Figure 11. Pressure and temperature profiles along loop and brief description of observed flow pat-
tern regimes.

lets

Filmwise condensation

.

of the flow patterns observed in the boiler, super heater, and vapor throttle. At the end of the corrosion test for which the loop was designed, the loop was dismantled. The preheater, boiler, superheater, vapor throttle, and condenser were sectioned and metallurgically examined. Corrosive attack and corrosion product deposits on the boiler wall were studied. The corrosion results of this loop are reported in reference 10. A motion-picture supplement (C-265) has been prepared to illustrate the flow patterns that were recorded with the X-ray image system. The motion pictures show details that are not readily reproduced by the half-tone photographs in this report.

The flow patterns are described in terms of the mercury liquid phase, which was opaque to X-rays. Droplets slightly less than 1/32 inch (1.2 mm) in diameter could be detected. Dispersed flow of a fine mist of mercury would have escaped detection. A film of mercury thicker than about 10-3 inch (2.5×10-3 cm), with the film front (edge) moving relative to the boiler-superheater wall, could be discerned. In the superheater, high-quality saturated vapor could not be differentiated from dry superheated vapor.

Only the most visible and, hence, the densest part of the liquid stream is described. The description is therefore essentially of the bulk liquid flow, that is, flow of liquid having a thickness or diameter detectable by the X-ray technique used. This description should not be construed as excluding a thin liquid film that covered the boiler surface irrespectively of the bulk stream motion. This type of film-like coverage was inferrable after the loop was sectioned and examined for surface deposits and attack patterns.

Boiler-Superheater Flow Patterns

An X-ray scan from the boiler inlet through the superheater revealed variations in the liquid flow pattern. For example, the boiler could be divided into four parts, each having a characteristic pattern as summarized in figure 11. There was no strong demarcation zone between the flow pattern regimes, and the divisions in figure 11 are therefore somewhat arbitrary. In general, the liquid phase flowed as a separate streamlet and appeared to be in full contact with the wall except for undetected suspension of droplets or mist flow.

The liquid mercury entering the boiler was constrained to flow

in a helical path around the insert. The inception of boiling occurred about halfway along the insert. Thereafter, because of centrifugal forces, the liquid streamlet flowed circumferentially in the helically swaged groove of the boiler wall. This helical streamlet persisted past the end of the insert, as indicated schematically in figure 12. Figure 13 shows X-ray images of the helical streamlet at different positions near the insert.

The streamlet was sharply defined and highly stable in this part of the boiler.

There was no evidence of or tendency for the streamlet to break or overflow the groove., Because of the fairly high velocity, it was not easy to tell if the streamlet was continuous or contained a train of gaps that moved with the stream; however, the streamlet appeared essentially continuous. The vapor phase flowed axially (possibly with some vortex motion) through the open core of the boiler tube. The sharply defined helical streamlet began to change about 12 inches (31 cm) from the boiler inlet. This change in flow pattern appeared to coincide with the slight change in groove depth described previously in the section Details of Construction.

Helical large-amplitude wave flow. - As indicated in figure 11, a second characteristic flow pattern appeared about 12 inches (31 cm) downstream of the boiler inlet. This

second pattern persisted to a point about 40 inches (100 cm) from the boiler inlet.

In the second flow pattern regime, the liquid streamlet still flowed essentially helically in the grooves, but it consisted of a succession of large-amplitude waves. This was the only regime where the continuity of the bulk stream appeared broken by a succession of gaps. These gaps appeared at the beginning of this regime. Also, there was slight intermittent axial motion of mercury along the bottom of the boiler tube at the start of this regime. But generally, the liquid appeared to flow within the helical groove under the influence of centrifugal forces. At the end of this regime, the liquid phase again formed a continuous helical stream let.

Diffuse helical streamlet. - The third flow pattern regime extended from about 40 to 70 inches (100 to 180 cm) downstream of the boiler inlet, as indicated in figure 11. In this regime, the liquid flowed in an apparently continuous, smooth helical streamlet. Compared with the well-defined streamlet near the insert, this streamlet was rather diffuse; that is, its bounds were fuzzy. As the liquid progressed farther along the boiler in this regime, the streamlet tended to spread axially in the groove. It appeared that the liquid was not confined to the deepest part of the groove but tended toward the crest. Here, the thickness of the bulk of the streamlet diminished markedly in the downstream direction because of evaporation. The drawing of figure 14 shows the diffuse helical streamlet regime.

Heat

Flow

CD-10685-12

Figure 14. Representation of bulk of liquid flow in region of boiler in which diffuse helical streamlet existed (il-
lustrated is general nature of bulk liquid flow pattern, showing axial spreading in downstream direction).

At fairly regular intervals, the helical pattern at the end of this regime exhibited a "'wave train' effect, which suggested a train of helically moving waves superimposed on the main pattern. The time interval between successive waves varied from fractions of a minute to tens of minutes.

Liquid film terminal. - The last boiler regime was the locus where ''wet wall'' boil

ing ended and essentially ''dry wall" began. The liquid film terminal was not strictly a separate flow regime. Instead, it was a locus where observations revealed a thinning of the bulk liquid streamlet into a uniform film covering the boiler wall. This evidence of rather uniform film coverage was confirmed by the corrosion product deposit pattern found on post-test examination of the boiler.

The locus of the liquid film terminal was from about 70 to over 90 inches (180 to 230 cm) downstream of the boiler inlet; that is, according to the X-ray observations, this regime ended near the boiler exit (all visual traces of the liquid phase vanished as the liquid film progressively thinned).

Superheater flow pattern.

From the superheater inlet to exit, repeated X-ray scans detected no liquid in the form of either a film or a mist. This was attributable to the detection limitations of the X-ray image system used. It is possible that very fine droplets of liquid were entrained in the vapor stream. Mass transfer evidence (i.e., deposits) indicated that a thin liquid film extended halfway into the superheater.

Flow Patterns During Metastable Intervals

Variations were observed in the flow pattern associated with off-design intervals in the loop. These metastable fluctuations became more frequent after several hundred hours of operation with the attendant increasing corrosive attack. When deviations from normal performance occurred, there were distinct changes in the flow pattern at certain locations.

During metastable intervals, there was usually a pronounced cyclic advancing and receding of the liquid film terminal near the boiler exit. The terminal film thickened and repeatedly extended into the superheater (at times, more than halfway toward the superheater exit). This was accompanied by a corresponding cyclic variation in boiler temperatures, boiler pressure drop, and flow rate. The automatic heater power control instruments responded by proportionately varying the boiler power level. This produced interactions that often caused sustained oscillations of flow, pressure, and temperature of the type shown in figure 10(a).

When the previously described perturbations in the liquid film terminal region occurred, the stratified helicoidal flow patterns in the upstream regimes remained unaltered. However, these perturbations seemed to be accompanied by changes in the locus of two-phase inception in the boiler insert zone. This could not be observed clearly because of the mode of construction of the insert zone. However, a cyclic increase and decrease in the thickness of the helical streamlet emerging from the insert zone were apparent. The most distinctly observable overall effect in the boiler consisted of helical wave trains superimposed on the normal flow pattern.