Data center energy saving system and method

Abstract

The invention discloses a data center energy saving system and method, and relates to the technical field of energy consumption saving.The system comprises a temperature prediction module used for predicting temperature distribution of a data center based on computational fluid dynamics simulation data and wireless sensor data; the active cooling control module is used for adjusting operation parameters of a cooling system and a fan in the data center according to the temperature prediction result of the temperature prediction module; the intelligent sensor monitoring module is used for collecting and monitoring operation data of the data center through an optical sensor, an acoustic sensor and a thermal sensor; the operation data are computational fluid dynamics simulation data and wireless sensor data; and the energy consumption evaluation and optimization module is used for evaluating the energy consumption of the cooling system in the data center in real time. The overall energy efficiency of the data center is improved, and the operation reliability is enhanced.

Description

Technical Field

[0001] The present application relates to the field of energy conservation technology, and in particular to a data center energy conservation system and method. Background Art

[0002] With the development of cloud computing and big data technology, data centers have become an important infrastructure in modern society. However, the energy consumption problem of data centers is becoming increasingly serious, and the power consumption of their cooling systems and fans accounts for a large proportion of the total energy consumption. Currently, most cooling systems use conservative static settings, resulting in energy waste. Summary of the invention

[0003] The purpose of this application is to provide a data center energy-saving system and method that integrates temperature prediction, active cooling control, and intelligent sensor monitoring. Through accurate prediction, dynamic optimization, and intelligent monitoring, it can effectively reduce the energy consumption of the data center, improve operating efficiency, and reduce the risk of equipment failure.

[0004] To achieve the above objectives, this application provides the following solutions: In a first aspect, the present application provides a data center energy saving system, comprising: The temperature prediction module is used to predict the temperature distribution of the data center based on computational fluid dynamics simulation data and wireless sensor data.

[0005] The active cooling control module is used to adjust the operating parameters of the cooling system and fans in the data center according to the temperature prediction results of the temperature prediction module.

[0006] An intelligent sensor monitoring module is used to collect and monitor the operation data of a data center through optical sensors, acoustic sensors and thermal sensors; the operation data are computational fluid dynamics simulation data and wireless sensor data; the intelligent sensor monitoring module includes: a heterogeneous sensor array; the heterogeneous sensor array is deployed in optical, acoustic and thermal sensor nodes of a cabinet / channel; a multimodal data fusion unit is used to use multimodal data fusion technology to fuse optical, acoustic and thermal data; a self-organizing wireless network is used to use a TDMA protocol to achieve low-power communication between sensor nodes.

[0007] Energy consumption assessment and optimization module, used to evaluate the energy consumption of cooling systems in data centers in real time.

[0008] Optionally, the temperature prediction module includes: CFD simulation unit, used to build a three-dimensional thermodynamic model based on the server layout and air flow path of the data center.

[0009] The sensor calibration unit is used to correct the output of the three-dimensional thermodynamic model through the wireless sensor data collected by the wireless sensor network; the wireless sensor data includes real-time temperature data and real-time wind speed data.

[0010] The multi-mode prediction model unit is used to establish a dedicated prediction model for normal operation and air conditioning failure / overload abnormal conditions.

[0011] Optionally, the CFD simulation unit automatically starts model parameter adjustment when the deviation between the measured temperature and the predicted value exceeds 0.5°C.

[0012] Optionally, the multi-mode prediction model unit includes a time series prediction subunit based on an LSTM neural network; the time series prediction subunit is used to integrate an abnormal operating condition simulation training data set; the abnormal operating condition simulation training data set is generated by CFD simulating temperature changes under air conditioning failure conditions.

[0013] Optionally, the active cooling control module includes: Dynamic fan control system that adjusts server fan speed based on CPU utilization and predicted inlet temperature.

[0014] CRAC optimization controller to dynamically adjust air supply rate and temperature set point based on temperature prediction results.

[0015] The partition control unit is used to divide the data center into multiple independent temperature control areas and implement differentiated control strategies.

[0016] Optionally, the dynamic fan control system is used to adjust the speed of the server fan based on a current temperature measurement value and a temperature prediction value at a set time according to a feedforward-feedback composite control strategy.

[0017] Optionally, the energy consumption assessment and optimization module includes: Real-time energy efficiency analysis unit, used to calculate the COP index of the cooling system and establish the energy consumption heat map.

[0018] Dynamic optimization suggestion unit, used to generate control parameter adjustment schemes based on reinforcement learning algorithms.

[0019] Long-term data storage unit, used to store historical operation data.

[0020] In a second aspect, the present application provides an energy saving method based on the data center energy saving system, including: Collect data center operation data based on intelligent sensor monitoring module.

[0021] Based on the operating data, the temperature changes in the data center within the set time are predicted to obtain the temperature prediction results.

[0022] Adjust the parameter settings of the cooling system based on the temperature prediction results.

[0023] Evaluate cooling system energy consumption in real time and optimize cooling strategies based on the evaluation results.

[0024] When an abnormality in the cooling system is detected, switch to the emergency control strategy.

[0025] Optionally, when an abnormality in the cooling system is detected, switching to an emergency control strategy specifically includes: Utilize multimodal sensor fusion data to identify abnormal conditions in the cooling system; the abnormal conditions include air conditioning failure or airflow obstruction.

[0026] When an abnormal situation occurs in the cooling system, the backup cooling equipment is activated or the operating power consumption of the high heat load area is preferentially reduced.

[0027] According to the specific embodiments provided in this application, this application discloses the following technical effects: The present application provides a data center energy-saving system and method, which includes a temperature prediction module for predicting the temperature distribution of the data center based on computational fluid dynamics simulation data and wireless sensor data; an active cooling control module for adjusting the operating parameters of the cooling system and fans in the data center according to the temperature prediction results of the temperature prediction module; an intelligent sensor monitoring module for collecting and monitoring the operating data of the data center through optical sensors, acoustic sensors and thermal sensors; the operating data are computational fluid dynamics simulation data and wireless sensor data; an energy consumption evaluation and optimization module for real-time evaluation of the energy consumption of the cooling system in the data center. The temperature prediction module of the present application combines CFD and real-time sensor data to improve the prediction accuracy and avoid the insufficient coverage and delay problems of traditional methods. The active cooling control module is prediction-driven and can adjust parameters in advance to reduce energy waste and prevent overheating. The intelligent sensor monitoring module uses a variety of sensors and non-invasive installation to reduce costs and improve reliability. Real-time monitoring and long-term analysis of the energy consumption evaluation module can help continuously optimize energy use. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0029] Figure 1 A schematic diagram of functional modules of a data center energy saving system provided in one embodiment of the present application.

[0030] Figure 2 A flowchart of a data center energy saving method provided in one embodiment of the present application.

[0031] Figure 3 A flowchart of a data center energy saving method that integrates temperature prediction, active cooling control, and intelligent sensor monitoring is provided in accordance with an embodiment of the present application. DETAILED DESCRIPTION

[0032] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0033] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0034] Embodiment 1 like Figure 1 As shown, this embodiment provides a data center energy saving system, including: The temperature prediction module is used to predict the temperature distribution of the data center based on computational fluid dynamics simulation data and wireless sensor data.

[0035] The active cooling control module is used to adjust the operating parameters of the cooling system and fans in the data center according to the temperature prediction results of the temperature prediction module.

[0036] An intelligent sensor monitoring module is used to collect and monitor the operation data of a data center through optical sensors, acoustic sensors and thermal sensors; the operation data are computational fluid dynamics simulation data and wireless sensor data; the intelligent sensor monitoring module includes: a heterogeneous sensor array; the heterogeneous sensor array is deployed in optical, acoustic and thermal sensor nodes of a cabinet / channel; a multimodal data fusion unit is used to use multimodal data fusion technology to fuse optical, acoustic and thermal data; a self-organizing wireless network is used to use a TDMA protocol to achieve low-power communication between sensor nodes.

[0037] Energy consumption assessment and optimization module, used to evaluate the energy consumption of cooling systems in data centers in real time.

[0038] The temperature prediction module comprises: The CFD simulation unit is used to establish a three-dimensional thermodynamic model according to the server layout and air flow path of the data center; the CFD simulation unit automatically starts model parameter adjustment when the deviation between the measured temperature and the predicted value exceeds 0.5°C.

[0039] The sensor calibration unit is used to correct the output of the three-dimensional thermodynamic model through the wireless sensor data collected by the wireless sensor network; the wireless sensor data includes real-time temperature data and real-time wind speed data.

[0040] The multi-mode prediction model unit is used to establish a dedicated prediction model for normal operation and air conditioning failure / overload abnormal conditions.

[0041] Among them, in some embodiments, temperature prediction by the temperature prediction module is the core link to achieve energy saving in data centers. The temperature prediction module of this embodiment solves the problems of low accuracy and limited coverage in traditional temperature monitoring methods by combining computational fluid dynamics (CFD) simulation with wireless sensor networks. The specific implementation includes the following aspects: 1) CFD simulation and calibration: Through computational fluid dynamics (CFD) simulation, the present embodiment can generate a temperature distribution map of the data center. During the simulation process, the present embodiment comprehensively considers the physical characteristics of the data center, such as the layout of the servers and the air flow path, and the operating conditions, such as the load of the servers.

[0042] This embodiment uses wireless sensors deployed inside the data center to collect temperature and air flow rate data in real time, and these data are used to calibrate the prediction results of the CFD model to ensure higher accuracy.

[0043] 2) Multi-modal prediction model unit: In response to the needs of the data center under different operating conditions (including normal operation and abnormal conditions, such as air conditioning failure or overload, etc.), this embodiment constructs a variety of prediction models. The multi-mode prediction model unit includes a time series prediction subunit based on an LSTM neural network; the time series prediction subunit is used to integrate abnormal working condition simulation training data sets; the abnormal working condition simulation training data sets are generated by CFD simulation of temperature changes under air conditioning failure conditions. Under abnormal conditions, CFD simulation can provide temperature change data under specific conditions, which are used to train the prediction model to ensure that the model can cover more potential operating scenarios.

[0044] 3) Efficient prediction and response: The system of this embodiment supports the prediction of temperature changes within 10 minutes, which provides sufficient time for the adjustment of cooling system parameters or timely intervention by the administrator. The accuracy of temperature prediction is controlled within 0.5°C, effectively avoiding overcooling or overheating problems caused by prediction errors.

[0045] In other words, the temperature prediction module achieves high-precision prediction of the global temperature distribution in the data center, providing crucial data support for the subsequent optimization of the cooling system.

[0046] In some embodiments, the active cooling control module may be specifically as follows: The active cooling control module of this embodiment aims to minimize cooling energy consumption while ensuring that the thermal safety requirements of the equipment are met by dynamically adjusting the operating parameters of the cooling system and the fan based on the temperature prediction data.

[0047] 1) Dynamic fan control system: The server fan speed is dynamically adjusted according to the CPU utilization and inlet temperature. The prediction algorithm is used to pre-set the fan operating parameters according to the temperature change trend in the next few minutes, thereby avoiding the response lag problem in traditional reactive control. Specifically, according to the feedforward-feedback composite control strategy, the server fan speed is adjusted based on the current temperature measurement value and the temperature prediction value at the set time.

[0048] 2) CRAC Optimization Controller: Dynamically adjust air flow rate and temperature set point based on temperature prediction results; optimize computer room air conditioning (CRAC) air flow rate and temperature set point to ensure that equipment cooling needs are met while minimizing energy consumption. Increase the flexibility of CRAC system temperature set point and avoid energy waste caused by set point being too low through accurate temperature prediction.

[0049] 3) Partition control unit: The CRAC system is divided into multiple independent zones through a zoning algorithm, and the control parameters of each zone are optimized separately according to its heat load characteristics.

[0050] The partitioned control approach significantly reduces the modeling and computational difficulty in dealing with complex thermodynamic characteristics in large-scale data centers.

[0051] The notable feature of the active cooling control module is its proactive predictive drive, which enables the system to take preventive measures before temperature changes occur, effectively avoiding thermal emergencies and significantly improving cooling efficiency.

[0052] In some embodiments, the intelligent sensor monitoring module is a set of intelligent sensor monitoring modules designed to support efficient operation of temperature prediction and cooling control. The module uses a multimodal sensor network to achieve real-time monitoring of the operating status of the data center through the combination of light, sound and thermal data. The specifics can be as follows: 1) Sensor deployment: In this system, a heterogeneous sensor array is deployed, which is deployed in the optical, acoustic, and thermal sensor nodes of the cabinet / channel. These sensors are installed in a non-invasive way, which means that their deployment does not require complex power transformation. In addition, the wireless sensor network is connected through a self-organizing mechanism, which greatly reduces the installation and maintenance costs of the system.

[0053] 2) Multimodal data fusion: The multimodal data fusion unit is used to use multimodal data fusion technology to fuse optical, acoustic and thermal data. By fusing multimodal data from different sensors, it effectively filters out environmental noise and false alarms, so that target events can be accurately extracted, such as changes in the operating status of the equipment.

[0054] This embodiment uses a lightweight algorithm to perform preliminary processing on the data at the local sensor node, which can significantly reduce the amount of data that needs to be transmitted, thereby extending the battery life of the sensor.

[0055] 3) Event detection and classification: This system uses unsupervised learning algorithms to perform cluster analysis on events captured by sensors and associate these events with the operating status of specific equipment.

[0056] The intelligent sensor monitoring module also provides a fast deployment and self-organizing wireless sensor network, using TDMA protocol to achieve low-power communication between sensor nodes.

[0057] Through automated event classification and device identification, the system is able to identify potential failure risks, which provides additional decision support for cooling control.

[0058] The flexibility and low-cost characteristics of the intelligent sensor monitoring module make it an indispensable part of this embodiment, and it can quickly adapt to data center environments of different sizes.

[0059] In some embodiments, the energy consumption assessment and optimization module is a key tool for real-time monitoring and dynamic adjustment of energy-saving effects. The specifics are as follows: 1) Real-time energy efficiency analysis unit: The energy consumption data of the cooling system (including CRAC and fans) is collected through sensors, and the energy consumption proportion of each component is recorded in real time.

[0060] Combined with the temperature prediction results, the energy utilization efficiency under the current operating state is evaluated.

[0061] 2) Dynamic optimization suggestion unit: Provides specific optimization recommendations based on the operating status of the cooling system and fans, such as increasing the CRAC temperature set point in a certain area or reducing the speed of a specific server fan.

[0062] When an abnormal situation occurs, the system can quickly switch to an emergency control strategy to prevent equipment overheating or cooling failure.

[0063] 3) Long-term data storage unit: The system can store and analyze long-term operating data, identify bottleneck areas of energy utilization, and further optimize system design and operating parameters.

[0064] Through the energy consumption assessment and optimization module, the system realizes real-time assessment of energy-saving effects and provides a basis for continuous improvement.

[0065] Embodiment 2 like Figure 2 As shown, this embodiment provides an energy saving method based on the data center energy saving system, including: Step 201: Collecting operation data of the data center based on the intelligent sensor monitoring module.

[0066] Step 202: Based on the operation data, predict the temperature change of the data center within a set time to obtain a temperature prediction result.

[0067] Step 203: Adjust the parameter settings of the cooling system according to the temperature prediction result.

[0068] Step 204: Evaluate the energy consumption of the cooling system in real time, and optimize the cooling strategy according to the evaluation result.

[0069] Step 205: When an abnormality in the cooling system is detected, switch to an emergency control strategy.

[0070] When an abnormality in the cooling system is detected, the emergency control strategy is switched to, specifically including: Utilize multimodal sensor fusion data to identify abnormal conditions in the cooling system; the abnormal conditions include air conditioning failure or airflow obstruction.

[0071] When an abnormal situation occurs in the cooling system, the backup cooling equipment is activated or the operating power consumption of the high heat load area is preferentially reduced.

[0072] like Figure 3As shown in FIG. 1 , a data center energy saving method that integrates temperature prediction, active cooling control, and intelligent sensor monitoring functions includes: Sensor selection and placement; sensor types include optical, acoustic, and thermal.

[0073] Determine the optimal installation location of the sensor based on the physical layout and heat load distribution of the data center.

[0074] The sensors are installed non-intrusively using wireless sensor nodes placed close to the equipment or in critical airflow paths.

[0075] Constructing a wireless sensor network; the configuration method involves: starting the sensor nodes, establishing communication links through the self-organizing network, and ensuring signal coverage between nodes to reduce data loss and communication delay.

[0076] Initialize the system and calibrate the sensors to ensure the accuracy of optical, acoustic, and thermal data. Configure the sampling frequency and data transmission rate to balance monitoring accuracy and energy consumption.

[0077] Specifically, the sensor data collected includes: real-time temperature, air flow rate, equipment operation status and other information. When collecting sensor data, the sensor node performs preliminary processing on the collected data (such as noise reduction and anomaly detection) to reduce the amount of transmitted data. Then, a multimodal fusion algorithm is used to associate optical, acoustic and thermal data to eliminate environmental noise and false alarms. The cleaned data is transmitted to the central processing unit (CPU) for further analysis.

[0078] Among them, the computational fluid dynamics (CFD) simulation and sensor data integration, the specific steps are as follows: First, CFD is used to simulate the temperature distribution inside the data center. This process requires input of key information such as equipment layout, heat source power consumption, and cooling system settings. Through simulation, this embodiment can obtain a temperature distribution map inside the data center, showing in detail the temperature conditions of different areas (such as rack entrances, exits, and aisles), thereby identifying the location and range of high-temperature hot spots. This step helps this embodiment identify hot spots and optimize the parameter settings of the cooling system (such as CRAC set points and fan speeds) to reduce energy waste.

[0079] Secondly, the real-time data collected by sensors (such as air velocity and temperature) is used to dynamically calibrate the CFD results to improve the prediction accuracy. Dynamic calibration is to continuously adjust the parameters and models in the CFD (computational fluid dynamics) simulation through real-time sensor data to ensure that the output results of the CFD simulation can more accurately reflect the actual operating status of the data center.

[0080] After completing the data integration work, the next step is the calibration process: Data collection: Real-time acquisition of sensor data (including equipment inlet and outlet temperatures, ambient temperature, wind speed in the cooling airflow path, equipment ID, server load, etc.), as well as preliminary output from CFD simulation (temperature distribution and airflow characteristics).

[0081] Data Verification: Compare temperature data at key points (server inlet and outlet temperatures), as well as speed and direction in the airflow path.

[0082] Parameter optimization: Based on the verification results, adjust the key input parameters of the CFD model (including heat source power consumption, boundary conditions, fluid properties, and geometric parameter correction) to make the simulation results closer to the actual situation.

[0083] Model retraining and calibration: Use sensor data to train and calibrate models to compensate for systematic deviations in CFD simulations and perform iterative optimization.

[0084] After outputting the calibrated simulation results, the specific steps are as follows: For different operating states (including normal, faulty, overloaded, etc.), the pre-trained model is used to predict temperature changes within the next 10 minutes. The predicted temperature distribution map and potential overheating risk points of key equipment are output. (The pre-trained model here refers to the data center temperature distribution model obtained through CFD simulation output.) The predicted results are compared with the data collected by the sensor in real time, and the model parameters are fine-tuned to ensure the accuracy and robustness of the prediction.

[0085] Implement dynamic fan speed adjustment; adjust the speed of server fans according to the predicted temperature change trend. Fan adjustment strategies include: prioritize reducing fan speeds in areas with lower operating loads to reduce fan energy consumption. Adjust the air supply rate and temperature set point of the Computer Room Air Condition (CRAC); use a partition optimization algorithm to divide the CRAC system into multiple zones, and each zone independently optimizes its cooling strategy; send the optimized control parameters to the fan and CRAC system to adjust their operating status in real time to ensure a smooth transition during the adjustment process and avoid system instability due to sudden changes. Collect energy consumption data of fans and CRACs through sensors, record their operating status and energy consumption share, analyze energy consumption trends, and evaluate the effectiveness of adjustment strategies.

[0086] Based on the energy consumption monitoring results, possible bottleneck areas are identified, such as fan speeds set too high or CRAC set points too low. Optimization algorithm parameters are adjusted to generate new cooling control strategies.

[0087] Anomaly detection of the cooling system; anomaly detection methods include: using multimodal sensor fusion data to identify abnormal conditions in the cooling system, such as air conditioning failure or airflow blockage; combining with temperature prediction models to locate potential overheating areas and issue timely alarms.

[0088] Conduct emergency response to abnormal cooling system conditions that have been alerted; emergency response methods include: starting backup cooling equipment, or prioritizing reducing operating power consumption in high heat load areas; triggering administrator notification mechanisms to prompt data center operators to perform manual intervention.

[0089] Store real-time collection and prediction data in a central database and analyze historical data regularly. Identify long-term energy consumption trends and cooling system performance issues to provide a basis for subsequent optimization. Use long-term data to retrain temperature prediction models and cooling optimization algorithms to improve their adaptability and accuracy; regularly update partition optimization strategies to adapt to dynamic changes in the data center.

[0090] In summary, this application has the following technical effects: 1) High efficiency and energy saving: Through active prediction and control, the energy consumption of the cooling system can be reduced by more than 30%.

[0091] 2) High precision and responsiveness: The temperature prediction error is controlled within 0.5°C, which can quickly respond to sudden abnormal conditions.

[0092] 3) Low-cost deployment: Smart sensor systems can be quickly deployed by non-professionals, significantly reducing the need for high-cost professional installation.

[0093] 4) Wide adaptability: Applicable to data centers of various sizes and types.

[0094] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0095] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. At the same time, for those skilled in the art, according to the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A data center energy saving system, characterized in that: The data center energy saving system comprises: A temperature prediction module is used to predict the temperature distribution of the data center based on computational fluid dynamics simulation data and wireless sensor data; An active cooling control module, used to adjust the operating parameters of the cooling system and fans in the data center according to the temperature prediction results of the temperature prediction module; An intelligent sensor monitoring module is used to collect and monitor the operation data of the data center through optical sensors, acoustic sensors and thermal sensors; the operation data is computational fluid dynamics simulation data and wireless sensor data; the intelligent sensor monitoring module includes: a heterogeneous sensor array; the heterogeneous sensor array is deployed in the optical, acoustic and thermal sensor nodes of the cabinet / channel; a multimodal data fusion unit is used to use multimodal data fusion technology to fuse optical, acoustic and thermal data; a self-organizing wireless network is used to use the TDMA protocol to achieve low-power communication between sensor nodes; Energy consumption assessment and optimization module, used to evaluate the energy consumption of cooling systems in data centers in real time.

2. A data center energy saving system according to claim 1, characterized in that: The temperature prediction module comprises: CFD simulation unit, used to build a three-dimensional thermodynamic model based on the server layout and air flow path of the data center; A sensor calibration unit, used to correct the output of the three-dimensional thermodynamic model through wireless sensor data collected by the wireless sensor network; the wireless sensor data includes real-time temperature data and real-time wind speed data; The multi-mode prediction model unit is used to establish a dedicated prediction model for normal operation and air conditioning failure / overload abnormal conditions.

3. A data center energy saving system according to claim 2, characterized in that: The CFD simulation unit automatically starts model parameter adjustment when the deviation between the measured temperature and the predicted value exceeds 0.5°C.

4. A data center energy saving system according to claim 3, characterized in that: The multi-mode prediction model unit includes a time series prediction subunit based on an LSTM neural network; the time series prediction subunit is used to integrate an abnormal operating condition simulation training data set; the abnormal operating condition simulation training data set is generated by CFD simulating temperature changes under air conditioning failure conditions.

5. A data center energy saving system according to claim 4, characterized in that: The active cooling control module includes: Dynamic fan control system to adjust server fan speed based on CPU utilization and predicted inlet temperature; CRAC optimization controller to dynamically adjust air supply rate and temperature set point based on temperature prediction results; The partition control unit is used to divide the data center into multiple independent temperature control areas and implement differentiated control strategies.

6. A data center energy saving system according to claim 5, characterized in that: The dynamic fan control system is used to adjust the speed of the server fan based on the current temperature measurement value and the temperature prediction value at a set time according to a feedforward-feedback composite control strategy.

7. A data center energy saving system according to claim 1, characterized in that: The energy consumption assessment and optimization module includes: Real-time energy efficiency analysis unit, used to calculate the COP index of the cooling system and establish the energy consumption heat map; Dynamic optimization suggestion unit, used to generate control parameter adjustment schemes based on reinforcement learning algorithms; Long-term data storage unit, used to store historical operation data.

8. An energy saving method for a data center energy saving system according to any one of claims 1 to 7, characterized in that: include: Collect data center operation data based on intelligent sensor monitoring module; Based on the operating data, the temperature change of the data center within the set time is predicted to obtain the temperature prediction result; Adjust the parameter settings of the cooling system according to the temperature prediction results; Evaluate the energy consumption of the cooling system in real time and optimize the cooling strategy based on the evaluation results; When an abnormality in the cooling system is detected, switch to the emergency control strategy.

9. The energy saving method according to claim 8, characterized in that: When an abnormality in the cooling system is detected, switching to an emergency control strategy specifically includes: Utilizing multimodal sensor fusion data to identify abnormal conditions in the cooling system; the abnormal conditions include air conditioning failure or airflow obstruction; When an abnormal situation occurs in the cooling system, the backup cooling equipment is activated or the operating power consumption of the high heat load area is preferentially reduced.

Images