Modeling And Optimization Of A Direct Air-Cooled System Based On Divisional Control

Large direct air-cooled systems with significant water-saving advantages are widely used in thermal power plants in areas rich in coal and short of water in China.However,under the influence of the ambient wind,the flow deformation and plume recirculation at the inlet of the axial flow fan array of the direct air-cooled system lead to the decline of the heat transfer performance of the system,which may cause the fluctuation of turbine back pressure,posing challenges to the stable operation of the direct air-cooled power plant.Therefore,to solve the problem of back pressure fluctuation of the direct air-cooled system under the influence of ambient wind,it is of great significance to establish an accurate model of the direct air-cooled system and realize accurate control of back pressure,so as to improve the system performance and reduce the operating cost of the direct air-cooled system.In this thesis,for the large-scale direct air-cooled systems under the influence of ambient wind,combined with the analysis of the heat transfer process mechanism,a data-driven model of the direct air-cooled systems is established and the accurate prediction of turbine back pressure is realized.The axial fan array performance experiment is conducted on an experiment system with a scaled-down air-cooled fan array,and then a divisional fan rotational speed control strategy of the direct air-cooled system is proposed.The main work of this thesis is as follows:1)Mechanism modeling and analysis of the direct air-cooled systemThe establishment of the process model of the direct air-cooled system and the deduction of the linear expression of the model is introduced.The influence of the ambient wind speed and direction on the system is expressed through the parameters of fan volumetric efficiency and plume recirculation rate respectively.The model was verified by the operation data and fan step test of a 330MW power plant.Through the simulation experiments,the influence of fan rotational speed,ambient temperature,turbine exhaust mass flow rate,fan volumetric efficiency,and plume recirculation on the turbine back pressure was analyzed.2)Data-driven model and back pressure prediction of direct air-cooled systemTo solve the problem of the inaccurate of mechanism model of the direct air-cooled system under the influence of equipment aging and environmental factors,a real-time updated back pressure data-driven model of the direct air-cooled system was established.Data preprocessing methods,including outlier detection based on global representative indicators,steady-state identification,and data standardization are introduced.Furthermore,based on the sparse least squares support vector regression model,a fast calculation method based on probability and a real-time model update strategy are proposed.Finally,the back pressure of a direct air-cooled system is predicted using the operating data in the DCS system of the power plant.The proposed algorithm has the advantages of good sparsity,high accuracy,and fast calculation speed,which can provide accurate operation guidance for operators and effective real-time information for the back pressure optimization control strategy in the following chapters.3)Experimental study on the direct air-cooled fan arrayTo solve the problem that the inlet airflow of the direct air-cooled fan array decreases under the effect of ambient wind,a 1:10 scaled-down experimental system of the direct air-cooled fan array was established.After the preprocessing of ambient wind speed and direction with empirical mode decomposition,the performance of the fan array and the identification and control of fan inlet airflow is studied.In the fan performance experimental study,the random forest regression algorithm is used to generalize the experiment results and estimate the volumetric efficiency of the fan.The relationship between the volumetric efficiency of each group of fans and the position,operation mode,fan rotational speed,and ambient wind speed is discussed.The energy-saving analysis of fan rotational speed divisional regulation is carried out.In the study of fan inlet airflow dynamic model identification and control,an integral-based model identification method is adopted to identify the fan inlet airflow process and disturbance model of ambient wind,and a PID control strategy with feedforward compensation is designed to suppress the disturbance of ambient wind on the fan inlet airflow.4)Fan divisional regulation of the direct air-cooled system based on multi-model predictive controlAiming at the problem of plume recirculation in the direct air-cooled system under the influence of ambient wind,a fan divisional regulation scheme of the direct air-cooled system based on the multi-model predictive control strategy is designed.Through the hierarchical clustering method with gap metric as the distance,the axial fan array divisional schemes are obtained by clustering analysis of the fan volumetric efficiency,and prediction model set under multiple working conditions are obtained by clustering analysis of the back pressure linearization models under different plume recirculation rate.The adaptive particle swarm optimization algorithm is used to calculate the optimal back pressure as the setting value of the back pressure control system.The control objectives of tracking back pressure setting value and minimizing the fan energy consumption are designed.With the proposed control strategy,the rotational speeds of fans in different regions are dynamically adjusted,reducing the actions of internal fans,and improving the control effect in the plume recirculation,system stability,and the economic benefits of the power plant.