A Hierarchical Framework-Based Coordinated Optimization of Building HVAC Systems and EVs

The demands of electric vehicles (EVs) and building heating, ventilation, and air conditioning (HVAC) systems have considerable flexibility. Their flexibility is influenced by occupants’ behavior resulting in huge complementary and dispatchable capability. Therefore, the coordination of EVs and HVAC systems holds significant potential for optimizing the building demand profiles and energy cost under time-of-use (TOU) tariffs. However, solving the coordinated problem in practice still faces the challenges in computational complexity and global information requirement due to the spatio-temperal coupling constraints. In this paper, a mixed-integer linear programming model is developed to formulate the multi-building energy system with EVs and the impact of occupants’ behavior on the demand and flexibility of the system. The problem is converted into a three-level structure using Lagrangian relaxation framework. A dynamic programming-based Lagrangian relaxation (DPLR) algorithm is developed to independently solve all sub-problems of the three-level structure in a decomposition and coordination way while avoiding the iterative computation between the middle and lower level. The numerical results show the developed method can obtain a near-optimal solution in an efficient way without perceivable degradation in accuracy, which is 4% worse but at least three times faster, compared to the existing centralized algorithm. Note to Practitioners—The escalating demand for EVs and HVAC systems results in increased energy costs and challenges to existing power systems, such as frequency deviations and higher peak loads. Therefore, this paper focuses on the coordinated optimization of the EV charging and building HVAC system operation, while considering rooftop photovoltaic generation supply within the system. Optimal coordination of the above system can effectively reduce energy costs under TOU tariffs and enhance the ability to utilize renewable energy sources. However, solving the coordinated optimization problem still faces difficulties since the computational complexity will exponentially grow with increasing problem scale due to the spatio-temperal coupling between the demand of EVs and HVAC systems. Therefore, a DPLR algorithm is developed. There is a central coordinator that collects the demand information of EVs and HVAC systems and broadcasts the coordination information calculated according to the demand information. Every single EV and HVAC system can make decisions based on the coordination information independently. The DPLR algorithm decouples EVs and HVAC systems and avoids the curse of dimensionality. It has great improvement on computational efficiency and global information requirement reduction. The computational efficiency and effectiveness of the developed method is verified through multi-scale case studies. The numerical results show that compared with independent optimization of EVs and HVAC systems, coordinated optimization can reduce over 44% of the energy costs.