Predictive control of multi-zone variable air volume air-conditioning system based on radial basis function neural network

The multi-zone variable air volume (VAV) air-conditioning system is a complex thermal system with large delay and nonlinearity. Due to the complex environment of multi-zone buildings and the complicated operation process of the VAV air-conditioning system, there are many difficulties in the room temperature control. This paper firstly establishes a multi-zone building model for room temperature using resistance-capacitance method. This investigation simulates and measures the dynamic response of room temperature in a three-floor building without/with air-conditioning for validation. Then a multizone VAV air-conditioning system room temperature predictive control model based on radial basis function (RBF) neural network (NN) is proposed. This study sets up a multi-zone VAV air-conditioning system experimental platform in the three rooms on the first floor of the building and implements the predictive control model based on the RBF neural network. The experimental results show that the predictive control model based on RBF NN is able to meet room temperature requirements. It also has strong antiinterference performance and ensures stable static pressure of the main air supply duct. The multizone building model can accurately simulate the temperature changes of each room when the air supply volume varies. (c) 2022 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).

Automated summary

This study investigates the difficulties in room temperature control in a multi-zone variable air volume (VAV) air-conditioning system and proposes a predictive control model based on radial basis function (RBF) neural network. Experimental results show that the proposed model can meet room temperature requirements and ensure stable static pressure of the main air supply duct.