Laserfiche WebLink
<br />insulated exhaust tube to release liquid propane. a <br />panicular rype of nozzle at Ihe end. and a remolely- <br />controlled on/off value in Ihe tubing. Inclusion of a <br />temperature sensor immediately beyond the n07.z1e <br />and a flowmeter in the exhaust lube provide <br />monitoring that propane is indero being dispensed <br />and at the desired rail:. Propane dispensers hav'e <br />proven to be very reliable because of their simplicity. <br /> <br />In cOnlrast. radio-conlrolled Agl generators are much <br />more complex. susceplible to clogging from highly <br />corrosive seeding solutions. They are seldom <br />equipped to monitor actual AgJ-in-soIUlion !low rate <br />bUI only llame temperature. Therefore. one cannot <br />be cenain from afar lhal Agl was actually relea~ed <br />which has somelimes been demonstrated to be a <br />!>erious problem. <br /> <br /> <br />\ -~ <br /> <br />Figure 3. The hIgh rcsolUlion precipitation gauge at the nonh control site ....Ith the \ie.... to....ard the nonh. In this pi~lurc. gaug.' <br />data logger dala arc being do....nloaded to a portabk Palm Pilot devil:e_ Tbe 12.ioch diameter gauge oritice i, ,urrounded by an <br />AltcHypc....indshield made of clear Lexan plaslie. <br /> <br />A Goodrich icing rate sensor. similar 10 Ihe <br />fn:ezing rain sensors us...-d in Ihe nalional nelwork of <br />Aulomated Surface Observing Syslem (ASOS) <br />weather stalions. was used 10 declare beginmngs of 2 <br />hour experimental periods containing a pair of 40 <br />minute EUs plus their 20 minUle buffers. This device <br />has a I inch high by 1/4 inch diameter sensing rod <br />ex lending venically from a cone-shaped structure. <br />Both were aUlomalically healed once rime ice <br />buildup on Ihe rod reduced ils frequency by a preset <br />amount. The liAS dala logger monilored I minute <br />frequency observations lrom the Goodrich sensor. <br />The frequency had 10 decrease O\-er a 10 minute <br />period by at least a specified amount. equivalent 10 <br />aboul 0.005 inch thickness of rime ice buildup from <br />SL\\' droplels freeztng on the sampling rod. in order <br />for an experimental period 10 be declared by Ihe dala <br />logger program. The 10 minute period provided a <br />reasonable time sample and minimized the impact of <br />any frequency noise spikes. After lhe first EU of a <br />pair was randomly declared as either seeded or <br /> <br />nonseeded. Ihe opposite decision was used for the <br />second EU of the pair. The icing sensor continued to <br />be moniton:d throughout the 2 hour period. <br />Ilowever. the sensor trequency output was not used <br />10 lest conditions lor stan of another EU pair until <br />two full hours had pa<;sed. A following pair would be <br />declared as soon the required rime ice buildup again <br />OCCUrrt.'<I. sometimes immediately after the firsl 2 <br />hour period if SL \\' was still present. At Ihe olher <br />e"\treme. if a stonn had ended. the next EU pair <br />would nOI be declared unlil the next stonn period <br />With sutlicient SL \\'. perhaps days or sometimes <br />weeks later. <br /> <br />The randomized seeding experiment was 10tal1y <br />automated. requiring no human decisions or <br />intervention except for routine equiplTl('nt <br />maintenance. Consequently. e:tperimentalion ....as <br />conducted on a 24,7 basis which maximized the <br />resuhing populalion size of seeJed and nonseeded <br />experimental units. The propane experiment was <br /> <br />5 <br />