Method, device and storage medium for PUE prediction tuning

Abstract

The application discloses a method and device for PUE prediction and optimization and a storage medium, and belongs to the technical field of equipment control. The application obtains load data of IT equipment of a data center, equipment performance data of refrigeration equipment and influence factors between refrigeration equipment, wherein the equipment performance data comprises an equipment performance curve, determines a refrigeration system according to the load data, determines system input data of a plurality of refrigeration systems and corresponding PUE results according to the equipment performance data, the influence factors and the refrigeration system, and generates a system recommendation report according to a plurality of system input data and corresponding PUE results, thereby achieving the beneficial effects of improving the operation efficiency of the data center and the use efficiency of resources.

Description

Technical Field

[0001] This invention relates to the field of equipment control technology, and in particular to a method, device and storage medium based on PUE prediction and optimization. Background Technology

[0002] With economic development, data centers are increasingly pursuing low-carbon emissions and reducing energy waste. Currently, data centers, due to their massive data computation, often require continuous operation of computing resources. During computation, they generate heat. To reduce the power consumption for cooling, highly efficient cooling equipment is typically installed. Alternatively, data centers are sometimes located underwater to improve heat dissipation. However, these cooling methods require significant upfront investment, and when computing demands fluctuate greatly, cooling redundancy can occur, leading to operational inefficiency and resource waste.

[0003] The above content is only used to help understand the technical solution of the present invention and does not represent an admission that the above content is prior art. Summary of the Invention

[0004] The main objective of this invention is to provide a method, device, and storage medium based on PUE prediction tuning, aiming to improve data center operational efficiency and resource utilization efficiency. To achieve the above objective, this invention provides a method based on PUE prediction tuning, which includes the following steps:

[0005] Obtain load data of IT equipment in the data center, equipment performance data of cooling equipment, and influencing factors between cooling equipment. The equipment performance data includes: equipment performance curves.

[0006] The refrigeration system is determined based on the load data;

[0007] Based on the equipment performance data, the influencing factors, and the refrigeration system, determine the system input data and corresponding PUE results of the refrigeration system;

[0008] A system recommendation report is generated based on multiple system input data and corresponding PUE results.

[0009] Optionally, the step of determining the system input data and corresponding PUE results of the refrigeration system based on the equipment performance data, the influencing factor, and the refrigeration system includes:

[0010] The system performance data is determined based on the equipment performance data and the refrigeration system, wherein the system performance data is the performance data of the refrigeration system.

[0011] The system input data is determined based on the system performance data;

[0012] The PUE results corresponding to multiple system input data are determined based on the system input data, the influence factor, and the system performance data.

[0013] Optionally, the step of determining the system input data based on the system performance data includes:

[0014] Based on the system performance data, a first type of system parameter and a second type of system parameter are determined. The first type of system parameter is a fixed setting parameter, and the second type of system parameter is a non-fixed setting parameter.

[0015] Multiple distinct second-type system datasets are generated based on a preset exhaustive algorithm and the parameters of the second-type system.

[0016] Multiple system input data are determined based on multiple second-class system datasets and first-class system data corresponding to first-class system parameters. Each system input data includes: first-class system data and second-class system datasets.

[0017] Optionally, the step of determining the PUE results corresponding to multiple system input data based on the system input data, the impact factor, and the system performance data includes:

[0018] The PUE prediction model is determined based on the system performance data and the influencing factors.

[0019] The corresponding PUE result is generated based on the PUE prediction model and each system input data.

[0020] Optionally, the step of determining the PUE prediction model based on the system performance data and the influencing factors includes:

[0021] The system performance data of each sub-device is fitted according to a preset fitting algorithm to obtain multiple performance functions, wherein each performance function is the performance function corresponding to the sub-device.

[0022] Multiple PUE data prediction values ​​are generated based on multiple preset system data, the performance function, and the influence factor;

[0023] A preset deep learning model is trained based on the predicted PUE data values ​​corresponding to the preset system data to obtain the PUE prediction model.

[0024] Optionally, the step of generating a system recommendation report based on multiple system input data and corresponding PUE results includes:

[0025] The system input data is sorted according to multiple PUE results to obtain a sorting result;

[0026] Select the corresponding system data as the target system input data based on the sorting results;

[0027] It also generates a corresponding system recommendation report based on the input data of the target system.

[0028] Optionally, the step of determining the refrigeration system based on the load data includes:

[0029] Based on the load data, a range of multiple devices is generated, including multiple different refrigeration devices, each of which differs in at least one of the following: parameters, functions, and manufacturers.

[0030] Display the range of devices to be selected and determine the device selection command;

[0031] The refrigeration system is constructed according to the device selection instructions.

[0032] Optionally, the refrigeration equipment includes: a chiller, a cooling tower, and a water pump, and the equipment performance curves include at least one of: a coefficient of performance curve, an energy efficiency ratio curve, and a partial load performance response curve.

[0033] Furthermore, to achieve the above objectives, the present invention also provides a device based on PUE prediction tuning, characterized in that the device includes: a memory, a processor, and a program based on PUE prediction tuning stored in the memory and executable on the processor, wherein the program based on PUE prediction tuning is configured to implement the steps of the method based on PUE prediction tuning described in any of the above claims.

[0034] In addition, to achieve the above objectives, the present invention also provides a storage medium storing a program based on PUE prediction tuning, wherein when the program based on PUE prediction tuning is executed by a processor, it implements the steps of the method based on PUE prediction tuning described above.

[0035] This invention proposes a PUE-based prediction and optimization method. This method uses load data of IT equipment in a data center, equipment performance data of cooling equipment, and influence factors between cooling equipment. Based on the load data, the method determines the cooling system, and then determines the system input data and corresponding PUE results of the cooling system based on the equipment performance data, the influence factors, and the cooling system itself. Compared with the traditional method of adjusting operation based on temperature, this method generates a system recommendation report based on multiple system input data and corresponding PUE results, thereby improving the resource utilization efficiency of the data center by pushing suitable system data. Attached Figure Description

[0036] Figure 1This is a schematic diagram of the structure of a device based on PUE prediction and optimization in the hardware operating environment involved in the embodiments of the present invention;

[0037] Figure 2 This is a flowchart illustrating the first embodiment of the PUE prediction and tuning method of the present invention;

[0038] Figure 3 This is a flowchart illustrating the second embodiment of the PUE prediction and tuning method of the present invention;

[0039] Figure 4 This is a flowchart illustrating the third embodiment of the PUE prediction and tuning method of the present invention.

[0040] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0041] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0042] Reference Figure 1 , Figure 1 This is a schematic diagram of the device structure based on PUE prediction and optimization of the hardware operating environment involved in the embodiments of the present invention.

[0043] like Figure 1 As shown, the device for PUE prediction tuning may include: a processor 1001, such as a central processing unit (CPU), a communication bus 1002, an interactive device 1003, a network interface 1004, and a memory 1005. The communication bus 1002 is used to enable communication between these components. The interactive device 1003 may include a display screen or an input unit such as a keyboard. Optionally, the interactive device 1003 may also be connected to the communication bus via standard wired or wireless interfaces. The network interface 1004 may optionally include standard wired or wireless interfaces (such as a Wi-Fi interface). The memory 1005 may be a high-speed random access memory (RAM) or a stable non-volatile memory (NVM), such as a disk drive. Optionally, the memory 1005 may also be a storage device independent of the aforementioned processor 1001.

[0044] Those skilled in the art will understand that Figure 1The structure shown does not constitute a limitation on the device based on PUE prediction tuning, and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0045] like Figure 1 As shown, the memory 1005, which serves as a storage medium, may include an operating system, a data storage module, a network communication module, a user interface module, and a program for PUE prediction and tuning.

[0046] exist Figure 1 In the device based on PUE prediction tuning shown, the network interface 1004 is mainly used for data communication with other devices; the interactive device 1003 is mainly used for data interaction with the user; the processor 1001 and memory 1005 in the device based on PUE prediction tuning of the present invention can be set in the device based on PUE prediction tuning, and the device based on PUE prediction tuning calls the program based on PUE prediction tuning stored in memory 1005 through processor 1001 and executes the method based on PUE prediction tuning provided in the embodiment of the present invention.

[0047] This invention provides a method for PUE prediction-based optimization, referring to... Figure 2 , Figure 2 This is a flowchart illustrating the first embodiment of a method for PUE prediction and optimization according to the present invention.

[0048] In this embodiment, the method for PUE prediction and optimization includes:

[0049] Step S1: Obtain load data of IT equipment in the data center, equipment performance data of cooling equipment, and influencing factors between cooling equipment. The equipment performance data includes: equipment performance curves.

[0050] In this embodiment, the cooling system is optimized based on Power Usage Effectiveness (PUE) to obtain the load data of the IT equipment. Since IT equipment often operates in different states, these states correspond to different heat generation effects. Furthermore, due to the large number of IT devices in the data, load data for all IT devices in the data center is typically collected. The performance data of the cooling equipment varies significantly due to differences in type, version, and manufacturer. In current data centers, due to redundancy design, multiple additional cooling devices are typically configured beyond the maximum operating power. A common practice is to use a dual-source cooling system, such as a chilled water and air-cooled system operating in parallel, to ensure system operation even in the event of a single source failure. The interaction between cooling devices leads to variations in cooling efficiency. The influencing factors include the first influencing factor: flow sharing and mutual interference effects, which refer to efficiency fluctuations caused by flow distribution among parallel devices. The second influencing factor corresponds to the efficiency superposition effect, which refers to the overall COP trend during multi-device combined operation. The third influencing factor corresponds to the redundancy configuration strategy, which refers to the standby state of backup equipment and energy burden sharing under N+1 configuration. The fourth influencing factor corresponds to the operation strategy, which refers to energy consumption differences under strategies such as rotational operation and load balancing. Optionally, by establishing a comprehensive energy consumption influencing factor matrix, the overall energy consumption output can be simulated to more closely resemble real-world operating conditions. The prediction results can cover indicators such as instantaneous power, total energy consumption, and energy efficiency, including system-level COP and EER.

[0051] Step S2: Determine the refrigeration system based on the load data;

[0052] Based on the load data, the selectable range of cooling equipment is determined, and users can build cooling systems combining multiple cooling devices. It should be noted that a cooling system here includes a collection of multiple cooling devices selected by the user. This function aims to support users in flexibly building cooling systems combining multiple cooling devices, achieving expansion from the device layer to the system layer. Optionally, a corresponding graphical interface is provided, displaying the corresponding selectable equipment types (chillers, cooling towers, water pumps, terminal equipment, etc.), brand models, and quantities. Preferably, when it is necessary to predict and build the cooling system corresponding to the data center, this graphical interface can also be configured in parallel or serial manner, completing system construction through drag-and-drop or parameter input.

[0053] Step S3: Determine multiple system input data and corresponding PUE results of the refrigeration system based on the equipment performance data, the influencing factors, and the refrigeration system;

[0054] Specifically, by setting a preset exhaustive search engine, the exhaustive search engine specifically includes the exhaustive search step size. Since most of the data is actually a continuous variable, the exhaustive search engine needs to determine the system input data. The input data can be: water supply temperature 6-12℃, load rate 30%-100%, and running time 0-24h.

[0055] The PUE result is calculated based on the system input data, the equipment performance data, and the impact factor.

[0056] Step S4: Generate a system recommendation report based on the multiple system input data and the corresponding PUE results.

[0057] In this embodiment, the PUE result of the refrigeration system under specific input conditions is output. Furthermore, the predicted output may include multiple indicators such as instantaneous power, energy consumption per unit time, and total daily operating energy consumption. Users can customize the output time granularity, commonly in minutes, hours, or days. The output results can be visually presented in front-end charts, including curves showing energy consumption changes with input variables and load sensitivity analysis to energy consumption. In addition, the system should support result export (e.g., Excel, PDF format) and automatic archiving of equipment energy consumption archives for subsequent statistical analysis and energy efficiency assessment. Further, to improve the accuracy of subsequent predictions, the system records the input and output data for each prediction task, using this data as training data or as a basis for model correction. This data is often used to correct the aforementioned influencing factors using subsequent measured data. Moreover, the deployment requirements for common 8KW, 12KW, and 16KW single-rack power air-cooled racks are generally dynamically adjustable. The system's recommended reports can be used to study the operating pressures related to cold aisle width, raised floor height, and airflow velocity.

[0058] In this embodiment, the load data of IT equipment in the data center, the performance data of cooling equipment, and the influence factors between cooling equipment are used to determine the cooling system based on the load data. Then, based on the equipment performance data, the influence factors, and the cooling system, the system input data and corresponding PUE results of multiple cooling systems are determined. Compared with the traditional method of adjusting operation based on temperature, a system recommendation report is generated based on multiple system input data and corresponding PUE results, thereby improving the resource utilization efficiency of the data center by pushing suitable system data.

[0059] Furthermore, based on the first embodiment, a second embodiment of the method for PUE prediction and optimization of the present invention is proposed. In this embodiment, reference is made to... Figure 3The step of determining the system input data and corresponding PUE results of multiple refrigeration systems based on the equipment performance data, the influencing factors, and the refrigeration system includes:

[0060] Step S31: Determine system performance data based on the equipment performance data and the refrigeration system, wherein the system performance data is the performance data of the refrigeration system;

[0061] The refrigeration system selects the corresponding system performance data from the equipment performance data. For example, if the refrigeration system selects refrigeration equipment of brand A, it can select data of brand A from the equipment performance data as one of the system performance data. Step S31 uses the instruction system to filter out the required data.

[0062] Step S32: Determine the system input data based on the system performance data;

[0063] Commonly, the parameter ranges of each refrigeration device in the refrigeration system are used as the data selection range, and multiple different system input data are generated within this range. This effectively avoids the generated system input data not matching the actual refrigeration devices. The devices and their corresponding power ratings in the refrigeration system can be represented in the following table:

[0064]

[0065] Table 1 Power consumption of equipment in the refrigeration system

[0066] This includes the operating power of different devices, as well as parameters such as corresponding cooling capacity. Multiple system input data can be generated based on the table above.

[0067] Step S33: Determine the PUE results corresponding to multiple system input data based on the system input data, the influence factor, and the system performance data.

[0068] In this embodiment, since the system performance data is generally obtained by the supplier through experimental data testing, it is necessary to obtain the corresponding energy consumption function through function fitting. The total power consumption is calculated by combining the energy consumption function, the system input data, and the influencing factor. The PUE result is obtained by dividing the IT equipment energy consumption by the total power consumption. Here, the IT equipment energy consumption can be calculated from the load data, commonly by integrating the load data.

[0069] In this embodiment, system performance data is determined by the device performance data and the refrigeration system, where the system performance data is the performance data of the refrigeration system. System input data is determined based on the system performance data, thereby obtaining multiple coefficient input data that can be used to analyze energy consumption. PUE results corresponding to multiple system input data are determined based on the system input data, the influencing factors, and the system performance data, thereby improving the accuracy of the data.

[0070] Furthermore, the step of determining the system input data based on the system performance data includes:

[0071] Based on the system performance data, a first type of system parameter and a second type of system parameter are determined. The first type of system parameter is a fixed setting parameter, and the second type of system parameter is a non-fixed setting parameter.

[0072] Multiple distinct second-type system datasets are generated based on a preset exhaustive algorithm and the parameters of the second-type system.

[0073] Multiple system input data are determined based on multiple second-class system datasets and first-class system data corresponding to first-class system parameters. Each system input data includes: first-class system data and second-class system datasets.

[0074] In this embodiment, considering that some input parameters may be immutable in real-world scenarios, such as generator brand, maximum load capacity, and building structure limitations, the first type of system parameters are manually locked before simulation optimization. The remaining second type of system parameters can be set to avoid invalid simulation results, thereby improving optimization efficiency and practicality. Correspondingly, the user interface should support intuitive locking settings, such as checkboxes, disabling drag bars, and graying out parameters. It should also provide prompts regarding the effective combination space after parameter locking. This ensures that the simulation results closely resemble practically executable operation and maintenance strategies, improving executability.

[0075] Furthermore, based on the first or second embodiment, a third embodiment of the method for PUE prediction and optimization of the present invention is proposed. In this embodiment, reference is made to... Figure 4 The step of determining the PUE results corresponding to multiple system input data based on the system input data, the influencing factor, and the system performance data includes:

[0076] Step S331: Determine the PUE prediction model based on the system performance data and the influencing factors;

[0077] It should be noted that after obtaining the PUE prediction model, it can be saved and recorded for future reference. Optionally, the PUE prediction model here can be a prediction model corresponding to a preset template, or it can be a deep learning model. It should also be noted that the preset template can be a system efficiency function, which specifically includes system variables. These system variables can be the performance data of each refrigeration device in the refrigeration system, the influencing factors, and the load data of the IT equipment. Different solutions will correspond to different PUE prediction models.

[0078] Step S332: Generate the corresponding PUE result based on the PUE prediction model and each system input data.

[0079] In this embodiment, the system input data is input into the PUE prediction model to generate the corresponding PUE result. In this embodiment, by constructing a prediction model and inputting the system input data into the prediction model, the PUE result can be predicted, and the method of early prediction can effectively select appropriate system input data.

[0080] Furthermore, the step of determining the PUE prediction model based on the system performance data and the influencing factors includes:

[0081] The system performance data of each sub-device is fitted according to a preset fitting algorithm to obtain multiple performance functions, wherein each performance function is the performance function corresponding to the sub-device.

[0082] Multiple PUE data prediction values ​​are generated based on multiple preset system data, the performance function, and the influence factor;

[0083] A preset deep learning model is trained based on the predicted PUE data values ​​corresponding to the preset system data to obtain the PUE prediction model.

[0084] The multiple performance functions obtained here each correspond to a sub-device. It should be noted that the sub-device here refers to the refrigeration equipment in the refrigeration system. This is used to distinguish refrigeration equipment that does not belong to the refrigeration system. It should be noted that in the step of generating multiple PUE data prediction values ​​based on multiple preset system data, the performance functions, and the influencing factors, the energy efficiency value of the sub-device determined by each preset system data and the corresponding performance function is adjusted using the influencing factors as coefficients. The PUE data prediction value is then calculated from the multiple energy efficiency values ​​based on the system efficiency function and the load data. Since there are multiple preset system data sets, PUE data prediction values ​​corresponding to the data volume can be obtained. Furthermore, multiple system efficiency functions with differences can be selected. It should be noted that to ensure the stability of the PUE data prediction values, the function structures of the different system efficiency functions need to be the same or approximately the same. In this embodiment, more training data is obtained through different system efficiency functions. The preset system data and the corresponding PUE data prediction values ​​are used as training data to obtain a training set including multiple training data sets. In some other embodiments, the IT operating load is ≤50% of full load, the chilled water system is a primary pump variable flow system, and the cold storage tank is connected in series in the system. Normally, the water temperature in the cold storage tank is equal to the chilled water supply temperature. The chilled water supply temperature is the system input data. In this scenario, multiple system efficiency functions can be preset, and the energy efficiency of this cooling method can be determined through simulation. Different temperatures correspond to different energy efficiencies, i.e., the PUE results.

[0085] Furthermore, based on any of the above embodiments, a fourth embodiment of the method for PUE prediction and optimization of the present invention is proposed. In this embodiment, the step of generating a system recommendation report based on multiple system input data and corresponding PUE results includes:

[0086] The system input data is sorted according to multiple PUE results to obtain a sorting result;

[0087] Select the corresponding system data as the target system input data based on the sorting results;

[0088] Specifically, the system input data is sorted from highest to lowest efficiency, and the system input data with the highest sorting result is selected as the target system input data.

[0089] It also generates a corresponding system recommendation report based on the input data of the target system.

[0090] In addition, the system should support result export, such as to Excel and PDF formats, and automatic archiving with the equipment energy consumption archive for subsequent statistical analysis and energy efficiency assessment. To facilitate subsequent algorithm iterations, the system should record the input and output data for each prediction task, which can be used as training data accumulation or as a basis for model correction.

[0091] Furthermore, based on any of the above embodiments, a fifth embodiment of the method for predictive tuning based on PUE of the present invention is proposed. In this embodiment, the step of determining the cooling system based on the load data includes:

[0092] Based on the load data, a range of multiple devices is generated, including multiple different refrigeration devices, each of which differs in at least one of the following: parameters, functions, and manufacturers.

[0093] Display the range of devices to be selected and determine the device selection command;

[0094] The refrigeration system is constructed according to the device selection instructions.

[0095] In this embodiment, an operation interface can be set up to select the equipment type, such as: chiller unit, cooling tower, water pump, terminal equipment, etc. In addition, the brand, model, quantity and series-parallel configuration method can be selected, and the system can be assembled by dragging and dropping or parameter input.

[0096] Furthermore, the refrigeration equipment includes: a chiller, a cooling tower, and a water pump, and the equipment performance curves include at least one of: a coefficient of performance curve, an energy efficiency ratio curve, and a partial load performance response curve.

[0097] Furthermore, embodiments of the present invention also propose a device based on PUE prediction tuning, the device comprising: a memory, a processor, and a program based on PUE prediction tuning stored in the memory and executable on the processor, the program based on PUE prediction tuning being configured to implement the steps of the method based on PUE prediction tuning described above.

[0098] Furthermore, embodiments of the present invention also propose a storage medium storing a program based on PUE prediction tuning, wherein when the program based on PUE prediction tuning is executed by a processor, it implements the steps of the method based on PUE prediction tuning described above.

[0099] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or system. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or system that includes that element.

[0100] The sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0101] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0102] The above are merely preferred embodiments of the present invention and do not limit the scope of the patent. Any equivalent structural or procedural transformations made based on the description and drawings of the present invention, or direct or indirect applications in other related technical fields, are similarly included within the scope of patent protection of the present invention.

Claims

1. A method for PUE prediction and optimization, characterized in that, The method for PUE prediction and optimization includes the following steps: Obtain load data of IT equipment in the data center, equipment performance data of cooling equipment, and influencing factors between cooling equipment. The equipment performance data includes: equipment performance curves. The refrigeration system is determined based on the load data; Based on the equipment performance data, the influencing factors, and the refrigeration system, determine multiple system input data and corresponding PUE results for the refrigeration system; A system recommendation report is generated based on multiple system input data and corresponding PUE results; Among them, the influencing factors include: the first influencing factor corresponding to the traffic sharing and mutual interference effect, the second influencing factor corresponding to the efficiency superposition effect, the third influencing factor corresponding to the redundancy configuration strategy, and the fourth influencing factor corresponding to the operation strategy. The traffic sharing and mutual interference effect refers to the efficiency fluctuation caused by traffic allocation among parallel devices. The efficiency superposition effect refers to the overall COP change trend when multiple devices are combined and running. The redundancy configuration strategy refers to the standby state of the backup device and the energy burden sharing under the N+1 configuration. The impact of the operation strategy refers to the energy consumption difference under the rotation operation and load balancing strategy.

2. The method for PUE prediction and optimization as described in claim 1, characterized in that, The step of determining multiple system input data and corresponding PUE results of the refrigeration system based on the equipment performance data, the influencing factors, and the refrigeration system includes: The system performance data is determined based on the equipment performance data and the refrigeration system, wherein the system performance data is the performance data of the refrigeration system. The system input data is determined based on the system performance data; The PUE results corresponding to multiple system input data are determined based on the system input data, the influence factor, and the system performance data.

3. The method for PUE prediction-based optimization as described in claim 2, characterized in that, The step of determining the system input data based on the system performance data includes: Based on the system performance data, a first type of system parameter and a second type of system parameter are determined. The first type of system parameter is a fixed setting parameter, and the second type of system parameter is a non-fixed setting parameter. Multiple distinct second-type system datasets are generated based on a preset exhaustive algorithm and the parameters of the second-type system. Multiple system input data are determined based on multiple second-class system datasets and first-class system data corresponding to first-class system parameters. Each system input data includes: first-class system data and second-class system datasets.

4. The method for PUE prediction and optimization as described in claim 2, characterized in that, The step of determining the PUE results corresponding to multiple system input data based on the system input data, the influencing factor, and the system performance data includes: The PUE prediction model is determined based on the system performance data and the influencing factors. The corresponding PUE result is generated based on the PUE prediction model and each system input data.

5. The method for PUE prediction and optimization as described in claim 4, characterized in that, The step of determining the PUE prediction model based on the system performance data and the influencing factors includes: The system performance data of each sub-device is fitted according to a preset fitting algorithm to obtain multiple performance functions, wherein each performance function is the performance function corresponding to the sub-device. Multiple PUE data prediction values ​​are generated based on multiple preset system data, the performance function, and the influence factor; A preset deep learning model is trained based on the predicted PUE data values ​​corresponding to the preset system data to obtain the PUE prediction model.

6. The method for PUE prediction and optimization as described in claim 1, characterized in that, The step of generating a system recommendation report based on multiple system input data and corresponding PUE results includes: The system input data is sorted according to multiple PUE results to obtain a sorting result; Select the corresponding system data as the target system input data based on the sorting results; It also generates a corresponding system recommendation report based on the input data of the target system.

7. The method for PUE prediction-based optimization as described in claim 1, characterized in that, The step of determining the refrigeration system based on the load data includes: Based on the load data, a range of multiple devices is generated, including multiple different refrigeration devices, each of which differs in at least one of the following: parameters, functions, and manufacturers. Display the range of devices to be selected and determine the device selection command; The refrigeration system is constructed according to the device selection instructions.

8. The method for PUE prediction-based optimization as described in any one of claims 1 to 7, characterized in that, The refrigeration equipment includes: a chiller, a cooling tower, and a water pump. The equipment performance curves include at least one of: a coefficient of performance curve, an energy efficiency ratio curve, and a partial load performance response curve.

9. A device based on PUE prediction and tuning, characterized in that, The device includes: a memory, a processor, and a PUE prediction tuning program stored in the memory and executable on the processor, the PUE prediction tuning program being configured to implement the steps of the PUE prediction tuning method as described in any one of claims 1 to 8.

10. A storage medium, characterized in that, The storage medium stores a program based on PUE prediction tuning, which, when executed by a processor, implements the steps of the method based on PUE prediction tuning as described in any one of claims 1 to 8.

Images