Research On Control Strategy Of Central Air-conditioning System Based On Active Disturbance Rejection-iterative Learning Control

The central-air conditioning(CAC)system is widely used in China and its energy consumption accounts for up to 50% of the total energy consumption in large buildings.Thus the research of energy saving control of the central air conditioning system is of great significance.The central air conditioner is a complex system aiming at air temperature and humidity processing.The system,which can operate in diversified modes,is composed of many equipments and can be connected in different ways.From a control point of view,the central air conditioning system is a complicated system with the characteristics including high nonlinearity,strong coupling,large delay and multiple disturbances.The energy-saving control strategy of CAC is studied,and the various subsystems of CAC are comprehensively considered,including the chilled/cooling water temperature setting,the variable flow control,the device start-stop strategy and control strategy.of the variable air volume(VAV)system.In order to overcome the problems of central air-conditioning system modeling difficulties and complex control strategies,an accurate building-air conditioning model was established through the TRNSYS platform,and a TRNSYS-MATLAB co-simulation system was built to facilitate the study of the effects of control strategies and meanwhile,improve the practicality of the designed control system.The active disturbance rejection based iterative learning control(ADR-ILC)algorithm is studied.ADR-ILC classifies all uncertainties of the system as total disturbances(including internal and external disturbances of the system),and extends the extended state observer(ESO)to the iterative domain.For periodic disturbances,it is possible to accelerate the convergence of estimation error of the total disturbance term by iteration.Design the control law to compensate for the total disturbance of the system to simplify the control problem of nonlinear and uncertain systems into a known linear system control problem.First of all,this paper deeply studies the active disturbance rejection based iterative learning control(ADR-ILC)algorithm.ADR-ILC extends the extended state observer(ESO)from time domain to the iterative domain.Similar to ADRC classifying all uncertainties of the system as total disturbances,including internal and external disturbances and designing the control rate to compensate for the total disturbance term,ADR-ILC accelerate theconvergence of the total disturbance term estimation error by iterative mechanism,disturbances,so that repeated control processes or system with periodic disturbance can obtain better dynamics response performance and stronger anti-interference ability.Then according to the time-varying law of air-conditioning load,the ADR-ILC algorithm is used to modify the linear regression prediction value.The prediction accuracy of air-conditioning demand is as high as 3.1%,which provides an important foundation for air-conditioning group control,parameter optimization and system energy-saving control.Then ADR-ILC is applied to the terminal temperature control.For this repeated control process,the ADR-ILC algorithm effectively solves the difficulties of control such as periodic interference,strong coupling and nonlinearity of the VAV system.After enough iterations,the area temperature can quickly track the target temperature,obtaining smaller static error and stronger disturbance rejection ability with the ADR-ILC controller,whose control effect is significantly better than other algorithms such as fuzzy PID and ILC.Finally,a CAC intelligent management system is designed for the production workshop of a factory in Guangzhou.Control strategies including device group control,chilled/cooling water flow control,chilled/cooling water temperature reset and other control methods are designed to ensure energy-saving and efficient operation of the system.