Bilayer stochastic optimization model for smart energy conservation systems

Energy conservation is a critical decision among industries with intensive energy usage. This study investigates a smart energy conservation system that consists of a central makeup unit (MAU) and a set of dry cooling coils (DCCs) in a manufacturing shop floor. The system is empowered by the proposed adaptive optimization control for MAU and DCCs. A bilayer stochastic optimization model is presented for saving energy and damping temperature against the uncertainties of atmospheric temperature and indoor heating sources. In our modeling strategy, the MAU for optimal control is constructed as the upper layer model, while the set of DCCs for distributed optimal control is considered the lower layer model. The two models correlate with each other. To achieve stochastic optimality, a scenario-based sample average approximation solution algorithm coupled with a genetic algorithm is developed for dynamically making optimal valve opening decisions for MAU and DCCs over time. Experiment results indicate that the proposed model and solution algorithm effectively control manufacturing temperature damping against uncertain outdoor and indoor temperatures while consuming less energy. • Bilayer stochastic optimization of saving energy and damping temperature for MAU and DCC. • Deal with the uncertainties of atmospheric temperature and indoor heating sources. • Scenario-based sample average approximation algorithm coupled with a genetic algorithm is developed.