On the Feasibility of Modeling Two-Phase Flow-Induced Fluidelastic Instability in Tube Bundles

In the 80's a number of theoretical models were developed to model fluidelastic instability, primarily for single phase flows. The models ranged from purely analytical models to semi-empirical models requiring considerable experimental data as input. While these models were very successful in uncovering the nature of fluidelastic instability and the underlying mechanisms in single phase flow, this work seemed to stop short of getting to the next step of practical application to two-phase flows. During the same period, Connors formula became 'entrenched' in industry to the extent that the formula now forms part of the design norms against fluidelastic instability. In an ongoing research program the quasi-steady model has been chosen as a possible candidate for modeling fluidelastic instability in two-phase flows. This paper discusses the challenges associated with accurate modeling of fluidelastic instability in two phase flows using this and other models. The unsteady model is shown to have limitations when it comes to measuring accurately the necessary unsteady fluid force coefficients. A comparison of the stability analysis results with experimental measurements shows that the quasi-steady model can give a reasonable estimate of the instability velocity as well as the inter-tube dynamics. Finally, the remaining challenges, before the quasi-steady model and possibly other models can be fully implemented for prototypical conditions are discussed. In particular the need for more work to understand the flow itself is highlighted.