by the mercury entering from the superheater. Some of the conclusions herein relative to metastable performance depend on X-ray observations of the throttle.

Instrumentation

Thermocouples were attached to the outside surface of the loop tubing at regular intervals. From the preheater through the superheater, the thermocouples were attached to the tube in the essentially adiabatic spaces between heaters. The thermocouples were spot welded to the tube and covered with quartz-fiber insulation. The locations of the more important thermocouples are identified in figure 1. The principal thermocouples on the boiler, superheater, and vapor throttle were platinum- platinum13-percent rhodium. The rest were Chromel-Alumel thermocouples, or iron-constantan thermocouples. Additional thermocouples were included for heater power control.

Pressure measurements were obtained with strain-gage-type pressure transducers. The transducers were connected to the loop with small-bore tubing. Pressure was

transmitted from the loop to the transducer diaphragm by liquid mercury. The previously mentioned insert in the boiler inlet was an extension of one of the pressure transmission tubes and was used to measure the boiler inlet pressure. Pressure tap locations are indicated in figure 1.

At the condenser exit (over the condensate column), the pressure was in the microtorr range and, therefore, was sensed by a vacuum gage. This measurement was made intermittently by opening the vacuum port indicated in figures 1 and 3.

The flowmeter consisted of 12 feet (360 cm) of tubing with a 1/16-inch (0.16-cm) inside diameter. The mercury flow rate was determined by the pressure difference between the flowmeter inlet and exit. The relation between pressure drop in the flowmeter tube and flow rate was obtained by direct calibration.

OPERATING CONDITIONS

The operating conditions in the mercury loop were set to meet the experimental requirements of the corrosion experiment for which the loop was designed. A prerequisite condition was that the mercury wet the boiler surface to a sufficient degree. This would help to produce good corrosion results as well as better heat transfer. Measures were therefore taken to induce wetting by the procedure described in appendix A. There was no way to determine if optimum wetting had been attained in the boiler, but there were indications that the conditioning procedure substantially improved the relative degree of wetting.

The loop operated semiautomatically with proportional feedback control for a total of 1147 hours. Periodic manual adjustments were required to compensate for drifts from normal conditions. After steady conditions were established, they usually persisted without a need for manual adjustments for intervals of 10 to 25 hours under automatic control. These stable conditions were sometimes interrupted because of line voltage perturbations.

Loop control was based on the wall temperatures along the preheater and boilersuperheater. The desired temperature levels and permitted temperature ranges at the end of each of the five heated zones were preset. When any disturbance occurred in the preset temperature pattern, a compensating change in the heater power levels automatically readjusted the heat input. The boiler inlet pressure and vapor throttle opening remained fixed so that all corrections were made solely through boiler power adjustments. The normal operating conditions are summarized in table I. The normal mercury flow rate was about 78 pounds per hour (9.8×10 ̄3 kg/sec). Corresponding to this flow rate was a liquid velocity of about 3 feet per second (90 cm/sec) in the liquid-filled helical channel around the insert at the boiler inlet. The downstream half of the insert zone

a

b

All values are approximate: pressures are rounded to two significant figures except for pump developed pressure. Temperatures are rounded to three significant figures except for insert zone peak temperature. Condenser exit pressure was in torr range.

Last 2 ft (62 cm) of condenser were cooled by water jacket and upstream of this, condenser was cooled by natural air convection and radiation.

dpump inlet pressure was produced by "static" head of mercury in condensate column (see fig. 1). ePump developed pressure was independent of flow rate in flow rate range of loop experiment.

had concurrent liquid and vapor flow. The tangential velocity of the liquid phase near the end of the insert was estimated to be about 8 feet per second (240 cm/sec).

METASTABLE FLUCTUATIONS

During the latter part of the 1147-hour test, there were more frequent intervals when the conditions in the loop became metastable. During the metastable intervals, there were cyclic flow, pressure, and temperature oscillations and excursions. Typical

Figure 10. Covariation of flow rate, pressure, and temperature during intervals involving oscillations and excursions.

traces of flow, pressure, and temperature oscillations and excursions that occurred during the metastable intervals are presented in figure 10. Usually, the amplitudes were small, and the normal operating conditions constituted the mean about which the deviations ranged.

Figure 10(a) typifies the flow rate, pressure, and temperature oscillations that occurred during some of the metastable intervals. During these intervals, the pumpdeveloped pressure and pump-outlet pressure were constant and showed no tendency to oscillate or drift. The type of metastable condition represented by figure 10(a) usually prevailed for several hours and sometimes for a period of 10 hours during times when the loop was unattended and manual adjustments could not be made. As shown in fig

ure 10(a), the flow rate varied about 14 pounds per hour (17×10-4 kg/sec), boiler saturation pressure varied about 25 psi (1.7×105 N/m2), and saturation temperature varied about 40° F (22° C). This type of condition usually required slight manual control adjustments to dampen and eliminate the oscillations. The most useful adjustment was to change the temperature level and gradient near the insert zone.

A typical minor metastable excursion is represented in figure 10(b). These excursions occurred after the loop had been operating in a normal steady mode for long intervals of the order of tens of hours. After the excursion in saturation pressure and flow rate (bottom of fig. 10(b)), the loop oscillated for about 11 hours and then automatically returned to normal. The type of pattern shown in figure 10(b) sometimes repeated itself five to six times during an interval of 24 hours.

A more pronounced excursion is represented in figure 10(c). This type of excursion was associated with a marked drift in the mean flow rate and mean saturation pressure. The particular pattern in figure 10(c) repeated approximately every 5 hours during an interval of 20 hours. Each time the excursion occurred, it was followed by oscillations that dampened automatically. During the dampening, the mean flow rate and saturation pressure drifted back to their previously set values. The flow and pressure traces were then quiescent until the next excursion. After adjustments were made to eliminate the pressure and flow drifts, the excursions ceased. The manual adjustments were increasingly difficult to make successfully as the test progressed and corrosive attack and mass transfer increased.

FLOW PATTERN OBSERVATIONS

The normal operating conditions for the loop are given in table I and are illustrated in figure 11. Also included in this figure are abbreviated descriptions of the two-phase flow patterns in the boiler, superheater, and condenser. These flow patterns were observed by means of the X-ray image system. A detailed description is given hereinafter