In this study, a novel air conditioning system combining package air conditioner(Package Air Conditioner) and membrane ceiling is proposed. This report presents the results of a full-scale cooling experiment. The temperature profiles of this air conditioning system indicated comfortable a temperature range, and the air conditioning performance was suitable for heat loads. The energy efficiency was as high as the rated performance of the package air conditioner. With the increase in the amount of air circulation between the attic and the lower side of the ceiling, the heat load removal performance was improved.

This study aims to generate training data that can demonstrate high diagnostic performance in fault detection and diagnosis (FDD) using convolutional neural networks (CNNs) in a heat source system. Previous studies have proposed an FDD method using CNNs and have reported that faults can be detected and diagnosed with high accuracy using this method. However, methods to generate training data that can demonstrate high diagnostic performance have not been sufficiently examined. Herein, we examine the method of generating training data in a target heat source system. In this study, we focused on a heat source system of a real building. This system consisted of centrifugal liquid chillers, primary-chilled water pumps, secondary-chilled water pumps, condenser water pumps, and direct-contact cooling towers. Actual operation data were collected by the building energy management system (BEMS) at 15-min intervals, and in this study, the data of 2013 and 2014 were used. First, we developed a simulation model that could calculate the target system behavior assumed in the design. Using this simulation, four studies were conducted with the aim of suppressing the learning time of CNNs and improving the diagnostic performance as follows: Study 1. Selection of the faults to be diagnosed Study 2. Setting of the severity level of the faults Study 3. Setting of simultaneous faults Study 4. Setting of the imaged data amount of each fault In Study 1, we selected the faults that were likely to occur in the target system. In addition, we selected only those faults that affected system efficiency and generated faulty behavior data for 56 types of faults to be diagnosed through simulation. In Study 2, we analyzed the relationship between the setting of the severity level of the faults in the training data and the diagnostic performance. As a result of the analysis, we found that three severity levels need to be set. In Study 3, we analyzed the result of diagnosing simultaneous faults, and found that either fault was generally diagnosed when faulty behavior data with a single fault was prepared in the training data. In Study 4, we analyzed the relationship between the imaged data amount of each fault and the diagnostic performance on the representative days in summer, winter, and the middle season. As a result of the analysis, we clarified the setting of the imaged data amount that improves the diagnostic performance in each season. By examining the above procedure, we generated training data with a reduced amount of imaged data and high diagnostic performance in the target system. We expected high-performance FDD by CNNs in other heat source systems by examining and creating training data based on this procedure. Further studies are required to establish a more detailed method of generating training data.

Rectangular metal ducts for air conditioning are mainly based on flange connection methods, such as the angle method and corner bolt method. As a simple connection method, slips and drives in the United States, and socket and spigot joints in the United Kingdom have been standardized. Slip-joint connections, such as socket and spigot joints are popular in Japan as a method for connecting round ducts and rectangular ducts in ships. In this study, to apply the slip-joint connection to rectangular metal ducts for Heating, Ventilation, and Air Conditioning (HVAC) system in Japan, the safety strength and air leakage flow rate of the test ducts were evaluated.

This study experimentally evaluates a dehumidifying and reheating air conditioning unit with a built-in heat pipe designed to process outside air without using primary energy for pre-cooling or reheating. In our previous report, in addition to evaluating the reduction effect of the cooling of the cold water coil from performance tests in the laboratory, we examined the operational results in a building where the system was in actual operation to confirm the normal operation of the dehumidifying and reheating cycle. In this report, we evaluate the heat pipe operation and the energy-saving effect over a year, based on measurement data from a building in which the system is in operation. Based on these measurements, we confirm that while the heat pipe operates when the pre-cooling coil inlet temperature is 15℃ or more, it does not operate when the air conditioning is stopped or during the heating operation. Considering the resistance loss due to installation of the heat pipe, the annual air conditioner load is reduced by 34.4% for the Western Japan grid and 39.9% for the Eastern Japan grid, and is this considered to have an energy-saving effect over the year.

The Building Energy Simulation Tool(BEST) Program was established as a new energy simulation tool with the cooperation of the industry, government, and academia in 2005, and the development committee has been established since then. Although several types of BEST programs exist, the professional version enables the most detailed examination and is believed to be useful in detailed design and research. However, we cannot be certain that it is fully utilized. One of the reasons is that the calculation results are not easy to handle and it is difficult to input data. Therefore, we developed a tool that can easily evaluate various air conditioning performances using the result of the BEST professional edition. In the developed tool, seven types of evaluation items were set, including the evaluation index for the water thermal storage type air conditioning system. This tool was used for evaluation when the input contents were varied in a water thermal storage type air conditioning system. Changes in the calculation results according to the input conditions could be easily confirmed, demonstrating efficiency in analysis. In the future, we intend to conduct the evaluation considering the partial load characteristics of the heat source and the evaluation method of the individual distributed system.

This study aims to propose a hybrid drainage system for newly-built and existing buildings, which is adaptable to changes to or extension of plumbing spaces due to conversion or renewal of these buildings; and to acquire knowledge that is conductive to the performance evaluation of the said hybrid drainage system and to formulating its planning/design method. This hybrid drainage system combines a “force-feed drainage system” which employs a force-feed pump unit and a small-diameter force-feed drainage pipe, and a “gravity drainage horizontal pipe system” in which drainage flows by gravity. The system is compatible with many plumbing variations in newly-built and existing buildings, and therefore, drainage performance is verified on each plumbing variation. Based on the knowledge acquired from the results of the basic experiments in the previous report, this study investigates the effect of the hybrid drainage system on drainage performance by applying the system to a real office building, subsequent to the conversion thereof, and determines the actual drainage load flow rates and internal pipe pressure variations for approximately one year while actually in use. Moreover, as a short-term experiment, this study examines the noise generated by the force-feed pump operating in the system, and investigates the effect of the noise on living space, etc. The effectiveness of this system in the real building is subsequently validated on the basis of the results of these experiments.