NCAR/RAP Flare Test Facility
The recent model study by Cooper et al. (1997) gives good insights into the theory behind hygroscopic cloud seeding and provides guidance on the necessary steps to optimize the cloud seeding flares. If the CCN that are introduced into the cloud from the seeding flare are larger in size than the natural CCN, the introduced CCN will activate preferentially over the natural CCN and change the character of the droplet size distribution to favor coalescence and the formation of rain. In addition, the modeling study indicated that the larger particles, if present in sufficient concentrations, would further enhance the transformation to precipitation.
Recent airborne measurements in Texas, designed to evaluate the output particle spectra from the different hygroscopic flares, found that the different flares produced different size spectra that could have significant impacts on the evolution of the warm rain process. According to previous modeling studies, the differences in the particle size spectra could have significantly different effects on the condensation coalescence process. It is also important to consider the natural background aerosol spectra in the evolution of these differences.
The majority of the hygroscopic cloud seeding flares currently in use are based on the formula of Hindman (1978) that was developed to initiate fog for cover of military vessels. These flares incorporate sodium and lithium salts along with potassium perchlorate, magnesium powder and an organic binder as the fuel. When burned, the flares produce a plume of sodium, lithium and potassium salt particles along with magnesium oxide. All of these species are incorporated into each of the flare particles that are produced.
Modeling studies on hygroscopic seeding show that the size of the particles introduced by the seeding flare influences the effectiveness of seeding on rainfall enhancement. Under the conditions modeled, the most effective seeding occurred when the introduced CCN had a size of approximately 1 mm. Little information is available on the particle sizes generated by the flares currently in use. A size distribution measurement of the dry particles indicates that a majority of the particles are in the 0.2-0.4 mm size range (Cooper et al., 1997). These measurements were made using flares manufactured with the original Hindman formula. Other airborne studies and ground-based wind tunnel studies confirm that a majority of the particles produced from the magnesium/potassium perchlorate flares are less than 0.5 mm. However, the large particle tail that is produced by the flares is largely unknown at this stage. The modeling studies also suggest that the use of calcium salts in the seeding material could enhance seeding effectiveness.
It is difficult to obtain measurements of the particle sizes produced by the flares in a field environment. This generally requires two aircraft, one to generate the seeding material, and the second to make the measurements. A further difficulty is the ability to reproduce these measurements, since the environment changes. To study hygroscopic cloud-seeding flares, a test facility has been designed and constructed that simulates the burning of flares from an aircraft.
The test facility has been designed to provide a reproducible environment for combustion of flares and measurement of the resultant particles. The facility is also designed to simulate the burning of flares on the wing of an aircraft. Figure 1 (on left) details a schematic diagram of the test facility that was constructed.
Photo of Flare Mouted in
(click to see enlargement)
A 156,000 lpm (liters per minute) blower is used to provide an airflow of 45-50 m s-1 through the combustion section, which is a 25.4 cm diameter and 1.5 meters long section of pipe. The flare is mounted near the head of the combustion section with the flare aligned axially with the airflow. The burning end of the flare is on the downstream side. This provides an environment similar to burning a flare on a rack mounted on an aircraft wing. The large airflow is sufficient to disperse and cool the particles as they are produced.
Figure 2 (on left) shows a photograph of the combustion section with a flare mounted. The airflow would be from left to right, and the right hand end of the flare would be ignited. In operation, the door to the combustion section is closed, and the flare is ignited using an electrical igniter
Instead of sampling the smoke directly from the high velocity combustion section, a deceleration section was added. The smoke now flows at approximately 12 m s-1 at the exit of this section. There are no molten salt particles, and the potential for plugging the sample tube has been eliminated. A second smoke dilution stage has been added providing a smoke dilution option up to a factor of 250,000. It will be possible to tailor this dilution to optimize the concentration for the particle measuring instrumentation.
To obtain relative information on the combustion temperature of the flare, two platinum resistance temperature probes have been mounted in the combustion section. The first probe measures the temperature immediately ahead of the flare, and a second measures the temperature near the end of the combustion section. From this temperature differential, we will be able to obtain a relative measure of the amount of heat generated in the combustion of the flare.
Some Flare Facility Results...
Sample distribution from a flare manufactured
by ICE containing 65% potassium perchlorate
Based on numerous tests with different formulations of flares, a new flare was developed that produced larger particles than the original South African flare used both in South Africa and Mexico. Modeling studies have shown that larger particles would be more effective in hygroscopic seeding. Figure 3 (above) shows the particle size and volume spectra of the Ice Crystal Engineering (ICE) flare that is now used in the UAE program to enhance rainfall.Read About Hygroscopic Cloud Seeding Principles