Seismic Analysis of Large-Scale Piping Systems for the JNES ... - NRC
Seismic Analysis of Large-Scale Piping Systems for the JNES ... - NRC
Seismic Analysis of Large-Scale Piping Systems for the JNES ... - NRC
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
code stress by using a damping ratio <strong>of</strong> 4% ra<strong>the</strong>r than 5% decreases from 8% to 2%. However, <strong>the</strong> case<br />
<strong>for</strong> <strong>the</strong> US2-1 shows just <strong>the</strong> opposite.<br />
To evaluate <strong>the</strong> impact <strong>of</strong> <strong>the</strong> damping ratio in a slightly different way, <strong>the</strong> reactions to <strong>the</strong> piping systems<br />
are utilized in this appendix <strong>for</strong> comparison. For <strong>the</strong> DM model, <strong>the</strong>re are 17 reaction <strong>for</strong>ces and 9<br />
reaction moments. For <strong>the</strong> US2-1 model, <strong>the</strong>re are 16 reaction <strong>for</strong>ces and 9 reaction moments. The<br />
comparison was made by ei<strong>the</strong>r considering all reactions or considering only <strong>the</strong> significant ones, which<br />
are defined as <strong>the</strong> top half reactions when ranked by magnitude.<br />
Figure A-5 and Figure A-6 show histograms <strong>for</strong> <strong>the</strong> ratio <strong>of</strong> <strong>the</strong> reaction <strong>for</strong>ces and <strong>the</strong> ratio <strong>of</strong> <strong>the</strong><br />
reaction moments using all data. The mean and <strong>the</strong> standard deviation <strong>of</strong> <strong>the</strong> percentage increase in <strong>the</strong>se<br />
reactions (PIR) using 4% damping versus 5% damping are about 11~12% and 4~5%, respectively. If <strong>the</strong><br />
top half significant reactions (ranked by magnitude) are used in <strong>the</strong> comparison to merit more weight on<br />
<strong>the</strong> large reactions, as shown in Figure A-7 and Figure A-8, <strong>the</strong> mean <strong>of</strong> <strong>the</strong> same percentage increase<br />
decreases slightly to 10~11% while <strong>the</strong> standard deviation remains about <strong>the</strong> same. The four comparisons<br />
shown in <strong>the</strong>se figures also show that <strong>the</strong> maximum percentage increase is about 20%. The PIR appear to<br />
be somewhat larger than those <strong>for</strong> <strong>the</strong> maximum code stress, partly because <strong>the</strong> internal pressure and<br />
gravity load used in <strong>the</strong> maximum code stress calculation do not change with damping. It should be noted<br />
that <strong>the</strong> internal pressure, gravity load, and seismic load in o<strong>the</strong>r directions are not considered in <strong>the</strong><br />
analyses to obtain <strong>the</strong> reactions.<br />
The effect <strong>of</strong> spectrum broadening on <strong>the</strong> PIR is demonstrated through Figure A-9 through Figure A-16.<br />
By using broadened input spectra instead <strong>of</strong> <strong>the</strong> unbroadened input spectra, <strong>the</strong> mean <strong>of</strong> <strong>the</strong> PIR<br />
uni<strong>for</strong>mly increases by 3% <strong>for</strong> all cases, <strong>the</strong> standard deviation <strong>of</strong> <strong>the</strong> PIR slightly decreases, and <strong>the</strong><br />
maximum <strong>of</strong> <strong>the</strong> PIR increase in various degrees but remains close to 20%. There<strong>for</strong>e, using <strong>the</strong><br />
broadened spectra as in <strong>the</strong> current design practice is unambiguously conservative in assessing <strong>the</strong><br />
difference in dynamic responses due to 4% damping versus 5% damping.<br />
In summary, compared to 5% damping, <strong>the</strong> reactions <strong>of</strong> piping systems subjected to just seismic<br />
excitations can be 10~16% [mean ~ mean + standard deviation] higher using 4% damping, while <strong>the</strong><br />
maximum increase in reactions can be 20%. These observations should be applicable to <strong>the</strong> internal<br />
<strong>for</strong>ces/moments in <strong>the</strong> piping systems as <strong>the</strong> linear analyses implies. Comparing cases using 4% damping<br />
versus 5% damping, <strong>the</strong> slightly lower increase in <strong>the</strong> maximum code stress (9%) is believed to be caused<br />
by <strong>the</strong> loads o<strong>the</strong>r than <strong>the</strong> subject seismic load. Broadening <strong>of</strong> <strong>the</strong> input spectra can add an additional<br />
3% percentage increase in reactions, compared to using unbroadened input spectra.<br />
145