NASA Tech Briefs Create the Future Design Contest Category Winner, 2010
In fighting wildfires, and in many other applications, there is a need to lift water a few feet, say from a creek or river, to fill a header tank or mobile water tender. Particularly in firefighting, turnaround time is what measures success at the end of the day.
My Lift Pump design comprises an axial flow impeller pump section located at the inlet end of the transfer hose, a power head at the flex-fitting located near the destination (tank or truck), and a flexible drive shaft running inside the transfer hose. Sections of transfer hose may be daisy-chained, with the internal drive shafts automatically coupling together as the hose is assembled.
The impeller is made of a custom-molded thermoplastic engineering resin, and is not only highly resistant to wear and trash, but is inexpensive, and easily field-replaceable.
Our first prototype successfully operated under test conditions. Flow rates of 500 to 1000 gallons per minute were achieved at up to 12 feet of head. For one series of tests, the pump and power head were fitted to a Boeing/Vertol helicopter, successfully filling a 1000 gallon drop tank during hover. Over the approximately one-hour flight trial, the pump successively drained a 5,000 gallon tender tank to empty
In addition to flight applications, this pump shows promise in making quick turnarounds, drawing water from small creeks and ponds with minimal accessibility to fill tender trucks. By linking a number of 20-foot hose sections together, a 5,000 gallon tender could be located on a road up to 1,000 feet from the source, and filled in a matter of a few minutes.
The Motor
A number of different motors were considered to power this pump, and choosing the right one seemed daunting. So I took the quick, easy path: find something cheap, replaceable, and readily available. So I quickly chose the Detroit Diesel V8 starter motor. Spare parts would be easy to get...
The series-wound 24VDC motor has a characteristic which is not necessarily desirable for this application, namely, that it tends to run away under no-load conditions; it has no inherent running RPM for any given voltage. It just does it's best to keep turning faster. Thus, we had to impose some operational criteria: Basically the rule was, "Don't start the pump until you know it is submerged. Thank you."
The Impeller
I wanted to make the impeller out of glass-reinforced nylon, because of its amazing durability with respect to both foreign object damage (FOD) and erosion due to cavitation (to which aluminum props are known to be susceptible). My experience years prior, working for Piranha Propellers, gave me great insight on this.
The first prototype therefore used actual retail Piranha Propeller hubs and blades for the impeller. We turned their outside diameter down on the lathe to fit inside our housing, and we ran flow rate vs. head tests using various blade pitches. We also ran various two stage configurations, usually with a steeper pitch installed for the second stage.
Data collected included the following:
- Flow rate in gallons per minute (gpm)
- Head in inches (In H2O)
- Current draw (AMPS)
- Actual voltage at the motor terminals (VDC)
- RPM of the impeller
After extensive testing and experimentation, I decided that I really needed to design and build my own impeller, so we built the mold for a four blade, single piece hub-and-blades, and molded our own custom pump impeller. With what we learned from this whole process, I now know what I would do next, in terms of design, to make the next one even better.
As is seen in the above renderings, the design is of a "reflex belt drive" arrangement, which would accommodate various ratios of underdrive or overdrive of the impeller to the motor. We never got to the point of trying various belt ratios; all tests and operations were performed at a 1:1 drive ratio. I'm not sure about this, but I suspect that the nature of a series-wound DC motor will cause it to "settle in" at any given RPM which corresponds to its basic horsepower capability. As the load increases, the RPMs drop, but the torque increases (and the current draw also increases). So I think it was a good starting point for experimentation.
I also ran some tests using the Briggs & Stratton ETEK motor, which is a permanent magnet DC motor that turns a given RPM on the basis of the input voltage. This motor actually produced quite favorable results, with the added advantage of never having to worry about the load—whether the pump head is submerged or not.
The Next Generation
The next generation pump, the Twinhead, has been designed, and prototype fabrication begun. Although most of the machining has been completed, the project unfortunately was put on the back burner as priorities changed and funding dried up. This model used smaller, lighter weight, higher RPM motors, also adapted from series-wound diesel starting motors, re-engineered to replace all bushings with precision sealed ball bearings. This pump employs water cooling of the motors by scavenging a small portion of the pump flow and channeling it through the custom motor housings.
