A Research Of Predictive Control In Indoor Temperature Control System

As acceleration of urbanization in China, the development of energy-saving building has become one of the most important portions in sustainable development strategy. The indoor temperature control system receives more and more attentions as an important subsystem of building facilities,which consists of the architectural design, electrical design, automatic control design and so on. The indoor temperature control system is researched in this paper.The step response curve of temperature is acquired by an experiment of the radiant floor heating system. For the large inertia of the system, the dynamic matrix control(DMC) based on the step response of controlled plant is applied to radiant floor heating control system. The DMC has some advantages such as simpler algorithm, less calculation, strong robustness and so on, without considering of the structure and order of model. Simulation of an example shows the validity of this control method.Against the large inertia, parameter uncertainty and stochastic perturbation of the radiant floor heating system, a constrained generalized predictive control(GPC) based on quantum genetic algorithm(QGA) is proposed. In order to circularly optimize the incremental quantity of GPC, the control incremental quantity is used as the quantum chromosomes. The method of multi-step forecasted is used in GPC, which remarkablely enhances anti-disturbance performance and resistance to random noise. Controll accuracy is improved with the diversity of particles and a better search ability on the global solution space of QGA. Simulation of an example showed that the proposed control method provides satisfactory performance and strong robustness, comparing with GPC control. Simulation of an example showed that the proposed control method provides satisfactory performance and strong robustness, comparing with GPC control. In this paper, fractional order PIλDμcontroller is used in a MIMO fractional nonlinear air-condition system model, and the parameters of controller are optimized by QGA. Statements range and Control Plane of system were expanded enormously by fractional order system theory, when system features were described by fractional order differential equations, which is an efficient and accurate theoretical tool. In order to reduce the influence of stochastic perturbation, outputs are multi-step forecasted by gray prediction in feedback channel. Simulation showed that the outputs of the air-condition system are desired.