Abstract

In spite of the remarkable development of technology, most studies for building energy controls to evaluate or estimate the energy performance have not accurately reflected actual building’s energy consumption patterns. For this issue, several techniques, such as simulation and calibration, comprehensive survey system, smart metering, and commissioning, have been attempted.However, in most studies, some factors in thermal systems derived from occupant behavior were perceived as fixed objects, and the factors were converted into simple numbers as parts of inputs into simulation templates. There was lack of studies on considerations that unpredictable responses derived from human anti-logic or common sense could deteriorate energy efficiency in theoretical analyses even though the systems were properly operated.This research proposes integrated energy supply models based on artificial intelligence responding to anti-logic or common sense that can reduce machine’s energy saving effects. By use of design scenarios assuming some unusual situations, a decision making model determines the extent to which the cause of the abnormal situations are associated with the occupant behavior. After the five-step phases in the decision making model, the actual outputs of the energy supply model for the buildings are determined, and the reciprocal communication between the thermal and decision making models mitigates thermal dissatisfaction and energy inefficiency. Comparative analysis describes the decision making model’s effectiveness that it improves thermal comfort levels by about 2.5% for an office building and about 10.2% for residential buildings, and that it reduces annual energy consumption by about 17.4% for an office building and about 25.7% for residential buildings. As a consequence, the integrated energy control model has advantages that it noticeably improves thermal comfort and energy efficiency, and that it properly respond to abnormal and abrupt indoor situations derived from human anti-logic or common sense.

Full Text
Paper version not known

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call