**4. Conclusions**

This study proposes an HVAC system optimization control strategy involving FCU temperature control. This strategy involves three main techniques: dynamic FCU temperature setting, indoor heat demand conversion, energy consumption optimization control strategy for a single chiller unit of HVAC systems. The algorithm for dynamic FCU temperature setting estimates a comfortable indoor temperature according to outdoor and sensed indoor temperatures, and automatically adjust the di fference between the indoor comfort temperature (*Tc*) and the indoor sensing temperature (*Ti*) every 30 minutes, and achieve the indoor comfort temperature (*Tc*) within the indoor comfortable range (23–26 ◦C). The di fference in temperature is derived and then converted into an RT for a HVAC system by using fuzzy rules and weight rules on the basis of the change the temperature and wind speed of the FCU by the user. The derived RT is considered the actual cooling air demand of an area for the chilled water system. A genetic algorithm is then applied to adjust chilled water and cooling water flow and calculate the operating frequency of the pumps that are required to meet the cooling air demand and stipulated minimal energy consumption. The e ffectiveness of the proposed strategy was verified in a field experiment. The study compared the operating status levels of the chilled water pump, cooling water pump, and cooling tower fans in the chosen building during the summer periods (i.e., from June–August) of 2016 and 2017. In 2016, the machines operated at full load, whereas in 2017, the proposed strategy was applied, which enabled the adjustment of the operating frequencies of chilled water system and cooling water pump and airflow of the cooling tower fans of the HVAC system to reduce unnecessary power consumption. Therefore, compared with that in 2016, the actual total power consumption of the HVAC water system in 2017 was reduced by 39.7%, on average. The operating e fficiency of the chilled water system was enhanced to 0.68 from 0.8 observed when the system operated at full load. Energy consumption per RT was reduced by approximately 15.6%. According to the baseline energy consumption obtained by a regression model used to calculate the power consumption of a chilled water system, the proposed strategy can achieve energy conservation when temperatures increase.

**Author Contributions:** Formal analysis, H.-Y.L.; Methodology, C.-M.L.; Software, K.-Y.T.; Validation, S.-F.L.

**Funding:** Bureau of Energy, Ministry of Economic A ffairs, Taiwan: 108-E0203.

**Acknowledgments:** We appreciate National Chiao Tung University for their generous assistance of computing time and Bureau of Energy, Ministry of Economic A ffairs, R.O.C. (Taiwan) for the finical support.

**Conflicts of Interest:** The authors declare no conflict of interest.
