Primitive model for the random excitation of a tube under two-phase cross flow

Flow-induced vibration of heat-exchangers tubes is particularly studied in the nuclear industry for safety and cost reasons. It implies to have, among others, relevant characterizations of the random buffeting forces the cross-flow applies to the tube bundle. Work is still needed in this domain, particularly for two-phase flow, to improve the available data as the ones for PWR steam generator, currently very envelope. In parallel to get new experimental data, using 'real' or substitutional mixtures (e.g. air-water instead of steam-water for PWR), it is essential to understand the basic excitation mechanisms which induce the vibrations under two-phase flow, as e.g. the influence of flow regimes. In this general framework, what can be learnt from deliberately simple models may be a contributive help. As a first attempt on this issue, the paper deals with the elementary case of a single rigid tube under air-water cross flow. This case is part of experiments carried out at CEA-Saclay with bundles where both tube support reactions and flow characteristics are measured, with respectively piezo-electrical sensors and bi-optical probes (BOP). The information provided by the BOP (mean interface velocity, statistical distribution, etc.) feeds a primitive model of water 'droplet' impulses on the tube, based on a lot of crude assumptions about impact velocity, momentum conservation, impulse shape, statistical independence, etc., and which uses analytical results of random processes constructed from the superposition of random pulses. The 'equivalent' excitation force, obtained in terms of dimensional PSD, is compared to the one measured in the drag and lift direction with an acceptable agreement, at least in order of magnitude. Comments and lessons are drawn from this first attempt, and some paths are advanced to improve this kind of primitive models, especially for treating rigid square bundles under air-water cross flow. (authors)