Integrated Computational and Experimental Method for Flow Problems

点击翻译按钮获取中文摘要

Numerical analysis and experiment are common tools in flow analysis. Each of these methods, however, has own shortcomings. The result of numerical analysis inherently depends on the mathematical model and the computational grid. A consistent way of validation of numerical solutions has not been established yet. Evaluation of result is relatively easy in experiment, but it usually gives only limited information of the whole flow field. It is, therefore, desirable to establish a new analysis methodology in which the numerical and experimental methods are integrated to compensate their shortcomings.In the present research the concept of "observer" in automatic control theory is applied to flow problems. In the observer the state of the physical system in asymptotically estimated from the mathematical model to which the measurement data are fed back in real time. The SIMPLER-based numerical simulation algorithm is used as the mathematical model and the boundary condition and/or external force terms are modified based on the difference between the experimental measurement and computational result. Validity of the present flow observer was examined through numerical analysis for the turbulent flow in a square duct. Possibility in application to turbulent flow control was also studied. The fundamental grid convergence property of numerical methods, and characteristics of the collapsible tube flow system, in which the flow and the tube deformation interact, were investigated as the basic understandings of the future research.These results give fundamental information in establishing the new methodology to deal with especially complex flow problems to which conventional methods in experiment or computation do not apply effectively.