Co2 emission amount estimation system and co2 emission amount estimation method
The CO2 emission estimation system uses machine learning to estimate emissions based on exhaust temperature and power consumption, addressing the cost and disruption issues of existing methods, enabling efficient and accurate emissions tracking and air conditioning management.
Patent Information
- Application Number
- JP2024027281
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2024-02-27
- Publication Date
- 2025-09-08
AI Technical Summary
Existing methods for determining power consumption and CO2 emissions of individual information processing devices in data centers are costly, disruptive, or restricted, making it difficult to accurately and inexpensively estimate emissions for multiple users.
A CO2 emission estimation system using machine learning models trained with exhaust temperature and power consumption data to estimate emissions, allowing for non-invasive installation and low-cost calculation of emissions without disrupting data center operations.
Enables inexpensive and easy estimation of CO2 emissions for multiple data center users by leveraging existing infrastructure, facilitating accurate emissions tracking and optimal air conditioning management.
Smart Images
Figure 2025130246000001_ABST
Abstract
Description
[Technical Field]
[0001] The present invention relates to a CO2 emission amount estimation system and a CO2 emission amount estimation method. [Background technology]
[0002] In recent years, investors and financial institutions have increasingly called on businesses to track and disclose their greenhouse gas (GHG) emissions. The Task Force on Climate-related Financial Disclosures (TCFD) has also strongly urged businesses to disclose GHG emissions. Given this background, some believe that data center operators, who have management authority, should be responsible for GHG emissions from power-consuming equipment within their data centers. Therefore, data center operators should disclose the power consumption figures for their users' information processing equipment. Currently, less than 10% of users measure power consumption, and measurement methods are not widely known. Furthermore, even if power consumption is apportioned based on the total power consumption of the data center, such as the number of information processing devices, it is impossible to accurately determine the power consumption of each individual data center user.
[0003] A known example of this type of technology is the facility maintenance management support system described in Patent Document 1. This facility maintenance management support system uses multiple internal temperature sensors to individually detect the temperatures of the internal components of electric power equipment, uses an external temperature sensor to detect the outside air temperature of the electric power equipment, calculates a temperature pattern that eliminates the influence of the outside air temperature from the temperature of each component, and checks the value of the temperature pattern to monitor the possibility of a fire. [Prior art documents] [Patent documents]
[0004] [Patent Document 1] Patent No. 7340902 Summary of the Invention [Problem to be solved by the invention]
[0005] Even if the technology of Patent Document 1 described above is used, it is not possible to acquire the amount of power consumed by each information processing device in a DC.
[0006] The following methods are envisioned for determining the power consumption of individual information processing devices within a DC using existing technology. The first method is to incorporate dedicated hardware for power measurement into information processing equipment. For example, by incorporating a board that measures power consumption into a server device, it is possible to grasp the power consumption. However, since a dedicated server device is used, there is a problem that the introduction cost is high when replacing an already used server device.
[0007] The second method is to introduce power measurement equipment. Power distribution units (PDUs) that can measure power consumption are widely used as power measurement equipment, but when a data center operator introduces a measuring device such as a PDU, they must turn off the power to all information processing equipment within the data center and shut down the systems of DC users. Another issue is that installing PDUs or other devices throughout the data center requires construction work, which is expensive.
[0008] The third method is to install dedicated software on server equipment. However, there are restrictions on the server equipment on which dedicated software can be installed, and it is not always possible to install dedicated software on a DC user's server equipment. If a server equipment on which dedicated software cannot be installed is being used, there is a problem that the introduction costs for replacing the server equipment will be high.
[0009] The present invention has been made in consideration of the above-mentioned problems, and aims to provide a CO2 emission estimation system and a CO2 emission estimation method that can inexpensively and easily calculate CO2 emissions for information processing devices of a large number of data center users installed within a data center. [Means for solving the problem]
[0010] (1) One aspect of the present invention is a CO2 emission estimation system including: a power consumption detection unit that detects power consumption information indicating the power consumption of a target information processing device among multiple information processing devices arranged in a data center facility; an exhaust temperature detection unit that detects exhaust temperature information indicating the exhaust temperature of the information processing device; a learning unit that trains a machine learning model to estimate the power consumption information when the exhaust temperature information is input; and an estimation unit that inputs the exhaust temperature information detected by the exhaust temperature detection unit to the machine learning model when estimating the CO2 emission of the information processing device, and estimates the CO2 emission of the information processing device based on the power consumption information estimated by the machine learning model.
[0011] (2) In one aspect of the present invention, a control unit may be provided that controls an air conditioning unit that performs air conditioning in the data center facility based on exhaust temperature information detected by the exhaust temperature detection unit when estimating the CO2 emissions of the information processing device.
[0012] (3) In one aspect of the present invention, the learning unit may train the machine learning model to estimate the exhaust temperature information when the power consumption information is input, the estimation unit may input the power consumption information detected by the power consumption detection unit to the machine learning model and estimate the exhaust temperature information using the machine learning model, and the control unit may control the air conditioning device based on the exhaust temperature information estimated by the estimation unit.
[0013] (4) In one aspect of the present invention, the estimation unit may estimate the CO 2 emissions of the information processing device based on power consumption information detected by the power consumption detection unit.
[0014] (5) In one aspect of the present invention, the information processing device may include a usage rate detection unit that detects usage rate information indicating CPU usage rate of the information processing device, and the learning unit may train the machine learning model to estimate the power consumption information when the usage rate information is input. The estimation unit may input the usage rate information detected by the usage rate detection unit to the machine learning model when estimating the CO2 emissions of the information processing device, and estimate the CO2 emissions of the information processing device based on the power consumption information estimated by the machine learning model.
[0015] (6) One aspect of the present invention includes a utilization detection unit that detects utilization information indicating CPU utilization of the information processing device, wherein the learning unit trains the machine learning model to estimate the exhaust temperature information when the utilization information is input, and the estimation unit inputs the utilization information detected by the utilization detection unit to the machine learning model when controlling air conditioning in the data center facility, and estimates the exhaust temperature information using the machine learning model, and the control unit controls the air conditioning device based on the exhaust temperature information estimated by the estimation unit.
[0016] (7) One aspect of the present invention is an estimation method including the steps of: detecting power consumption information indicating the power consumption of a target information processing device among a plurality of information processing devices arranged in a data center facility; detecting exhaust temperature information indicating the exhaust temperature of the information processing device; training a machine learning model to estimate the power consumption information when the exhaust temperature information is input; and inputting exhaust temperature information detected when estimating CO2 emissions from the information processing device into the machine learning model, and estimating the CO2 emissions of the information processing device based on the power consumption information estimated by the machine learning model. [Effects of the Invention]
[0017] According to one aspect of the present invention, it is possible to inexpensively and easily calculate CO2 emissions for information processing devices of many data center users installed in a data center. [Brief explanation of the drawings]
[0018] [Figure 1] 1 is a block diagram showing an example of the configuration of a CO2 emission amount estimation system 1 according to an embodiment. [Figure 2] FIG. 2 is a perspective view showing an example of an information processing device 210 and a rack 200a according to the embodiment. [Figure 3] 2 is a schematic diagram showing an example of an information processing device 210 and a rack 200a according to an embodiment. FIG. [Figure 4] 1 is a block diagram illustrating an example of a data acquisition system according to an embodiment. [Figure 5] FIG. 10 is a diagram illustrating an example of a machine learning model for the relationship between power consumption, exhaust heat temperature, and CPU utilization rate in the embodiment. [Figure 6] FIG. 10 is a diagram showing the relationship between the actual measured value and the predicted value of power consumption according to a machine learning model (A) in the embodiment. [Figure 7] FIG. 2 is a diagram illustrating an example of a use case of the CO 2 emission amount estimation system 1 according to the embodiment. [Figure 8] FIG. 10 is a diagram illustrating another example of a use case of the CO 2 emission amount estimation system 1 according to the embodiment. [Figure 9] FIG. 10 is a diagram illustrating another example of a use case of the CO 2 emission amount estimation system 1 according to the embodiment. [Figure 10] 10 is a flowchart illustrating an example of a model creation process according to an embodiment. [Figure 11] 10 is a flowchart showing an example of a process for calculating CO2 in the embodiment. [Figure 12] 10 is a flowchart illustrating an example of data preprocessing according to an embodiment. DETAILED DESCRIPTION OF THE INVENTION
[0019] The CO2 emission estimation system and the CO2 emission estimation method to which the present invention is applied estimate the CO2 (carbon dioxide) emission amount for a target information processing device among a plurality of information processing devices arranged in a data center. This estimation system and estimation method estimate the CO2 emission amount of an information processing device for each data center user by using the facilities installed in the data center in a situation where, for example, an information processing device owned by a data center user is arranged in the data center, and present data for visualizing the estimated CO2 emission amount to the data center user. The estimation system and estimation method enable the data center user to easily and at low cost confirm the CO2 emission amount of the information processing device that the user owns. Hereinafter, the estimation system and estimation method to which the present invention is applied will be described.
[0020] <Configuration of CO2 Emission Estimation System 1> First, the overall configuration of the CO2 emission estimation system 1 in the embodiment will be described. FIG. 1 is a block diagram showing a configuration example of the CO2 emission estimation system 1 in the embodiment. The CO2 emission estimation system 1 includes, for example, an arithmetic unit 100, a data center facility 200, an air conditioner 220, and a data center user terminal 300. The arithmetic unit 100, the data center facility 200, and the data center user terminal 300 are connected to, for example, a communication network. Each device connected to the communication network includes a communication interface such as a NIC (Network Interface Card) or a wireless communication module (not shown in FIG. 1). The communication network includes, for example, the Internet, a WAN (Wide Area Network), a LAN (Local Area Network), a cellular network, and the like.
[0021] The computing device 100 includes an exhaust gas temperature detector 102, a power consumption detector 104, a learning unit 106, an estimation unit 108, a visualization unit 110, and a control unit 112. The functional units, such as the exhaust gas temperature detector 102, the power consumption detector 104, the learning unit 106, the estimation unit 108, the visualization unit 110, and the control unit 112, are implemented by a processor, such as a central processing unit (CPU), executing a program stored in a program memory. Some or all of these functional units may be implemented by hardware, such as a large-scale integration (LSI), an application-specific integrated circuit (ASIC), or a field-programmable gate array (FPGA), or may be implemented by a combination of software and hardware. The exhaust gas temperature detector 102, the power consumption detector 104, the learning unit 106, the estimation unit 108, the visualization unit 110, and the control unit 112, may be integrated into a single device, but are not limited to this and may be distributed across multiple devices.
[0022] The exhaust temperature detection unit 102 and the power consumption detection unit 104 have a communication interface for the data center facility 200 and acquire signals from the data center facility 200. The power consumption detection unit 104 detects power consumption information indicating the power consumption of a target information processing device 210 among multiple information processing devices 210 arranged in the data center facility 200. The exhaust temperature detection unit 102 detects exhaust temperature information indicating the exhaust temperature of the information processing device 210.
[0023] FIG. 2 is a perspective view showing an example of an information processing device 210 and a rack 200a according to an embodiment, and FIG. 3 is a schematic diagram showing an example of an information processing device 210 and a rack 200a according to an embodiment. In the information processing device 210 and data center facility 200 according to the embodiment, racks 200a for installing a large number of information processing devices 210 are arranged. A temperature sensor 212 and a temperature sensor 214 are installed on one end side (hot aisle) and the other end side (cold aisle) of the information processing device 210 in the rack 200a, respectively. The temperature sensor 212 detects the temperature of exhaust air from the information processing device 210, and the temperature sensor 214 detects the temperature of air taken in by the information processing device 210. In FIG. 3, temperature sensors 212A and 214A are provided corresponding to the information processing device 210A, temperature sensors 212B and 214B are provided corresponding to the information processing device 210B, and temperature sensors 212C and 214C are provided corresponding to the information processing device 210C. As a result, the data center facility 200 transmits a signal indicating the temperature of each information processing device 210 to the computing device 100, and the exhaust temperature detection unit 102 can detect exhaust temperature information for each information processing device 210.
[0024] The learning unit 106 trains the machine learning model 120 to estimate power consumption information when exhaust temperature information is input. The learning unit 106 is supplied with learning data from a storage unit (not shown). The learning data is information in which the exhaust temperature of the information processing device 210 is associated with power consumption. The learning unit 106 inputs exhaust temperature information to the machine learning model 120 and obtains an output result of the machine learning model 120. The learning unit 106 learns processing parameters of the machine learning model 120 so that the information input to the machine learning model 120 becomes power consumption corresponding to the exhaust temperature, and generates learning result data. The processing parameters of the machine learning model 120 are, for example, filters (also called weights or biases) included in a neural network. The learning result data is stored in a storage unit (not shown). The learning unit 106 supplies the processing parameters to the estimation unit 108.
[0025] The estimation unit 108 inputs exhaust temperature information detected by the exhaust temperature detection unit 102 to the machine learning model 108a when estimating the CO2 emission amount of the information processing device 210, and estimates the CO2 emission amount of the information processing device 210 based on the power consumption information estimated by the machine learning model 108a. The estimation unit 108 performs processing using the machine learning model 108a with the processing parameters supplied from the learning unit 106. The estimation unit 108 supplies the estimated CO2 emission amount to the visualization unit 110.
[0026] The visualization unit 110 generates data for a customer corresponding to the information processing device 210 to visualize the CO2 emissions of the information processing device 210, and transmits the data to the data center user terminal 300. The data center user terminal 300 can grasp the CO2 emissions of the information processing device 210 based on the data received from the computing device 100.
[0027] The control unit 112 controls the air conditioner 220 that performs air conditioning in the data center facility 200 based on the exhaust temperature information detected by the exhaust temperature detection unit 102 when estimating the CO 2 emission amount of the information processing device 210 .
[0028] The data center user terminal 300 is a terminal device such as a personal computer, smartphone, or tablet device operated by a data center user. In the embodiment, the data center user is a person who installs information processing devices such as a server device in the data center facility 200 and provides web services and the like using the server device. The data center user installs their own server device in a rack installed in the data center facility 200. The data center user terminal 300 can obtain visualized data for viewing the CO2 emissions of their own server device and display the CO2 emissions.
[0029] In the data center facility 200, information processing devices of many data center users are installed. FIG. 2 is a perspective view showing an example of a rack and information processing device 210 installed in a data center facility 200 according to an embodiment, and FIG. 3 is a schematic diagram showing an example of a rack and information processing device 210 installed in a data center facility 200 according to an embodiment.
[0030] FIG. 4 is a block diagram showing an example of a data acquisition system according to an embodiment. In order to train a machine learning model, the CO2 emission estimation system 1 constructs a data acquisition system as shown in FIG. 4 and acquires training data. The data acquisition system includes, for example, temperature sensors 212A, 212B, and 212C corresponding to information processing devices 210A, 210B, and 210C, respectively, a temperature sensor 214, watt monitors 216A, 216B, and 216C, a power / temperature measuring device 230, a load measuring device 232, and a storage device 240. The temperature sensors 212A, 212B, and 212C detect the exhaust temperatures of the information processing devices 210A, 210B, and 210C, respectively, and the temperature sensor 214 detects the temperature around the information processing device 210. The watt monitors 216A, 216B, and 216C detect the power consumption of the information processing devices 210A, 210B, and 210C, respectively. The power and temperature measuring device 230 receives signals detected by the temperature sensors 212A, 212B, and 212C, the temperature sensor 214, and the watt monitors 216A, 216B, and 216C, and measures the power consumption, ambient temperature, and exhaust temperature for each information processing device 210. The load measuring device 232 measures the CPU utilization rate for each information processing device 210. The storage device 240 stores the exhaust temperature, ambient temperature, and power consumption measured by the power and temperature measuring device 230, and the CPU utilization rate measured by the load measuring device 232 for each information processing device 210 in the storage device 240. The information stored in the storage device 240 is used for learning processing by the learning unit 106.
[0031] FIG. 5 is a diagram illustrating an example of a machine learning model for the relationship between the amount of power consumption, the exhaust heat temperature, and the CPU utilization rate according to the embodiment. The CO2 emission estimation system 1 constructs a machine learning model (A) using the relationship between power consumption and exhaust heat temperature. The learning unit 106 trains the machine learning model (A) so that it estimates power consumption information when exhaust temperature information is input. FIG. 6 is a diagram showing the relationship between the actual measured value and the predicted value of power consumption by the machine learning model (A) in the embodiment. The learning unit 106 trains the machine learning model (A) so that the power consumption (prediction result) output from the machine learning model (A) tracks the time change of power consumption (actual measured value), while the time change of power consumption (actual measured value) tracks the time change of exhaust temperature (actual measured value). The learning unit 106 may train the machine learning model (A) so that it estimates exhaust temperature information when power consumption information is input.
[0032] The CO2 emission amount estimation system 1 constructs a machine learning model (B) using the relationship between power consumption and CPU utilization rate. The learning unit 106 trains the machine learning model (B) so that it estimates the CPU utilization rate when the power consumption amount is input. The learning unit 106 may also train the machine learning model (B) so that it estimates the power consumption amount when the CPU utilization rate is input.
[0033] The CO2 emission estimation system 1 constructs a machine learning model (C) using the relationship between CPU utilization rate and exhaust heat temperature. The learning unit 106 trains the machine learning model (C) so that it estimates the exhaust heat temperature when the CPU utilization rate is input. The learning unit 106 may also train the machine learning model (C) so that it estimates the CPU utilization rate when the exhaust heat temperature is input.
[0034] FIG. 7 is a diagram showing an example of a use case of the CO 2 emission amount estimation system 1 according to the embodiment. The exhaust gas temperature detection unit 102 detects the exhaust gas heat temperature, and the estimation unit 108 inputs the detected exhaust gas heat temperature into a machine learning model (A) and acquires an estimated value of the power consumption based on the output of the machine learning model (A). The estimation unit 108 performs a process of calculating the CO2 emissions from the estimated power consumption. Furthermore, the control unit 112 controls the air conditioner 220 based on the exhaust gas heat temperature detected by the exhaust gas temperature detection unit 102.
[0035] FIG. 8 is a diagram showing another example of a use case of the CO 2 emission amount estimation system 1 according to the embodiment. The power consumption detection unit 104 detects the amount of power consumption, and the estimation unit 108 inputs the detected amount of power consumption into a machine learning model (A) and acquires an estimated value of the exhaust heat temperature based on the output of the machine learning model (A). The estimation unit 108 performs a process to calculate the amount of CO2 emissions from the estimated exhaust heat temperature. Furthermore, the control unit 112 calculates the amount of CO2 emissions based on the amount of power consumption detected by the power consumption detection unit 104.
[0036] FIG. 9 is a diagram showing another example of a use case of the CO 2 emission amount estimation system 1 according to the embodiment. The CO2 emission estimation system 1 may include a usage rate detection unit that detects usage rate information indicating the CPU usage rate of the information processing device 210, and the learning unit 106 may train the machine learning model 120 to estimate power consumption information when usage rate information is input. The CO2 emission estimation system 1 detects the CPU usage rate, and the estimation unit 108 inputs the usage rate information detected by the usage rate detection unit when estimating the CO2 emission amount of the information processing device 210 to the machine learning model 120, and estimates the CO2 emission amount of the information processing device 210 based on the power consumption information estimated by the machine learning model 120.
[0037] Furthermore, the learning unit 106 may train the machine learning model (C) so as to estimate exhaust temperature information when utilization rate information is input. The estimation unit 108 inputs utilization rate information detected by the utilization rate detection unit to the machine learning model (C) when controlling the air conditioning of the data center facility 200, and estimates exhaust temperature information using the machine learning model (C). The control unit 112 controls the air conditioner 220 based on the exhaust temperature information estimated by the estimation unit 108.
[0038] FIG. 10 is a flowchart showing an example of a model creation process according to the embodiment. When creating the machine learning model 120, the exhaust temperature detection unit 102 measures the exhaust heat temperature of the information processing device 210 (step S100), and the power consumption detection unit 104 measures the amount of power consumption of the information processing device 210 (step S102). The storage device 240 accumulates the measured exhaust heat temperature information and power consumption information (step S104). The learning unit 106 trains the machine learning model 120 using the data accumulated in the storage device 240 (step S106). The learning unit 106 updates the machine learning model 108a in the estimation unit 108 based on the processing parameters of the trained machine learning model 120 (step S108).
[0039] FIG. 11 is a flowchart showing an example of a process for calculating CO2 according to the embodiment. First, the exhaust heat temperature detection unit 102 measures exhaust heat temperature information of the information processing device 210 (step S200). The estimation unit 108 performs data preprocessing (step S202), inputs the preprocessed exhaust heat temperature information to the machine learning model 108a, calculates power consumption using the machine learning model 108a (step S204), and calculates CO2 emissions based on the calculated power consumption (step S206). The visualization unit 110 creates data for visualizing the exhaust heat temperature information, power consumption information, and CO2 emission information for each information processing device 210, and transmits the created data to the data center user terminal 300 corresponding to the information processing device 210 (step S208).
[0040] FIG. 12 is a flowchart showing an example of data preprocessing according to the embodiment. First, the exhaust gas temperature detection unit 102 acquires temperature data from the temperature sensor 212 (step S300), and the estimation unit 108 standardizes the exhaust gas temperature information (step S302). The standardization process for the exhaust gas temperature information involves, for example, setting the average value to 0 and converting the values into values dispersed in the range of 0 to 1. The estimation unit 108 determines whether the standardized exhaust gas temperature is higher than the minimum temperature (step S304) and whether the standardized exhaust gas temperature is lower than the maximum temperature (step S306). The minimum temperature and maximum temperature are values for eliminating outliers in the detected values, and are set based on the specifications of the temperature sensor 212 and the information processing device 210. If the standardized exhaust gas temperature is higher than the minimum temperature or lower than the minimum temperature (step S304: YES, step S306: YES), the estimation unit 108 performs the processes from step S300 onward. If the standardized exhaust temperature is not higher than the minimum temperature and not lower than the minimum temperature (step S304: NO, step S306: NO), the estimation unit 108 proceeds to calculate the power consumption amount using the standardized exhaust temperature.
[0041] (Effects of the embodiment) As described above, according to the CO2 emission estimation system 1 of the embodiment, the machine learning model 120 is trained to estimate power consumption information when exhaust temperature information is input, the exhaust temperature information is input to the machine learning model 108a when estimating the CO2 emission amount of the information processing device 210, and the CO2 emission amount of the information processing device 210 can be estimated based on the power consumption information estimated by the machine learning model 108a. This CO2 emission estimation system 1 makes it possible to inexpensively and easily calculate the CO2 emissions for the information processing devices 210 of many data center users installed in the data center facility 200. In other words, according to the CO2 emission estimation system 1, by installing the temperature sensors 212 without touching the server devices of the data center users, it is possible to inexpensively estimate the CO2 emissions for each data center user regardless of the configuration of the information processing device 210.
[0042] Furthermore, the CO2 emission estimation system 1 can control the air conditioning devices that perform air conditioning in the data center facility 200 based on exhaust temperature information when estimating the CO2 emission amount of the information processing device 210. As a result, the CO2 emission estimation system 1 can perform air conditioning in the data center facility 200 based on the exhaust temperature detected when estimating the CO2 emission amount of the information processing device 210, and can maintain the inside of the data center facility 200 at an optimal temperature in conjunction with the operation of the information processing device 210.
[0043] Furthermore, according to the CO2 emission estimation system 1, the machine learning model 120 is trained to estimate exhaust temperature information when power consumption information is input, the power consumption information is input to the machine learning model 108a, and the exhaust temperature information is estimated by the machine learning model 108a, thereby making it possible to control the air conditioner based on the estimated exhaust temperature information. In this way, if the data center facility 200 already has a configuration for detecting the power consumption of the information processing device 210, the CO2 emission estimation system 1 can maintain an optimal temperature inside the data center facility 200 in conjunction with the power consumption of the information processing device 210. Furthermore, according to the CO2 emission estimation system 1, it is possible to estimate the CO2 emissions of the information processing device 210 based on the power consumption information.
[0044] Furthermore, according to the CO2 emission estimation system 1, the machine learning model 120 is trained to estimate power consumption information when a CPU utilization rate is input, and when estimating the CO2 emission amount of the information processing device 210, the CPU utilization rate is input to the machine learning model 108a, and the CO2 emission amount of the information processing device 210 can be estimated based on the power consumption information estimated by the machine learning model 108a. According to the CO2 emission estimation system 1, if a configuration for detecting the CPU utilization rate is already in place, the CO2 emission amount can be easily estimated based on the detected CPU utilization rate.
[0045] Furthermore, according to the CO2 emission amount estimation system 1, the machine learning model 120 is trained to estimate exhaust temperature information when CPU utilization rate information is input, and when controlling the air conditioning of the data center facility 200, the CPU utilization rate is input to the machine learning model 108a, the machine learning model 108a estimates the exhaust temperature, and the air conditioner 220 can be controlled based on the estimated exhaust temperature. According to the CO2 emission amount estimation system 1, if the data center facility 200 already has a configuration for detecting CPU utilization rate, it is possible to easily control the air conditioning based on the detected CPU utilization rate.
[0046] Although each embodiment and each variant have been described, these are merely examples and are not intended to limit the scope of the present invention. For example, one aspect of the present invention may be realized by combining any of the embodiments or variants, or a part of each embodiment or a part of each variant, with one or more other embodiments or one or more other variants.
[0047] In addition, the various processes described above related to the CO2 emission estimation system 1 may be performed by recording a program for executing each process of the CO2 emission estimation system 1 in this embodiment on a computer-readable recording medium, and having a computer system read and execute the program recorded on the recording medium.
[0048] Note that the term "computer system" here may include hardware such as the OS and peripheral devices. Furthermore, if a WWW system is used, the term "computer system" also includes the homepage provision environment (or display environment). Furthermore, "computer-readable recording media" refers to storage devices such as flexible disks, magneto-optical disks, ROMs, and writable non-volatile memory such as flash memory, portable media such as CD-ROMs, and hard disks built into computer systems.
[0049] Furthermore, the term "computer-readable recording medium" also includes a storage medium that stores a program for a certain period of time, such as a volatile memory (e.g., DRAM (Dynamic Random Access Memory)) within a computer system that serves as a server or client when the program is transmitted via a network such as the Internet or a communication line such as a telephone line. The program may also be transmitted from a computer system that stores the program in a storage device or the like to another computer system via a transmission medium or by transmission waves in the transmission medium.
[0050] Here, the "transmission medium" for transmitting the program refers to a medium that has the function of transmitting information, such as a network (communication network) such as the Internet or a communication line (communication line) such as a telephone line. The program may also be one that realizes part of the above-mentioned functions. Furthermore, it may be a so-called differential file (differential program) that can realize the above-mentioned functions in combination with a program already recorded in a computer system.
[0051] Although the embodiments of the present invention have been described in detail above with reference to the drawings, the specific configuration is not limited to this embodiment, and the present invention also includes designs within the scope of the gist of the present invention. [Explanation of symbols]
[0052] 1. Emissions estimation system 100 Computing equipment 102 Exhaust temperature detector 104 Power consumption detector 106 Learning Department 108a, 120 Machine Learning Models 108 Estimation part 110 Visualization section 112 Control section 200 Data Center Facilities 200 200a rack 210, 210A, 210B, 210C Information processing device 212, 212A, 212B, 212C, 214, 214A, 214B, 214C Temperature Sensors 216A, 216B, 216C Watt Monitor 220 Air conditioning equipment 230 Power and temperature measurement device 232 Load Measuring Device 240 Storage device 300 Data Center User Terminals
Claims
1. a power consumption detection unit that detects power consumption information indicating power consumption of a target information processing device among a plurality of information processing devices arranged in a data center facility; an exhaust temperature detection unit that detects exhaust temperature information indicating an exhaust temperature of the information processing device; a learning unit that trains a machine learning model so as to estimate the power consumption information when the exhaust gas temperature information is input; an estimation unit that inputs exhaust temperature information detected by the exhaust temperature detection unit into the machine learning model when estimating the CO2 emission amount of the information processing device, and estimates the CO2 emission amount of the information processing device based on power consumption information estimated by the machine learning model; A CO2 emission estimation system comprising:
2. 2. The CO2 emission estimation system of claim 1, further comprising a control unit that controls an air conditioning device that performs air conditioning in the data center facility based on exhaust temperature information detected by the exhaust temperature detection unit when estimating the CO2 emission amount of the information processing device.
3. the learning unit trains the machine learning model to estimate the exhaust gas temperature information when the power consumption information is input; the estimation unit inputs the power consumption information detected by the power consumption detection unit into the machine learning model, and causes the machine learning model to estimate exhaust gas temperature information; The CO2 emission estimation system according to claim 2 , wherein the control unit controls the air conditioning device based on the exhaust gas temperature information estimated by the estimation unit.
4. The CO2 emission estimation system according to claim 3 , wherein the estimation unit estimates the CO2 emission amount of the information processing device based on the power consumption information detected by the power consumption detection unit.
5. a utilization rate detection unit that detects utilization rate information indicating a CPU utilization rate of the information processing device; the learning unit causes the machine learning model to learn so as to estimate the power consumption information when the utilization rate information is input; the estimation unit inputs usage rate information detected by the usage rate detection unit into the machine learning model when estimating the CO2 emissions of the information processing device, and estimates the CO2 emissions of the information processing device based on power consumption information estimated by the machine learning model. The CO2 emission estimation system according to claim 1 .
6. a utilization rate detection unit that detects utilization rate information indicating a CPU utilization rate of the information processing device; the learning unit trains the machine learning model to estimate the exhaust temperature information when the usage rate information is input; the estimation unit inputs utilization rate information detected by the utilization rate detection unit into the machine learning model when controlling the air conditioning of the data center facility, and estimates exhaust gas temperature information using the machine learning model; The CO2 emission estimation system according to claim 2 , wherein the control unit controls the air conditioning device based on the exhaust gas temperature information estimated by the estimation unit.
7. detecting power consumption information indicating power consumption of a target information processing device among a plurality of information processing devices arranged in a data center facility; detecting exhaust temperature information indicating an exhaust temperature of the information processing device; a step of training a machine learning model to estimate the power consumption information when the exhaust gas temperature information is input; inputting exhaust temperature information detected when estimating the CO2 emission amount of the information processing device into the machine learning model, and estimating the CO2 emission amount of the information processing device based on power consumption information estimated by the machine learning model; A method for estimating CO2 emissions, comprising:
Citation Information
Patent Citations
Monitoring methods for power equipment
JP7340902B1