Research On Data Driven Modeling And Control Of Thermal Comfort Under The Internet Of Things

Buildings provide us proper environments for living and working. With the development of economy, the buildings should not only stress external artistic beauty, but also meet the basic requirements on energy saving and comfort. However, there are some conflicts between the comfort and energy consumption objectives. Both objectives need to be coordinated and optimized with the corresponding strategy. This paper proposes a data-driven modeling and control strategy of thermal comfort under the IOT environment. This study mainly involves the following aspects:Firstly, environmental data are collected by the Internet of Things for Building Equipments, and then preprocessed using statistical methods. The preprocessed data will be used in the following research.Secondly, This paper proposes a method to build data-driven type-2 fuzzy set model. Compared with the traditional fuzzy method, the ability of the type-2 fuzzy set on dealing with uncertainties is obviously enhanced. This method is applied to model the thermal comfort in indoor environment. And, based on the type-2 fuzzy thermal comfort model, a control strategy for the indoor environment control is put forward. Simulation results demonstrate the effectiveness of the proposed strategy through modeling the thermal comfort of a specific room.Thirdly, in this paper, a data-driven single-input-rule-modules (SIRMs) connected neural-fuzzy method is presented for multivariate fuzzy systems. The proposed method can effectively solve the fuzzy rule explosion problem that usually faced in traditional fuzzy systems. The proposed SIRMs connected neural-fuzzy system is applied to forecast the cooling and heating load. Simulation results showed the efficiency and superiority of the proposed method in the energy consumption prediction problem.Fourthly, a multi-objective optimization strategy is proposed to realize the low-energy and high-comfort objectives. This strategy uses the genetic algorithm with strong search capabilities to obtain a set of optimal solution sets (also called Pareto optimal solution set). And, based on the priori knowledge and the practical application background, an optimal solution is provided to balance the conflicting objectives.