Correlation optimization energy-saving method and system based on calculation and thermal management energy consumption and related device

Abstract

The invention provides a correlation optimization energy-saving method and system based on calculation and thermal management energy consumption and a related device, and the method comprises the following steps: inputting the operation data of collection equipment in a current time period and under a load condition into a multi-source information fusion processing unit, and carrying out the cascade optimization and joint deduction, and outputting to obtain a thermal management system regulation and control parameter set; the temperature and humidity set value of the machine room and the angle of an air supply guide blade are adjusted through the thermal management system regulation and control parameter set; according to the method, the thermal management system regulation and control parameter set is constructed and the machine room air treatment device is directly driven to execute dynamic linkage regulation, so that quick response and accurate execution of the control instruction are realized, closed-loop intelligent control from sensing, decision making to execution is formed, the real-time performance and stability of the system are improved, and the real-time performance of the system is improved. And the service life of equipment is effectively prolonged.

Description

Technical Field

[0001] This invention belongs to the field of data centers, and specifically relates to a method, system and related device for energy saving based on the correlation optimization of computing and thermal management energy consumption. Background Technology

[0002] With the rapid development of global information technology, data centers, as key facilities supporting applications such as cloud computing, big data analytics, and artificial intelligence, are increasingly facing challenges in energy consumption. Data center energy consumption primarily stems from equipment and thermal management systems, and temperature and humidity control is not only a crucial factor affecting energy consumption but also impacts the normal operation of the server room. Equipment generates significant heat during task processing, which must be dissipated through a thermal management system to ensure its proper functioning. However, existing temperature and humidity control solutions have some shortcomings: (1) Temperature and humidity settings are usually static values, which are difficult to adapt to dynamic loads and environmental changes; (2) The lack of intelligent control in humidity regulation can easily lead to over-dehumidification or over-humidification, increasing unnecessary energy consumption; (3) Most current solutions fail to comprehensively optimize the coupled effects of temperature and humidity on energy consumption, making it difficult to achieve optimal overall energy efficiency; (4) The influence of airflow organization is often overlooked, and point-to-point cooling cannot be carried out according to the load distribution of the equipment; (5) The control parameters of the data center thermal management system, such as air supply temperature, relative humidity, and wind direction angle, are mutually influential and mutually restrictive, and traditional control strategies cannot take them into account at the same time.

[0003] Therefore, how to optimize the energy efficiency of data centers by precisely adjusting temperature and humidity while ensuring the normal operation of equipment has become a problem that needs to be solved. Summary of the Invention

[0004] The purpose of this invention is to provide a method and system for energy saving based on the correlation optimization of energy consumption through calculation and thermal management, thereby overcoming the shortcomings of the prior art.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows: In a first aspect, the energy-saving method based on the correlation optimization of energy consumption through calculation and thermal management provided by the present invention includes the following steps: The operating data of the acquisition equipment under the current time period and load conditions is input into the multi-source information fusion processing unit. Through cascade optimization and joint simulation, the set of control parameters of the thermal management system is output. The temperature and humidity setpoints of the machine room and the angle of the air supply guide vanes are adjusted using the set of control parameters of the thermal management system.

[0006] Preferably, through cascaded optimization and joint deduction, a set of control parameters for the thermal management system is output, specifically including: Construct a global energy consumption objective function, which integrates both equipment energy consumption and thermal management system energy consumption; A cluster intelligent search algorithm is used to optimize the global energy consumption objective function to obtain the optimal supply air temperature; The optimal supply air temperature and the aforementioned operating data are input into the time-series simulation model, and the set of thermal management system control parameters under the condition of minimum total system energy consumption is output.

[0007] Preferably, the expression for the energy consumption objective function is:

[0008] in, Total energy consumption of all devices; For thermal management systems to provide Performance coefficient when the temperature is cold air.

[0009] Preferably, the time-series simulation model is constructed using LSTM and is used to predict the optimal relative humidity and the optimal air supply direction angle.

[0010] Preferably, the set of control parameters of the thermal management system includes the optimal supply air temperature, the optimal relative humidity, and the optimal supply air guide angle.

[0011] Secondly, the energy-saving system based on the correlation optimization of energy consumption through calculation and thermal management provided by the present invention includes: The multi-source information fusion processing unit is used to input the operating data of the acquisition equipment under the current time period and load conditions, and output the set of control parameters of the thermal management system through cascade optimization and joint deduction. An air handling control unit is used to adjust the temperature and humidity setpoints of the computer room air handling unit and the angle of the air supply guide vanes according to the set of control parameters of the thermal management system.

[0012] Preferably, the multi-source information fusion processing unit includes: An objective function construction component is used to construct a global energy consumption objective function, which integrates both the device's computational energy consumption and the thermal management system's energy consumption. The optimization solution component is used to solve the global energy consumption objective function using a cluster intelligent search algorithm to obtain the optimal supply air temperature under the condition of minimum total system energy consumption; The parameter extrapolation component is used to input the optimal supply air temperature and the operating data into the time series extrapolation model, and output the optimal relative humidity and optimal supply air guide angle under the condition of minimum total system energy consumption. The parameter set generation component is used to construct the optimal supply air temperature, optimal relative humidity, and optimal supply air directional angle into a set of control parameters for the thermal management system.

[0013] Preferably, the time-series inference model used by the parameter inference component is an LSTM network, whose inputs include the spatial distribution of equipment load, air supply temperature, and environmental parameters.

[0014] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method as described.

[0015] Fourthly, the present invention provides an electronic device including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method as described above.

[0016] Compared with the prior art, the beneficial effects of the present invention are: The energy-saving method based on the correlation optimization of computing and thermal management energy consumption provided by this invention adopts a multi-source information fusion processing unit and a cascaded optimization joint deduction mechanism to achieve unified and coordinated control of data center computing energy consumption and thermal management system energy consumption. Specifically: This method constructs a global energy consumption objective function for the integrated device's computational energy consumption and the thermal management system's energy consumption, and uses a cluster intelligent search algorithm for efficient solution. It can accurately locate the optimal supply air temperature when the total system energy consumption is at its lowest, fundamentally breaking the limitation of fixed temperature setpoints in traditional control strategies. This enables the system to dynamically adapt to real-time load and environmental changes, significantly improving the accuracy and adaptability of energy efficiency control.

[0017] Meanwhile, by introducing a time-series extrapolation model, using the optimal supply air temperature as the dominant variable, and collaboratively predicting the optimal relative humidity and supply air guidance angle, the problem of multi-parameter coupled control of temperature, humidity, and airflow organization is effectively solved. This avoids excessive dehumidification or humidification energy consumption caused by improper humidity control. Furthermore, dynamic guided airflow optimizes airflow distribution, enhances targeted heat dissipation efficiency in areas with high local heat loads, and reduces hot and cold air mixing and heat recirculation, thereby further reducing overall system energy consumption while ensuring equipment operational reliability. This application, by constructing a thermal management system control parameter set and directly driving the data center air handling unit to perform dynamic linkage adjustment, achieves rapid response and precise execution of control commands, forming a closed-loop intelligent control from perception, decision-making to execution. This not only improves the system's real-time performance and stability but also effectively extends equipment lifespan, providing a comprehensive, collaborative, and adaptive energy-saving solution for data centers, with significant energy-saving benefits and application value. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of the process involved in an embodiment of the present invention; Figure 2 This is a schematic diagram of the angle adjustment of the air supply guide vanes in the computer room according to an embodiment of the present invention. Detailed Implementation

[0019] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0020] Example 1 This embodiment provides a correlation optimization energy-saving method based on computation and thermal management energy consumption. This method achieves coordinated optimization control of the supply air temperature, relative humidity, and supply air guide angle of the thermal management system driven by multi-source data fusion. Specifically, it includes the following steps: Step 1: Input the operating data of the acquisition equipment under the current time period and load conditions into the multi-source information fusion processing unit. Through cascade optimization and joint simulation, output the set of control parameters for the thermal management system. Specifically: Construct a global energy consumption objective function, which integrates both equipment energy consumption and thermal management system energy consumption; The global energy consumption objective function is optimized. The optimization results and the operating data are input into the time series simulation model, and the set of thermal management system control parameters under the condition of minimum total system energy consumption is output.

[0021] Step 2: Adjust the temperature and humidity setpoints of the machine room and the angle of the air supply guide vanes using the set of control parameters of the thermal management system.

[0022] Example 2 This embodiment provides a unified calculation and correlation optimization energy-saving system based on data center computing energy consumption and thermal management system energy consumption, including: The multi-source information fusion processing unit is used to input the operating data of the acquisition equipment under the current time period and load conditions, and output the set of control parameters of the thermal management system through cascade optimization and joint deduction. An air handling control unit is used to adjust the temperature and humidity setpoints of the computer room air handling unit and the angle of the air supply guide vanes according to the set of control parameters of the thermal management system.

[0023] In this embodiment, the multi-source information fusion processing unit includes: An objective function construction component is used to construct a global energy consumption objective function, which integrates both the device's computational energy consumption and the thermal management system's energy consumption. The optimization solution component is used to solve the global energy consumption objective function using a cluster intelligent search algorithm to obtain the optimal supply air temperature under the condition of minimum total system energy consumption; The parameter extrapolation component is used to input the optimal supply air temperature and the operating data into the time series extrapolation model, and output the optimal relative humidity and optimal supply air guide angle under the condition of minimum total system energy consumption. The parameter set generation component is used to construct the optimal supply air temperature, optimal relative humidity, and optimal supply air directional angle into a set of control parameters for the thermal management system.

[0024] Example 3 This embodiment provides a unified calculation and correlation optimization energy-saving method based on data center computing energy consumption and thermal management system energy consumption. To achieve efficient and coordinated adjustment of multiple control parameters in the thermal management system, such as supply air temperature, relative humidity, and supply air guide angle, this application first performs global optimization of the supply air temperature by constructing a total equipment energy consumption model to obtain the optimal supply air temperature value with minimum system energy consumption. Then, using this optimal supply air temperature as the dominant variable, combined with the current system load and environmental parameters, it predicts and outputs the matching optimal relative humidity and supply air guide angle, thereby realizing dynamic coordinated control of multiple parameters in the thermal management system. Specifically, it includes the following steps: Step 1: Sensor Deployment and Data Acquisition Targets: Total device load data: The device load is measured by the number of user task requests every minute on the device monitoring platform. Equipment load spatial distribution: Number the equipment according to the rack position and build a spatial coordinate index; the load of each piece of equipment corresponds to its physical coordinates, forming a load spatial distribution map; Inlet and outlet air temperatures: High-precision temperature sensors are installed at the front and rear of each cabinet section. The temperature sampling period is 1 minute by default, and the error range is controlled within ±0.3℃. Ambient temperature: Temperature sensors are installed at regular intervals, with a temperature sampling period of 1 minute and an error range controlled within ±0.3℃; Ambient relative humidity: Humidity sensors are installed at regular intervals, with a humidity sampling period of 1 minute and an error range controlled within ±2%RH; Step 2, Data Processing Data integration: All collected raw data is uploaded to the edge controller; each parameter is simultaneously tagged with a timestamp and spatial label for subsequent model feature construction.

[0025] Data synchronization and update frequency: The default data synchronization cycle is 1 minute, which can be configured to 10 seconds to 5 minutes; if connected to a multi-source information fusion processing unit, the inference cycle is... Then the sampling period Should meet ≤ .

[0026] Step 3: Construct a total energy consumption model for the equipment and perform global optimization of the supply air temperature to obtain the optimal supply air temperature value that minimizes system energy consumption. Specifically: (1) Solving for the dominant variable (i.e., supply air temperature): Constructing a device energy consumption model: Device energy consumption includes static power, which is the basic energy consumption generated in the idle state, and dynamic power, which is the additional energy consumption caused by the increased utilization of key components such as CPU and GPU during computing tasks. In addition, the heat generated by the devices will create a heat recirculation effect in the data center, where hot air exhausted by some devices will be reabsorbed by neighboring devices. This phenomenon can be described by a heat recirculation matrix (HRM).

[0027]

[0028] in, This indicates the intake air temperature of the equipment. This indicates the supply air temperature from the thermal management system. It is the identity matrix. Represents the heat reflux matrix. It is the heat exchange coefficient between the equipment and the air. This indicates the device's computational power consumption.

[0029] Data center energy consumption includes equipment energy consumption and thermal management system energy consumption. Assume that at time t, the data center receives a task request from user j. Let Pj(t) represent the power consumed by the device processing user j's task, Lj(t) represent the number of task requests from user j at time t, and mj(t) represent the number of devices providing services to user j. The power consumption of a single device can be expressed as:

[0030] in, It is the marginal power consumption of the device's CPU. It is the device's basic power consumption, including the power consumption of components such as power supply and memory, and is unrelated to the device's workload.

[0031] Therefore, the total energy consumption P(t) of all equipment in the data center can be expressed as:

[0032] Constructing an energy consumption model for the thermal management system: The energy consumption of the thermal management system for cooling depends on its sensible heat load (i.e., the load generated by cooling) and the coefficient of performance (COP) of the thermal management system. Since non-IT equipment generates relatively little heat, the sensible heat load is usually approximated by the total power of IT equipment. Therefore, the cooling power of the thermal management system can be calculated by dividing the sensible heat load by the COP. The COP of the thermal management system is usually a function of the supply air temperature and can be approximated as:

[0033] At this point, the energy consumption of the thermal management system can be expressed as:

[0034] Increase supply air temperature While this can improve COP and reduce the cooling energy consumption C(t) of the thermal management system, it also increases device temperature and power consumption, thus affecting the reliability of task execution. Therefore, the cooling strategy of the thermal management system must find a balance between device energy consumption and the cooling energy consumption of the thermal management system to minimize total energy consumption, while keeping the CPU temperature below a safe threshold and meeting SLA (Service Level Agreement) (QoS) constraints.

[0035] in, Control variables: At time step t, these represent control operations or decisions that adjust the system state, such as temperature setting, humidity setting, refrigeration equipment scheduling, and fan control. State variables: At time step t, these represent the state or performance metrics of the system, such as service quality-related metrics like latency, response time, bandwidth, and task completion time.

[0036] The objective function for the total energy consumption of the system is as follows:

[0037] Therefore, the optimal supply air temperature It is determined by the characteristics of the equipment load and the COP curve. There exists an optimal [condition / condition]. This balances the power consumption of the equipment with the cooling power consumption of the thermal management system. This is achieved by solving... This allows us to obtain the optimal supply air temperature for the data center under the current load conditions:

[0038] Cluster Intelligent Search Algorithm: To solve the problem of calculating the optimal air supply temperature of a data center under current load conditions, in this embodiment, the cluster intelligent search algorithm is a particle swarm optimization (PSO) algorithm. This optimization algorithm iteratively optimizes the positions of particles in the search space, gradually approaching the optimal solution. Specifically: The constructed fitness function is:

[0039] in, For particles The supply air temperature; The current ambient relative humidity (constant); This represents the energy consumption of the equipment at this supply air temperature. Thermal management system to maintain and humidity The cooling energy consumption of the thermal management system.

[0040] The objective is to minimize the fitness function.

[0041] (2) Solving for the response variables (i.e., relative humidity and air supply direction angle): The impact of supply air temperature on the prediction of optimal relative humidity and optimal supply air steering angle: An LSTM model was used to predict optimal relative humidity and supply air steering angle based on time series data. To ensure the accuracy and stability of the time series model, the obtained supply air temperature was used as the dominant variable in the construction of the input variables for the LSTM model. Supply air temperature, relative humidity, and supply air steering angle have a significant coupling relationship in the control process of the thermal management system. Changes in supply air temperature directly affect the cooling and dehumidification effect of the air, as well as the density and flow characteristics of the outlet airflow, thus affecting the system's requirements for humidity and airflow organization. Therefore, from a physical mechanism perspective, supply air temperature is an important prerequisite for humidity and airflow direction adjustment.

[0042] 1) Optimal relative humidity prediction In data center server rooms, it is necessary not only to regulate the temperature to control system energy consumption, but also to control the ambient humidity. Both of these factors will affect system energy consumption. When the air humidity is too high, the thermal management system must consume additional energy for dehumidification to ensure the safe operation of the server room, thereby increasing the cooling energy consumption of the thermal management system.

[0043] Since humidity affects not only the ratio of sensible heat to latent heat but also the cooling energy consumption of fans and thermal management systems, its impact on data center performance is inherently dynamic and complex. Therefore, this application employs a Long Short-Term Memory (LSTM) network to predict the optimal relative humidity value. The LSTM model is trained based on historical operational data from the data center, including variables such as equipment load, power efficiency (PUE), and supply air temperature, to predict the optimal relative humidity value that minimizes energy consumption.

[0044] 2) Prediction of optimal air supply guidance angle To further reduce the cooling energy consumption of the thermal management system, this embodiment introduces an air supply guide angle control mechanism. Adjusting the air supply guide angle directly affects the airflow organization in the computer room and is a crucial factor influencing hot air return and the cooling efficiency of the thermal management system. If the air supply guide angle is unreasonable, cold air may fail to cover high heat load areas, or hot air may stagnate near the equipment, causing the equipment intake temperature to rise, thereby increasing the energy consumption of the thermal management system and the power consumption of the equipment.

[0045] The supply air temperature affects the momentum and velocity of the cold airflow. Cold air with a relatively high temperature has lower kinetic energy due to its smaller pressure and density difference. This results in weaker penetrating power of the cold air and makes it more difficult to penetrate the hot air zone, thus leading to poor cooling effect. Therefore, it is necessary to adjust the air supply guide angle to guide the cold air directly into the high-load area and penetrate the hot air zone.

[0046] Based on the distribution of equipment in the data center, the data center is divided into zones. After receiving a user's task request, the load of each zone is determined according to the equipment call strategy. Areas with high loads need to increase cooling efforts, so more cold air needs to be diverted to areas with high loads. This requires adjusting the air delivery angle of the thermal management system.

[0047] Similarly, the LSTM model is used to predict the optimal air supply direction angle. The LSTM model is trained based on historical operating data of the computer room, including variables such as spatial distribution of equipment load, current air supply temperature of the thermal management system, air supply direction angle, and power efficiency (PUE) to predict the optimal air supply direction angle that minimizes energy consumption.

[0048] By inputting the optimal supply air temperature obtained through the solution and the real-time equipment load distribution, the system outputs the optimal supply air guide angle under the current conditions. This supply air guide angle refers to the angle of the outlet blades (e.g., θ is 30°, 60°, 90°, 120°, or 150°).

[0049] 3) Input / Output The input includes the optimal supply air temperature as the dominant variable, as well as the total equipment load, the spatial distribution of the equipment load, and the current relative humidity value.

[0050] Output the optimal relative humidity value and the optimal air supply guiding angle of all thermal management systems under the current working conditions.

[0051] Step 4, data integration: Integrate the optimization results and the prediction results uniformly to form the state matrix of the thermal management system of the control system.

[0052] Obtain the optimal air supply temperature 、optimal relative humidity and optimal air supply guiding angle respectively from the multi-source information fusion processing unit, and construct the state matrix of the thermal management system in a time synchronization manner in combination with the operating state parameters such as load information and equipment inlet air temperature. This state matrix of the thermal management system is used to drive the adjustment of the set values of the thermal management system and realize the coordinated control of multiple parameters.

[0053] Step 5, use the obtained state matrix of the thermal management system to adjust the temperature and humidity set values of the thermal management system and the angle of the outlet air duct blades, so as to realize the dynamic coordinated adjustment of the temperature and humidity set values and the angle of the air supply guiding blades, and thus realize the joint optimization of the system energy efficiency and the local heat dissipation efficiency.

[0054] Based on the state matrix of the thermal management system, dynamically adjust the operating parameters of the thermal management system to achieve the optimal coordinated control of the air supply temperature, relative humidity and air supply guiding angle, ensure the stability of the computer room thermal environment while reducing the system energy consumption. Specifically, it includes: 1) Extract the optimal control parameters of each air handling device in the thermal management system at the current moment, including: optimal air supply temperature 、optimal relative humidity and optimal air supply guiding angle .

[0055] 2) Receive the optimal control parameters output by the multi-source information fusion processing unit, and send the optimal air supply temperature, relative humidity and air supply guiding angle to the control interface of the thermal management system through the standard industrial control protocol to control the temperature setting module and the humidity control module of the thermal management system: Send as the refrigeration target temperature of the thermal management system to the control panel of the thermal management system; send as the target humidity to the humidity control loop; at the same time, keep the control frequency to be dynamically refreshed once every cycle to adapt to the load change.

[0056] Adjust the angle of the outlet air duct blades of the thermal management system to to cooperate with the load space distribution and improve the cold air refrigeration efficiency of the thermal management system.

[0057] Embodiment 4 This embodiment also provides a computing device. The computing device includes a bus, a processor, a memory, and a communication interface. The processor, memory, and communication interface communicate with each other via the bus. The computing device can be a server or a terminal device. It should be understood that this application does not limit the number of processors and memory in the computing device.

[0058] A bus can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of representation, a bus can include a path for transmitting information between various components of a computing device (e.g., memory, processor, communication interfaces).

[0059] The processor may include any one or more of the following: central processing unit (CPU), graphics processing unit (GPU), tensor processing unit (TPU), application specific integrated circuit (ASIC), field-programmable gate array (FPGA), microprocessor (MP), or digital signal processor (DSP).

[0060] Memory can include volatile memory, such as random access memory (RAM). Processors can also include non-volatile memory. volatile memory, such as read-only memory (ROM). ROM (memory only), flash memory, hard disk drive (HDD), or solid state drive (SSD).

[0061] The memory stores executable program code, which the processor executes to implement the functions of the aforementioned units, thereby achieving, for example, the method described in Embodiment 1. That is, the memory may store instructions for the methods and functions relating to the computing device in any of the above embodiments.

[0062] The communication interface uses transceiver modules such as, but not limited to, network interface cards and transceivers to enable communication between computing devices and other devices or communication networks.

[0063] Example 5 This embodiment also provides a computer-readable storage medium storing computer instructions that, when executed by a processor, cause the processor to perform the methods and functions of the computing device involved in any of the above embodiments.

[0064] Generally, the various embodiments of this disclosure can be implemented in hardware or dedicated circuitry, software, logic, or any combination thereof. Some aspects can be implemented in hardware, while others can be implemented in firmware or software, which can be executed by a controller, microprocessor, or other computing device. Although various aspects of the embodiments of this disclosure are shown and described as block diagrams, flowcharts, or represented using some other illustration, it should be understood that the blocks, apparatuses, systems, techniques, or methods described herein can be implemented as, as non-limiting examples, in hardware, software, firmware, dedicated circuitry or logic, general-purpose hardware or controllers or other computing devices, or some combination thereof.

[0065] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A correlation optimization energy-saving method based on computational and thermal management energy consumption, characterized in that, Includes the following steps: The operating data of the acquisition equipment under the current time period and load conditions is input into the multi-source information fusion processing unit. Through cascade optimization and joint simulation, the set of control parameters of the thermal management system is output. The temperature and humidity setpoints of the machine room and the angle of the air supply guide vanes are adjusted using the set of control parameters of the thermal management system.

2. The method according to claim 1, characterized in that, Through cascaded optimization and joint simulation, a set of control parameters for the thermal management system is output, specifically including: Construct a global energy consumption objective function, which integrates both equipment energy consumption and thermal management system energy consumption; A cluster intelligent search algorithm is used to optimize the global energy consumption objective function to obtain the optimal supply air temperature; The optimal supply air temperature and the aforementioned operating data are input into the time-series simulation model, and the set of thermal management system control parameters under the condition of minimum total system energy consumption is output.

3. The method according to claim 2, characterized in that, The expression for the energy consumption objective function is: in, Total energy consumption of all devices; For thermal management systems to provide Performance coefficient when the temperature is cold air.

4. The method according to claim 2, characterized in that, The time-series simulation model is constructed using LSTM and is used to predict the optimal relative humidity and the optimal air supply direction angle.

5. The method according to claim 2, characterized in that, The set of control parameters for the thermal management system includes the optimal supply air temperature, optimal relative humidity, and optimal supply air guide angle.

6. An energy-saving system based on the correlation optimization of energy consumption through computation and thermal management, characterized in that, include: The multi-source information fusion processing unit is used to input the operating data of the acquisition equipment under the current time period and load conditions, and output the set of control parameters of the thermal management system through cascade optimization and joint deduction. An air handling control unit is used to adjust the temperature and humidity setpoints of the computer room air handling unit and the angle of the air supply guide vanes according to the set of control parameters of the thermal management system.

7. The system according to claim 6, characterized in that, The multi-source information fusion processing unit includes: An objective function construction component is used to construct a global energy consumption objective function, which integrates both the device's computational energy consumption and the thermal management system's energy consumption. The optimization solution component is used to solve the global energy consumption objective function using a cluster intelligent search algorithm to obtain the optimal supply air temperature under the condition of minimum total system energy consumption; The parameter extrapolation component is used to input the optimal supply air temperature and the operating data into the time series extrapolation model, and output the optimal relative humidity and optimal supply air guide angle under the condition of minimum total system energy consumption. The parameter set generation component is used to construct the optimal supply air temperature, optimal relative humidity, and optimal supply air directional angle into a set of control parameters for the thermal management system.

8. The system according to claim 7, characterized in that, The time-series inference model used by the parameter inference component is an LSTM network, whose inputs include spatial distribution of equipment load, supply air temperature, and environmental parameters.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the method as described in any one of claims 1 to 5.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the method as described in any one of claims 1 to 5.

Images