THE UBIQUITOUS COMPRESSOR: THE EXIT

How large must the diameter of a compressor be? Every element in the stage affects the diameter, but the exit element has a peculiar impact. In recent articles, we have looked closely at the impeller and the diffuser; now we have the task of getting the flow out of the compressor and doing that efficiently. The common choices are a volute, a collector, or a return channel which eventually discharge into a pipe. Each has its own advantages and disadvantages.

The volute is simply a duct of increasing area that is wrapped around the exit of the diffuser; its geometry is rather simple, but the flowfield can be complex, especially in the region of the tongue, or cutwater, as it is also known. The aerodynamic loading in this region can be high,with flowfield separation common at off-design point operation. Mechanically, noise and vibration can be generated in this region, and cracks are not uncommon in the tongue region unless carefully designed.

Collectors are a crude variation of the volute, and while they are simple and inexpensive, they are not often used due to the typical loss in performance.

The return channel is used for multistage compressors, where the flow from one stage must enter the eye of the next stage, which is located in very close proximity. Two major styles are in use: 1) the common return channel, comprised of a radial diffuser, a vaneless crossover, and a return or deswirl cascade; and 2) the continuous crossover vane system that is a very compact cascade that starts at the impeller discharge and then bends over and down to the eye of the next stage. The latter is fully vaned and is made up of all the elements of the first system in a compact vane row. The first type is widely used in industrial compressors, whereas the second type is rarely found in compressors, but is extremely common in pumps where virtually every vertical or submerged pump uses multiple sets. Photographs of each are found in Figure 1.

The performance of the continuous crossover systems for diverse pumps is shown in Figure 2. It is clear that low levels of loss can be attained and that considerable pressure recovery is possible. Furthermore, it is seen that the minimum loss of approximately 0.2 occurs at a radius of about 40% above the impeller tip. This is both small and efficient compared to the classical return channel system, which has a minimum loss of 0.45 or more and a radius ratio of about 1.7 to 1.9 or more. Recall that between 25 and 50% of the power delivered by an impeller to the flow remains as kinetic energy at the impeller exit, and hence, using these loss coefficients, each stage loses from 12 to 25% of its work input for the traditional return channel and only 5 to 10% for the continuous crossover. Therefore,the latter can save from 7 to 15% if one uses it instead of the conventional system. Hence, for an 8- stage machine, it may be possible to achieve the same head rise with one less stage and with a nice savings in energy and much less metal (about a 30% reduction in diameter).

While the continuous crossover is rare in compressor designs, one commercial company has employed just such a system in a 2-MW gas turbine, and it performed very well. Concepts NREC has also tested such a system for a similar application and found it to be just as satisfactory as when used for pumps. Hence, we can see that industrial designers have another option to consider, one which reduces the machine diameter, increases the efficiency, and considerably reduces the machine weight and therefore its cost. The continuous vane system may have a length issue, as the flow path wraps back to form the eye of the next stage. This is at least partially offset by the fact that the losses are reduced and fewer stages are required for a machine. Sadly, this has not been seriously considered for industrial compressor applications, even though it is a centuries old practice for vertical pumps and even the venerable space shuttle hydrogen fuel pump.

Author

Dr. Japikse will be teaching severalweeklong seminars in 2012. “Experimental Techniques for T u r b o m a c h i n e r y Development”, is the first, on March 19-22, 2012 in White River Junction,VT. Hewill also be teaching some courses in SE Asia as well. Please visit www.ConceptsNREC.com /education.