Study Of Indoor Ventilation Characteristics Based On CFD Simulation And Multi-Objective Coupled Preference Decision Optimization Method

Ventilation systems are used to create an acceptable indoor environment for residents in buildings.Although they consume a large amount of energy to ensure thermal comfort(TC)and indoor air quality(IAQ),dissatisfaction with the thermal environment is still widespread among indoor personnel.In recent years,the optimization of ventilation systems based on computational fluid dynamics(CFD)has received widespread attention.However,current optimization of ventilation systems has certain limitations,which cannot meet the requirements of maximizing energy savings while achieving TC and good IAQ.This work conducts a yearround optimization study on the energy-saving stratified air distribution ventilation system with theoretical analysis,experiments,simulation technology,and intelligent optimization methods to meet the residents’ requirements for TC and IAQ while saving energy as much as possible.Firstly,a parametric study based on simulations,experiments,and sensitivity analysis is conducted for a floor standing air conditioner(FSAC)with forced cooling capacity in cooling applications.The comparison between the individual and combined effects of operating parameters on ventilation performance shows that each ventilation performance index can be determined in a larger proportion by the corresponding most important operating parameters.Based on the result,a control strategy is proposed to meet the requirements of TC and IAQ while saving energy to a certain extent.Secondly,the prediction performance of a new neural network algorithm——Extreme learning machine(ELM)for the CFD simulation database of the impinging jet ventilation(IJV)system with energy saving potential and suitable for heating is explored in heating application.By comparative analysis with the modeling algorithms of response surface methods widely used in the field,it is concluded that data-driven prediction models based on ELM neural network models have higher accuracy and are more reliable.Finally,this work develops a model combining multi-objective optimization(MOO)and multi criteria decision-making(MCDM)to determine the optimal operating parameters.Meanwhile,control the ventilation performance indicators with TC or IAQ international standards within an appropriate range;For indicators without relevant TC or IAQ standards,they are weighed against energy consumption indicators through the MOO method.Compared with the TOPSIS method,the proposed method reduces the EC by 16.99% on average,while still meeting TC and IAQ requirements.In addition,this method can obtain the optimal return vent and supply mode for different decision-making preferences,also,can maximize energy saving when meeting the TC and IAQ requirements.In summary,this work provides a framework for the accurate design and dynamic operation control of ventilation system parameters,and also contributes to the practical application of ANN as a data-driven prediction model in different ventilation scenarios with lower design and calculation costs.In addition,this study found that MCDM technology embedded with the MOO method can further save energy while meeting indoor air quality and personnel thermal comfort.