The Chilled Water Reduced-Scale Advanced Demonstrator (CW-RSAD) is a small-scale replica of DDG-51 Class chilled water system and is located at the Naval Surface Warfare Center in Philadelphia, PA. The RSAD was originally constructed to investigate the use of component-level intelligent distributed control system (CLIDCS) technology to achieve reliable, unmanned operation of shipboard auxiliary systems. The RSAD features a vertically offset main loop (1-inch nominal pipe size) that distributes chilled water to 15 cooling coils via 8 branches. Each of the two air-conditioning plants contains a 30-gallon expansion tank and has the capacity to deliver 20 gpm of chilled water flow at 120 psig. Seven pairs of electrically actuated ball valves installed in the main loop divide the main loop into 6 zones. Fairmount Automation developed and integrated the DLSM for the 100+ programmable LonWorks control nodes that monitor and control the more than 50 sensors and 80 actuators (i.e., smart valves and pumps) present in the RSAD system. The RSAD DCN consists of 14 LonWorks Free Topology subnets and a single Ethernet channel that serves as a high-speed backbone. Fairmount also developed the HLSM-DLSM interface for communicating feedback and commands between the LonWorks control nodes and a high-level control system executing on a Unix workstation.
Fairmount Automation used the RSAD facility as a testing ground for its rupture detection and isolation algorithms. Fairmount Automation upgraded selected valves on the RSAD main loop with an enhanced flow inventory algorithm capable of rapidly detecting and isolating piping system damage. Testing demonstrated the ability of the smart valves to detect and isolate leaks and rupture within 10 seconds (this includes the 5 second valve stroke time) of the damage event. During this project, Fairmount Automation developed a novel method for determining the smart valve flow estimation model, which is a set of mathematical expressions used by a smart valve to compute an estimate of the volumetric flow rate and to calculate the corresponding uncertainty in the flow estimate. The enhanced flow inventory logic uses the uncertainty predicted by the flow estimation model to reduce the probability of false detection without compromising detection performance. Additionally, this method may be applied without the need for manual biasing and tuning after the valves are installed in the shipboard piping system. This addresses a critical weakness in other smart valve systems, such as those implementing hydraulic resistance logic, which require significant in-place tuning of pressure and flow setpoints so that the activation of fluid services (e.g., heat exchangers, fire plugs, or water mist nozzles) does not prompt the false detection of a rupture condition.