Take Transient Startup And Shutdown Data - Rule Number One.
Abstract
The thrust herein is to encourage more people to take transient data during startups and rundowns. The common complaint is that it takes too much time and effort to prepare for this information. That argument is "weak" within today's instrumentation capabilities. Further, the more forms of information on the same transient data gives the person saddled with the responsibility of eliminating the problem a better confidence level in the final conclusion. The rotor system, whatever its configuration, forces all the system's responses during the run up and rundown. The ability to look at plots of time waveforms, orbits, polars, Bode plots, cascades, eccentricities, spectra, raw vs synchronous vs compensated information can be valuable. The ability to define accurately the "resonances" in the system and the corresponding ability to direct logical in-situ balancing. The ability to identify rubs when they occur. The initiation of "instabilities" is near breathtaking. The ability to correlate running speed vibration values vs subrotative (whirl or whip) values can be very meaningful and normally with unexpected results, i.e., qualifying the "damping" or "lack of' in a system. In all this, the FM tape recorder should never be overlooked; however, the digital storage capability with computer attached is extremely powerful and simple to manipulate. Speed tracking offers the ability of saving memory during operational delays, as the next speed threshold has not been reached. Three case histories are used to illustrate the advantages of transient measure. The three descend in "actual data" presented to simplify the process. The more complicated/involved issue is addressed first, and it is a 22 MW compressor turbine gas generator, power turbine, generator system. This equipment suffered from overfiring the combustors, over acceleration, and over heating the gas turbine system, rubs on compressor blades and heavy destructive rubs on hot turbine shroud bands. The unit whirls and whips. For about 18 hr after repairs, it suffered other problems of "surge/stall" due to IGV control problems (not covered in this text). The second case discusses a simple low horsepower, medium speed, five stage back pressure steam turbine with problems which existed for over one and one-half years. The basic trending, which showed varying amplitudes and phase, just simply did not define the problem. One set of transient startup/shutdown data on the soloed turbine brought to bear, "quickly," the main problem. Several options were open for correction; the simplest was performed, and the problem totally eliminated. Further, stress corrosion cracking failures were better understood under high stress. The third is a very complicated hot gas turbine single shaft, expander compressor steam turbine, unit. It goes through three (two rigid and one bending) criticals (=resonances) to reach an operating speed of -18,000 rpm. The failure of the first wreck is better investigated via transient data. The rebuilding and rebalancing involved is better understood. The comparison through a "resonant whirl" region that allowed correction is better understood. The original OEM went out of business in 1984.
Description
LecturePg. 69-80
Subject
TurbomachinesCollections
Citation
Jackson, Charles (1994). Take Transient Startup And Shutdown Data - Rule Number One.. Texas A&M University. Turbomachinery Laboratories. Available electronically from https : / /hdl .handle .net /1969 .1 /163482.