Waterhammer Simulation and Mitigation for a Fire Protection Network at a Nuclear Power Plant
A waterhammer risk assessment and computer analysis of the fire protection system at the Edwin I. Hatch Power Plant (Plant Hatch) in Baxley, Georgia was performed. The activation of the highest demand deluge system with and without various surge mitigation devices is presented. Vapor void formation and high pressure void collapse can occur at high elevation risers in fire protection loops as a result of the sudden flow demand from the activation of dry pipe fire suppression systems (such as preaction or deluge sprinkler systems). This paper demonstrates that all pressurized piping should be considered in a fire protection system, especially localized high points and longer lengths of piping to dead ends or remote hydrants. This paper demonstrates a comprehensive computer analysis of the entire pressurized fire protection system should be the preferred method of waterhammer risk assessment.
Dylan Witte, PE, Brown and Caldwell, USA (formerly Purple Mountain Technology Group, USA)
Presented at the 14th International Conference on Pressure Surges | April 12-14, 2023, Eindhoven, Netherlands
The addition of surge vessels on the risers and at four critical network locations at Plant Hatch was shown to be an effective mitigation against void formation and collapse within the entirety of the fire protection system. Detailed computer simulations and field testing should be performed at all similar fire suppression systems to ensure their safe operation.
Fire suppression systems are dynamic in nature with sudden demand for high flow from previously stagnant conditions. The sudden change from no to high flow will drop pressure and can be significant enough to result in void formation and collapse. Pressure spikes resulting from void collapse can be significant.
These high pressures may be of particular concern in aging cast iron pipe which is the majority of fire protection piping at Plant Hatch and many other nuclear plants. Previous studies have reported on this phenomenon but have focused on high-elevation risers (3, 4). This study shows that low pressures and significant voids can form throughout the network and are a probable cause of multiple observed piping failures at Plant Hatch. As a result, a more comprehensive analysis that considers the entire fire protection system is required.
Previous analyses have shown vacuum breaker valves to provide effective mitigation against void formation in risers (4) but were demonstrated to be ineffective at Plant Hatch due to poor dynamic response which resulted in significant air compression and high pressures. Additionally, the vacuum breaker valves had little to no effect on remote system pressures where void formation and collapse still occurred.
Both preaction and deluge piping networks are initially dry pipe and isolated from the pressurized fire protection network by a control valve that is activated from a heat or smoke source. The primary difference between the two is deluge systems have entirely open sprinkler heads and preaction systems limit flow to individual sprinkler heads that have been activated at a specific temperature. Deluge systems are intended to extinguish large scale fires and preaction systems are intended for use in water sensitive areas. Deluge systems typically have higher long term flow demands, but both deluge and preaction systems have a similar initial response as the dry pipe is rapidly filled with water.
During a fire event, preaction or deluge control valves can actuate in less than 0.1 seconds after being triggered by a heat or smoke source to allow water through empty piping and sprinklers to extinguish the fire. The sudden introduction of flow from the pressurized fire protection network into the dry pipe system can drop network pressures significantly. If the pressure drops below the vapor pressure of water, it will result in a localized phase change and the formation of a vapor void or pocket. These vapor voids are likely to form at high points in the system where the local hydrostatic pressure is lower, and within complex network piping with dead end branches where pressure wave reflections are amplified. After a short period of void formation, the dynamic response of the system and the start of the main fire pumps will raise the local pressure at the voids causing them to collapse, potentially resulting in extremely high pressures.
This void formation and collapse behavior was discussed thoroughly in Dr. Samuel Martin’s 2012 waterhammer report on Plant Hatch (1). Dr. Martin concluded severe pressure spikes only occur in risers with preaction valves. However, this study demonstrates a more complete analysis of the entire system is warranted.
The void formation and collapse phenomenon were responsible for a catastrophic valve failure at WNP-2, a nuclear power plant Washington State (now Columbia Generating Station) (2, 3, 4). Activation of the preaction system caused voiding in the reactor risers, and the subsequent void collapse generated a large pressure wave. Repeated waterhammer events lead to weakened supports, and the valve at the base of the riser split in half due to increased force during the final waterhammer event which flooded an estimated 617,000 liters.
Plant Hatch and many other nuclear power plants share a similar fire protection system design to WNP-2. These plants are potentially susceptible to the same void formation phenomenon and should have both testing and waterhammer computer analysis performed to ensure safe operation during the activation of a dry pipe fire suppression system.
Use the links above to view the full papers.