruction based on the real-time temperature data of the computer room and a temperature control plan corresponding to the computer room;

[0040] The adjustment module is configured to automatically adjust the operating parameters of the air conditioner according to the temperature control instruction to ensure that the temperature in the computer room is maintained within a preset range.

[0041] In a third aspect, the present disclosure provides an electronic device comprising a memory and a processor, wherein the memory stores a computer program. When the processor runs the computer program stored in the memory, the processor executes the computer room temperature control method based on data-driven and AI analysis as described in any one of the first aspects.

[0042] In a fourth aspect, the present disclosure provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the computer room temperature control method based on data-driven and AI analysis as described in any one of the first aspects above is implemented.

[0043] Beneficial effects:

[0044] The present disclosure provides a computer room temperature control method based on data-driven and AI analysis, a computer room temperature control system based on data-driven and AI analysis, an electronic device and a storage medium; by real-time monitoring and analysis of temperature data inside and outside the computer room, air-conditioning operating parameters and other dynamic environment data, and through an algorithm model, predicting the temperature change trend in the computer room, thereby analyzing how the computer room should perform air-conditioning control under various circumstances, generating a suitable control plan for the computer room, and ensuring the stable operation of the equipment while improving the energy efficiency of the air-conditioning operation, quickly responding to changes in the computer room environment through data-driven intelligent control, and providing precise temperature control. Furthermore, temperature control models with different strategies have been developed for different computer rooms, tailored to individual differences (such as room size, business volume, and number of devices) as well as seasonal variations. This allows for a "one control strategy for each computer room" approach, enabling dynamic, real-time management. Based on real-time data and predictive models, the operating status of air conditioning equipment is proactively adjusted to reduce over- or under-temperature events and ensure long-term stable operation. AI-powered control significantly reduces the frequency of manual inspections and adjustments, while intelligent management enhances the system's automation and responsiveness. This not only reduces the risks of manual operation but also enables timely prevention and resolution of temperature control issues, improving the efficiency and reliability of computer room management. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 A schematic diagram of a flow chart of a method for controlling temperature in a computer room based on data-driven and AI analysis, provided in the first embodiment of the present disclosure;

[0046] Figure 2 An architectural diagram of an overall system for computer room temperature control based on data-driven and AI analysis, provided in an embodiment of the present disclosure;

[0047] Figure 3 A temperature adjustment statistics page diagram of a system provided in an embodiment of the present disclosure;

[0048] Figure 4 An energy-saving statistics page diagram of a system provided by an embodiment of the present disclosure;

[0049] Figure 5 An overview page diagram of air conditioning equipment of a system provided by an embodiment of the present disclosure;

[0050] Figure 6 A diagram of a command report page of a system provided in an embodiment of the present disclosure;

[0051] Figure 7 A control logic diagram of a temperature control model provided in an embodiment of the present disclosure;

[0052] Figure 8 A control logic diagram of an energy-saving model provided in an embodiment of the present disclosure;

[0053] Figure 9 This is an architectural diagram of a computer room temperature control system based on data-driven and AI analysis, provided in Example 3 of the present disclosure;

[0054] Figure 10 This is an architectural diagram of an electronic device provided in Example 4 of the present disclosure. DETAILED DESCRIPTION

[0055] To enable those skilled in the art to better understand the technical solutions of the present disclosure, the present disclosure is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments and drawings described herein are only used to explain the present disclosure, rather than to limit the present disclosure.

[0056] It should be noted that the terms "first", "second", etc. in the specification and claims of the present disclosure and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence; and, in the absence of conflict, the embodiments and features in the embodiments of the present disclosure can be arbitrarily combined with each other.

[0057] The terms used in the embodiments of the present disclosure are for the purpose of describing specific embodiments only and are not intended to limit the present disclosure. The singular forms "a," "an," "the," and "the" used in the embodiments of the present disclosure and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise.

[0058] In the subsequent description, suffixes such as "module," "component," or "unit" used to represent elements are used only to facilitate the description of the present disclosure and have no specific meaning. Therefore, "module," "component," or "unit" may be used interchangeably.

[0059] In computer room management, temperature and humidity control are very important. CPU overheating may cause the system to slow down or freeze, while memory overheating may cause data loss. In addition, high temperature environments will accelerate the aging of electronic components and shorten the service life of equipment. On the other hand, low temperature environments may also have an adverse effect on equipment. For example, it may cause the equipment to run slower or even malfunction. The current static settings and manual management of various parameters (such as air conditioning supply temperature, power on / off operation, working mode, etc.) set once and then almost permanently change (waiting for the next manual inspection or adjustment). The management is extensive and relies heavily on manual labor. The temperature stability in the computer room cannot be guaranteed, making it difficult to quickly respond and adjust the temperature control settings.

[0060] Therefore, there is an urgent need for a computer room temperature control system that can achieve data-driven, dynamic response and intelligent regulation to address the shortcomings of the existing technology.

[0061] The following is a detailed description of the technical solutions of the present invention and how the technical solutions of the present invention solve the technical problems in the prior art with specific embodiments. It will be appreciated that, in the embodiments of the present application, the execution subject may perform some or all of the steps in the embodiments of the present application, and these steps or operations are merely examples. The embodiments of the present application may also perform other operations or variations of various operations. In addition, the various steps may be performed in different orders as presented in the embodiments of the present application, and it may not be necessary to perform all the operations in the embodiments of the present application. Furthermore, the following specific embodiments may be combined with each other, and the same or similar concepts or processes may not be described in detail in certain embodiments.

[0062] Figure 1 A flow chart of a method for controlling the temperature of a computer room based on data drive and AI analysis is provided in the first embodiment of the present disclosure. Figure 1 As shown, the method includes:

[0063] Step S101: Collecting historical data corresponding to the computer room within a preset time period, wherein the historical data includes computer room environment data, air conditioning operation data, and external environment data;

[0064] Step S102: Collating and analyzing the collected historical data, predicting the temperature change trend in the computer room through a preset algorithm model, and generating a temperature control plan corresponding to the computer room;

[0065] Step S103: collecting real-time temperature data of the computer room;

[0066] Step S104: generating a temperature control instruction based on the real-time temperature data of the computer room and the temperature control plan corresponding to the computer room;

[0067] Step S105: Automatically adjust the operating parameters of the air conditioner according to the temperature control instruction to ensure that the temperature in the computer room is maintained within a preset range.

[0068] This disclosed embodiment implements intelligent O&M of a computer room through dynamic adjustment based on the computer room's environmental parameters and air conditioner operating status. By real-time monitoring of multiple data types, including temperature and humidity inside and outside the computer room, power equipment status, and alarm information, and taking into account factors such as the number and power of air conditioners, adaptively controlling key parameters such as air conditioner on / off, temperature, and air flow rate, the system optimizes air conditioner operation, extends air conditioner life, saves energy, and enhances the system's adaptability.

[0069] The embodiments of the present disclosure are based on Figure 2In the implementation of the overall system, the intelligent temperature control system for the computer room fully utilizes the "four remote" capabilities of the dynamic environment monitoring system (telemetry, telesignaling, remote adjustment, and remote control). Through data collection, data processing, and command control, it builds "temperature adjustment" and "energy saving" models, outputs refined and intelligent air conditioning adjustment solutions, and realizes "one station, one policy" for computer room air conditioning energy saving. The computer room temperature is self-adaptive and self-adaptive, maximizing air conditioning energy saving. In this overall system, the implementation process of each function includes:

[0070] ① Collection dimension:

[0071] Temperature sensor: Responsible for collecting ambient temperature data, which is crucial for monitoring and regulating indoor temperature.

[0072] Air conditioning: The air conditioning equipment itself may have data collection capabilities and can provide its working status, such as operating mode, wind speed, etc.

[0073] Smart meters: used to monitor and record energy consumption data, such as electricity consumption, of air conditioners and other equipment.

[0074] FSU monitoring unit: As the central node for data acquisition, it collects data from devices such as temperature sensors, air conditioners and smart meters, and transmits this information to the dynamic environment monitoring system.

[0075] ②Control dimension:

[0076] The dynamic environment monitoring system receives real-time data from the FSU monitoring unit and monitors ambient temperature, air conditioning status, switching power supply, smart meter, battery temperature, etc. Based on this data, the dynamic environment monitoring system can make decisions and send control instructions to the FSU monitoring unit.

[0077] FSU monitoring unit: Receives control instructions from the dynamic environment monitoring system, such as air conditioning on / off, air supply temperature, etc., and transmits these instructions to the corresponding equipment.

[0078] Air conditioning: According to the instructions of the FSU monitoring unit, perform corresponding operations, such as adjusting temperature, wind speed, etc.

[0079] The intelligent computer room temperature control system, serving as the brain of the entire system, receives basic computer room parameters (such as room type, area, and structure) and equipment data (such as air conditioners, switching power supplies, and batteries) from the resource management system, as well as outdoor weather and temperature data provided by the weather forecast interface. Based on this information, the intelligent control system generates a control model to guide the dynamic environment monitoring system for precise temperature regulation.

[0080] Through this acquisition and control architecture, the entire system realizes intelligent monitoring and adjustment of the computer room temperature, ensuring the stability of the computer room environment and the safe operation of the equipment.

[0081] Detailed explanation of system composition:

[0082] a. Data acquisition module: real-time collection of data such as ambient temperature, air conditioning working status, supply and return air temperature, switching power supply, smart meter, etc. in the computer room.

[0083] b. Data processing module: Receives and analyzes data from the data acquisition module, predicts the temperature change trend through the preset algorithm model, and generates an optimized control plan.

[0084] c. Air conditioning control module: Connected to the air conditioning system in the computer room, it receives control instructions from the data processing module, issues control instructions, and automatically adjusts the operating parameters of the air conditioner, such as air supply temperature, power on / off status, and operating mode, to maintain a suitable temperature in the computer room.

[0085] e. User interface module: provides a visual interface that displays real-time temperature data, control schemes and fault warning information for users to view and adjust system parameters.

[0086] The screenshot of the system operation is as follows: Figure 3 For the temperature control statistics page, Figure 4 For the energy saving statistics page, Figure 5 For the air conditioning equipment overview page, Figure 6 This is the instruction report page.

[0087] The system uses the data acquisition module to collect real-time power and environmental data from the computer room. The data processing module analyzes this data, identifies patterns in temperature fluctuations, and, combined with a pre-set optimization algorithm, generates an optimal air conditioning control plan. The air conditioning control module automatically adjusts the operating status of the air conditioner based on this generated control plan to ensure that the temperature in the computer room remains within an appropriate range.

[0088] Specifically, after obtaining the historical data of the computer room, the temperature change trend in the computer room is predicted through a preset algorithm model, and a temperature control plan corresponding to the computer room is generated; the preset algorithm model can be an LSTM (long short-term memory) neural network, relying on LSTM for predictive analysis and dynamic optimization. Through deep learning analysis of time series data such as the computer room ambient temperature, air conditioning operating status and external environment, the system can predict temperature change trends, dynamically adjust the air conditioning operation strategy, and achieve precise temperature control and efficient energy saving; specifically, through historical data, the temperature change trend of the computer room under various conditions can be analyzed, and it can be obtained what measures should be taken for the air conditioning in various situations to ensure the stability of the computer room temperature, thereby obtaining the corresponding temperature control plan for the computer room, and then according to the current situation of the computer room, the corresponding control instructions are obtained from the temperature control plan, so as to automatically adjust the working parameters of the air conditioner according to the instructions to ensure that the temperature in the computer room is maintained within the preset range, such as 18-27°C.

[0089] The disclosed embodiments monitor and analyze real-time data such as temperature data inside and outside the computer room, air conditioning operating parameters, and other dynamic environmental data. Using an algorithmic model, they predict temperature trends within the computer room, thereby analyzing how the computer room's air conditioning should be controlled in various situations. This generates appropriate control plans for the computer room, ensuring stable equipment operation while improving air conditioning energy efficiency. Data-driven intelligent control rapidly responds to changes in the computer room environment and provides precise temperature control. Furthermore, temperature control models with different strategies are developed for different computer rooms based on individual differences within the room (such as room size, business volume, number of devices, etc.) as well as seasonal variations, achieving a "one control strategy for each computer room" approach. This enables dynamic, real-time management, proactively adjusting the operating status of air conditioning equipment based on real-time data and predictive models, reducing the occurrence of over- or under-temperature events and ensuring long-term stable operation of the equipment. AI-powered control significantly reduces the frequency of manual inspections and adjustments, and intelligent management improves the system's automation level and response speed. This not only reduces the risk of manual operation but also enables timely prevention and resolution of temperature control issues, improving the efficiency and reliability of computer room management.

[0090] Furthermore, the method further comprises:

[0091] Based on historical overheating and overcooling data and outdoor weather, and using a preset algorithm model, possible overheating and overcooling failures are predicted, warning signals are issued in advance, and guidance is provided on setting temperature control plans.

[0092] Through AI prediction of historical data, the system can identify potential extreme risks, such as overcooling or overheating, in advance and make preventive adjustments to further improve the operational safety and stability of the equipment.

[0093] For example, if a potential high-temperature fault is detected, the system will promptly issue an early warning and act on the temperature control model to dynamically adjust model parameters. Simultaneously, the system will notify the administrator through the user interface module, providing guidance on temperature control settings and preventing faults. This entire process requires no human intervention, achieving intelligent and automated temperature management in the computer room.

[0094] Further,

[0095] The computer room environmental data includes: indoor temperature, humidity, return air temperature, and supply air temperature;

[0096] The air conditioning operation data includes: air conditioning operation mode, power on / off status, air supply temperature, and air supply speed;

[0097] The external environment data includes: outdoor temperature, humidity, and weather conditions.

[0098] Historical data comes from various monitoring and sensing systems, including: computer room environment data: indoor temperature, humidity, return air temperature, supply air temperature; air conditioning operation data: operation mode, power on / off status, supply air speed; external environment data: outdoor temperature, humidity, and weather conditions.

[0099] The current management system fails to fully utilize historical data from the computer room for analysis and prediction, failing to proactively prevent potential high-temperature failures. This results in a lack of scientific basis for temperature control settings. It simply relies on real-time data for responses and lacks the ability to predict future temperature changes.

[0100] The disclosed embodiment uses historical data as samples and inputs them into the LSTM model for training, thereby predicting future temperature changes in the computer room. The data sample time span is 6-12 months, set according to actual conditions, and the data frequency is sampled every 10 minutes; the number of collected data samples is about 100,000 sample data; each data contains the following information: timestamp, indoor temperature, humidity, return air temperature, supply air temperature, air conditioner on / off status, outdoor temperature, humidity, air conditioner operating mode (cooling / heating / standby). During the data collection stage, the computer room can be regulated according to the regulation scheme of the same type of computer room combined with manual intervention to maintain the stability of the computer room temperature while reducing energy consumption as much as possible.

[0101] Before input, the collected data needs to be preprocessed, including: missing value processing: due to network instability or hardware problems, some devices may cause some sensor data to be missing, and these missing values ​​are filled by interpolation; normalization processing: normalize each data to the [0,1] interval to facilitate the training of the LSTM neural network; time window processing: construct a time window, use the data of the past 1 hour (that is, 6 data) as input features, and predict the temperature changes in the next 1 hour.

[0102] Preferably, the specific situation of the LSTM model is:

[0103] 1. Model architecture, including:

[0104] Input layer: includes data of 6 time steps, with feature dimensions of [indoor temperature, humidity, return air temperature, supply air temperature, air conditioning operation mode, outdoor temperature].

[0105] LSTM layer: A two-layer LSTM network with 64 units in the first layer and 32 units in the second layer to avoid overfitting.

[0106] Output layer: predict the temperature change value within the next hour.

[0107] 2. Training Process

[0108] Data split: 70% of the data is used for training, 20% of the data is used for validation, and 10% of the data is used for testing.

[0109] Loss function: Mean Squared Error (MSE).

[0110] Optimizer: Adam optimizer with a learning rate of 0.001.

[0111] Training epochs: 50, batch size 64.

[0112] 3. Training process example

[0113] During training, the LSTM model continuously optimizes its weights and biases to learn the temporal dependencies of data and the patterns of temperature changes. For example, if the input data at a certain moment is the room's ambient temperature, humidity, and air conditioning status over the past hour, the LSTM can use this data to predict the temperature trend over the next hour.

[0114] Data Validation

[0115] 1. Verification indicators

[0116] Mean Squared Error (MSE): used to evaluate the error between the predicted value and the true value.

[0117] Coefficient of determination (R2): used to measure the fit of the model to the temperature change trend.

[0118] 2. Training Results

[0119] After 50 rounds of training, the model’s MSE was 0.03 and the R2 value was 0.89, which indicates that the model can capture the trend of temperature change well.

[0120] 3. Comparison of prediction results

[0121] Assume that a data input in the test set is shown in Table 1.

[0122] Table 1 Test set data samples

[0123]

[0124] Comparing the model's predictions with the actual values, the forecast indicates a temperature rise of 26.2°C within the next hour, while the actual observed value is 26.1°C. This comparison shows that the model's predictions are very close to the actual values, indicating a relatively accurate prediction.

[0125] 4. Visual Verification

[0126] By comparing the actual temperature and predicted temperature curves, it can be concluded that the temperature change trends are completely consistent. Especially in high temperature periods, the LSTM model can predict the temperature increase in advance.

[0127] Furthermore, the prediction of the temperature change trend in the computer room and the generation of a temperature control plan corresponding to the computer room include:

[0128] If the room temperature or the air conditioner return air temperature exceeds the threshold value a°C, the air conditioner supply air temperature will be automatically adjusted to the set value X°C. If this adjustment process is triggered N times continuously and the room temperature continues to be higher than the threshold value a°C, control will be stopped and an air conditioner performance degradation notice will be issued.

[0129] If the room temperature or the air conditioner return air temperature is within the threshold range [b°C, a°C], the air conditioner supply air temperature is adjusted downwards to (x1°C, x2°C, x3°C, ...). If the room temperature remains within the threshold range after M consecutive adjustments, the air conditioner supply air temperature is set to the absolute threshold value Y°C, control is stopped, and an air conditioner performance degradation notice is issued.

[0130] Among them, a>b>x1>x2>x3≥X, x3>Y, a, b, X, x1, x2, x3, Y, N, M are confirmed based on the historical data of the computer room and the predicted results of temperature change trend.

[0131] Through data-driven intelligent model control, including temperature control model and energy-saving model;

[0132] The goal of the temperature control model is to ensure the normal operation of the computer room without high temperature alarms. Its core logic is to dynamically adjust the air supply temperature based on the comparison between the room temperature and the air conditioner return air temperature and the set threshold value. The specific control process is as follows:

[0133] When the room temperature or the air conditioner return air temperature exceeds a threshold of a°C, the system automatically adjusts the air conditioner supply air temperature to a set value of X°C. If this adjustment process is triggered N times consecutively and the room temperature remains above the threshold of a°C, the system stops control and issues an air conditioner performance degradation notice, prompting equipment inspection and maintenance.

[0134] When the room temperature or the air conditioner return air temperature falls within the threshold range [b°C, a°C], the system adjusts the air conditioner supply air temperature downwards to (x1°C, x2°C, x3°C, ...). If the room temperature remains within this range after M consecutive adjustments, the air conditioner supply air temperature is set to the absolute threshold value Y°C, the system stops control, and issues an air conditioner performance degradation notice.

[0135] Among them, a, b, X, x1, x2, x3, Y, N, and M are confirmed based on the computer room's historical data and temperature trend prediction results. For each computer room, they can be set to corresponding values ​​based on the external environment or season. For a computer room in summer, for example, a is set to 32, b is set to 28, X can be set to 20, x1, x2, and x3 are set to 25, 22, and 20 respectively, Y is set to 20, and N and M are both set to 2. Although the setting logic is the same for each computer room, the adjustment parameters set for each room may vary. The LSTM model can analyze the subsequent temperature changes and the parameters that need to be adjusted under different conditions, thereby setting the corresponding room adjustment parameters and thresholds.

[0136] The LSTM model can predict future room temperature changes based on certain air conditioning operation data, external environmental data, and the current room ambient temperature. This allows for accurate control of air conditioning settings. For example, in a room with a summer temperature of 35°C, three air conditioners are operating, the set temperature is 25°C, and the room temperature is 30°C. The prediction is that the temperature will rise to 32°C. However, when the air conditioner temperature is set at 22°C, the room temperature remains stable at 26°C. By analyzing room temperature variations under various conditions, parameter settings can be optimized.

[0137] Furthermore, these parameter values ​​and threshold values ​​can be dynamically adjusted based on the temperature prediction results of the LSTM model.

[0138] The control logic of the temperature control model is as follows Figure 7 As shown;

[0139] The system controls the temperature by judging the conditions and executing actions based on the return air or ambient temperature and the air conditioning temperature setting. Taking a computer room as an example, the following is a detailed logic description:

[0140] Condition 1: If the return air or ambient temperature reaches 32 degrees, the air conditioner will adjust the temperature to 20 degrees. If it is triggered twice and the temperature is not controlled, the system will stop control and dispatch an order.

[0141] Condition 2: If the return air or ambient temperature is between 28 degrees (inclusive) and 32 degrees, the system will trigger temperature adjustment. Temperature judgment: the system will determine whether the current temperature meets the condition that the return air or ambient temperature is ≥28 degrees and <32 degrees; Temperature control: all air conditioners will lower the supply air temperature to 25 degrees, 22 degrees, and 20 degrees in turn, one level each time; if the adjustment is ≥2 times, the system will check whether at least two temperature adjustments have been made; if two adjustments have not been made, continue to adjust the temperature; if two adjustments have been made, proceed to the next step.

[0142] Determine the return air or ambient temperature. If it is greater than 32 degrees, perform temperature control and adjust all air conditioners to 20 degrees. Determine whether the adjustment is greater than or equal to 2 times. If so, determine whether the temperature is still greater than 32 degrees. If so, the temperature adjustment fails, stop control, and issue an air conditioner performance degradation notice. If it is less than 2 times, continue to determine whether the temperature is greater than 32 degrees. If so, repeat the above steps. If the temperature is between 28 and 32 degrees, control the temperature and lower the supply air temperature of all air conditioners to 25, 22, and 20 degrees in turn, one level each time. After a period of time, determine that the return air or ambient temperature is between 28 and 32 degrees, and the issued temperature is not 20 degrees, continue to adjust the temperature. If the return air or ambient temperature is between 28 and 32 degrees, and the issued temperature has been 20 degrees twice, it is determined that the air conditioner performance has deteriorated, a prompt is given, and control is stopped.

[0143] Through the above logic, the system can intelligently adjust the air conditioning temperature according to different temperature ranges and temperature adjustment times, and prompt maintenance when necessary to ensure that the ambient temperature remains within a reasonable range.

[0144] Furthermore, when generating the temperature control plan for the computer room, the LSTM model can make proactive predictions by analyzing historical data (temperature, equipment load, energy consumption, external environment, etc.) to predict the temperature change trend over a period of time in the future (such as 1 hour). This allows the threshold to be dynamically optimized, and the prediction results are input into the temperature control model to dynamically adjust the preset temperature threshold or air conditioning parameter setting value, rather than relying on a fixed threshold.

[0145] In addition, collaborative energy-saving decisions are made, combining the prediction results with the current temperature to determine whether the air conditioner needs to be started / shut down in advance to avoid ineffective starts and stops in a short period of time.

[0146] Due to the complexity of temperature changes in the computer room, temperature fluctuations are affected by multiple factors, such as sudden changes in equipment load, changes in the external environment (such as outdoor temperature and humidity), and air conditioning response delays. Relying solely on current data is not enough to cope with it. For example, when the server suddenly experiences a high load, the temperature may rise rapidly within a few minutes, and the inertia of air conditioning cooling may cause overshoot (temperature exceeding the standard).

[0147] Through model prediction, you can:

[0148] Early intervention: If it is predicted that the temperature will exceed the threshold in the future, the air conditioning supply temperature can be lowered in advance (rather than waiting until it actually exceeds the threshold before responding), reducing the temperature fluctuation range.

[0149] Avoid ineffective actions: If it is predicted that the temperature will continue to drop in the future, even if the current temperature is slightly higher, you can reduce the cooling capacity or delay starting the air conditioner to avoid overcooling.

[0150] Achieve a balance between energy saving and stability: By predicting when to "reserve cooling capacity in advance" or "release cooling capacity", we can ensure temperature stability while saving energy.

[0151] In some cases, dynamic threshold adjustments can also be made. Traditional static thresholds are set within a fixed temperature range (e.g., 22-25°C). With predictive dynamic thresholds, if a rapid temperature rise is predicted, the lower threshold can be temporarily lowered (e.g., 20°C) to reserve cooling capacity in advance. If the temperature is predicted to fluctuate slowly, the threshold range can be widened (e.g., 23-26°C) to reduce the frequency of air conditioning starts and stops.

[0152] (3) Control strategy optimization example

[0153] Scenario 1: If the server load is predicted to surge in one hour (e.g., when a video rendering task is started), you can take action by lowering the supply air temperature in advance to offset future heat input and avoid overloading the air conditioner after a sudden temperature rise.

[0154] Scenario 2: The outside temperature is predicted to drop at night. Action: Reduce air conditioning capacity to allow natural heat dissipation through cool outside air, reducing energy consumption.

[0155] By setting a temperature control plan for each computer room or similar computer room, and developing temperature control models with different strategies based on individual room differences (such as room size, business volume, number of devices), as well as seasonal variations, this system enables dynamic, real-time management. The system proactively adjusts the operating status of air conditioning equipment based on real-time data and predictive models, reducing overheating and underheating events and ensuring long-term stable operation. It also dynamically adjusts various air conditioning equipment parameters to ensure stable operation while improving energy efficiency. Data-driven intelligent control rapidly responds to changes in the computer room environment, providing precise temperature control and ensuring successful command execution, thereby preventing equipment from overheating or overcooling. Furthermore, the control plan significantly reduces the frequency of manual inspections and adjustments, and intelligent management improves the system's automation and responsiveness. This not only reduces the risk of manual operation but also enables timely prevention and resolution of temperature control issues, improving the efficiency and reliability of computer room management.

[0156] Furthermore, the temperature control scheme also includes:

[0157] When the ambient temperature is higher than the threshold value c℃, the system will turn on all air conditioners;

[0158] When the ambient temperature is not higher than the threshold value c°C, the energy-saving mode is activated. The energy-saving mode includes:

[0159] When the ambient temperature is within the threshold value [d℃, c℃] and the difference δ between the current temperature value and the last collected value is greater than a predetermined value, the system will determine the current number of air conditioners. If the number of air conditioners is less than the threshold value T, K air conditioners will be turned off. If the number of air conditioners is greater than the threshold value W, L air conditioners will be turned off.

[0160] When the ambient temperature is lower than the threshold value d℃, all air conditioners are turned off;

[0161] Among them, b>c>d, L≥K, c, d, K, W, L are confirmed based on the historical data of the computer room and the prediction results of the temperature change trend.

[0162] The goal of the energy-saving model is to ensure the normal operation of the computer room while minimizing energy consumption. Its core logic is to turn off the air conditioner as much as possible at an appropriate temperature, based on the room temperature and air conditioner status. The specific control process is as follows:

[0163] Model start and stop judgment: When the number of times the air conditioner is turned on and off exceeds the threshold value S within a certain time range, the system will turn off the energy-saving mode.

[0164] Turn off air conditioners at low temperatures: When the ambient temperature is lower than the threshold value d℃, the system will turn off all air conditioners.

[0165] Adjust the number of air conditioners turned off at medium temperatures: When the ambient temperature is within the threshold range [d°C, c°C] and the difference δ between the current temperature and the last collected value is greater than a predetermined value, the system determines the number of air conditioners currently in use. If the number of air conditioners is less than the threshold value T, K air conditioners are turned off; if the number of air conditioners is greater than the threshold value W, L air conditioners are turned off.

[0166] Enable air conditioning at high temperature: When the ambient temperature is higher than the threshold value c℃, the system will turn on all air conditioners.

[0167] c, d, K, W, and L are set according to the situation of the computer room. They are determined by combining historical data analyzed by the LSTM model with the situation of the computer room. For example, in an energy-saving mode, c is set to 25, d is set to 21, K is set to 1, W is set to 2, and L and W are set to 2. The preset value can be set to 1°C or 2°C.

[0168] Energy saving mode control logic is as follows Figure 8 shown.

[0169] Turn on the energy-saving mode (implemented through the energy-saving model), determine whether the ambient temperature is at ambient temperature 1 (≤21 degrees), if so, turn off all air conditioners; when the ambient temperature is at ambient temperature 2 (between 21 degrees and 25 degrees), determine whether the difference with the upper abnormal temperature is above ±2 degrees, if so, determine the number of air conditioners, if the number of air conditioners is ≤2, ensure that one is turned off, if the number of air conditioners is ≥3, ensure that two are turned off; when the ambient temperature is at ambient temperature 3 (≥25.1 degrees), turn on all air conditioners.

[0170] The disclosed embodiments can implement data-driven intelligent model control. Through the dynamic environment monitoring system, resource management system, weather forecast interface, and other means, the system can obtain real-time data on multiple types of data, including the temperature inside and outside the computer room, the operating status of power equipment, alarms, and computer room area. The system analyzes the correlation between each data type, dynamically assesses cooling demand, and, based on additional factors such as the number and power of air conditioners in the computer room, adaptively controls key parameters such as the on / off status, temperature, and air flow rate of the air conditioners, achieving intelligent operation and maintenance without human intervention. Different models are bound to different computer rooms at different times to achieve refined logical control in specific scenarios, thereby achieving the goal of reasonably controlling the operating time of the air conditioners, extending their lifespan, and saving electricity and energy.

[0171] For LSTM models, dynamic optimization processes can also be performed, including the following aspects.

[0172] 1. Temperature prediction drive control strategy

[0173] Based on the predictions of the LSTM model, the air conditioning system can adopt the following strategies: High temperature response: When the predicted temperature is about to exceed the set threshold, the system will start the air conditioning in advance, adjust the supply air temperature or increase the number of air conditioners to ensure that the temperature in the computer room remains within a safe range.

[0174] Energy-saving control: If the temperature is predicted to remain within the set range for a period of time in the future, the system will gradually shut down some air conditioners to reduce energy consumption.

[0175] 2. Model Self-Optimization

[0176] Real-time feedback: When the actual temperature change of the system deviates significantly from the predicted value (for example, >1.5°C), the system will adjust the parameters of the LSTM model based on the deviation to ensure a more accurate prediction next time.

[0177] Regular retraining: As time goes by and data accumulates, the system will regularly retrain the LSTM model to adapt to new computer room environments and external weather changes.

[0178] 3. Dynamic Optimization Example

[0179] Assuming that the system predicts that the temperature will exceed the standard at a certain moment, the system will activate the air conditioner to cool down the room in advance. The data verification is shown in Table 2.

[0180] Table 2 Verification result examples

[0181]

[0182]

[0183] By comparing the predicted and actual temperature changes, it can be seen that the system successfully predicted the temperature increase and took timely air conditioning startup measures to ensure the stability of the temperature control system.

[0184] The LSTM-based temperature prediction and dynamic optimization mechanism achieves the following functions: Accurate prediction: The LSTM model can accurately predict temperature changes over the next hour, with an MSE value stable at 0.03 and an R² value of 0.89. Automatic adjustment: Based on the temperature prediction, the air conditioning system can be adjusted in advance to avoid the risks posed by high temperatures. Energy saving and high efficiency: By optimizing and adjusting the temperature prediction results, the energy efficiency of the air conditioning system has been significantly improved, reducing energy consumption. Continuous optimization: Based on real-time data feedback, the system can continuously optimize the prediction model to ensure efficient operation under different environmental conditions. This process provides an intelligent and automated adjustment solution for the computer room air conditioning system, significantly improving the system's response speed, adjustment accuracy, and energy saving effects, and enhancing the stability and safety of the computer room environment.

[0185] AI technology is used to continuously optimize the temperature control model. By utilizing historical data and conducting deep learning analysis on time series data such as the ambient temperature of the computer room, air conditioning operating status, and external environment, the system can predict temperature change trends, optimize and adjust model parameters, and achieve precise temperature control and efficient energy saving.

[0186] Combining the temperature control model with the LSTM model prediction, through real-time monitoring of multiple types of data such as temperature, humidity, power equipment status, and alarm information inside and outside the computer room, and taking into account factors such as the number and power of air conditioners, adaptive control of key parameters such as air conditioner on and off, temperature, and air supply intensity is implemented to achieve the goals of optimizing air conditioner operation, extending air conditioner life, saving energy consumption, and improving the system's adaptability.

[0187] Furthermore, the temperature control scheme also includes:

[0188] When the number of times the air conditioner is turned on and off exceeds the threshold value S within the preset time range, the energy-saving mode is turned off.

[0189] The preset time range can be set to 1 hour, 2 hours, or 3 hours, and S can be set to 3 or 4 times. If the air conditioner is turned on and off ≥ 3 times within 2 hours, the energy-saving mode is turned off; by setting the switching conditions for the energy-saving mode, mechanical wear caused by frequent starting and stopping of the air-conditioning compressor can be effectively prevented.

[0190] Furthermore, the method further comprises:

[0191] Collect performance indicators related to the operation of air-conditioning equipment, conduct real-time monitoring and analysis of the collected performance indicators, and combine the issuance of control instructions and status changes to determine whether the air-conditioning equipment has executed the instructions as expected.

[0192] The disclosed embodiment also provides an air-conditioning state change feedback mechanism;

[0193] In actual production, the FSU in the dynamic environment system is solely responsible for executing remote commands and does not provide feedback on successful execution. This design makes it impossible to verify in real time whether the commands have successfully reached the air conditioning equipment, resulting in reduced system control accuracy. This is especially true in computer room environments, where air conditioning equipment must accurately regulate environmental parameters such as temperature and humidity. Failure to successfully execute commands can lead to equipment overheating, failure, and even direct disruption to the normal operation of the data center.

[0194] The main issues that FSU cannot directly provide feedback on are as follows:

[0195] Inconsistency between command execution and status feedback: After receiving a command, the FSU simply sends it to the air conditioner but cannot guarantee successful execution. For example, an air conditioner's on / off command or temperature adjustment command may not be executed due to a device malfunction, communication issues, or other factors.

[0196] Unclear device status: The status of the air conditioning equipment (such as temperature, wind speed, and humidity) may not meet the expected values, but the FSU cannot obtain accurate device status information in real time, making it difficult to determine the cause of the problem.

[0197] To address the FSU's inability to provide feedback on successful execution, this disclosure incorporates a state-based feedback mechanism. By monitoring and analyzing collected performance indicators in real time, the system can indirectly determine whether the air conditioning equipment is executing instructions as expected, ensuring the effectiveness and success rate of the adjustment process. This can be done in the following scenarios.

[0198] Performance indicator collection and monitoring

[0199] The system collects a series of performance indicators related to the operation of air conditioning equipment, such as temperature, humidity, voltage, current, wind speed, return air temperature, etc., and analyzes and verifies the data by combining the issuance of switch commands and status changes. Specific key indicators include:

[0200] Voltage / current indicators: including AC input A / B / C phase voltage, AC input A / B / C phase current, total active energy, etc. These data reflect whether the air conditioning equipment is powered normally and whether there is an electrical fault or overload problem.

[0201] Ambient temperature and humidity indicators: such as return air temperature, supply air temperature, indoor humidity, etc. These indicators can effectively determine whether the air conditioner is regulating the indoor ambient temperature normally.

[0202] Fault and abnormality monitoring: such as fan failure, filter blockage, water immersion, humidifier failure, air conditioner failure, compressor overheating, etc. Any abnormal signal can indicate that the air conditioning equipment fails to operate as instructed.

[0203] Remote control status: Monitors the air conditioner's remote control on / off, remote cooling / heating mode, remote temperature setting, etc., and determines whether the instructions achieve the expected results by comparing with the expected values.

[0204] For example, it can continuously monitor changes in temperature and humidity in the environment 24 hours a day, collecting data every 1-5 minutes to ensure real-time and accurate data. After issuing a command, the air conditioner status changes provide timely feedback to confirm whether the command was successful.

[0205] Data analysis and judgment mechanism

[0206] By collecting and analyzing these performance indicators in real time, the system can promptly detect command execution failures when data deviations occur. For example, if the AC input voltage, current, return air temperature, and other data remain unchanged after the air conditioner receives a power-on command, the system can infer that the air conditioner failed to start or adjust the temperature successfully.

[0207] Furthermore, the system uses big data analysis and AI algorithms, combined with historical data, to calculate the expected behavior of the air conditioner under given conditions, and then determine whether the current status meets expectations. If a certain indicator (such as return air temperature or humidity) remains unchanged for a long time and the system has issued an adjustment command, it can be determined that the command was not successfully executed.

[0208] State change feedback mechanism

[0209] In order to ensure the successful execution of commands and timely feedback, the system has designed a state change feedback mechanism, that is, by monitoring the state changes of the air-conditioning equipment after receiving the command, to determine whether the command execution is successful or not. The specific implementation method is as follows:

[0210] Temperature control model: remote control combined with remote measurement

[0211] In the temperature control model, remote control is used to set the air conditioner's temperature parameters, while remote sensing provides feedback on the air conditioner's operating status through the return air temperature. The system compares the set temperature in the command with the actual return air temperature to determine whether the temperature control has achieved the desired effect. If the return air temperature does not change according to the set value, it can be inferred that the air conditioner has not successfully executed the control command, and the system will automatically reissue the command for adjustment.

[0212] Energy-saving model: combination of remote control and remote signaling

[0213] In the energy-saving model, the remote control is used to turn the air conditioner on and off, while the telecontrol system uses the air conditioner's on / off status to provide feedback on whether the air conditioner has successfully responded to the control command. When the system issues a power-on / off command, the telecontrol system verifies the air conditioner's on / off status. If the status does not change as expected (for example, the air conditioner fails to start or shut down successfully), the system immediately determines that the command was not executed and resends the command or activates the emergency plan.

[0214] Actual data support and verification

[0215] To verify the effectiveness of this system solution, simulations and tests can be performed based on historical data. In a typical computer room environment, assuming that the air conditioning equipment receives a temperature adjustment command, the ambient temperature should be adjusted to the set range within 10 minutes. By analyzing the following data:

[0216] Return air temperature change: From 24°C before the command is issued to 20°C (i.e., the set temperature). The system needs to verify whether the temperature change curve meets the expectations within 10 minutes.

[0217] Current and voltage monitoring: If the air conditioning equipment fails to receive the command successfully, the voltage and current values ​​may be abnormal. The system can use these abnormal signals to determine that the equipment has not performed the adjustment.

[0218] Fault monitoring data: If problems such as fan failure and filter blockage occur, the system should be able to detect them in time after the command is issued and provide feedback on the problem.

[0219] By collecting this data and comparing it with the control strategy, the system can verify whether the instruction execution is successful and provide real-time feedback on the results.

[0220] in conclusion

[0221] By introducing a state change feedback mechanism, the system effectively addresses the FSU's inability to provide timely feedback on command execution success and enables precise control of air conditioning equipment. By collecting and analyzing performance metrics in real time, combined with AI algorithms, the system can promptly detect and adjust command execution failures, ensuring stable operation of the computer room's temperature control system, reducing failures, and improving equipment availability and safety. This mechanism not only enhances the system's automation and intelligence, but also provides strong support for refined management of the computer room environment.

[0222] The disclosed embodiments can predict temperature change trends and dynamically adjust air conditioning operation strategies, thereby reducing overheating events in the computer room while achieving energy conservation and consumption reduction. The following objectives can be achieved:

[0223] 1. Data-driven intelligent control: By real-time monitoring and analysis of temperature data inside and outside the computer room, air conditioning operating parameters, and other dynamic environmental data, the system dynamically adjusts the various parameters of the air conditioning equipment, ensuring stable operation while improving air conditioning energy efficiency. Data-driven intelligent control can quickly respond to changes in the computer room environment, provide precise temperature control, ensure successful command execution, and prevent equipment from overheating or overcooling.

[0224] 2. Dynamic Response and Real-Time Management: This invention develops temperature control models with different strategies based on individual differences within the computer room (such as room size, business volume, number of devices, etc.) as well as seasonal variations, enabling dynamic, real-time management. The system proactively adjusts the operating status of air conditioning equipment based on real-time data and predictive models, reducing over- and under-temperature events and ensuring long-term stable operation.

[0225] 3. Monitoring and evaluating the success of issued commands: This invention incorporates a feedback mechanism based on state changes. By monitoring and analyzing collected performance indicators in real time, the system can indirectly determine whether the air conditioning equipment has executed commands as expected, ensuring the timeliness, effectiveness, and success rate of the adjustment process.

[0226] 4. Use AI to analyze historical data: This invention uses AI technology to continuously optimize the temperature control model. By using historical data and conducting deep learning analysis on time series data such as the ambient temperature of the computer room, the operating status of the air conditioner, and the external environment, the system can predict temperature change trends, optimize and adjust model parameters, and achieve precise temperature control and efficient energy saving.

[0227] 5. Reduced manual intervention and improved automation: The frequency of manual inspections and adjustments has been significantly reduced, while intelligent management has enhanced the system's automation and response speed. This not only reduces the risk of manual operation but also enables timely prevention and resolution of temperature control issues, improving the efficiency and reliability of computer room management.

[0228] In summary, the present disclosure realizes intelligent adjustment of the temperature environment of the computer room through data-driven and AI optimization, effectively solving the problems of static settings, lack of dynamic response, reliance on manual inspections, and failure to effectively utilize historical data in the existing technology, thereby improving the efficiency and reliability of computer room environment management.

[0229] The second embodiment of the present disclosure also provides a computer room temperature control system based on data drive and AI analysis;

[0230] The disclosed system mainly obtains the output data of the computer room environment, air conditioning temperature, smart meters, etc. in real time through the dynamic environment monitoring system, combines the outdoor weather data, obtains the data of the computer room category, area, structure, etc. from the asset management system to calculate the heat load of the computer room, and builds a computer room temperature control model in combination with the location information. It generates a one-stop one-scene refrigeration equipment control strategy through the supply and demand relationship of cooling capacity, and automatically issues it for execution. At the same time, it predicts the temperature change trend through the AI ​​algorithm, and dynamically adjusts the air conditioning operation strategy, so as to reduce the overheating incidents in the computer room while achieving energy conservation and consumption reduction. The disclosed system mainly obtains the output data of the computer room environment, air conditioning temperature, smart meters, etc. in real time through the dynamic environment monitoring system, combines the outdoor weather data, obtains the data of the computer room category, area, structure, etc. from the asset management system to calculate the heat load of the computer room, and builds a computer room temperature control model in combination with the location information. It generates a one-stop one-scene refrigeration equipment control strategy through the supply and demand relationship of cooling capacity, and automatically issues it for execution. At the same time, it predicts the temperature change trend through the AI ​​algorithm, and dynamically adjusts the air conditioning operation strategy, so as to reduce the overheating incidents in the computer room while achieving energy conservation and consumption reduction.

[0231] Specific technical implementation:

[0232] One-stop, one-scenario strategy: Obtain computer room structural data through the asset management system, combine location information to build a thermal field model, and dynamically match the cooling equipment combination (such as precision air conditioning and natural cooling sources) to achieve "one computer room, one control strategy."

[0233] Status feedback mechanism: This system monitors performance indicators such as the air conditioning unit's current, voltage, and compressor status, and uses abnormal fluctuations to determine the execution status of commands. For example, if a command requires the supply air temperature to be adjusted to 22°C, and the system detects that the actual temperature has not changed and the compressor current has not fluctuated, it will determine that the command has not been executed, automatically resend the command, and trigger an alarm.

[0234] AI prediction model: This model uses an LSTM network to analyze historical data (such as temperature, load, and energy consumption over the past 30 days) to predict temperature trends over the next 24 hours and adjust air conditioning operating parameters in advance. For example, if a nighttime load increase is predicted, the system will increase cooling power in advance to prevent a sudden temperature rise.

[0235] Solution Overview

[0236] By real-time monitoring of multiple types of data such as temperature, humidity, power equipment status, alarm information inside and outside the computer room, and combining factors such as the number and power of air conditioners, adaptive control of key parameters such as air conditioner on and off, temperature and air supply intensity is implemented to optimize air conditioner operation, extend air conditioner life, save energy and improve system adaptability.

[0237] 1.1 System Architecture

[0238] The intelligent temperature control system for the computer room fully utilizes the "four remotes (telemetry, telesignaling, remote adjustment and remote control)" capabilities of the dynamic environment monitoring system. Through data collection, data processing, command issuance and control, it builds "temperature control" and "energy saving" models, outputs refined and intelligent air conditioning adjustment solutions, and realizes "one station, one policy" for energy saving of computer room air conditioning. The computer room temperature is self-adaptive and self-adaptive, maximizing energy saving of air conditioning. The system architecture diagram is shown below. Figure 2 .

[0239] Figure 2 System architecture diagram

[0240] ① Collection dimension:

[0241] Temperature sensor: Responsible for collecting ambient temperature data, which is crucial for monitoring and regulating indoor temperature.

[0242] Air conditioning: The air conditioning equipment itself may have data collection capabilities and can provide its working status, such as operating mode, wind speed, etc.

[0243] Smart meters: used to monitor and record energy consumption data, such as electricity consumption, of air conditioners and other equipment.

[0244] FSU monitoring unit: As the central node for data acquisition, it collects data from devices such as temperature sensors, air conditioners and smart meters, and transmits this information to the dynamic environment monitoring system.

[0245] ②Control dimension:

[0246] The dynamic environment monitoring system receives real-time data from the FSU monitoring unit and monitors ambient temperature, air conditioning status, switching power supply, smart meter, battery temperature, etc. Based on this data, the dynamic environment monitoring system can make decisions and send control instructions to the FSU monitoring unit.

[0247] FSU monitoring unit: Receives control instructions from the dynamic environment monitoring system, such as air conditioning on / off, air supply temperature, etc., and transmits these instructions to the corresponding equipment.

[0248] Air conditioning: According to the instructions of the FSU monitoring unit, perform corresponding operations, such as adjusting temperature, wind speed, etc.

[0249] The intelligent computer room temperature control system, serving as the brain of the entire system, receives basic computer room parameters (such as room type, area, and structure) and equipment data (such as air conditioners, switching power supplies, and batteries) from the resource management system, as well as outdoor weather and temperature data provided by the weather forecast interface. Based on this information, the intelligent control system generates a control model to guide the dynamic environment monitoring system for precise temperature regulation.

[0250] Through this acquisition and control architecture, the entire system realizes intelligent monitoring and adjustment of the computer room temperature, ensuring the stability of the computer room environment and the safe operation of the equipment.

[0251] 1.2 Functional Description

[0252] Detailed explanation of system composition:

[0253] a. Data acquisition module: real-time collection of data such as ambient temperature, air conditioning working status, supply and return air temperature, switching power supply, smart meter, etc. in the computer room.

[0254] b. Data processing module: Receives and analyzes data from the data acquisition module, predicts the temperature change trend through the preset algorithm model, and generates an optimized control plan.

[0255] c. Air conditioning control module: Connected to the air conditioning system in the computer room, it receives control instructions from the data processing module, issues control instructions, and automatically adjusts the air supply temperature, power on / off status, and working mode of the air conditioner to maintain a suitable temperature in the computer room.

[0256] d. Fault prediction module: Based on historical high temperature data, outdoor weather, etc., using machine learning algorithms, it predicts possible high temperature faults and issues early warning signals to guide temperature control settings.

[0257] e. User interface module: provides a visual interface that displays real-time temperature data, control schemes and fault warning information for users to view and adjust system parameters.

[0258] The system screenshot is as follows: Figure 3 For the temperature control statistics page, Figure 4 For the energy saving statistics page, Figure 5 For the air conditioning equipment overview page, Figure 6 This is the instruction report page.

[0259] 1.3 Working Principle

[0260] The system uses a data acquisition module to collect real-time power and environmental data from the computer room. The data processing module analyzes this data, identifies patterns in temperature fluctuations, and, combined with a pre-set optimization algorithm, generates an optimal air conditioning control plan. The air conditioning control module automatically adjusts the operating status of the air conditioner based on this generated control plan, ensuring the room temperature remains within an appropriate range.

[0261] When the fault prediction module detects a potential high-temperature fault, it issues a timely warning and acts on the temperature control model to dynamically adjust model parameters. It also notifies the administrator through the user interface module, guiding temperature control settings and preventing faults. This entire process requires no human intervention, achieving intelligent and automated temperature management in the computer room.

[0262] 2. Key technologies and methods

[0263] 2.1 Data-driven intelligent model control

[0264] Using the environmental monitoring system, resource management system, and weather forecast interfaces, the system collects real-time data on various types, including room temperature, power equipment operating status, alarms, and room area. It analyzes correlations between these data types, dynamically assesses cooling needs, and, taking into account additional factors such as the number and power of air conditioners in the room, adaptively controls key parameters such as air conditioner on / off, temperature, and air flow, enabling unmanned intelligent operation and maintenance. Different models are bound to different rooms at different times to implement refined logic control in specific scenarios. This helps optimize air conditioner operating time, extend air conditioner life, and conserve energy. Temperature and humidity are monitored 24 / 7, with data collected every 1-5 minutes to ensure real-time and accurate data. After issuing a command, the system provides timely feedback confirming the success of the command through changes in air conditioner status.

[0265] 2.1.1 Temperature Control Model

[0266] The goal of the temperature control model is to ensure the normal operation of the computer room without high temperature alarms. Its core logic is to dynamically adjust the air supply temperature based on the comparison between the room temperature and the air conditioner return air temperature and the set threshold value. The specific control process is as follows:

[0267] When the room temperature or the air conditioner return air temperature exceeds a threshold of a°C, the system automatically adjusts the air conditioner supply air temperature to a set value of X°C. If this adjustment process is triggered N times consecutively and the room temperature remains above the threshold of a°C, the system stops control and issues an air conditioner performance degradation notice, prompting equipment inspection and maintenance.

[0268] When the room temperature or the air conditioner return air temperature is within the threshold value [b℃, a℃], the system will adjust the air conditioner supply air temperature to (x1℃, x2℃, x3℃, ...) in sequence. If the room temperature is still within this range after M consecutive adjustments, the air conditioner supply air temperature will be set to the absolute threshold value Y℃, the system will stop control and issue an air conditioner performance degradation notice. The control logic is as follows: Figure 7 shown.

[0269] 2.1.2 Energy-saving model

[0270] The goal of the energy-saving model is to ensure the normal operation of the computer room while minimizing energy consumption. Its core logic is to turn off the air conditioner as much as possible at an appropriate temperature, based on the room temperature and air conditioner status. The specific control process is as follows:

[0271] Model start and stop judgment: When the number of times the air conditioner is turned on and off exceeds the threshold value S within a certain time range, the system will turn off the energy-saving mode.

[0272] Turn off air conditioners at low temperatures: When the ambient temperature is lower than the threshold value d℃, the system will turn off all air conditioners.

[0273] Adjust the number of air conditioners turned off at medium temperatures: When the ambient temperature is within the threshold range [d°C, c°C] and the difference δ between the current temperature and the last collected value is greater than a predetermined value, the system determines the number of air conditioners currently in use. If the number of air conditioners is less than the threshold value T, K air conditioners are turned off; if the number of air conditioners is greater than the threshold value W, L air conditioners are turned off.

[0274] Enable air conditioning at high temperature: When the ambient temperature is higher than the threshold value c℃, the system will turn on all air conditioners.

[0275] Control logic such as Figure 8 shown

[0276] 2.3 LSTM-based prediction and dynamic optimization

[0277] To further improve the air conditioning system's accuracy and responsiveness, this system leverages an LSTM (Long Short-Term Memory) neural network for predictive analysis and dynamic optimization. Through deep learning analysis of time-series data such as room ambient temperature, air conditioning operating status, and the external environment, the system predicts temperature trends and dynamically adjusts air conditioning operation strategies, achieving precise temperature control and efficient energy conservation.

[0278] 2.3.1 Data Collection and Preprocessing

[0279] 1. Data Source

[0280] The data comes from the following monitoring and sensing systems: Computer room environmental data: indoor temperature, humidity, return air temperature, and supply air temperature.

[0281] Air conditioning operation data: operation mode, power on / off status, and air supply speed.

[0282] External environment data: outdoor temperature, humidity, and weather conditions.

[0283] 2. Data sample description

[0284] Time span: 6 months

[0285] Data frequency: Sampling every 10 minutes

[0286] Number of collected data samples: approximately 100,000 sample data

[0287] Each data contains the following information: timestamp

[0288] Indoor temperature and humidity

[0289] Return air temperature, supply air temperature

[0290] Air conditioner power status

[0291] Outdoor temperature and humidity

[0292] Air conditioning operation mode (cooling / heating / standby)

[0293] 3. Preprocessing Steps

[0294] Missing value processing: Due to network instability or hardware problems, some devices may cause some sensor data to be missing. Interpolation is used to fill these missing values.

[0295] Normalization: Normalize all data to the range [0,1] to facilitate the training of LSTM neural network.

[0296] Time window processing: Construct a time window and use the data of the past hour (i.e. 6 data items) as input features to predict the temperature change in the next hour.

[0297] 2.3.2 Model Training

[0298] 1. Model Architecture

[0299] Input layer: includes data of 6 time steps, with feature dimensions of [indoor temperature, humidity, return air temperature, supply air temperature, air conditioning operation mode, outdoor temperature].

[0300] LSTM layer: A two-layer LSTM network with 64 units in the first layer and 32 units in the second layer to avoid overfitting.

[0301] Output layer: predict the temperature change value within the next hour.

[0302] 2. Training Process

[0303] Data split: 70% of the data is used for training, 20% of the data is used for validation, and 10% of the data is used for testing.

[0304] Loss function: Mean Squared Error (MSE).

[0305] Optimizer: Adam optimizer with a learning rate of 0.001.

[0306] Training epochs: 50, batch size 64.

[0307] 3. Training process example

[0308] During training, the LSTM model continuously optimizes its weights and biases to learn the temporal dependencies of data and the patterns of temperature changes. For example, if the input data at a certain moment is the room's ambient temperature, humidity, and air conditioning status over the past hour, the LSTM can use this data to predict the temperature trend over the next hour.

[0309] 2.3.3 Data Verification

[0310] 1. Verification indicators

[0311] Mean Squared Error (MSE): used to evaluate the error between the predicted value and the true value.

[0312] Coefficient of determination (R 2 ): It is used to measure the model's fit to the temperature change trend.

[0313] 2. Training Results

[0314] After 50 rounds of training, the model’s MSE is 0.03 and R 2 The value is 0.89, which indicates that the model can capture the trend of temperature change well.

[0315] 3. Comparison of prediction results

[0316] Assume that a data input in the test set is shown in Table 1.

[0317] Table 1 Test set data samples

[0318]

[0319] Comparing the model's predictions with the actual values, the forecast indicates a temperature rise of 26.2°C within the next hour, while the actual observed value is 26.1°C. This comparison shows that the model's predictions are very close to the actual values, indicating a relatively accurate prediction.

[0320] 4. Visual Verification

[0321] By comparing the actual temperature and predicted temperature curves, we can see that the temperature change trends are completely consistent. Especially during high temperature periods, the LSTM model can predict the temperature increase in advance and make corresponding adjustments in time.

[0322] 2.3.4 Dynamic Optimization Process

[0323] 1. Temperature prediction drive control strategy

[0324] Based on the predictions of the LSTM model, the air conditioning system can adopt the following strategies: High temperature response: When the predicted temperature is about to exceed the set threshold, the system will start the air conditioning in advance, adjust the supply air temperature or increase the number of air conditioners to ensure that the temperature in the computer room remains within a safe range.

[0325] Energy-saving control: If the temperature is predicted to remain within the set range for a period of time in the future, the system will gradually shut down some air conditioners to reduce energy consumption.

[0326] 2. Model Self-Optimization

[0327] Real-time feedback: When the actual temperature change of the system deviates significantly from the predicted value (for example, >1.5°C), the system will adjust the parameters of the LSTM model based on the deviation to ensure a more accurate prediction next time.

[0328] Regular retraining: As time goes by and data accumulates, the system will regularly retrain the LSTM model to adapt to new computer room environments and external weather changes.

[0329] 3. Dynamic Optimization Example

[0330] Assuming that the system predicts that the temperature will exceed the standard at a certain moment, the system will activate the air conditioner to cool down the room in advance. The data verification is shown in Table 2.

[0331] Table 2 Verification result examples

[0332] Time (minutes) Indoor temperature (℃) Predicted temperature (℃) Air conditioning status 12:00 25.4 26.1 Open 12:10 25.6 26.3 Open 12:20 26.0 26.5 Open 12:30 26.4 26.7 Open

[0333] By comparing the predicted and actual temperature changes, it can be seen that the system successfully predicted the temperature increase and took timely air conditioning startup measures to ensure the stability of the temperature control system.

[0334] 2.3.5 Summary

[0335] Through the LSTM-based temperature prediction and dynamic optimization mechanism, this system achieves the following functions:

[0336] Accurate prediction: The LSTM model can accurately predict the temperature change in the next hour, with the MSE value stable at 0.03 and R 2 The value reaches 0.89.

[0337] Automatic adjustment: Based on temperature forecasts, the air conditioning system can be adjusted in advance to avoid risks caused by high temperatures.

[0338] Energy saving and high efficiency: Through the optimization and adjustment of temperature prediction results, the energy efficiency of the air-conditioning system has been significantly improved, reducing energy consumption.

[0339] Continuous optimization: Based on real-time data feedback, the system can continuously optimize the prediction model to ensure efficient operation under different environmental conditions.

[0340] This process provides an intelligent and automated adjustment solution for the computer room air conditioning system, significantly improving the response speed, adjustment accuracy and energy-saving effect of the air conditioning system, and enhancing the stability and safety of the computer room environment.

[0341] 3. Effectiveness

[0342] This system is an intelligent operation and maintenance solution that dynamically adjusts the computer room's environmental parameters and air conditioner operating status. Combining real-time monitoring and intelligent analysis of multi-source data, it optimizes air conditioner operation, adaptively controls air conditioners, and comprehensively improves operation and maintenance efficiency. Currently, it has been implemented in all of the company's access communication computer rooms, connecting 2,562 computer rooms, 3,729 air conditioners, and over 950,000 measurement points. The following is a summary of the overall results:

[0343] 1. Precision environmental control

[0344] Dynamic adjustment: By collecting real-time data on indoor and outdoor temperature, humidity, air conditioning status, and power equipment operation, the air conditioner's on / off status, temperature setting, and air supply intensity can be precisely adjusted, reducing the fluctuation range of indoor environmental parameters from ±3°C to ±0.5°C.

[0345] Adaptable to complex scenarios: By combining the number and power of air conditioners in different computer rooms, adaptive operation is achieved in high-load and low-load scenarios to meet diverse needs.

[0346] 2. Significant improvement in energy efficiency

[0347] Energy-saving optimization: Based on temperature control and energy-saving models, air conditioning energy consumption is optimized, resulting in a 20% improvement in overall energy efficiency and a 15% reduction in data center energy consumption. This energy-saving model shuts down the air conditioning in a single data center for an average of 18 hours per day, saving approximately 5.5 kWh of electricity per day. This is expected to generate an annual economic benefit of 1.1 million yuan in electricity savings, over 1.5 million kWh of electricity, and 1.44 million tons of CO2 emissions.

[0348] Intelligent energy-saving strategy: The LSTM prediction model predicts temperature changes in advance, optimizes air conditioner startup timing and operating mode, avoids unnecessary frequent power on and off, and further reduces energy consumption.

[0349] 3. Enhanced equipment reliability and operation and maintenance efficiency

[0350] Reduced Failure Rate: Through data-driven intelligent monitoring, the failure rate of air conditioning equipment has dropped by 30%, extending equipment life by over 20%. The success rate of temperature control models exceeds 90%, and overheating failures in computer rooms have decreased by 47% year-on-year. Furthermore, deterioration of over 300 air conditioners in 287 computer rooms has been identified, and on-site operations and maintenance personnel have been assisted to complete these repairs, effectively extending the service life of these air conditioners.

[0351] Improved operation and maintenance efficiency: Automated management of air conditioning operation and maintenance has been achieved, with daily monitoring and fault response efficiency increased by 50% and the need for manual intervention reduced by 70%.

[0352] 4. System intelligence and visualization upgrade

[0353] Data-driven management: Introducing LSTM-based prediction and dynamic optimization models to deeply mine historical and real-time data, accurately identify temperature control modes, and optimize operating strategies.

[0354] Visualization capability: Build a full-process visualization interface to intuitively display the computer room environmental parameters, air conditioning operating status and alarm information, supporting rapid problem location and resolution.

[0355] 5. Comprehensive value realization

[0356] Overall stability: Through dynamic adjustment and intelligent optimization, the stability of the computer room temperature control system in sudden environmental changes has been improved by 40%.

[0357] Intelligent Management: Achieve a transformation from a passive response to an active prediction management model, providing comprehensive support for intelligent O&M of the computer room.

[0358] Extendibility: The system design is highly adaptable and can be extended to other intelligent management scenarios of power equipment, further expanding its application value.

[0359] This disclosure not only improves the accuracy and stability of computer room temperature control, but also provides innovative solutions for energy saving and life management of air-conditioning systems. At the same time, through data-driven and intelligent model optimization, it sets an advanced example in the field of smart operation and maintenance.

[0360] The third embodiment of the present disclosure also provides a computer room temperature control system based on data drive and AI analysis, such as Figure 9 As shown, the system includes:

[0361] A collection module 11 is configured to collect historical data corresponding to the computer room within a preset time period, wherein the historical data includes computer room environment data, air conditioning operation data, and external environment data;

[0362] The control module 12 is configured to organize and analyze the collected historical data, predict the temperature change trend in the computer room through a preset algorithm model, and generate a temperature control plan corresponding to the computer room;

[0363] The acquisition module 11 is also configured to collect real-time temperature data of the computer room;

[0364] A generating module 13, which is configured to generate a temperature control instruction based on the real-time temperature data of the computer room and the temperature control plan corresponding to the computer room;

[0365] The adjustment module 14 is configured to automatically adjust the operating parameters of the air conditioner according to the temperature control instruction to ensure that the temperature in the computer room is maintained within a preset range.

[0366] Furthermore, the control module 12 is further configured to:

[0367] Based on historical overheating and overcooling data and outdoor weather, and using a preset algorithm model, possible overheating and overcooling failures are predicted, warning signals are issued in advance, and guidance is provided on setting temperature control plans.

[0368] Further,

[0369] The computer room environmental data includes: indoor temperature, humidity, return air temperature, and supply air temperature;

[0370] The air conditioning operation data includes: air conditioning operation mode, power on / off status, air supply temperature, and air supply speed;

[0371] The external environment data includes: outdoor temperature, humidity, and weather conditions.

[0372] Furthermore, the control module 12 is specifically configured to:

[0373] If the room temperature or the air conditioner return air temperature exceeds the threshold value a°C, the air conditioner supply air temperature will be automatically adjusted to the set value X°C. If this adjustment process is triggered N times continuously and the room temperature continues to be higher than the threshold value a°C, control will be stopped and an air conditioner performance degradation notice will be issued.

[0374] If the room temperature or the air conditioner return air temperature is within the threshold range [b°C, a°C], the air conditioner supply air temperature is adjusted downwards to (x1°C, x2°C, x3°C, ...). If the room temperature remains within the threshold range after M consecutive adjustments, the air conditioner supply air temperature is set to the absolute threshold value Y°C, control is stopped, and an air conditioner performance degradation notice is issued.

[0375] Among them, a>b>x1>x2>x3≥X, x3>Y, a, b, X, x1, x2, x3, Y, N, M are confirmed based on the historical data of the computer room and the predicted results of temperature change trend.

[0376] Furthermore, the control module 12 is specifically configured to:

[0377] When the ambient temperature is higher than the threshold value c℃, the system will turn on all air conditioners;

[0378] When the ambient temperature is not higher than the threshold value c°C, the energy-saving mode is activated. The energy-saving mode includes:

[0379] When the ambient temperature is within the threshold value [d℃, c℃] and the difference δ between the current temperature value and the last collected value is greater than a predetermined value, the system will determine the current number of air conditioners. If the number of air conditioners is less than the threshold value T, K air conditioners will be turned off. If the number of air conditioners is greater than the threshold value W, L air conditioners will be turned off.

[0380] When the ambient temperature is lower than the threshold value d℃, all air conditioners are turned off;

[0381] Among them, b>c>d, L≥K, c, d, K, W, L are confirmed based on the historical data of the computer room and the prediction results of the temperature change trend.

[0382] Furthermore, the temperature control scheme also includes:

[0383] When the number of times the air conditioner is turned on and off exceeds the threshold value S within the preset time range, the energy-saving mode is turned off.

[0384] Furthermore, the system further includes a feedback module 15;

[0385] The feedback module 15 is configured to collect performance indicators related to the operation of the air-conditioning equipment, and to determine whether the air-conditioning equipment has executed the instructions as expected by monitoring and analyzing the collected performance indicators in real time, combined with the issuance of control instructions and state changes.

[0386] The computer room temperature control system based on data drive and AI analysis in the embodiment of the present disclosure is used to implement the computer room temperature control method based on data drive and AI analysis in method embodiment 1, so the description is relatively simple. For details, please refer to the relevant description in the previous method embodiment, which will not be repeated here.

[0387] In addition, if Figure 10 As shown, the fourth embodiment of the present disclosure further provides an electronic device, including a memory 100 and a processor 200, wherein the memory 100 stores a computer program. When the processor 200 runs the computer program stored in the memory 100, the processor 200 executes the above-mentioned various possible methods.

[0388] The memory 100 is connected to the processor 200 . The memory 100 may be a flash memory, a read-only memory, or other memory. The processor 200 may be a central processing unit or a single-chip microcomputer.

[0389] In addition, an embodiment of the present disclosure further provides a computer-readable storage medium, on which a computer program is stored, and the computer program is used by a processor to execute the above-mentioned various possible methods.

[0390] The computer-readable storage medium includes volatile or nonvolatile, removable or non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, computer program modules or other data). Computer-readable storage media include, but are not limited to, RAM (Random Access Memory), ROM (Read-Only Memory), EEPROM (Electrically Erasable Programmable read only memory), flash memory or other memory technology, CD-ROM (Compact Disc Read-Only Memory), Digital Versatile Disk (DVD) or other optical disk storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store the desired information and can be accessed by a computer.

[0391] It is understood that the above embodiments are merely exemplary embodiments for illustrating the principles of the present disclosure, and the present disclosure is not limited thereto. Those skilled in the art may make various modifications and improvements without departing from the spirit and substance of the present disclosure, and such modifications and improvements are also considered to be within the scope of protection of the present disclosure.

Claims

1. A computer room temperature control method based on data drive and AI analysis, characterized in that: The method comprises: Collect historical data corresponding to the computer room within a preset time period, the historical data including computer room environment data, air conditioning operation data and external environment data; Organize and analyze the collected historical data, predict the temperature change trend in the computer room through the preset algorithm model, and generate the corresponding temperature control plan for the computer room; Collect real-time temperature data of the computer room; Generate temperature control instructions based on the real-time temperature data of the computer room and the corresponding temperature control plan of the computer room; Automatically adjust the air conditioning operating parameters according to temperature control instructions to ensure that the temperature in the computer room is maintained within the preset range.

2. The method according to claim 1, characterized in that The method further comprises: Based on historical overheating and overcooling data and outdoor weather, and using a preset algorithm model, possible overheating and overcooling failures are predicted, warning signals are issued in advance, and guidance is provided on setting temperature control plans.

3. The method according to claim 1, characterized in that The computer room environmental data includes: indoor temperature, humidity, return air temperature, and supply air temperature; The air conditioning operation data includes: air conditioning operation mode, power on / off status, air supply temperature, and air supply speed; The external environment data includes: outdoor temperature, humidity, and weather conditions.

4. The method according to claim 1, wherein The method of predicting the temperature change trend in the computer room and generating a temperature control plan corresponding to the computer room includes: If the room temperature or the air conditioner return air temperature exceeds the threshold value a°C, the air conditioner supply air temperature will be automatically adjusted to the set value X°C. If this adjustment process is triggered N times continuously and the room temperature continues to be higher than the threshold value a°C, control will be stopped and an air conditioner performance degradation notice will be issued. If the room temperature or the air conditioner return air temperature is within the threshold range [b°C, a°C], the air conditioner supply air temperature is adjusted downwards to (x1°C, x2°C, x3°C, ...). If the room temperature remains within the threshold range after M consecutive adjustments, the air conditioner supply air temperature is set to the absolute threshold value Y°C, control is stopped, and an air conditioner performance degradation notice is issued. Among them, a>b>x1>x2>x3≥X, x3>Y, a, b, X, x1, x2, x3, Y, N, M are confirmed based on the historical data of the computer room and the predicted results of temperature change trend.

5. The method according to claim 4, characterized in that The temperature control scheme also includes: When the ambient temperature is higher than the threshold value c℃, the system will turn on all air conditioners; When the ambient temperature is not higher than the threshold value c°C, the energy-saving mode is activated. The energy-saving mode includes: When the ambient temperature is within the threshold value [d℃, c℃] and the difference δ between the current temperature value and the last collected value is greater than a predetermined value, the system will determine the current number of air conditioners. If the number of air conditioners is less than the threshold value T, K air conditioners will be turned off. If the number of air conditioners is greater than the threshold value W, L air conditioners will be turned off. When the ambient temperature is lower than the threshold value d℃, all air conditioners are turned off; Among them, b>c>d, L≥K, c, d, K, W, L are confirmed based on the historical data of the computer room and the prediction results of the temperature change trend.

6. The method according to claim 5, characterized in that The temperature control scheme also includes: When the number of times the air conditioner is turned on and off exceeds the threshold value S within the preset time range, the energy-saving mode is turned off.

7. The method according to claim 1, characterized in that The method further comprises: Collect performance indicators related to the operation of air-conditioning equipment, conduct real-time monitoring and analysis of the collected performance indicators, and combine the issuance of control instructions and status changes to determine whether the air-conditioning equipment has executed the instructions as expected.

8. A computer room temperature control system based on data drive and AI analysis, characterized in that: The system comprises: A collection module configured to collect historical data corresponding to the computer room within a preset time period, wherein the historical data includes computer room environment data, air conditioning operation data, and external environment data; The control module is configured to organize and analyze the collected historical data, predict the temperature change trend in the computer room through a preset algorithm model, and generate a corresponding temperature control plan for the computer room; The acquisition module is further configured to collect real-time temperature data of the computer room; A generation module configured to generate a temperature control instruction based on the real-time temperature data of the computer room and a temperature control plan corresponding to the computer room; The adjustment module is configured to automatically adjust the operating parameters of the air conditioner according to the temperature control instruction to ensure that the temperature in the computer room is maintained within a preset range.

9. An electronic device, characterized in that: It includes a memory and a processor, wherein a computer program is stored in the memory. When the processor runs the computer program stored in the memory, the processor executes the computer room temperature control method based on data drive and AI analysis as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, which, when executed by a processor, implements the computer room temperature control method based on data drive and AI analysis according to any one of claims 1 to 7.

Images