3.1. Flow rate test

Owing to the opaque nature of liquid PbBi, especially under high temperature and chemical reaction conditions, traditional measuring technology for water or other flowing medium cannot be applied to the PREKY loop. Therefore, the flow meter should be calibrated in advance. Based on previous experience and test results, the Venturi-nozzle flow meter (VNFM) was chosen for flow rate test of the PREKY loop.

The scheme of VNFM is shown in Fig.2. In principle, VNFM can be used in a wide range of flow rates based on the measuring principle of differential pressures. The flow rate (Q) of VNFM is calculated by Eq.(1) ^[13]:

Fig. 2Full-screen View

Schematic diagram of VNFM

where, μ is flow coefficient of the VNFM, Δh is the pressure loss between pressure ports located at the entrance 1-1 and throat 2-2 (see Fig.2); and K is a constant, determined by Eq.(2) ^[13]:

where, d₁ and d₂ are inner diameters of the pipe at the entrance and throat (see Fig.2), respectively; and g is the acceleration of gravity.

The flow coefficient μ is determined by Eq.(3):

where (see Fig.2), l is the length between the 1-1 and 2-2; λ is the drag coefficient; a₁ and a₂ are the correction coefficients of kinetic energy at the 1-1 and 2-2, respectively; ζ is the local head loss coefficient from 1-1 to 2-2. In this experiment, the μ is evaluated at 0.9316.

Calibration experiment of the VNFM was implemented to obtain the relationship between flow rate Q and pressure loss △h. The measurement data were compared with the calculation results using the above equations. At beginning of the experiment, the output signal under differential pressures was unstable and changed due to the less wetting of internal surface of the flow channel ^[14,15]. This, however, can be sufficiently improved by flowing liquid PbBi at large flow rate and high temperature for tens of hours before the calibration experiment ^[16].

The flow rate calibration was carried out at different flow rates. Gravity was the driving force of liquid PbBi flow, and different flow rates were obtained by adjusting the valve opening at the calibration tank exit. The calibration procedures are as follows:

(1) The system is fully preheated, and evacuated for at least 3 times;

(2) Liquid PbBi is driven to the safety level in the calibration tank; and

(3) Open the valve quickly to a certain degree, and record the experiment data.

In the calibration, the calibration vessel and storage vessel were kept at the same gas pressures. To improve the value precision and extend the range of flow rate, the experiment was repeated for seven times. The maximum flow rate was up to 2 m/s, which was about the maximum flow rate of the KYLIN-II loop, for corrosion and thermal hydraulic experiments.

Typically, the experiment was performed at about 350°C and flow rate of 2.0 m/s. Fig.3 shows that basically, the VNFM data agree with the level meter measurement results. The flow rate of VNFM increases stably each time, while the level meter curve fluctuates a little comparing with the VNFM output, and slope of the level meter curve is a little bigger than that of VNFM.

Fig. 3Full-screen View

Calibration results (a) of the VNFM and the relative error (b)

The relative errors between VNFM and level meter measurement are shown in Fig.3(b). At flow rates of below about 0.6 m/s, the maximum measurement error is about 20% or even more, so the VNFM is not suitable for measurements at low flow rates. It can also be seen that the flow rate calibration has an uncertainty of ±5% in the range of 0.6‒2.0 m/s. This is because that: (1) the driving force for liquid PbBi to flow through the calibration vessel is the gravity, hence the non-steady flowing state between two level electrodes; and (2) changes in diameter and height of the calibration vessel under different temperatures were not considered in the experiment.

3.2. Local temperature test

Temperature controlling and monitoring are of importance for safe, reliable and stable operation of molten PbBi facilities. Accurate measurement of temperature distribution is an important emergency control for safe operation.

For studying heat transfer characteristics of PbBi in the channel, temperature tests were implemented by two quick response thermocouples of K-type (Table 2 and Fig. 4) installed inside and outside the flow channel located at the outlet of air-cooled heat exchanger. The inside thermal-couple was directly contacted with flowing PbBi itself under high temperature.

Fig. 4Full-screen View

Schematic diagram of thermocouple

The heat exchanger was used to cool down the PbBi flowing out of the test section. Its inlet and outlet were opened manually. A few minutes were needed to reach stable state of flowing PbBi. The thermal-couples were contacted with the pipe surface, and the heating and controlling system would need a thermal response time of over 60 s. In other words, the heat exchanger would start to stop working 60 seconds later when the temperature of outer surface of flow channel reached the preset value, while the temperature of liquid PbBi had exceeded the preset value at this moment. Therefore there were temperature difference at the same position and time between flowing PbBi and the outer surface of flow channel.

Thermal response time of the thermal-couple, which depends on outer diameter of the measuring terminal, can be defined by Eq.(4) ^[13]:

where, h is heat transfer coefficient; and ρ, V, c and S are density, volume, specific heat and surface area of the thermal-couple, respectively. From Eq. (1), supposed L is the terminal diameter of the thermal-couple, the volume V∝L³, the surface area S∝L², similarly the thermal response time τ∝(V/S) ∝L. Therefore, the thermal response time τ is proportional to L. So, the terminal diameter is 3mm, thus response time is about 10 s.

The results of temperature experiments were shown in Fig.5. It can be seen that (1) the working period of the cooler was about 10 min, the temperature fluctuations of flowing PbBi and the pipe surface were about 35°C and 18°C, respectively; (2) the maximum temperature difference between the PbBi and pipe surface was roughly 8°C； and (3) the temperature peak positions of pipe surface was about 1 min later than that of the PbBi, and the maximum and minimum temperatures of PbBi, and of the pipe surface, had a time difference of about 4.5 min.

Fig. 5Full-screen View

Curve of temperature measurement

These can be explained as follows:

(1) The two thermocouples were located at the outlet of the heat exchanger. When the heat exchanger stopped working, output of the two thermocouples increased for a short period of time due to continuous heat release from the electromagnetic pump. After a few seconds, temperature went down quickly by heat exchange in the working state. At this time, a working period for the heat exchanger had been completed.

(2) When temperature of liquid PbBi began to drop, due to heat transfer of liquid PbBi and thermal response time of thermocouples, temperature of pipe surface started to drop about 60 s later. The maximum temperature difference (about 8°C) between PbBi and pipe surface did not appear in the inflexion of the operating state for the heat exchanger.

Following the above results, thermocouples could be in direct contact with flowing PbBi to obtain the accurate and real-time temperature of liquid PbBi.