In rooms with large windows, the thermal environment tends to deteriorate owing to thermal radiation, cold draft, and vertical temperature difference during air-conditioned heating in winter. To improve the thermal environment near windows with low insulation performance, equipment such as perimeter fans and natural convection type heaters are used. In this study, the characteristics of airflow from a fan outlet and heating element by means of PIV (Particle Image Velocimetry) and the temperature distribution around the equipment were measured. A non-isothermal jet model and a plume model, which evaluate the airflow in the vertical temperature distribution prediction model, were used to reproduce the experimental results.

In recent years, data centers (DCs) have become increasingly important, thereby necessitating their highly efficient and reliable operation and management. This study aims to enable the highly efficient operation and management of DCs using data collected from a server room and machine learning. We attempted to estimate the baseline which is estimation technology of power consumption of an air conditioner for evaluating the validity of energy saving effects using a machine learning model, targeting an air conditioner in the server room. First, compared the baseline models using multiple methods, and determined the influence of explanatory variables and learning period on the accuracy (Coefficient of determination, RMSE, Peak error). Additionally, we constructed a model to estimate the baseline power consumption of the air conditioner with high accuracy. Second, to study the versatility of the baseline model, we verified the accuracy by applying the same baseline model to air conditioners in different server rooms, and confirmed that it was possible to estimate power consumption with high accuracy. Finally, using the baseline model that was subjected to these accuracy verifications, we verified the effect of energy saving measures on energy consumption and confirmed the effectiveness of the proposed baseline model.

The energy consumption of building equipment accounts for 40% of the total energy consumption, with air conditioning systems accounting for majority of the total energy consumption. Variable refrigerant flow (VRF) systems have made significant progress in recent years owing to their high energy efficiency and improved design flexibility. These systems have been installed not only in small-sized buildings but also in large-sized buildings instead of central air conditioning systems. Performance evaluation of VRF systems has been performed by the Japan Industrial Standards committee. However, because fixed conditions have been applied in their tests, the energy performance of the systems is dissociated from the actual situations, including low load operations and short cycle effects in outdoor unit areas. Furthermore, as refrigerants directly exchange heat with the ambient air in VRF systems, it is generally difficult to evaluate their performance accurately. In the present study, an onsite performance evaluation system was developed to determine the operational performance of VRF systems. The system realized the real-time calculation of operational performance using a high-performance microcomputer board and accumulated data in the cloud server via the Internet. This system was installed nationwide in various buildings with different uses and meteorological conditions, and the obtained data were analyzed. As a result, the actual operations were revealed. Additionally, the performance visualization of the VRF systems on the user side was realized.

Because businesses are required to save energy in accordance with the Strategic Energy Plan proposed by the Agency for Natural Resources and Energy of Japan, it is necessary to reduce the energy utilized for building operation as much as possible. Therefore, a new rousing system using lighting that conveys to occupants the suitability of the present state of the outer air for natural ventilation and operates the mechanical air conditioning system has been developed and introduced in a new office building at Takamatsu. In this paper, the actual energy consumption report and the indoor environment survey conducted in this building for a floor with a total of 23 windows on all four sides, are discussed. First, the rousing system was turned on frequently from March to June and from September to December. Energy consumption for air conditioning was reduced by 50 - 70 % in these periods, and particularly in May, June and October in the second year. In other words, the rousing system helped this building become net zero energy building (ZEB) ready. Second, the natural ventilation efficiency of this building was investigated. Fresh air was found to enter through the north window and exit through the south window, when all the windows were open. Furthermore, at this time, the ambient air velocity was under 0.5 m/s and the ventilation volume was such that the indoor air was replaced 5 to 15 times per hour. Third, the survey on indoor environment during natural ventilation indicated that the carbon dioxide concentration was 400 to 600 ppm. Although only half the number (10 - 12) of windows were opened, this value was lower by 150 to 200 ppm than that during mechanical air conditioning. Lastly, the results of the occupant questionnaire indicate that there was no significant difference in thermal sensation, thermal comfort, and overall satisfaction between natural ventilation in autumn or spring and mechanical cooling in summer. However, the air flow sensation exhibited a significant difference between spring and summer, and the average of the subjective intellectual productivity in autumn was slightly lower than that in summer. In summary, this report suggests that it is possible for manually operated natural ventilation to not only enable energy saving but also maintain the same levels of indoor thermal sensation, thermal comfort, and satisfaction as those during mechanical air conditioning. However, further studies must be conducted so as to not reduce intellectual productivity.