住宅の空調利用における多様な居住者の行動の究明：統計分析と予測モデリング

概要

九州大学学術情報リポジトリ
Kyushu University Institutional Repository

Exploring the Diverse Occupants' Behaviors of
Residential Air-Conditioning Use: Statistical
Analysis and Predictive Modeling
呂, 嘉俊

https://hdl.handle.net/2324/7165098
出版情報：Kyushu University, 2023, 博士（工学）, 課程博士
バージョン：
権利関係：

氏

名 ：

呂 嘉俊

Name

論 文 名 ： Exploring the Diverse Occupants' Behaviors of Residential Air-Conditioning
Use: Statistical Analysis and Predictive Modeling

Title

（住宅の空調利用における多様な居住者の行動の究明：統計分析と予測モデリング）

区

分 ： 甲

Category

論 文 内 容 の 要 旨
Thesis Summary
Building-related carbon emissions have reportedly accounted for 28% of global energy-related carbon
emissions, reaching an all-time high of approximately 10 Gt in 2022. Among various residential appliances,
air-conditioning (AC) has been confirmed as one major contributor to household energy use, as well as a
critical factor in realizing a comfortable indoor thermal environment. It was confirmed by substantial
studies that OB features in the usage of the heating, ventilating, and air-conditioning (HVAC) system have a
dominant influence on the total electricity consumption. The occupant behaviours reflected in AC use are
stochastic and complex, especially in the residential sector and for buildings equipped with split AC units
rather than central systems because their operation states are simply decided by the occupants’ thermal
preferences and occupancy schedules.
In recent year, So-called bottom-up approach based on building-energy simulations coupled with stochastic
modelling of occupant behaviours has been intensively developed to properly estimate the effect of diverse
and stochastic occupant behaviours on energy loads. Most of these studies included the modeling of
stochastic occupancy schedules and OB, which were mainly derived from a statistical analysis of AC usage
observation data. To model OB related to AC use, various factors have been adopted, including the ambient
temperature, indoor air temperature, time of day, and residents’ demographic information.
However, the comprehensive analysis of the characteristics of residential AC use in the real community has
been hampered by the limited availability of appliance level interval consumption data. Validation of the
models based on long-term observation data with large samples has not been fully conducted. Moreover,
current research on the stochastic modeling of occupants’ AC use has mostly rarely considered the
stochastic nature of individual behavioral preferences among different households or occupants. Most of the
previous statistical analyses and stochastic AC use models were derived from measurements with a limited
number of samples from several to dozens of households Therefore, the characteristics of the diverse OB
and AC energy demand patterns of a community consisting of diverse people were difficult to reproduce
using the existing models.
With this background, this thesis proposes a sequence of studies on the statistical analysis and predictive
modeling of diverse behaviour of residential air-conditioning use derived from energy demand data. It was
conducted based on the appliance-level electricity demand data measured by smart meters for 586
dwellings in a real community during two consecutive cooling seasons in Osaka, Japan.
Firstly, statistical analysis of the energy data was conducted to examine the stochastic features and
contribution of residential air conditioning loads to the total electricity demand. After processing the raw
data, the statistical analysis of the ACL and household-level load revealed a significant contribution of ACL

on the annual peak electricity demand of the community. Furthermore, the effectiveness of the two demand
response scenarios, with maximum and limited deferral potential, were examined to assess the potential
reduction in the annual peak demand. The target households were classified into six groups based on their
dependence on AC cooling to evaluate the DR effectiveness with different participation ratios. The results
indicate that a reduction of up to 4.9% can be achieved by applying the proposed DR strategy in the
evaluated community. Additionally, considerable deferral and reduction were confirmed even with the
engagement of a small portion of households by proper identification of target households that intensively
use AC for a long-time duration.
Secondly, the authors intend to grasp the stochastic features of residential HP use based on the electricity
demand data of a large number of dwellings and examine the validity of the state transitional probabilities
of HP use widely used in the bottom-up approach. A BES model coupled with the state transitional
probability modelling of occupants’ cooling use was proposed to consider seasonal effects on stochastic
modelling of HP use behaviour based on the findings. The simulated daily cooling hours showed a seasonal
trend of the difference against the observed data, implying that the relationship between the occupants’
thermal tolerance, namely the probability of switching on AC units and the indoor thermal condition at that
moment is affected by occupants’ thermal exposure in previous days. To reproduce the seasonal change of
the occupants’ thermal tolerance, an improved model was further proposed and performed with better
agreement compared to the observation model.
Thirdly, this work proposed a prediction method for the stochastic AC on/off state in a residential building
considering the inter-occupant diversity of AC use behaviour. Analysis of the energy data reveals the
inter-occupant diversity of OBs in the measured dwellings. In particular, individual preferences regarding
occupancy schedules, daily cooling schedules, and thermal sensitivity were found to show great variability
across the community. Clustering analysis was then applied to classify the dwellings into different
schedules and thermal preference patterns. The XGBoost model was applied to predict the hourly AC
on/off state and showed satisfactory performance. Results shown three and four types of households for the
occupants’ behaviours related to their cooling schedule and thermal sensitivity patterns, respectively.
Furthermore, the proposed model considering diverse OBs, showed satisfactory prediction performance,
with an AUC score of 0.845, indicating a high chance of accurate distinguishment of AC operation states. In
addition, instead of the outdoor temperature, the behaviours of the occupants were found to have a crucial
impact on a household’s AC operation. Feature importance scores of occupants’ schedule preference and
thermal preference in AC state prediction were found to be 0.384 and 0.263, respectively.

この論文で使われている画像