This report describes the results from the questionnaire and detailed survey conducted in 1998 and 1999. The 17 items of questionnaire included in facilities management, maintenance management, saving energy, aquarium facilities. The number of aquariums investigated by questionnaire survey counted for about 57 aquariums in Japan. From the survey results, I discuss the tendency and characteristics about each item. This paper has emphasised the importance of having the facilities・maintenance management in aquariums.

View full abstract

This paper describes a modeling procedure for an environmental chamber as a typical example of an air conditioning system taking a mathematical analysis into account by using the experimental data of a real plant. In these air conditioning systems, the heat-sources have been by far cold-water, heat wire and hot-water. Since these systems overcool room air below the setpoint temperature at first and then warm room air up the setpoint, it will be a waste of energy. The control problem is of economic significance since improved control scheme can reduce heating/cooling costs without lowering the performance of controlling room temperature. We made the air-conditioning machine system of the constant temperature clean room in the semiconductor manufacturing process. This system controlled the clean room temperature variation minutely. This clean room system was composed of clean room insulated from temperature variation in environment and the air conditioner that used only cooling water as heat-source to control clean room temperature. The air conditioner was composed of double cooling water circulation including a cooling coil, and a heat exchanger. As the experimental results, We have understood that it was important to increase the number of cooling water circulation for controlling the clean room temperature variation minutely, and the clean room temperature variation was less than about ±0.01 degree by using this system. In this study, we have made the simulation model for this air-conditioning system including automatic control system by using the experimental data. We have simulated the step response of room temperature, and the room temperature was influenced by change of temperature of cold water. We have compared the simulation results with the experimental results, and we have understood that the simple simulation model of air-conditioning system could predict the clean room temperature. At last, we have simulated the room temperature at various conditions of the air-conditioning system. And we have evaluated the application of the simulation model to the air-conditioning system design.

View full abstract

On a parent and child type level regulating valve for low water of a diaphragm, with a function of controling flow rate into a receiving tank, it was examined how the opening of the parent valve changes when the opening of the ball tap as the chaild valve changes, and The next conclusions were obtained (1) There are two cases that the opening of the parent valve changes stably or suddenly. (2) The range in which the opening of the parent valve stably decreases expands if the largest opening was limited, when this opening stably increases in the small opening range and afterwards it suddenly increases to the largest opening.

View full abstract

We developed the model that could simulate one-day flow field in a few km^2 scale, and we carried out the simulation for the developing area in northern Minoo City and verified the validity of the model. To take in a synoptic influence in a developing area (a few km^2 scale), the scale-down of four stage (large, middle, small and developing area scale) is performed, and the nesting technique is used at each scale-down. The calculated results in the large and the middle scale were compared with AMeDAS data, and the calculated results in the developing area scale were compared with data at two observatories in the developing area. The calculated flow fields at each scale were reasonable agreement with the observed one in general. It makes clear that the multiple nesting technique is useful method to simulate one-day flow field in a few km^2 scale.

View full abstract

This report deals with the simulation of discharge characteristics of a thermal energy storage column packed with cross-linked polymer rods when a heat exchanger is installed between the column and the heat load units. The transient response of the inlet and the outlet temperatures of the heat exchanger and the temperature distribution of the heat transfer fluid in the column were calculated. The calculations took into account; the varying initial temperature of the column, height of the column, mass velocity of heat transfer fluid and ratio of flow rate of heat transfer fluid, the number of heat transfer units of heat exchanger and inlet temperature of the load. In addition, a similar study is described on controlling mass velocity of the heat transfer fluid in the column to keep the outlet temperature of the heat exchanger at a constant value.

View full abstract

For evaluating domestic fan noise, sound power levels of 26 fans are measured by in-duct and reverberant-field methods. There are 12 forward-curved fans, four backward-curved fans, four axial fans and six mixed-flow fans. Specific sound power levels for these domestic fans are calculated from sound power levels measured by the reverberant-field methods. The applications of correction factor C and blade frequency increment BFI described in the ASHRAE Handbook to domestic fans are discussed. The averages and the standard deviations of the specific sound power levels for four kinds of domestic fans are also calculated and compared with those of ASHRAE.

View full abstract

This paper constitutes Part 2 in the series of studies on hybrid utilization of renewable energy for a low energy house. The objective is to evaluate annual operating results of an experimental house. First of all, it was experimentally found that a ground source heat pump system utilizing about 10℃-constant earth temperature layer was efficient for cooling. System COP in the 4 hour-intermittent cooling was 9.1. Experimental results of the heating operation with the ground source heat pump showed that COP and System COP were quite high; 4.0 and 3.1, respectively. Next, it was shown that approximately 80% of the total energy was provided from photovoltaic modules, solar collectors, underground and exhaust heat. Annual energy consumption was 12.5% of typical house's one in Hokkaido. From the viewpoint of environmental protection, CO_2 reduction rate was 77%.

View full abstract

This paper concerns the application of an artificial neural network (ANN) to serve as a controller for HVAC systems. ANN is one of the optimization approaches for any systems and it simulates neurons as a mathematical model by expressing a multi-layered network. One of the best advantages of an ANN is to express the nonlinearities of input/output functions by learning feedback errors. Usually PID (proportional plus integral plus derivative) controllers are used to control HVAC systems, and parameter tuning of controllers is one of the most important and difficult problem. The parameters depend on the delay time and dead time of the control system. The elements of the control system consist of control valve, cooling coil, duct, room and sensors. Complex models are required to express such a control system. Instead of using PID controllers, the authors propose an ANN-based controller for HVAC systems. The previous paper already reported that the control model using a multi-layered ANN shows accuracy and stability in heating and humidifying mode without tuning control parameters. Such a controller may minimize jobs for tuning control parameters and it may stabilize the control systems. Errors are back-propagated to revise the connecting loads of ANN by using descent gradient. The method minimizes the error squares between the setpoints and the process values. The inputs of the ANN are the errors and the time integrals of errors of the room temperature, and the output is the valve opening. ANN-based controller gives the exact openings in order to follow the set point of the room temperature and to compensate for the control loop offset. This paper shows experimental data in cooling mode. As results, it is proved that the control of HVAC systems using ANN is stable and accurate also in cooling mode without tuning any control parameters. The ANN-based controller works every 10 seconds for discrete control. The comparison is made among 4 cases for ε=0.01/0.1 and α=0.3/0.6, what are called learning rate and learning factor. The RMS errors of the room temperature and the valve openings demonstrate the stability and controllability of the ANN-based controller. The experiments are performed in the states of partial thermal loads. The conditions of the experiments are strict, which make occasionally control valves unstable because of their nonlinear characteristics. Discrete FFT analysis shows that the dominant cycles of the room temperature and chilled water supply temperature are nearly half an hour in all cases. Dominant cycle of the room temperature depends on the changes of chilled water supply temperature caused by ON/OFF motion of the compressors of chiller. ANN-based controller mitigates the influences of the low frequency disturbance by the cyclic changes of the chilled water supply temperature. It is illustrated that ANN is a really effective tool to control HVAC systems without tuning control parameters in cooling mode and heating and humidifying mode. The ANN controller is feasible and is expected to use actual buildings.

View full abstract

In order to accurately evaluate the dynamic behavior of HVAC system, it is necessary to use dynamic simulation program with accurately verification. In the present paper, the reproducibility of a dynamic simulation program called HVACSIM^+(J) was verified by comparing the actual measurements and simulated results in two kinds of states, normal and fault, for a HVAC system with single duct and constant air volume. As the result, it was verified that the program could reproduce the dynamic behavior of the system with sufficient accuracy, if users correctly input characteristic values of the real system and properly adjust some parameters in the component model as well.

View full abstract