By Jaehyok Lim, Ph.D., Basar Ozar. Ph.D., Christopher E. Henry, Ph.D., and Kevin Ramsden, Fauske and Associates
A study to understand the effects of geometry and water supply pressure on the void transport has been performed using RELAP5. Two different piping configurations were considered for a hypothetical nuclear power plant. The cases that were analyzed considered switchover between two different water supplies, i.e. water storage tank and SX for a safety system that acted as the ultimate heat sink.
Two different pressures were considered for the pressure of SX to also investigate the effect of supply water pressure on void transport. Results were interpreted based on the differences in the geometries of the piping configurations and supply water pressures.
Each boiling or pressurized light-water nuclear power reactor fueled with uranium oxide pellets within cylindrical zircaloy or ZIRLO cladding must be provided with a safety system that must be designed so that its calculated cooling performance following postulated loss-of-coolant accidents conforms to the criteria set by the U.S. Nuclear Regulatory Commission (NRC).
A typical safety system is a reactor cooling system for safe removal of residual heat in the case of an interruption of the heat transfer between reactor and heat sink. These systems are designed so that even in the case of reactor coolant Floss – e.g. when both ends of a live-steam pipe are broken – the reactor is cooled and the decay heat can be removed over weeks. Due to its important task, safety systems should be in operable condition without any defects at all times. Degradation or malfunctioning of this system is not a desired condition.
Gas voiding in a safety system is a phenomenon that may cause a degradation of this system. Such gas voiding may result because of several reasons. Some of these causes are identified as degassing due to depressurization, maintenance activities and operation maneuvers. The gasses, which exist in the system because of the above reasons, usually accumulate at an elevated location or trapped between two normally closed valves in a horizontal header. This header is connected to the pumps related with the safety system through a vertical section. Once the flow is initiated, the gas-water mixture flows along the horizontal header and then downward through this vertical suction and then reaches the pump suction. The void fraction, flow pattern and the duration of the void transport at the suction of the pump are the critical parameters that will affect pump performance. High void fraction for a prolonged time may cause degradation or malfunctioning of an injection pump causing degradation of the safety system.
The Nuclear Energy Institute has published the extent of non-condensable gas that the industry has considered could be transmitted to specific pump types without challenging the pump operability.
In this paper, a RELAP5 model was built for a hypothetical system that serves as the ultimate heat sink. An automatic switchover scenario of the water supply was investigated from a water storage tank to the essential services water system (SX). The air was trapped between two normally closed valves, which would open during switchover to SX. The effects of piping configuration and the SX pressure were discussed.
Description of the model
A RELAP5 model was built for two trains with different geometries for a hypothetical plant. The nodalization diagrams for the two models with different configurations are presented in Figure 1 and Figure 2. The air was trapped between valve components V804 and V805 by initially specifying 100 percent air in the pipe component P803. The remaining pipe components in the model were filled with water at 90°F. The valves 804 and 805 were initially closed, while water was being supplied by the water storage tank (TDV815) to the pump (PP904) through the pipe components 817, 812 and 801. The pump model was built using homologous curves. The initial flow rate of the system from the water storage tank was approximately at 1000 gpm based on the pump and system characteristics. PP904 discharged into a time dependent volume (TDV), which simulated the steam generator, via the piping configuration on the discharge side.
The SX water supply was also modeled with a time dependent volume (TDV). In order to understand the sensitivity of void transport with respect to supply water pressure, TDV was set to 87 psia and 100 psia. The models were first kept at stagnant condition for 40 s so that the model reached to hydrostatic equilibrium. PP904 was started at 40 s. Finally, the transient was initiated by opening the valves 804 and 805 simultaneously at 50 s with a stroke time of approximately 13 s.
Both models had approximately similar geometries downstream of the branch, which connects both the piping from the water storage and the SX to the pump, to the pump suction. Also, the piping from the water storage tank to the branch was similar. The major difference between models A and B was the piping configuration from the SX to the branch. Model A had a long voided section (12.8 ft) with a short distance (1.7 ft) to the branch. On the other hand, model B had a short voided section (3.3 ft) and a long distance (54.6 ft) to the branch with several elevation changes in the piping.
Results and Discussion
Figure 3 shows the mass flow rates for model A at the TEE junction (B813) that connects both the SX and the CST water supplies to the downcomer that feeds the pump suction when the SX pressure is set to 87 psia. The reference direction for the flow rates is shown in Figure 1. If the flow is going in the same direction as the arrow, then this is presented as a positive mass flow rate.
Once both V804 and V805 started opening, there was a small amount of back flow from the TEE junction toward the voided section. At the initiation of the valve opening, the void was at 14.7 psia, so it was being compressed by flow from both sides until it attained a pressure that overcame the supply pressure from the water storage tank. At that time, the flow from the water storage tank started decreasing and eventually reaches to “0”, at which time the check valve V816 closes in response to the relatively high pressure being supplied by the SX system.
However this is a highly transient progression. Simultaneously, the flow from the SX increased and started pushing the void toward the branch (B813) and into the downcomer. This is also a highly transient response.
It is emphasized that, during the period at approximately 52 seconds when the mass flow from the SX source to the branch approached zero, this does not mean that the flow is stopping. It means that the flow was predominantly low-density air, so while the total mass flow (variable mflowj) was low, the volumetric flow of air was large.
Furthermore, at this time, most of the liquid flow to the pump is being supplied by the flow from the water storage tank suction, as demonstrated by the similar values for mflowj-813020000 and mflowj-813030000.
Figure 4 shows the void fraction time histories along the downcomer. Void fractions as high as 80 percent were estimated at the top of the vertical downcomer (Node 801-03). Meanwhile, the void fraction reached to almost 98 percent at the bottom (Node 801-11). The reason for the void fraction being higher at the bottom was due to different two-phase flow regimes estimated at the top and the bottom of the downcomer. The slip between the two-phase regimes was different at the bottom compared to the top. Therefore, higher void fractions were computed at the bottom. The void transport occurred within a period of ~3 seconds.
The void fraction time histories along the downcomer for model B are presented in Figure 5. Peak void fractions are less than 5 percent, and the entire void transport through the pump persists for approximately 3 seconds. The void fraction that is transported to the pump suction is significantly lower for model B compared to model A.
This is due to several reasons. The first reason is because of the fact that the volume of trapped void is smaller in model B. The second reason is related with the time for the void to be transported from the initial location to the pump suction. Since the void gets transported over a much longer distance until it reaches the pump in model B, a longer time is allowed for the air volume to be compressed by the SX pressure, hence reducing the void fraction.
The effect of SX pressure on void transport is also investigated. The results for maximum and average void fractions are provided in Table 1 for 100 psia SX pressure and compared with the results of the 87 psia SX pressure calculations. Comparisons show that the amount of void transported into the pump suction decreases slightly as the SX pressure is increased. This is somewhat an expected result since the air bubbles will compress more and occupy a smaller volume with increasing system pressure.
The cases that were analyzed considered switchover between two different water supplies, i.e. water storage tank and SX for a safety system that acted as the ultimate heat sink. Two different pressures (87 psia and 100 psia) were considered for the pressure of SX to also investigate the effect of supply water pressure on void transport.
The void that was transported to the pump intake had larger values for the configuration with longer voided section and shorter traveling distance (model A) compared to the model with shorter voided section and longer traveling distance (model B). The main reasons for this were a)Model A had more air than model B and b) The traveling distance of the void, and therefore, the transport time was longer in model B letting the air to compress more during this duration until it reached the pump intake.
In addition to these, the effect of SX showed that the void fraction of air that gets transported into the pump is slightly less with increased pressure. This was attributed to the fact that the air bubbles compressed more with increased SX pressure.
Power Engineerng Issue Archives
View Power Generation Articles on PennEnergy.com