Number of people prediction system, number of people estimation system, power consumption prediction system, number of people prediction method and program

The system uses user count prediction and power consumption models to enhance air conditioner efficiency by accurately forecasting user presence and energy needs, addressing the inaccuracy of existing methods.

JP7591906B2Active Publication Date: 2024-11-29MITSUBISHI HEAVY IND THERMAL SYST
View PDF 6 Cites 0 Cited by

Patent Information

Application Number
JP2020186358
Authority / Receiving Office
JP · JP
Patent Type
Patents
Current Assignee / Owner
Filing Date
2020-11-09
Publication Date
2024-11-29
Estimated Expiration
2040-11-09

AI Technical Summary

Technical Problem

Existing methods fail to accurately predict the number of users in an air-conditioned area, which hinders precise power demand forecasting for air conditioners, leading to inefficiencies in energy consumption.

Method used

A system that utilizes a number of people prediction model, powered by human sensor data, attendance schedules, and reservation information to estimate current and future user counts, coupled with a power consumption estimation model to predict energy usage based on user presence.

Benefits of technology

Accurately predicts user numbers and corresponding power consumption, enabling efficient operation of air conditioners and reducing energy waste.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 0007591906000001
    Figure 0007591906000001
  • Figure 0007591906000002
    Figure 0007591906000002
  • Figure 0007591906000003
    Figure 0007591906000003
Patent Text Reader

Abstract

To provide a person number prediction system that can accurately predict a user number.SOLUTION: A person number prediction system comprises: a learning unit that learns a person number estimation model which outputs a person number estimation value of users existing in a predetermined area at predetermined time, a person number prediction model which acquires person number estimation values of a plurality of pieces of time to output a person number prediction value of users existing in the area at future time, and a power consumption amount estimation model which acquires a predetermined variable number including the person number prediction value to estimate a power consumption amount of an air conditioner in the future, such that an estimation value of the power consumption amount is an actual value of the power consumption amount; and a person number prediction unit that predicts a user number in the future on the basis of the learned person number prediction model.SELECTED DRAWING: Figure 1
Need to check novelty before this filing date? Find Prior Art

Description

[Technical field]

[0001] The present disclosure relates to a number of people prediction system, a number of people estimation system, a power consumption prediction system, a number of people prediction method, and a program. [Background technology]

[0002] From the viewpoint of energy saving, it is necessary to predict the power demand of an air conditioner with high accuracy in order to operate the air conditioner efficiently with less power consumption. In general, predicted values ​​of outside temperature and days of the week are often used to predict the power demand of an air conditioner. Patent Document 1 discloses a method of estimating the number of users on a floor based on the number of people entering and leaving the floor to be air-conditioned, and estimating the load of the air conditioner using the estimated number of users in addition to the amount of solar radiation, temperature, etc. This method makes it possible to operate the air conditioner while taking into account the actual flow of people that changes from moment to moment. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] International Publication No. 2017 / 216833 Summary of the Invention [Problem to be solved by the invention]

[0004] It is believed that prediction accuracy can be improved if the number of users can be taken into consideration when predicting future power demand for air conditioners. To achieve this, it is necessary to predict the number of users in an air-conditioned area with high accuracy. However, no method for predicting the number of users with high accuracy has been provided to date. Furthermore, Patent Document 1 does not disclose any technology for predicting the future number of users and predicting future power demand.

[0005] The present disclosure aims to provide a number of people prediction system, a number of people estimation system, a power consumption prediction system, a number of people prediction method, and a program that can solve the above-mentioned problems. [Means for solving the problem]

[0006] The present disclosure relates to a number of users prediction system, and a number of users estimation model that acquires a value of a predetermined variable related to the number of users present in a predetermined area at a predetermined time and outputs an estimated value of the number of users present in the area at the time, a number of users prediction model that acquires the estimated number of users at multiple times estimated by the number of users estimation model and outputs a predicted value of the number of users present in the area at a prediction target time in the future than the multiple times, and a number of users prediction model that acquires the estimated number of users at multiple times and outputs a predicted value of the number of users present in the area at a prediction target time in the future than the multiple times, and Value a power consumption estimation model that acquires the power consumption and outputs an estimated value of the power consumption of the air conditioner during the prediction target time, a learning unit that learns the estimated value of the power consumption output by the power consumption estimation model to become an actual value of the power consumption of the air conditioner during the prediction target time, and a number of people prediction unit that predicts the number of users present in the area based on the learned number of people prediction model. The specified variables related to the number of users present in the specified area at the specified time include a detection value of the number of users present in the area by a human presence sensor, user attendance data at the specified time, a reservation schedule for the area, and a detection value of whether or not a user is present at each seat in the area.

[0007] The number of people estimation system of the present disclosure includes the learning unit, and a number of people estimation unit that estimates the number of users present in the area based on the learned number of people estimation model.

[0008] The power consumption prediction system of the present disclosure includes the learning unit, the number of people prediction unit, and a power consumption prediction unit that predicts the power consumption of the air conditioner based on the learned power consumption estimation model.

[0009] The number of users prediction method of the present disclosure includes a number of users estimation model that acquires a value of a predetermined variable related to the number of users present in a predetermined area at a predetermined time and outputs an estimated value of the number of users present in the area at the time, a number of users prediction model that acquires the estimated number of users at a plurality of times estimated by the number of users estimation model and outputs a predicted value of the number of users present in the area at a prediction target time in the future than the plurality of times, and a number of users prediction model that acquires the estimated number of users at a plurality of times estimated by the number of users estimation model and outputs a predicted value of the number of users present in the area at a prediction target time in the future than the plurality of times. Valuea step of learning a power consumption estimation model that acquires and outputs an estimated value of the power consumption of the air conditioner during the prediction target time, so that the estimated value of the power consumption output by the power consumption estimation model becomes an actual value of the power consumption of the air conditioner during the prediction target time; and a step of predicting the number of users present in the area based on the trained number-of-users prediction model. The specified variables related to the number of users present in the specified area at the specified time include a detection value by a human presence sensor of the number of users present in the area, user attendance data at the specified time, a reservation schedule for the area, and a detection value of whether or not a user is present at each seat in the area.

[0010] The program of the present disclosure includes a computer that includes a number of users estimation model that acquires a value of a predetermined variable related to the number of users present in a predetermined area at a predetermined time and outputs an estimated value of the number of users present in the area at the time, a number of users prediction model that acquires the estimated number of users at a plurality of times estimated by the number of users estimation model and outputs a predicted value of the number of users present in the area at a prediction target time in the future than the plurality of times, and a number of users prediction model that acquires the estimated number of users at a plurality of times estimated by the number of users estimation model and outputs a predicted value of the number of users present in the area at a prediction target time in the future than the plurality of times. Value a step of learning an energy consumption estimation model that acquires and outputs an estimated value of the energy consumption of the air conditioner during the prediction target time, so that the estimated value of the energy consumption output by the energy consumption estimation model becomes an actual value of the energy consumption of the air conditioner during the prediction target time; and a step of predicting the number of users present in the area based on the learned number-of-users prediction model; a process in which the predetermined variables relating to the number of users present in the predetermined area at the predetermined time include a detection value of the number of users present in the area by a human presence sensor, attendance data of the users at the predetermined time, a reservation schedule for the area, and a detection value of whether or not a user is present for each seat in the area; Execute the command. Effect of the Invention

[0011] According to the above-mentioned number of users prediction system, number of users prediction method, and program, the number of users can be predicted with high accuracy. According to the above-mentioned number of users estimation system, the number of users can be estimated with high accuracy. According to the above-mentioned power consumption prediction system, the power consumption can be predicted with high accuracy. [Brief description of the drawings]

[0012] [Figure 1] FIG. 1 is a functional block diagram illustrating an example of a prediction system according to an embodiment. [Figure 2A]FIG. 11 is a diagram showing a relationship between air-conditioning areas and the number of people according to an embodiment. [Figure 2B] FIG. 11 is a diagram showing the relationship between the number of people in an air-conditioned area and the amount of power consumption according to an embodiment. [Figure 3A] FIG. 13 is a diagram illustrating an example of a number of people estimation model according to an embodiment. [Figure 3B] FIG. 13 is a diagram illustrating an example of a number of people prediction model according to an embodiment. [Figure 3C] FIG. 4 is a diagram illustrating an example of a power consumption prediction model according to an embodiment. [Figure 4] FIG. 4 is a diagram illustrating an example of functions of a learning unit according to an embodiment. [Diagram 5] 11 is a flowchart illustrating an example of a construction process for each model according to an embodiment. [Figure 6A] 11 is a flowchart illustrating an example of a number of people estimation process according to an embodiment. [Figure 6B] 13 is a flowchart illustrating an example of a number of people prediction process according to an embodiment. [Figure 6C] 10 is a flowchart illustrating an example of a power consumption prediction process according to an embodiment. [Figure 7] FIG. 2 is a diagram illustrating an example of a hardware configuration of a prediction system according to an embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0013] <Embodiment> Hereinafter, a prediction system according to an embodiment of the present disclosure will be described with reference to FIGS. (composition) FIG. 1 is a functional block diagram illustrating an example of a prediction system according to an embodiment. The prediction system 10 predicts the future number of users present in an air-conditioned area, and predicts the future power consumption of the air conditioner based on the predicted number of users.

[0014] As shown in the figure, the prediction system 10 includes a data acquisition unit 11, a learning unit 12, a number of people estimation unit 13, a number of people prediction unit 14, a power consumption prediction unit 15, and a memory unit 16. The data acquisition unit 11 acquires parameters required for predicting the number of users and power consumption, and learning data required for creating various models used for prediction (number of users estimation model 161, number of users prediction model 162, power consumption estimation model 163). The learning unit 12 uses the learning data acquired by the data acquisition unit 11 to create each of a number of people estimation model 161, a number of people prediction model 162, and a power consumption estimation model 163. The number of people estimation model 161 is a model that acquires parameters such as a detection value detected by a human presence sensor provided in an air conditioner and a reservation schedule for use of the air-conditioned area, and estimates the number of users present in the air-conditioned area at the same time as the time targeted by these parameters. The number of people prediction model 162 is a model that predicts the future number of users present in the air-conditioned area when the number of users present in the air-conditioned area at one or more past times is acquired. The power consumption estimation model 163 is an estimation model that acquires parameters that affect the power consumption of the air conditioner, such as the number of users present in the air-conditioned area, and estimates the power consumption consumed by the air conditioner at the same time as the time targeted by these parameters. The number of people estimation unit 13 estimates the number of users present in the air-conditioned area based on the detection value of the human presence sensor and the like and the number of people estimation model 161. The number of people prediction unit 14 predicts the number of users present in the air-conditioned area at a time in the future than the time at which the number of people was estimated, based on the number of people at one or more times estimated by the number of people estimation unit 13 and the number of people prediction model 162. The power consumption prediction unit 15 predicts the amount of power consumption that the air conditioner will use for air conditioning at that time based on the future number of users predicted by the number-of-users prediction unit 14 and the power consumption estimation model 163. The storage unit 16 stores the parameters and learning data acquired by the data acquisition unit 11, a number of people estimation model 161, a number of people prediction model 162, a power consumption estimation model 163, and the like.

[0015] Next, the relationship between the number of people in the air-conditioned area and the amount of power consumed by the air conditioner will be described with reference to FIGS. 2A and 2B. FIG. 2A is a diagram showing the relationship between air-conditioned areas and the number of people according to one embodiment. Air-conditioned areas 1A, 1B, 1C, and 1D are spaces that are conditioned by air conditioners (indoor units) 2A, 2B, 2C, and 2D, respectively. Air-conditioned areas 1A, 1B, 1C, and 1D have 9, 7, 10, and 4 users, respectively. Figure 2B shows an example of the relationship between the number of people and the amount of power consumed by an air conditioner. The vertical axis of the graph shown in FIG. 2B indicates the power consumption, and the horizontal axis indicates the number of people present in the air-conditioned area. The dots in FIG. 2B indicate actual values, and the dashed line indicates a graph showing the relationship between the number of users calculated based on the actual values ​​and the power consumption of the air conditioner. As shown in the figure, a certain relationship is established between the number of users present in the air-conditioned area and the power consumption of the air conditioner. Using this property, the prediction system 10 predicts the future number of users present in the air-conditioned areas 1A-1D, and predicts the power consumption of the air conditioner based on this number. By predicting the future power consumption, for example, an operation plan for the air conditioner can be created. For this purpose, the prediction system 10 predicts the future number of users present in the air-conditioned areas 1A-1D based on the past number of users assumed to be about the same number of users. In addition, since it is difficult to accurately grasp the number of users who actually existed in the air-conditioned areas 1A-1D in the past assumed to be about the same number of users, the number of users in the air-conditioned areas 1A-1D is estimated based on the human sensor equipped in the air-conditioned areas 2A-2D, the use reservation schedule of the air-conditioned areas 1A-1D, and the like.

[0016] (Number of people estimation model) Next, a number of people estimation model 161 for estimating the number of users in an air-conditioned area will be described with reference to Fig. 3A. Fig. 3A is a diagram showing an example of a number of people estimation model according to an embodiment. For example, when estimating the number of users present in the air-conditioned area 1A at time T1, the input parameters of the number of people estimation model 161 are detection values ​​at time T1 of n human sensors provided in the air-conditioned area 1A, attendance data of users at time T1, reservation schedule for use of the air-conditioned area 1A, detection values ​​for each seat in the air-conditioned area 1A (for example, log-on information for a PC provided for each seat, detection history of a seating / leaving detection system using short-range wireless communication provided for each seat, detection value of a seating sensor), etc. Note that time T1 is a predetermined length of time during which the actual value of each input parameter can be acquired, that is, a predetermined length of time in the past. The number of people estimation model 161 is constructed using various input parameters related to the number of users in the past as learning data. The number of people estimation model 161 is a function with each input parameter as a variable. As a simple example, the number of people estimation model 161 is a function expressed by, for example, the following formula (1). y=A1x1+A2x2+A3x3+···+B···(1) Here, x1, x2, ... are input parameters, and y is the estimated number of users. 1、 A 2、 A3 etc. coefficient 、 The value of constant B is set by a learning process that will be described with reference to Fig. 4. Upon acquiring input parameters for time T1, the number of people estimation model 161 outputs an estimate of the number of users present in the air-conditioned area 1A at time T1.

[0017] (Number of people prediction model) Next, a number-of-people prediction model 162 for predicting the number of users in an air-conditioned area will be described with reference to Fig. 3B. Fig. 3B is a diagram showing an example of a number-of-people prediction model according to an embodiment. While the number of people estimation model 161 estimates the number of users at the same time as the input parameters, the number of people prediction model 162 predicts the future number of users. For example, when estimating the future number of users in air-conditioned area 1A, the input parameters of the number of people prediction model 162 are the number of users in air-conditioned area 1A at past times T1 to Tn. The number of people prediction model 162 is a function that uses each input parameter as a variable. As a simple example, the number of people prediction model 162 is a function expressed by, for example, the following formula (2). y=C1x1+C2x2+C3x3+···+D···(2) Here, x1, x2, ... are the past number of users, and y is the predicted number of users in the future. C 1、 C 2、 Coefficients of C3 etc. 、 The value of the constant D is set by a learning process described with reference to FIG. 4. The number of users prediction model 162 acquires the number of users in the air-conditioned area 1A in the past as an input parameter, and outputs a predicted value of the number of users in the air-conditioned area 1A in the future. When an accurate number of users in the past can be obtained, this may be input to x1, x2, ... in formula (2). However, when an accurate number of users in the past cannot be obtained, the estimated value of the number of users at times T1 to Tn estimated by the number of users estimation model 161 can be used. Note that it is preferable that the times T1 to Tn are past times at which the number of users is expected to be about the same as that at a certain time Tn+1 in the future to be predicted. For example, when the air-conditioned area 1A is used according to a regular weekly schedule, the number of users prediction model 162 predicts the future number of users based on the number of users at the same time on the same day 1 to n weeks ago.

[0018] (Power consumption estimation model) Next, the power consumption estimation model 163 for estimating the power consumption of the air conditioner will be described with reference to Fig. 3C. Fig. 3C is a diagram showing an example of the power consumption estimation model according to one embodiment. For example, when estimating the total amount of power consumption of air conditioners 1 to m (for example, air conditioners 2A to 2D), the input parameters of the power consumption estimation model 163 are at least the number of users in air-conditioned areas 1 to m (for example, air-conditioned areas 1A to 1D). The input parameters may further include the temperature and weather at the time of estimating the amount of power consumption, and past actual values ​​of the amount of power consumption when air conditioners 1 to n were operated under similar conditions. The power consumption estimation model 163 acquires these input parameters and outputs an estimate of the total amount of power consumption of the air conditioners 1 to m. For example, the power consumption estimation model 163 is constructed by regression analysis, machine learning, etc., using the actual values ​​of the amount of power consumption for each number of users as learning data. In this embodiment, learning data is created by combining the total actual values ​​of the amount of power consumption of air conditioners 1 to n at a certain time T and the number of users in air-conditioned areas 1 to n at time T predicted based on the number of users prediction model 162. In addition, based on the number of users in air-conditioned areas 1 to n at future time Tf predicted based on the number of people prediction model 162 and the power consumption estimation model 163, the total value of the power consumption of air conditioners 1 to n at future time Tf is predicted.

[0019] (Construction of each model) Next, a process of constructing the number of people estimation model 161, the number of people prediction model 162, and the power consumption estimation model 163 will be described with reference to Fig. 4, taking as an example a case where learning data related to the air-conditioned area 1A is used. Fig. 4 is a diagram illustrating an example of the functions of a learning unit according to an embodiment. It is assumed that the actual power consumption value of the air conditioner 2A at time Tn+1, the detection values ​​of the human presence sensor in the air-conditioned area 1A at times T1 to Tn before time Tn+1, and a record of the planned usage schedule for the air-conditioned area 1A are available. The learning unit 12 inputs the detection values ​​of the human sensor in the air-conditioned area 1A at past time T1, the planned usage schedule, etc. (actual values) to the number of people estimation model 161. The number of people estimation model 161 outputs the number of users in the air-conditioned area 1A at time T1. Similarly, the learning unit 12 inputs the detection values ​​of the human sensor at past times T2 to Tn, etc. to the number of people estimation model 161 for each time. The number of people estimation model 161 outputs the number of users in the air-conditioned area 1A at each of times T2 to Tn. The learning unit 12 inputs the number of users at times T1 to Tn to the number of people prediction model 162. The number of people prediction model 162 outputs a predicted value of the number of users in the air-conditioned area 1A at a future time Tn+1. The learning unit 12 inputs the predicted value of the number of users in the air-conditioned area 1A at time Tn+1 to the power consumption estimation model 163. The power consumption estimation model 163 outputs an estimated value of the amount of power consumption consumed by the air conditioner 2A at time Tn+1. The learning unit 12 calculates the difference between the actual value of the amount of power consumption of the air conditioner 2A at time Tn+1 and the estimated value of the amount of power consumption output by the power consumption estimation model 163, and adjusts the coefficients and constants of the number of people estimation model 161, the coefficients and constants of the number of people prediction model 162, and the coefficients and constants of the power consumption estimation model 163 so that the difference becomes zero, thereby learning appropriate parameters for each model.

[0020] The flow of the above construction process is shown in FIG. FIG. 5 is a flowchart illustrating an example of a process for constructing each model according to an embodiment. The data acquiring unit 11 acquires input parameters for the number of people estimation model 161 (step S1). For example, the data acquiring unit 11 acquires detection values ​​of the human presence sensor in the air-conditioned area 1A for times T1 to Tn, utilization reservation schedules, and other performance values. Next, the learning unit 12 estimates the number of users in the past based on the number of users estimation model 161 (step S2). For example, the learning unit 12 inputs the learning data acquired in step S1 to the number of users estimation model 161. The number of users estimation model 161 outputs an estimated value of the number of users from time T1 to Tn. Next, the learning unit 12 predicts the number of users at time Tn+1, which is future than times T1 to Tn, based on the number-of-users prediction model 162 (step S3). For example, the learning unit 12 inputs the number of users at times T1 to Tn estimated in step S2 to the number-of-users prediction model 162. The number-of-users prediction model 162 outputs a predicted value of the number of users at time Tn+1. Next, the learning unit 12 predicts future power consumption based on the power consumption estimation model 163 (step S4). For example, the learning unit 12 inputs the number of users at time Tn+1 predicted in step S3 to the power consumption estimation model 163. The power consumption estimation model 163 outputs a predicted value of the number of users at time Tn+1. Next, the data acquiring unit 11 acquires input parameters of the power consumption estimation model 163 (step S5). For example, the data acquiring unit 11 acquires the actual value of the power consumption at time Tn+1. Next, the learning unit 12 learns the parameters of the number of people estimation model 161, the number of people prediction model 162, and the power consumption estimation model 163 (step S6). The learning unit 12 calculates the difference between the predicted value of the power consumption predicted in step S4 and the actual value of the power consumption acquired in step S5, and adjusts the parameters (coefficients, constants) of each model of the number of people estimation model 161, the number of people prediction model 162, and the power consumption estimation model 163 so that this difference becomes zero. For example, the learning unit 12 adjusts the values ​​of A1, A2, A3,...,B in formula (1) and C1, C2, C3,...,D in formula (2). For example, with regard to formula (1), if multiple human presence sensors are provided in the air conditioner 2A and one of them is broken, the value of the coefficient related to the broken human presence sensor is adjusted to be smaller. This makes it possible to construct each model. Next, the process of predicting the power consumption will be described with reference to Figures 6A to 6C, taking as an example the case of predicting the power consumption of the air conditioner 2A for the next day.

[0021] (Estimated number of past users) FIG. 6A is a flowchart illustrating an example of a number of people estimation process according to an embodiment. First, the data acquisition unit 11 acquires, for example, hourly detection values ​​of the human sensor in the air-conditioned area 1A on the same day of the week as the prediction target day (next day) up to 1 to n weeks ago, and a reservation schedule for use (step S11). Next, the number of people estimation unit 13 estimates the number of users on the same day 1 to n weeks ago (step S12). For example, the number of people estimation unit 13 inputs the detection values ​​of the human sensor on an hourly basis 1 to n weeks ago to the number of people estimation model 161. The number of people estimation model 161 outputs the number of users on the same day 1 to n weeks ago for each hour. This makes it possible to obtain an estimate of the number of users who were in the air-conditioned area 1A at any time in the past. Next, the number of people estimation unit 13 outputs the estimated number of people to the number of people prediction unit 14 (step S13).

[0022] (Prediction of future number of users) FIG. 6B is a flowchart illustrating an example of a number-of-people prediction process according to an embodiment. The number of people prediction unit 14 acquires the number of users for each hour on the same day from 1 to n weeks ago that was estimated by the number of people estimation unit 13 (step S21). The number of people prediction unit 14 predicts the number of users for the next day (step S22). For example, the number of people prediction unit 14 inputs the number of users for each hour from 1 to n weeks ago to the number of people prediction model 162. The number of people prediction model 162 outputs a predicted value of the number of users for each hour on the next day. This makes it possible to obtain a predicted value of the number of users present in the air-conditioned area 1A on a future prediction target date. Next, the number of people prediction unit 14 outputs the predicted number of users to the power consumption prediction unit 15 (step S23).

[0023] (Prediction of future power consumption) FIG. 6C is a flowchart illustrating an example of a power consumption prediction process according to an embodiment. The power consumption prediction unit 15 acquires the number of users for each hour of the next day predicted by the number of people prediction unit 14, the predicted values ​​of temperature and weather for each hour of the next day, and the actual value of past power consumption as input parameters (step S31). The power consumption prediction unit 15 predicts the power consumption of the air conditioner 2A for each hour of the next day (step S32). For example, the power consumption prediction unit 15 inputs each input parameter acquired in step S31 to the power consumption estimation model 163. The power consumption estimation model 163 outputs an estimated value of the power consumption for each hour of the next day. This makes it possible to obtain a predicted value of the power consumption for each hour of the prediction target day required for air conditioning the air-conditioned area 1A. Next, the power consumption prediction unit 15 outputs the predicted power consumption to the control device of the air conditioner 2A (step S33). For example, the control device of the air conditioner 2A creates an operation plan for the next day based on the acquired predicted value of the power consumption for each hour of the next day.

[0024] As described above, the prediction system 10 estimates the past number of users based on the number of users estimation model 161, and predicts the future number of users based on the past number of users and the number of users prediction model 162. In this way, the prediction system 10 can predict the power consumption of the air conditioner based on the future number of users and the power consumption estimation model 163. Furthermore, since the number of users estimation model 161, the number of users prediction model 162, and the power consumption estimation model 163 can be constructed using data that is actually easy to obtain, such as the detection value of a human sensor equipped in the air conditioner, attendance data, the reservation schedule for use of the air conditioner, and the power consumption of the air conditioner, the prediction system 10 can be introduced in any air conditioner area where an air conditioner is installed.

[0025] In the above embodiment, the number of users in the air-conditioned area 1A per hour is predicted using an estimated value of the number of users at the same time on the same day of the week in the past. However, an estimated value of the number of users at the same time on the previous n days may be used, or an estimated value of the number of users at the same time on the same day 1 to n years ago may be used. Each model may be constructed so that the number of users and the amount of power consumption can be predicted for all areas of the air-conditioned areas 1A to 1D. Each model may be constructed so that the amount of power consumption is predicted not on an indoor unit basis but on an air conditioner (outdoor unit basis). The unit of time for predicting the number of users and the amount of power consumption is not limited to one hour, but may be three hours or one day.

[0026] 7 is a diagram showing an example of a hardware configuration of a prediction system according to an embodiment. A computer 900 includes a CPU 901, a main memory device 902, an auxiliary memory device 903, an input / output interface 904, and a communication interface 905. The prediction system 10 described above is implemented in a computer 900. Each of the above-described functions is stored in the auxiliary storage device 903 in the form of a program. The CPU 901 reads the program from the auxiliary storage device 903, loads it in the main storage device 902, and executes the above-described processing in accordance with the program. The CPU 901 also reserves a storage area in the main storage device 902 in accordance with the program. The CPU 901 also reserves a storage area in the auxiliary storage device 903 for storing data being processed in accordance with the program.

[0027] In addition, a program for realizing all or part of the functions of the prediction system 10 may be recorded on a computer-readable recording medium, and the program recorded on the recording medium may be read into a computer system and executed to perform processing by each functional unit. The term "computer system" as used herein includes hardware such as an OS and peripheral devices. In addition, if a WWW system is used, the term "computer system" also includes a homepage providing environment (or display environment). In addition, the term "computer-readable recording medium" refers to a portable medium such as a CD, DVD, or USB, or a storage device such as a hard disk built into a computer system. In addition, when the program is distributed to the computer 900 via a communication line, the computer 900 that receives the program may expand the program in the main storage device 902 and execute the above processing. In addition, the above program may be for realizing part of the above-mentioned functions, and may further be capable of realizing the above-mentioned functions in combination with a program already recorded in the computer system.

[0028] As described above, several embodiments according to the present disclosure have been described, but all of these embodiments are presented as examples and are not intended to limit the scope of the invention. These embodiments can be implemented in various other forms, and various omissions, substitutions, and modifications can be made without departing from the gist of the invention. These embodiments and their modifications are included in the scope of the invention and its equivalents as described in the claims, as well as in the scope and gist of the invention.

[0029] <Additional Notes> The number of people prediction system, number of people estimation system, power consumption prediction system, number of people prediction method, and program described in each embodiment can be understood, for example, as follows.

[0030] (1) A number of people prediction system (prediction system 10) according to a first aspect includes a number of people estimation model 161 that acquires values ​​of predetermined variables (detection values ​​of a human presence sensor, attendance data, and reservation schedule) related to the number of users present in a predetermined area (air-conditioned area 1A) at a predetermined time and outputs an estimated value of the number of users present in the area at the time, and a number of people prediction model 162 that acquires the estimated number of people for a plurality of times (times T1 to Tn) estimated by the number of people estimation model 161 and outputs a predicted value of the number of users present in the area at a prediction target time (time Tn+1) in the future than the plurality of times. the power consumption estimation model 163 that acquires predetermined variables (number of users, weather, temperature, past actual power consumption value, etc.) including the number of people predicted value for the prediction target time and outputs an estimated value of the power consumption of an air conditioner (e.g., air conditioner 2A) for the prediction target time, a learning unit 12 that learns so that the estimated value of power consumption output by the power consumption estimation model 163 becomes the actual value of the power consumption of the air conditioner for the prediction target time, and a number of users prediction unit 14 that predicts the number of users present in the area based on the learned number of users prediction model 162. This allows the number of future users to be predicted with high accuracy.

[0031] (2) A second aspect of the number of people prediction system (prediction system 10) is the number of people prediction system (prediction system 10) of (1), wherein the number of people prediction model 162 predicts the number of people using a linear combination model (equation (2)) of the number of people estimated values ​​at each of the multiple times, and the learning unit 12 learns coefficients (C1, C2,...) related to the number of people estimated values ​​at each time of the linear combination model.

[0032] (3) A third aspect of the present invention is a number of people prediction system (prediction system 10) according to (1) to (2), in which the variables input to the number of people estimation model 161 include detection values ​​of multiple human presence sensors, and the learning unit 12 learns weighting for each of the human presence sensors.

[0033] (4) A fourth aspect of the number of people prediction system (prediction system 10) is a number of people prediction system (prediction system 10) of (1) to (3), in which the variables input to the number of people estimation model 161 include a human presence sensor and planned usage information of the area, and the learning unit 12 learns the weighting for each of the variables.

[0034] (5) A fifth aspect of the present invention is a number of people prediction system (prediction system 10) according to any one of (1) to (4), further comprising a number of people estimation unit 13 that estimates the number of users present in the area based on the trained number of people estimation model 161. This makes it possible to obtain past population estimates necessary for predicting future numbers of people.

[0035] (6) A sixth aspect of the present invention relates to a number of people estimation system (prediction system 10) that includes a number of people estimation model 161 that acquires values ​​of predetermined variables (detection values ​​of a human presence sensor, attendance data, and reservation schedule) related to the number of users present in a predetermined area (air-conditioned area 1A) at a predetermined time and outputs an estimated value of the number of users present in the area at the time, and a number of people prediction model 162 that acquires the estimated number of people for a plurality of times (times T1 to Tn) estimated by the number of people estimation model 161 and outputs a predicted value of the number of users present in the area at a time (time Tn+1) in the future than the plurality of times. the power consumption estimation model 163 that acquires predetermined variables (number of users, weather, temperature, past actual power consumption value, etc.) including the predicted number of people at the future time and outputs an estimated value of the power consumption of an air conditioner (e.g., air conditioner 2A) at the future time, a learning unit 12 that learns so that the estimated power consumption value output by the power consumption estimation model 163 becomes the actual power consumption value of the air conditioner at the future time, and a number of users estimation unit 13 that estimates the number of users present in the area based on the trained number of users estimation model 161. This makes it possible to estimate the number of users at any time when the value of a predetermined variable related to the number of users can be obtained. Also, if the value of a predetermined variable related to the number of users and the actual value of the power consumption of the air conditioner in the future can be obtained, a number of users estimation model 161, a number of users forecast model 162, and a power consumption estimation model 163 can be constructed.

[0036] (7) The energy consumption system (prediction system 10) according to the seventh aspect includes a number of users estimation model 161 that acquires values ​​of predetermined variables (detection values ​​of a human presence sensor, attendance data, and a reservation schedule) related to the number of users present in a predetermined area (air-conditioned area 1A) at a predetermined time and outputs an estimated value of the number of users present in the area at the time, a number of users prediction model 162 that acquires the estimated number of users for a plurality of times (times T1 to Tn) estimated by the number of users estimation model 161 and outputs a predicted value of the number of users present in the area at a time (time Tn+1) in the future than the plurality of times, and a number of users prediction model 162 that acquires values ​​of predetermined variables (detection values ​​of a human presence sensor, attendance data, and a reservation schedule) related to the number of users present in the predetermined area (air-conditioned area 1A) at a predetermined time and outputs an estimated value of the number of users present in the area at the time, The system comprises a power consumption estimation model 163 that acquires information such as the number of users, weather, temperature, and past actual power consumption values, and outputs an estimated value of the power consumption of an air conditioner (e.g., air conditioner 2A) at the future time, a learning unit 12 that learns so that the estimated value of power consumption output by the power consumption estimation model 163 becomes the actual value of the power consumption of the air conditioner at the future time, a number of users prediction unit 14 that predicts the number of users in the area based on the learned number of users prediction model, and a power consumption prediction unit 15 that predicts the power consumption of the air conditioner based on the predicted number of users and the learned power consumption estimation model. This allows the prediction of the power consumption of the air conditioner with high accuracy. Also, if the values ​​of certain variables related to the number of users and the actual values ​​of the power consumption of the air conditioner in the future can be obtained, it is possible to construct a number of people estimation model 161, a number of people prediction model 162, and a power consumption estimation model 163.

[0037] (8) A number of users prediction method according to an eighth aspect includes a step of learning a number of users estimation model that acquires values ​​of predetermined variables related to the number of users present in a predetermined area at a predetermined time and outputs an estimated number of users present in the area at the time, a number of users prediction model that acquires the estimated number of users for a plurality of times estimated by the number of users estimation model and outputs a predicted number of users present in the area at a prediction target time that is in the future than the plurality of times, and a power consumption estimation model that acquires predetermined variables including the predicted number of users at the prediction target time and outputs an estimated value of power consumption of an air conditioner at the prediction target time, so that the estimated power consumption output by the power consumption estimation model becomes an actual value of power consumption of the air conditioner at the prediction target time, and a step of predicting the number of users present in the area based on the learned number of users prediction model.

[0038] (9) A program according to a ninth aspect causes a computer 900 to execute the following steps: a number of users estimation model that acquires values ​​of predetermined variables related to the number of users present in a predetermined area at a predetermined time, and outputs an estimated number of users present in the area at the time; a number of users prediction model that acquires the estimated number of users for multiple times estimated by the number of users estimation model, and outputs a predicted number of users present in the area at a prediction target time that is in the future than the multiple times; and a power consumption estimation model that acquires predetermined variables including the predicted number of users at the prediction target time, and outputs an estimated value of power consumption of an air conditioner at the prediction target time, so that the estimated power consumption output by the power consumption estimation model becomes an actual value of power consumption of the air conditioner at the prediction target time; and a step of predicting the number of users present in the area based on the trained number of users prediction model. [Explanation of symbols]

[0039] 10. Prediction system 11 Data acquisition section 12. Learning Section 13...Number of people estimation section 14. Population Prediction Section 15 Power Consumption Prediction Unit 16...Storage section 161...Number of people estimation model 162. Number of people prediction model 163...Power Consumption Estimation Model 900...Computer 901···CPU 902...Main memory 903...Auxiliary storage device 904 Input / Output Interface 905 Communication Interface

Claims

1. a learning unit that learns a number of users estimation model that acquires values ​​of a predetermined variable related to the number of users present in a predetermined area at a predetermined time and outputs an estimated number of users present in the area at the said time, a number of users prediction model that acquires the estimated number of users for multiple times estimated by the number of users estimation model and outputs a predicted number of users present in the area at a prediction target time in the future than the multiple times, and a power consumption estimation model that acquires the predicted number of users at the prediction target time and outputs an estimated value of power consumption of an air conditioner at the prediction target time, so that the estimated power consumption output by the power consumption estimation model becomes an actual value of power consumption of the air conditioner at the prediction target time; a number-of-users prediction unit that predicts the number of users present in the area based on the trained number-of-users prediction model; Equipped with The predetermined variables related to the number of users present in the predetermined area at the predetermined time include a detection value of the number of users present in the area by a human presence sensor, attendance data of the users at the predetermined time, a reservation schedule for the area, and a detection value of whether or not a user is present for each seat in the area. Population forecasting system.

2. The number of people prediction model predicts the number of people using a linear combination model of the number of people estimated values ​​at each of the multiple times; The learning unit learns coefficients of the linear combination model related to the estimated number of people at each time. The number of people prediction system according to claim 1 .

3. The detection value by the human presence sensor includes detection values ​​by a plurality of human presence sensors, The learning unit learns a weighting for each of the human presence sensors. The number of people prediction system according to claim 1 or 2.

4. The power consumption estimation model acquires, in addition to the predicted number of people during the prediction target time, at least one of the temperature and weather at the time of estimating the power consumption, and a past actual value of the power consumption when the air conditioner was operated under the same conditions as the prediction target time. The number of people prediction system according to any one of claims 1 to 3.

5. a number of people estimation unit that estimates the number of users present in the area based on the trained number of people estimation model; The number of people prediction system according to claim 1 , further comprising:

6. a learning unit that learns a number of users estimation model that acquires values ​​of a predetermined variable related to the number of users present in a predetermined area at a predetermined time and outputs an estimated number of users present in the area at the said time, a number of users prediction model that acquires the estimated number of users for multiple times estimated by the number of users estimation model and outputs a predicted number of users present in the area at a prediction target time in the future than the multiple times, and a power consumption estimation model that acquires the predicted number of users at the prediction target time and outputs an estimated value of power consumption of an air conditioner at the prediction target time, so that the estimated power consumption output by the power consumption estimation model becomes an actual value of power consumption of the air conditioner at the prediction target time; a number of people estimation unit that estimates the number of users present in the area based on the trained number of people estimation model; Equipped with The predetermined variables related to the number of users present in the predetermined area at the predetermined time include a detection value of the number of users present in the area by a human presence sensor, attendance data of the users at the predetermined time, a reservation schedule for the area, and a detection value of whether or not a user is present for each seat in the area. People estimation system.

7. a learning unit that learns a number of users estimation model that acquires values ​​of a predetermined variable related to the number of users present in a predetermined area at a predetermined time and outputs an estimated number of users present in the area at the said time, a number of users prediction model that acquires the estimated number of users for multiple times estimated by the number of users estimation model and outputs a predicted number of users present in the area at a prediction target time in the future than the multiple times, and a power consumption estimation model that acquires the predicted number of users at the prediction target time and outputs an estimated value of power consumption of an air conditioner at the prediction target time, so that the estimated power consumption output by the power consumption estimation model becomes an actual value of power consumption of the air conditioner at the prediction target time; a number-of-users prediction unit that predicts the number of users present in the area based on the trained number-of-users prediction model; a power consumption prediction unit that predicts the power consumption of the air conditioner based on the predicted number of users and the trained power consumption estimation model; Equipped with The predetermined variables related to the number of users present in the predetermined area at the predetermined time include a detection value of the number of users present in the area by a human presence sensor, attendance data of the users at the predetermined time, a reservation schedule for the area, and a detection value of whether or not a user is present for each seat in the area. Power consumption prediction system.

8. a step of learning a number of users estimation model that acquires values ​​of a predetermined variable related to the number of users present in a predetermined area at a predetermined time and outputs an estimated number of users present in the area at the said time, a number of users prediction model that acquires the estimated number of users for multiple times estimated by the number of users estimation model and outputs a predicted number of users present in the area at a prediction target time in the future than the multiple times, and a power consumption estimation model that acquires the predicted number of users at the prediction target time and outputs an estimated value of the power consumption of an air conditioner at the prediction target time, so that the estimated value of power consumption output by the power consumption estimation model becomes an actual value of the power consumption of the air conditioner at the prediction target time; predicting the number of users present in the area based on the trained number-of-users prediction model; having The predetermined variables related to the number of users present in the predetermined area at the predetermined time include a detection value of the number of users present in the area by a human presence sensor, attendance data of the users at the predetermined time, a reservation schedule for the area, and a detection value of whether or not a user is present for each seat in the area. How to predict the number of people.

9. On the computer, a step of learning a number of users estimation model that acquires values ​​of a predetermined variable related to the number of users present in a predetermined area at a predetermined time and outputs an estimated number of users present in the area at the said time, a number of users prediction model that acquires the estimated number of users for multiple times estimated by the number of users estimation model and outputs a predicted number of users present in the area at a prediction target time in the future than the multiple times, and a power consumption estimation model that acquires the predicted number of users at the prediction target time and outputs an estimated value of the power consumption of an air conditioner at the prediction target time, so that the estimated value of power consumption output by the power consumption estimation model becomes an actual value of the power consumption of the air conditioner at the prediction target time; predicting the number of users present in the area based on the trained number-of-users prediction model; having A process in which the predetermined variables relating to the number of users present in the predetermined area at the predetermined time include a detection value of the number of users present in the area by a human presence sensor, attendance data of the users at the predetermined time, a reservation schedule for the area, and a detection value of whether or not a user is present for each seat in the area; A program that executes the following.

Citation Information

Patent Citations

  • Air conditioner

    JP2014109415A

  • Thermal load estimation device and air conditioning control system

    JP2015152179A

  • Human flow prediction method and human flow prediction system

    JP2020001860A

  • Air conditioner management device, heat source equipment management device, air conditioner management method and heat source equipment management method

    WO2017216833A1

  • Digital smart energy saving system, method, and program

    WO2018047328A1