STUDY ON MULTIPLE INPUT MULTIPLE OUTPUT TYPE MODEL PREDICTIVE CONTROL IN ORDER TO REDUCE MUTUAL INTERFERENCE OF AIR CONDITIONING INDOOR UNITS

Abstract

 In recent years, Multi-split packaged air-conditioning systems have recently become commonplace in office buildings in Japan, and in many cases one space is served by several indoor air-conditioning units. The advantage of a multi-split packaged air-conditioning system is that each indoor unit can be operated and controlled individually, with users being able to change the temperature settings of the indoor units as desired. However, this can cause air-conditioning control problems such as a deterioration in mutual interference between adjacent indoor units. In particular, when each air-conditioned zone has a different internal load and set temperature, the sensor temperatures are influenced by disturbances, the indoor units are driven harder than they need be, energy consumption increases, and indoor air quality worsens.

 In this study, we carried out CFD analyses incorporating MIMO MPC that controls each zone while considering the thermal environment between adjacent zones in an office space in which the set temperature differs between zones. Furthermore, we showed the effectiveness of this method by comparing it with SISO MPC.

 Based on the results, the following was found out.

 1. When comparing the outlet air temperatures of the respective cases, the fluctuation range of the outlet air temperatures of the periWS and inW02 zones where the set temperature is 26°C is large in CASE 1 but small in CASE 2.

 2. The averaged room air temperature in the zone where the set temperature was 27 °C showed almost the same behavior in both control methods, but comparing the behavior of the occupancy-area sensor temperature at each point, the temperature fluctuation range in SISO MPC was large but that in MIMO MPC was small. The reason is presumed to be that the fluctuation of the room temperature was suppressed in MIMO MPC in accordance with the difference in the behavior of the outlet air temperature of the adjacent zone.

 3. From the distribution of temperature and flow velocity vectors, it was confirmed that not only the sensor temperature but also the entire work space was controlled to the set temperature in each zone.