US 12,228,303 B2
HVAC system using interconnected neural networks and online learning and operation method thereof
Ye Eun Jang, Pohang-si (KR); and Young Jin Kim, Pohang-si (KR)
Assigned to POSTECH RESEARCH AND BUSINESS DEVELOPMENT FOUNDATION, Pohang-si (KR)
Filed by POSTECH RESEARCH AND BUSINESS DEVELOPMENT FOUNDATION, Pohang-si (KR)
Filed on Dec. 30, 2021, as Appl. No. 17/566,587.
Claims priority of application No. 10-2021-0023592 (KR), filed on Feb. 22, 2021.
Prior Publication US 2022/0268479 A1, Aug. 25, 2022
Int. Cl. F24F 11/64 (2018.01); F24F 11/47 (2018.01); G05B 13/02 (2006.01); G06N 3/08 (2023.01); F24F 110/12 (2018.01); F24F 110/32 (2018.01); F24F 130/20 (2018.01); F24F 140/20 (2018.01); F24F 140/50 (2018.01); F24F 140/60 (2018.01)
CPC F24F 11/64 (2018.01) [F24F 11/47 (2018.01); G05B 13/0265 (2013.01); G06N 3/08 (2013.01); F24F 2110/12 (2018.01); F24F 2110/32 (2018.01); F24F 2130/20 (2018.01); F24F 2140/20 (2018.01); F24F 2140/50 (2018.01); F24F 2140/60 (2018.01)] 14 Claims
OG exemplary drawing
 
1. A heating, ventilation, and air conditioning (HVAC) system, comprising:
air conditioning sensor units installed in or outside a building to detect environmental data;
an HVAC device configured to supply thermal energy into an inner space of the building using input power;
a predictive controller configured to generate operational data based on the environmental data and control the HVAC device by adjusting the input power; and
an artificial neural network (ANN) training unit configured to perform operations of predicting indoor temperatures of the building by using an integrated model to generate predictive operational data,
wherein the integrated model comprises:
a first sub-model configured to receive a temperature set point, an indoor temperature of a previous time, and an input power of the HVAC device of the previous time to calculate a next input power of the HVAC device;
a second sub-model configured to receive input power of the HVAC device, an ambient temperature, and an evaporator-side air temperature to calculate an output cooling power of the HVAC device; and
a third sub-model configured to receive the output cooling power of the HVAC device, atmospheric environment variables, and the indoor temperature of the previous time to calculate an indoor temperature of the building,
wherein the predictive controller is configured to control the HVAC device by adjusting the input power to reduce a difference between the indoor temperature of the detected environmental data and the predicted indoor temperature.