Multi-objective optimization of pressure regulators in buildings' HVAC systems

The control accuracy of the adjustable dynamic flow balance valve regulators is closely related to the energy consumption of the buildings' HVAC system. Therefore, optimizing the control accuracy of the pressure regulator plays an important role in reducing the energy consumption of the HVAC system and indoor cooling and heating comfort. In the study, the transient dynamics calculation method was used to simulate the movement process of the regulator diaphragm. The rebound force and fatigue life of the diaphragm are the optimization goals. An improved whale algorithm is proposed to optimize the Kriging model. The sensitivity analysis of the objective function surrogate model was carried out by the single factor analysis method. Based on the improved seagull algorithm, the multi-objective optimization of the diaphragm was carried out, and the corresponding MATLAB program was compiled. The optimized Kriging proxy model has higher fitting accuracy and smaller prediction error. The rebound force of the optimized diaphragm is reduced by 13.13% compared with that before optimization, and the fatigue life is increased by 87.89%. The dynamic characteristics and control accuracy of the adjustable dynamic flow balance valve after optimizing the diaphragm are verified by experiments. The maximum flow value is reduced by 2.03%, the transition time is reduced by 15.1%, and the stable flow value is closer to 20 t/h. This study contributes to the development of machine learning models with higher fitting accuracy and provides a new feasible design method for reducing energy consumption in buildings' HVAC systems.

Automated summary

Optimizing the control accuracy of pressure regulators is crucial for reducing energy consumption and improving indoor cooling and heating comfort in HVAC systems of buildings. This study used the transient dynamics calculation method to simulate the movement process of the regulator diaphragm and proposed an improved whale algorithm to optimize the Kriging model. The results showed that the optimized diaphragm had higher fitting accuracy, smaller prediction error, reduced rebound force by 13.13%, and increased fatigue life by 87.89%. The experiments confirmed the improved dynamic characteristics and control accuracy of the optimized diaphragm, resulting in a 2.03% reduction in maximum flow value, 15.1% reduction in transition time, and closer stable flow value to 20 t/h.