| Solid Oxide Fuel Cell(SOFC)power system is a clean and efficient energy conversion device with the advantages of high efficiency,no pollution,solid structure,and fuel flexibility.It has a promising solution in the fields of military,transportation,and power station.As the core component to generate electricity,SOFC stack must work safely,stably and efficiently,which can finally make sure that SOFC system works normally in practical application scenarios.However,the SOFC stack works under the temperature of 600~900 °C,which brings many challenges on high-efficiency and long-term running of the SOFC system.When the temperature is too low,the stack cannot reach the normal thermal activation state,which will affect its electrical performance.When the temperature is too high,the stack lifetime will be shortened due to the insufficient material durability.In addition,there exists temperature gradient due to the different local temperature inside the stack.When the temperature gradient is too large,the thermal stress on the stack material is non-uniform,and may result in stack deformation or even cell cracking.Therefore,it is necessary to manage the stack temperature and temperature gradient within a proper range for its good performance and long lifetime.The basis of the stack temperature management is to obtain the stack internal temperature.However,it is difficult to install thermo-couples to directly detect the temperature inside the stack due to the strict sealing requirement on the stack structure.As for cross-flow planar SOFC stack,its internal temperature is more complexly distributed on a two-dimensional plane,which is more difficult to obtain with thermo-couples.Thus,it is necessary to design a temperature observer with limited and easily measured input and output variables to fast and accurately estimate the stack internal temperature distribution,which has great significance in the stack temperature management.In this thesis,a mechanism model for a cross-flow planar SOFC stack is built and validated.Then through the analysis of the model inputs,outputs and its dynamic characteristics,the observer input parameters and the model discretization step size are provided for the subsequent observer design.Finally based on unscented kalman filter(UKF)algorithm,a temperature distribution observer is implemented to fast and accurately estimate the stack internal temperature distribution on different operating conditions.The detailed contents are proposed as the following.Firstly,a two-dimensional(2D)mechanism model for a cross-flow SOFC stack is built and the experiment data obtained from the long-term test of an integrated SOFC power generation system is utilized to verify the model and ensure its accuracy.Then,with the input and output parameters of the stack model,the observer input variables are analyzed and determined;With different power conditions,the dynamic stack characteristics are analyzed to obtain the model discretization step size for the observer design,which can provide a basis for high-precision temperature estimation under power switching conditions.Finally,the stack model is discretized and then combined with the UKF algorithm to develop a two-dimensional temperature distribution observer for the cross-flow SOFC stack.With the input step disturbance and power switching process,the designed observer can accurately and fast estimate the internal temperature distribution of the stack,and has a good performance in the dynamic temperature response.The researching work in the thesis has solved the problem that the cross-flow planar SOFC stack temperature distribution can't be directly detected by thermal-couples.To fast and accurately estimate the stack temperature distribution lays a solid foundation for the stack thermal management and finally ensures that the stack and SOFC power generation system can finally work safely?stably and efficiently in practical application scenarios. |
PDF Full Text Request