Modeling And Control Of Refrigeration Temperature In The Combined Cooling,heating And Power System Based On T-S Model

As one of the important directions of current energy development,the combined cooling,heating and power system has not only achieved a higher energy utilization rate,but also achieved the goal of environmental protection and economic development.However,the control complexity of the system is relatively increasing,and the transmission path of the energy supply medium is long,resulting in a large temperature delay,so it is difficult to achieve real-time temperature control.Among them,the refrigeration temperature is one of those important parameters which need to be monitored in the actual production process,so reducing the time delay of the system and achieving the best control effect of the refrigeration temperature are of great significance for the application and development of the combined cooling,heating and power system.In the combined cooling,heating and power system,the refrigeration temperature directly affects the refrigeration effect of the whole system.Due to the restriction of a variety of production processes,the system has the characteristics of many disturbances,large time-delay,time-varying,nonlinear,strong coupling,pure delay and so on.In this paper,the boiler-steam driven combined cooling,heating and power system of a power plant is taken as the research object,the technological process of the system is analyzed,and the relative influencing factors of the refrigeration temperature are analyzed from the main equipment of the system.On this basis,the modeling and control of refrigeration temperature are studied.The prediction model of refrigeration temperature is established by using T-S fuzzy algorithm.The combination of subtractive clustering and fuzzy C-means clustering algorithm is used in the model structure identification,which not only improves the accuracy of model,but also avoids the shortcomings of their respective algorithms.The hybrid learning algorithm which combines the least squares algorithm and the ordered derivative algorithm in fuzzy neural network is used for parameter recognition in order to obtain more accurate parameters.Real-time running data of cooling water supply temperature,cooling water flow rate,hot water supply temperature,hot water flow rate and frozen water supply temperature were collected from the DCS of a power plant,and multi-input,single-output modeling and simulation training is carried out on the MATLAB platform.The results show that the model obtained by this modeling method has better prediction results and can accurately fit the change trend of the freezing water temperature,which provides the basis for the control and optimization of the back refrigeration temperature.In this system,the refrigeration process is relatively complex,the parameters are changeable,the control requirement of refrigeration temperature is high,and the time delay is large,so it is difficult to realize precise control by traditional control.In this paper,a multi-step predictive controller based on the T-S state space model is designed on the basis of T-S model and combined with multi-step predictive control method,and the transformation from constrained nonlinear optimization to linear quadratic programming is realized.Thus,the calculation of system optimization is reduced.In order to study the predictive control method more conveniently in the process of research and analysis,the general form of T-S model is transformed into the state space expression form.Finally,the T-S fuzzy model of the temperature of frozen water is established by using the flow rate of cooling water as the input.Specifically,the T-S fuzzy model of chilled water temperature is established by taking cooling water flow as input,and the simulation of the cooling temperature control of the system is carried out by combining the designed control method.The results show that the system output response is relatively stable and the effect of controlling is better.