PUE checking and reducing method and system suitable for data center containing office area, and medium

Through nonlinear regression method and big data analysis, a cooling system benchmark model is built to accurately evaluate the cooling energy consumption in the office area of ​​the data center, solving the problem of inaccurate PUE measurement in the existing technology, and improving the energy utilization efficiency of data centers.

CN119048271APending Publication Date: 2024-11-29SOUTH CHINA UNIV OF TECH +1
View PDF 0 Cites 1 Cited by

Patent Information

Application Number
CN202410852546.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-06-28
Publication Date
2024-11-29

AI Technical Summary

Technical Problem

The existing technology is difficult to accurately evaluate the cooling energy consumption in the office area of ​​the data center, resulting in inaccurate PUE measurement and affecting the energy utilization efficiency of the data center.

Method used

The nonlinear regression method is used to build a baseline model of the cooling system, combining big data analysis and artificial intelligence algorithms, collect outdoor meteorological environment parameters and cold source system operating parameters, divide the working conditions and eliminate outliers, and accurately predict the energy without office load during the cold season. Consumption, calculate the real PUE value of the data center after excluding office electricity and cooling energy consumption.

Benefits of technology

No on-site construction is required, and the energy consumption generated by cooling in the office area is accurately eliminated, helping enterprises reduce the PUE of data centers and achieve the energy efficiency goals set by the state.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119048271A_ABST
    Figure CN119048271A_ABST
Patent Text Reader

Abstract

The invention discloses a PUE checking and reducing method and system suitable for a data center containing an office area and a medium. The method comprises the steps that outdoor meteorological environment parameters, cold source system operation parameters and data center energy consumption data are collected; preprocessing the acquired data, wherein the preprocessing comprises working condition division, abnormal value elimination, missing value interpolation and cold source energy consumption feature selection; by taking the preprocessed data as input and adopting a nonlinear regression method, constructing a cooling system reference model without office load in the cooling season, and predicting the energy consumption without office load under the working condition of the cooling season; dividing cold energy consumption of office areas; and calculating the real PUE value of the data center after the office electricity consumption and the office cold consumption energy consumption are removed. According to big data analysis and an artificial intelligence algorithm, the cooling system reference model is constructed, on-site construction is not needed, energy consumption generated by cooling in an office area is accurately eliminated, enterprises are helped to check and reduce real PUE of a data center, and the energy efficiency target formulated by the country is achieved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the technical field of building energy conservation, and in particular relates to a PUE reduction method, system and medium applicable to a data center including an office area. Background Art

[0002] PUE (Power Usage Effectiveness) is a commonly used indicator for evaluating the energy efficiency of data centers. It is defined as the ratio of the total energy consumption of data centers to the energy consumption of electronic information equipment, as shown in formula (1).

[0003]

[0004] Among them, E total is the total energy consumption of the data center, E IT is the energy consumption of IT equipment, E Power is the energy consumption of the power supply equipment, E cold is the energy consumption of the cooling system, else The closer the PUE value is to 1, the higher the greenness of a data center.

[0005] In actual data center projects, some data centers have integrated office areas. The electricity consumption for office area lighting and sockets, and office area cooling are included in the PUE metering range, resulting in the data center metered PUE being higher than the actual PUE of the computer room. Currently, smart meters can be installed to effectively eliminate the impact of office area lighting and socket electricity consumption on the data center PUE. However, the data center office area and IT computer room share a centralized air conditioning system, making it difficult to accurately assess the energy consumption generated by office area cooling.

[0006] Industry personnel use cooling meters to collect cooling energy used in office areas, and eliminate the impact of cooling energy consumption in office areas on PUE based on the proportion of cooling energy used in office areas to the total cooling capacity of the air conditioning system. However, this method has high construction costs and poor measurement accuracy, and it is difficult to accurately evaluate the true PUE of a data center. In addition, industry personnel estimate cooling energy used in office areas based on the power consumption of air conditioning systems in similar data centers without office areas. This method is more extensive and simple, and it is difficult to obtain the true PUE of a data center. Therefore, it is crucial to find a method to accurately reduce cooling energy consumption in office areas in order to measure the true PUE of a data center. Summary of the invention

[0007] The main purpose of the present invention is to overcome the shortcomings and deficiencies of the prior art and to provide a PUE reduction method, system and medium suitable for a data center including an office area.

[0008] In order to achieve the above object, the present invention adopts the following technical solutions:

[0009] In one aspect of the present invention, a PUE reduction method for a data center including an office area is provided, comprising the following steps:

[0010] Collect outdoor meteorological environment parameters, cold source system operating parameters and data center energy consumption data;

[0011] Preprocess the collected data, including dividing the working conditions, removing outliers and interpolating missing values, and selecting the characteristics of cold source energy consumption;

[0012] Using the preprocessed data as input, a nonlinear regression method is used to construct a benchmark model of the cooling system without office load in the cooling season, and to predict the energy consumption without office load under office conditions in the cooling season.

[0013] Divide the cooling energy consumption of office areas;

[0014] Calculate the actual PUE value of the data center after excluding office electricity and office cooling energy consumption.

[0015] As a preferred technical solution, the outdoor meteorological environment parameters, the cold source system operating parameters and the data center energy consumption data are collected, specifically:

[0016] The data center external environmental parameters collected include outdoor dry-bulb temperature, outdoor wet-bulb temperature, and outdoor relative humidity;

[0017] Collect the operating parameters of the data center air conditioning system, including chilled water supply and return water temperature, cooling water supply and return water temperature, chilled water supply and return water pressure difference, cooling water supply and return water pressure difference, chilled water flow, cooling water flow, host evaporation pressure, host condensation pressure and host load rate;

[0018] The energy consumption data collected from the data center include the total power consumption of the data center, the power consumption of the air conditioning system, the power consumption of office area lighting and sockets, and the power consumption of IT equipment.

[0019] As a preferred technical solution, the division of working conditions is specifically as follows:

[0020] The historical data of the data center are divided into cooling season with office load conditions, cooling season without office load conditions, and non-cooling season conditions according to actual office conditions; the cooling season is April to October, and the remaining months are non-cooling seasons; the office time period is 8:00-21:00, and the remaining time period is no office load time.

[0021] As a preferred technical solution, the data outlier removal is specifically as follows:

[0022] The quartile method is used for outlier detection, specifically:

[0023] IQR i =Q 3, -Q1,

[0024] T u,i =Q 3, +1.5*IQR i

[0025] T d,i =Q 1, -1.5*QR i

[0026] In the formula, Q 3, is the 75% quantile value of each attribute data in the data set at time i after being arranged from small to large; Q 1, It is the 25% quantile value of each attribute data in the data set at time i after being arranged from small to large; IQR i is the interquartile range of the data at time i; T u,i is the upper limit of abnormal threshold at time i; T d,i is the lower limit of the abnormal threshold at time i; the abnormal value judgment standard is that the sample value is less than the lower limit of the abnormal threshold at the time it belongs to or greater than the upper limit of the abnormal threshold at the time it belongs to;

[0027] Remove detected outliers.

[0028] As a preferred technical solution, as a preferred technical solution, the energy consumption characteristics of the cold source are specifically selected as follows:

[0029] By using correlation analysis, key influencing factors of air-conditioning system power consumption whose absolute values ​​of Pearson correlation coefficients are greater than preset values ​​are selected as input parameters.

[0030] As a preferred technical solution, the nonlinear regression method is used to construct a benchmark model of the air conditioning cooling system without office load in the cooling season, specifically:

[0031] The CatBoost algorithm is used to build a benchmark model of the cooling system without office load in the cooling season, and grid search and ten-fold cross validation are used to find the optimal hyperparameters.

[0032] As a preferred technical solution, the cooling energy consumption of the office area is specifically divided as follows:

[0033] Cooling energy consumption in office area: Cooling energy consumption in office area = actual collected air conditioning system energy consumption - air conditioning system benchmark energy consumption;

[0034] For the cooling season with office load conditions, the air conditioning system benchmark energy consumption is equal to the predicted energy consumption. For the cooling season without office load conditions, the air conditioning system benchmark energy consumption is equal to the actual collected air conditioning system energy consumption. For the non-cooling season conditions, the air conditioning system benchmark energy consumption is equal to the actual collected air conditioning system energy consumption.

[0035] The cooling energy consumption of the office area in the cooling season without office load conditions and in the non-cooling season is 0.

[0036] As a preferred technical solution, the calculation of the actual PUE value of the data center after excluding office electricity and office cooling energy consumption is specifically:

[0037] Set the actual PUE calculation formula in the data center energy consumption monitoring platform:

[0038]

[0039] In the formula, E total is the total energy consumption of the data center, E Of is the energy consumption of office area lighting and sockets, E of. is the cooling energy consumption of the office area, E IT Energy consumption of IT equipment;

[0040] Input the cooling energy consumption of the office area calculated by the energy consumption division module, the energy consumption of IT equipment, the energy consumption of office area lighting and sockets, and the total energy consumption of the data center collected by the energy supervision platform into the above-mentioned real PUE calculation formula;

[0041] Calculate the actual PUE value of the data center after excluding office electricity and office cooling energy consumption.

[0042] Another aspect of the present invention further provides a PUE reduction system for a data center including an office area, which is applied to the above-mentioned PUE reduction method for a data center including an office area, and includes a data acquisition module, an energy consumption division module, and a PUE calculation module;

[0043] The data acquisition module is used to collect outdoor meteorological environment parameters, cold source system operating parameters and data center energy consumption data;

[0044] The energy consumption division module is used to pre-process the collected data, including dividing the working conditions, removing outliers and interpolating missing values, and selecting the characteristics of cold source energy consumption; using the pre-processed data as input, a nonlinear regression method is used to construct a cooling system benchmark model without office load in the cooling season, and predict the energy consumption without office load under office working conditions in the cooling season; and dividing the cooling energy consumption of the office area;

[0045] The PUE calculation module is used to calculate the real PUE value of the data center after excluding office electricity and office cooling energy consumption.

[0046] Another aspect of the present invention further provides a storage medium storing a program, which, when executed by a processor, implements the above-mentioned PUE reduction method for a data center including an office area.

[0047] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0048] (1) The present invention constructs a cooling system benchmark model based on big data analysis and artificial intelligence algorithms. It does not require on-site construction and accurately eliminates the energy consumption generated by cooling in office areas, helping enterprises to reduce the actual PUE of data centers and achieve the energy efficiency targets set by the state. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 This is a flow chart of a PUE reduction method for a data center including an office area according to an embodiment of the present invention;

[0050] Figure 2 is a schematic diagram of the structure of data collected by the data collection module of an embodiment of the present invention;

[0051] Figure 3 It is a comparison chart of the measured energy consumption of the air conditioning system and the predicted energy consumption of the air conditioning system in an embodiment of the present invention; a comparison chart of the PUE of the machine building before and after excluding the electricity consumption of the office area and the cooling energy consumption of the office;

[0052] Figure 4 This is a comparison chart of office area lighting, sockets and cooling energy consumption in an embodiment of the present invention;

[0053] Figure 5 This is a comparison diagram of the PUE of the machine building before and after excluding the electricity consumption in the office area and the cooling energy consumption in the office area according to an embodiment of the present invention;

[0054] Figure 6 This is a structural diagram of a PUE reduction system for a data center including an office area according to an embodiment of the present invention;

[0055] Figure 7 It is a schematic diagram of the structure of the storage medium of an embodiment of the present invention. DETAILED DESCRIPTION

[0056] In order to enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without making creative work are within the scope of protection of the present application.

[0057] Example

[0058] like Figure 1 As shown, this embodiment provides a PUE reduction method for a data center including an office area, including the following steps:

[0059] S1. Collect outdoor meteorological environment parameters, cold source system operating parameters and data center energy consumption data;

[0060] Further, such as Figure 2 As shown, in this embodiment, the outdoor meteorological environment parameters, the cold source system operating parameters and the data center energy consumption data are collected, specifically:

[0061] The data center external environmental parameters collected include outdoor dry-bulb temperature, outdoor wet-bulb temperature, and outdoor relative humidity;

[0062] Collect the operating parameters of the data center air conditioning system, including chilled water supply and return water temperature, cooling water supply and return water temperature, chilled water supply and return water pressure difference, cooling water supply and return water pressure difference, chilled water flow, cooling water flow, host evaporation pressure, host condensation pressure and host load rate;

[0063] The energy consumption data collected from the data center include the total power consumption of the data center, the power consumption of the air conditioning system, the power consumption of office area lighting and sockets, and the power consumption of IT equipment.

[0064] S2. Preprocessing the collected data, including dividing the working conditions, removing abnormal values ​​and interpolating missing values, and selecting the characteristics of cold source energy consumption;

[0065] Furthermore, in this embodiment, the division working conditions are specifically:

[0066] The historical data of the data center are divided into cooling season with office load conditions, cooling season without office load conditions, and non-cooling season conditions according to actual office conditions; the cooling season is April to October, and the remaining months are non-cooling seasons; the office time period is 8:00-21:00, and the remaining time period is no office load time.

[0067] Furthermore, in this embodiment, the data outliers are removed specifically as follows:

[0068] The quartile method is used for outlier detection, specifically:

[0069] IQR i =Q 3, -Q 1,

[0070] T u,i =Q 3, +1.5*IQR i

[0071] T d,i =Q 1, -1.5*QR i

[0072] In the formula, Q 3,is the 75% quantile value of each attribute data in the data set at time i after being arranged from small to large; Q 1, It is the 25% quantile value of each attribute data in the data set at time i after being arranged from small to large; IQR i is the interquartile range of the data at time i; T u,i is the upper limit of abnormal threshold at time i; T d,i is the lower limit of the abnormal threshold at time i; the abnormal value judgment standard is that the sample value is less than the lower limit of the abnormal threshold at the time it belongs to or greater than the upper limit of the abnormal threshold at the time it belongs to;

[0073] Remove detected outliers.

[0074] Furthermore, in this embodiment, the missing value interpolation is specifically as follows:

[0075] After removing the abnormal data, the missing values ​​are interpolated.

[0076] Furthermore, in this embodiment, the cold source energy consumption characteristic is selected specifically as follows:

[0077] By using correlation analysis, the key influencing factors of air-conditioning system power consumption with the absolute value of Pearson correlation coefficient greater than the preset value are selected as the input parameters of the cooling system benchmark model.

[0078] S3. Using the preprocessed data as the cooling system benchmark model input, a nonlinear regression method is used to construct an air conditioning benchmark energy consumption model without office load in the cooling season, and the energy consumption without office load under office conditions in the cooling season is predicted;

[0079] Furthermore, in this embodiment, the historical data of the power consumption of the air conditioning system and related characteristic variables in the cooling season without office load conditions are used as the input of the cooling system benchmark model, and the cooling system benchmark model of the centralized air conditioning system of the data center (without office area cooling) is constructed by using the nonlinear regression method;

[0080] Furthermore, this embodiment uses the CatBoost algorithm to construct a cooling system benchmark model without office load in the cooling season, and uses grid search and ten-fold cross validation to find the optimal hyperparameters. The grid search range is set as follows: the number of iterations is 50 to 2000, the step size is 50, the learning rate is 0.001 to 0.1, the step size is 0.001, the maximum tree depth is 3 to 16, the step size is 1, the sample sampling ratio is 0.6 to 1, the step size is 0.1, the column sampling ratio is 0.6 to 1, the step size is 0.1, and the grid search result is: the number of iterations is 950, the learning rate is 0.012, the maximum tree depth is 8, the sample sampling ratio is 0.8, and the column sampling ratio is 0.9.

[0081] Energy consumption prediction without office load in office conditions during the cooling season: The above cooling system benchmark model is embedded in the data center energy monitoring platform, and the external environmental parameters and air conditioning system operating parameters at the current office time are input to predict the energy consumption of the air conditioning system without office load in the current operating state.

[0082] S4. Divide the cooling energy consumption of office areas;

[0083] Furthermore, in this embodiment, the cooling energy consumption of the office area is specifically divided as follows:

[0084] Cooling energy consumption in office area: Cooling energy consumption in office area = actual collected air conditioning system energy consumption - air conditioning system benchmark energy consumption;

[0085] For the cooling season with office load conditions, the air conditioning system benchmark energy consumption is equal to the predicted energy consumption. For the cooling season without office load conditions, the air conditioning system benchmark energy consumption is equal to the actual collected air conditioning system energy consumption. For the non-cooling season conditions, the air conditioning system benchmark energy consumption is equal to the actual collected air conditioning system energy consumption.

[0086] The cooling energy consumption of the office area in the cooling season without office load conditions and in the non-cooling season is 0.

[0087] S5. Calculate the actual PUE value of the data center after excluding office electricity and office cooling energy consumption.

[0088] Furthermore, in this embodiment, the actual PUE value of the data center after excluding office electricity and office cooling energy consumption is calculated as follows:

[0089] Set the actual PUE calculation formula in the data center energy consumption monitoring platform:

[0090]

[0091] In the formula, E total is the total energy consumption of the data center, E Of is the energy consumption of office area lighting and sockets, E Of. is the cooling energy consumption of the office area, E IT Energy consumption of IT equipment;

[0092] Input the cooling energy consumption of the office area calculated by the energy consumption division module, the energy consumption of IT equipment, the energy consumption of office area lighting and sockets, and the total energy consumption of the data center collected by the energy supervision platform into the above-mentioned real PUE calculation formula;

[0093] Calculate the actual PUE value of the data center after excluding office electricity and office cooling energy consumption.

[0094] In another embodiment of the present application, a data center in Guangdong Province is taken as an example to illustrate a method for reducing the PUE of a data center containing an office area in the above embodiment. The data center has currently undergone energy-saving transformation and installed an air-conditioning energy-saving control system and an energy supervision platform. The platform can monitor and record local meteorological environmental parameters, various operating parameters of each subsystem of the data center (air-conditioning system, terminal system, etc.), energy consumption of each subsystem and other data in real time, so as to realize real-time monitoring and intelligent control of the PUE of the data center. The data center contains a comprehensive office area, which uses a centralized air-conditioning system for cooling together with the computer room. Since the energy consumption of the centralized air-conditioning system has not been divided at present, during the cooling season, the cooling energy consumption of the office area will lead to high energy consumption of the air-conditioning system, which will in turn lead to high PUE of the data center.

[0095] The specific steps are as follows:

[0096] Step 1: Collect external environmental parameters of the data center including outdoor dry-bulb temperature, outdoor wet-bulb temperature, and outdoor relative humidity; collect operating parameters of the data center air-conditioning system including chilled water supply and return water temperature, cooling water supply and return water temperature, chilled water supply and return water pressure difference, cooling water supply and return water pressure difference, chilled water flow rate, cooling water flow rate, host evaporation pressure, host condensation pressure, and host load rate; collect various energy consumption data of the data center including total power consumption of the data center, power consumption of the air-conditioning system, power consumption of office area lighting and sockets, and power consumption of IT equipment; the time range for data collection is from 0:00 on July 1, 2022 to 23:00 on December 31, 2022.

[0097] Step 2: Preprocessing, including the following steps:

[0098] The quartile method is used to detect abnormal values ​​of data at different times. The specific formula is:

[0099] IQR i =Q 3, -Q 1,

[0100] T u,i =Q 3, +1.5*IQR i

[0101] T d,i =Q 1, -1.5*QR i

[0102] In the formula, Q 3, is the 75% quantile value of each attribute data in the data set at time i after being arranged from small to large; Q 1, It is the 25% quantile value of each attribute data in the data set at time i after being arranged from small to large; IQR i is the interquartile range of the data at time i; Tu,i is the upper limit of abnormal threshold at time i; T d,i is the lower limit of the abnormal threshold at time i; the judgment standard of the abnormal value is that the sample value is less than the lower limit of the abnormal threshold at the time it belongs to or greater than the upper limit of the abnormal threshold at the time it belongs to; finally, the detected abnormal value is eliminated.

[0103] Feature selection: The key influencing factors of the power consumption of the air-conditioning system under the condition of excluding office workload in the cooling season are selected using the Pearson correlation coefficient as the input parameters of the cooling system benchmark model under this condition for subsequent modeling. The key influencing factors are factors with an absolute value of the Pearson correlation coefficient with the power consumption of the air-conditioning system greater than 0.4. Finally, the outdoor dry-bulb temperature, outdoor wet-bulb temperature, chilled water supply temperature, and cooling water return temperature are selected as the cooling system benchmark model inputs.

[0104] Step 3: Using the historical data of air conditioning system power consumption and related characteristic variables in the cooling season without office load conditions as the input of the cooling system benchmark model, the nonlinear regression method is used to construct the cooling system benchmark model of the data center centralized air conditioning system (without office area cooling), specifically:

[0105] The CatBoost algorithm is used to establish a cooling system benchmark model for the centralized air-conditioning system in the data center (without cooling in the office area). Grid search and ten-fold cross validation are used to find the optimal hyperparameters of the cooling system benchmark model. The grid search range is set as follows: the number of iterations is 50-2000, the step size is 50, the learning rate is 0.001-0.1, the step size is 0.001, the maximum tree depth is 3-16, the step size is 1, the sample sampling ratio is 0.6-1, the step size is 0.1, and the column sampling ratio is 0.6-1, the step size is 0.1. The grid search result is: the number of iterations is 950, the learning rate is 0.012, the maximum tree depth is 8, the sample sampling ratio is 0.8, and the column sampling ratio is 0.9.

[0106] Step 4: Taking the data of the machine building on August 1 as an example, predict the energy consumption without office load in the office working condition during the cooling season. The results are as follows: Figure 3 As shown in the figure, the difference between the measured energy consumption of the air conditioning system during the working hours and the benchmark energy consumption of the air conditioning system is the cooling energy consumption of the office area;

[0107] Step 5: Cooling energy consumption E in office area Of. Calculation results and office area lighting and socket collection energy consumption E Of like Figure 4 shown.

[0108] Step 6: Calculate the actual PUE value of the data center after excluding office electricity and office cooling energy consumption. The result is as follows: Figure 5 shown.

[0109] like Figure 6 As shown, in another embodiment of the present application, a PUE reduction system for a data center including an office area is provided, the system comprising a data acquisition module, an energy consumption division module and a PUE calculation module;

[0110] The data acquisition module is used to collect outdoor meteorological environment parameters, cold source system operating parameters and data center energy consumption data;

[0111] The energy consumption division module is used to pre-process the collected data, including dividing the working conditions, removing outliers and interpolating missing values, and selecting the characteristics of cold source energy consumption; using the pre-processed data as the cooling system benchmark model input, using the nonlinear regression method, constructing the cooling system benchmark model without office load in the cooling season, and predicting the energy consumption without office load under the office working conditions in the cooling season; dividing the cooling energy consumption of the office area;

[0112] The PUE calculation module is used to calculate the real PUE value of the data center after excluding office electricity and office cooling energy consumption.

[0113] It should be noted here that the system provided in the above embodiment is only illustrated by the division of the above functional modules. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure is divided into different functional modules to complete all or part of the functions described above. The system is a PUE reduction method for a data center containing an office area applied to the above embodiment.

[0114] like Figure 7 As shown, in another embodiment of the present application, a storage medium is further provided, storing a program, and when the program is executed by a processor, a PUE reduction method for a data center including an office area in the above embodiment is implemented, specifically:

[0115] Collect outdoor meteorological environment parameters, cold source system operating parameters and data center energy consumption data;

[0116] Preprocess the collected data, including dividing the working conditions, removing outliers and interpolating missing values, and selecting the characteristics of cold source energy consumption;

[0117] The preprocessed data is used as the input of the cooling system benchmark model. The nonlinear regression method is used to construct the cooling system benchmark model without office load in the cooling season, and the energy consumption without office load under office working conditions in the cooling season is predicted.

[0118] Divide the cooling energy consumption of office areas;

[0119] Calculate the actual PUE value of the data center after excluding office electricity and office cooling energy consumption.

[0120] It should be understood that the various parts of the present application can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0121] The above embodiments are preferred implementation modes of the present invention, but the implementation modes of the present invention are not limited to the above embodiments. Any other changes, modifications, substitutions, combinations, and simplifications that do not deviate from the spirit and principles of the present invention should be equivalent replacement methods and are included in the protection scope of the present invention.

Claims

1. A PUE reduction method for data centers with office areas. It is characterized in that The steps include: Collect outdoor meteorological environment parameters, cold source system operating parameters and data center energy consumption data; Preprocess the collected data, including dividing the working conditions, removing outliers and interpolating missing values, and selecting the characteristics of cold source energy consumption; Using the preprocessed data as input, a nonlinear regression method is used to construct a benchmark model of the cooling system without office load in the cooling season, and to predict the energy consumption without office load under office conditions in the cooling season. Divide the cooling energy consumption of office areas; Calculate the actual PUE value of the data center after excluding office electricity and office cooling energy consumption.

2. According to claim 1, a PUE reduction method for a data center including an office area, It is characterized in that The collection of outdoor meteorological environment parameters, cold source system operating parameters and data center energy consumption data is specifically: The data center external environmental parameters collected include outdoor dry-bulb temperature, outdoor wet-bulb temperature, and outdoor relative humidity; Collect the operating parameters of the data center air conditioning system, including chilled water supply and return water temperature, cooling water supply and return water temperature, chilled water supply and return water pressure difference, cooling water supply and return water pressure difference, chilled water flow, cooling water flow, host evaporation pressure, host condensation pressure and host load rate; The energy consumption data collected from the data center include the total power consumption of the data center, the power consumption of the air conditioning system, the power consumption of office area lighting and sockets, and the power consumption of IT equipment.

3. According to the method for reducing PUE of a data center including an office area according to claim 1, It is characterized in that The division conditions are specifically as follows: The historical data of the data center are divided into cooling season with office load conditions, cooling season without office load conditions, and non-cooling season conditions according to actual office conditions; the cooling season is April to October, and the remaining months are non-cooling seasons; the office time period is 8:00-21:00, and the remaining time period is no office load time.

4. According to claim 1, a PUE reduction method for a data center including an office area, It is characterized in that The data outlier removal is specifically as follows: The quartile method is used for outlier detection, specifically: IQR i JQ 3,i -Q 1,i T u,i JQ 3,i +1.5*IQR i T d,i JQ 1,i -1.5*IQR o In the formula, Q 3,i is the 75% quantile value of each attribute data in the data set at time i after being arranged from small to large; Q 1,i It is the 25% quantile value of each attribute data in the data set at time i after being arranged from small to large; IQR i is the interquartile range of the data at time i; T u,i is the upper limit of abnormal threshold determination at time i; T d,i is the lower limit of the abnormal threshold at time i; the abnormal value judgment standard is that the sample value is less than the lower limit of the abnormal threshold at the time it belongs to or greater than the upper limit of the abnormal threshold at the time it belongs to; Remove detected outliers.

5. According to the method for reducing PUE of a data center including an office area according to claim 1, It is characterized in that The cold source energy consumption characteristic selection is specifically: By using correlation analysis, key influencing factors of air-conditioning system power consumption whose absolute values ​​of Pearson correlation coefficients are greater than preset values ​​are selected as input parameters.

6. According to claim 1, a PUE reduction method for a data center including an office area, It is characterized in that The nonlinear regression method is used to construct a cooling system benchmark model without office load in the cooling season, specifically: The CatBoost algorithm is used to build a benchmark model of the cooling system without office load in the cooling season, and grid search and ten-fold cross validation are used to find the optimal hyperparameters.

7. According to claim 1, a PUE reduction method for a data center including an office area, It is characterized in that The cooling energy consumption of the office area is specifically divided as follows: Cooling energy consumption in office area: Cooling energy consumption in office area = actual collected air conditioning system energy consumption - air conditioning system benchmark energy consumption; For the cooling season with office load conditions, the air conditioning system benchmark energy consumption is equal to the predicted energy consumption. For the cooling season without office load conditions, the air conditioning system benchmark energy consumption is equal to the actual collected air conditioning system energy consumption. For the non-cooling season conditions, the air conditioning system benchmark energy consumption is equal to the actual collected air conditioning system energy consumption. The cooling energy consumption of the office area in the cooling season without office load conditions and in the non-cooling season is 0.

8. According to claim 1, a PUE reduction method for a data center including an office area, It is characterized in that The actual PUE value of the data center after excluding office electricity and office cooling energy consumption is calculated as follows: Set the actual PUE calculation formula in the data center energy consumption monitoring platform: In the formula, E total is the total energy consumption of the data center, E Of is the energy consumption of office area lighting and sockets, E Of.cold is the cooling energy consumption of the office area, E IT Energy consumption of IT equipment; Input the cooling energy consumption of the office area calculated by the energy consumption division module, the energy consumption of IT equipment, the energy consumption of office area lighting and sockets, and the total energy consumption of the data center collected by the energy supervision platform into the above-mentioned real PUE calculation formula; Calculate the actual PUE value of the data center after excluding office electricity and office cooling energy consumption.

9. A PUE reduction system for data centers with office areas. It is characterized in that A PUE reduction method for a data center including an office area, applied to any one of claims 1-8, comprising a data acquisition module, an energy consumption division module and a PUE calculation module; The data acquisition module is used to collect outdoor meteorological environment parameters, cold source system operating parameters and data center energy consumption data; The energy consumption division module is used to pre-process the collected data, including dividing the working conditions, eliminating abnormal values ​​and interpolating missing values, and selecting the cold source energy consumption characteristics; Using the preprocessed data as input, the nonlinear regression method is used to construct a cooling system benchmark model without office load in the cooling season, and the energy consumption without office load under office conditions in the cooling season is predicted; the cooling energy consumption of office areas is divided; The PUE calculation module is used to calculate the real PUE value of the data center after excluding office electricity and office cooling energy consumption.

10. A storage medium storing a program, Features: When the program is executed by a processor, a PUE reduction method for a data center including an office area as described in any one of claims 1 to 8 is implemented.

Citation Information

Cited By

  • Data center PUE simulation method and platform based on dynamic load response

    CN121615381A