건물에너지 효율적 관리를 위한 실내온도 제어방법에 관한 연구

Abstract

In this dissertation, a building internal temperature control algorithm in cooling systems of a large glass-covered single zone building is proposed using control horizon method. An important performance in building thermal control is to ensure thermal comfort with minimal energy consumption. A control horizon switching method is considered to use occupancy schedule for optimal temperature control and the optimization problem is solved with LP(linear programming) or PSO(particle swarm optimization) algorithms. Both TOU(time of use) and MD(maximum demand) electricity rates are included to calculate the energy costs. Simulation results show that the reductions of energy cost and peak power can be obtained using proposed algorithms. Demand response in building energy systems is an approach to give incentives to customers to change their electric usage pattern from their regular practice, in response to the time-varying price of electricity. The use of building thermal storage has been recognized as an important mean to reduce the peak demand for decades. Model-based control algorithms are desirable for both building designers and operators in that they can be simulated and tested even before a building is actually built. Moreover, an effective control system for one building is relatively easier to be adjusted and applied to another. Nowadays, TOU electricity rates have been implemented on most smart grids for demand response. A typical model of a building can present some nonlinearity and include several hundreds of states. Though model reduction techniques can reduce this number, they require an initial detailed model of the building and are not effective for nonlinear models. Here the structure of the low-order model is found directly and the values of the parameters are estimated by using experimental identification techniques. Control horizon algorithms has several features that make it suitable for the problems encountered in intermittently heated buildings. This model-based predictive control algorithm optimizes not only the comfort but also an energy usage. As cooling systems generally consume energy to provide thermal comfort, it makes a trade-off between energy savings and thermal comfort. This dissertation focuses on the application of control horizon algorithm to cooling systems that are charged on TOU and MD rates. A control horizon approaches with LP and PSO algorithm are selected to model and simulate the cooling systems. An model-based predictive control strategy is selected, because its periodic reoptimization characteristic provides stability during external disturbances. The periodic reoptimization also compensates for inaccurate or simplified system models.