|
Gaining the edge
|
13/07/2008
|
| |
A high-speed compressor neatly combines the best features of rotary and reciprocating machines in a simple basic unit. Tom Shelley reports
How do you compress air efficiently at speed, without the need for seals or very elaborate valves? The answer would seem to be by having an eccentric cylindrical rotor within a chamber, rotating with an edge velocity that is supersonic.
While the edge may be moving faster than sound, the air is moving much more slowly and in a manner that avoids unnecessary changes of direction – so it’s expected to run more quietly and efficiently than conventional designs.
The first intended application is in engine supercharging, but the design is suitable for any kind of air or gas compression task.
The man behind the concept is John Hammerbeck, who invented the Scram compliant variable ratio coiled spring drive (Eureka, September 2000) and the hypocycloidal reducer (October 2003). His latest idea is a hypocycloidal air compressor, which has been extensively investigated by Imperial Consultants – the spin-off consulting arm of Imperial College’s mechanical engineering department – with the help of a government grant.
Its basic principle is that the rotor moves round the inside of the pumping chamber in such a way that air is forced outwards towards the periphery and through valves by the space between rotor and chamber narrowing down to a pinch point.
“The idea is to combine the low mechanical losses of a centrifugal compressor with the high gas compression efficiency of a reciprocating compressor” says Hammerbeck.
Originally, he imagined having the rotor ‘walk’ round the inside of the cylindrical chamber, so there would be no shear between the rotor surface and the air. But in the validation machine, recently commissioned at City University, the rotor spins off centre, with an internal counterbalance weight, to avoid the need to build a special gearbox. Since the 8kg rotor is intended to turn at up to 25,000 rpm in its 20 cm wide chamber, developing a gearbox capable of withstanding the required loads and speeds would have been costly in a project that is not backed by a very large amount of money. The validation machine is a ‘no flow’ device. However, in order to become a practical compressor, it is expected to need some kind of non-return valves.
The study used computational fluid dynamics (CFD) code specially developed for the task, using Matlab, in conjunction with Ansys CFX.
The report prepared for DTI (now DIUS) describes having a series of alternate inlet and outlet ports around the periphery of the compression chamber, each of which would have its own non-return valve. However, Hammerbeck describes an idea involving a ring of spring steel all the way around the outside of the compression chamber, closing a small gap within it. This would be of slightly greater internal diameter than the chamber, so that pressure would lift it clear of the gap as the pinch point passed round. Should this not work, “there are six different possible ways of valving it”, he adds. Among other experiments, the intention is “to put in a few one way outlet valves, to see how the machine would function as a vacuum pump”.
The main future of the idea is seen in supercharging small gas engines – gas meaning gas and not ‘gasoline’, in this instance. Natural gas is likely to greatly outlast oil as a fuel, with vast reserves available in the form of methane hydrates on sea beds, as well as gas from oil production, much of which is presently flared off and wasted, as well as gas from gas wells.
Hypocycloidal pumps are not a new idea. There is a lot of interest in using them in artificial hearts, since they don’t damage the blood cells in the same way as a centrifugal pump. A refrigeration compressor with an eccentrically mounted rolling piston – developed by Ron Driver and his company EA Technical Service, with support from the Carbon Trust – was described in the January 2007 edition of Eureka. There are also a number of patents in the literature, but none, to our knowledge, is as fundamentally simple as this idea.
Most hypocycloidal compressors and engines have foundered in the past, because of either sliding seal problems or, when there are no seals, on account of the need to machine and run rotors to very tight manufacturing tolerances. The new design overcomes this set of problems by having no seals and running fast enough so that air leakage past the pinch point does not matter. One of the main reasons for building a ‘no flow’ validation machine is to find out what happens when things get really hot. If the spring steel ring valve works, it could be the breakthrough, thanks to its simplicity. This should allow manufacturing at low cost and, as Hammerbeck explains: “My feeling is that a single component is not going to have as great a chance of failure as a lot of feather valves.”
The plan is to make the ring as a single piece, rather than as a joined strip, in order to avoid possible fatigue crack failures at the join. The rotor should be much cheaper to machine than the complex-shaped rotors used in conventional screw and scroll compressors. Also, as Hammerbeck observes: “The rotor could be ceramic coated, because it is such a simple shape that the coating is unlikely to fly off. We hope to get up to a five to one compression ration in a single stage, but, for the super charger application, there is not much point in producing a pressure ratio greater than 3.5 to one.”
Pointers
* The design of the compressor is fundamentally simple, with a single eccentric, cylindrical rotor either rolling round the inside of a cylindrical cylinder, but not touching its surface, or being driven round as an eccentric rotor
* Target compression ration is 5:1. But in the target application, which is gas engine superchargers, it is only required to deliver 3.5:1
* The aim of the development is to produce something that is fundamentally cheap to manufacture, compresses gas in the most efficient manner possible, and minimises friction losses in drive shafting and gearing
|
| |
Author
Tom Shelley
|
| |
| Email this article |
| |
|
|
|
|