Laserfiche WebLink
SFM BY = ~17-97 ; 15 ~ 30 ; I NDFPE7VDFNCE M 1 N Imo' ~ 30~. 832 8106 ; # 5 <br />meter wntinuea ty be used. The change in meter is evident (and, thou, recorded) in the record (Attachtrten[ 1) by <br />tits significantly larger cumulative flow recorded at the beginning of lts use (larger, again, because the meter had <br />been accttu,tdazing flow at Phase I prior to use az Phase II). <br />Using the data graphed in Attacht¢eue 2, the temporal changes of the two variables were examined. The graph of <br />tintC versus the flow and head measurements is providt~ as Attathatettt 4. This ~=~r^~s"ion confirmed [hat the flow <br />rtes respond relatively rapidly to dtanges in head Lt the Prelmant Solution Storage Area (terttted the PLS in the <br />text of this letter). This cotrelazion suggest (1) the imperfection allowing pregnant solution ty enter the LVSCS is <br />relatively near the LVSCS pumps (which are Irx:ated ty the side of the overlying footprint of the preg,taat solution <br />pump wissona), or (2) the LVSCS gravel is, as designed, very permeable, or (3) both. It is believed that the data <br />reflect a combinazien of these two possibilities. <br />The "EPA' estimating technique refert"¢ced above was used repeatedly ro obtain a better idea of the type of <br />"pcrttteabilily configuration' that may axis[ i¢ the as-boil[ Casson Phase U PLS based on the recent flaw and head <br />measurements. This computational approach was taken ty evaluate the adequacy of the sin of the submersible <br />pampa. "Permeability configuration" refers to the locatiotta(a) of pusaiblt imperfections th[oughout the lower <br />synthetic liner of the PLS. For the computations, this PLS litter surfsce was divided into area slices based on 5-ttxt <br />increments of elrvadon. The average clevatioa and cortesponding area of lined surface (corrected For slope) for <br />each slice was usal ty compare the fluid pressure for different depths of pregnant solution. Each increment was <br />assigned the assumed number (or siu) of imperfaxions, based on probabilities of occurrence of both the <br />itrrpetfection accurrtnce and size. Calculations were first pertotrard with the "standard' asautr~tion of one 2 <br />millimeter (mm) diameter imperfection per acre (i.e., about 13 such imperfections for the ca. 13 ecTe area of the <br />Phase II P1.5), az the median (5096) confidence level (i.e., a diameter of 2mm t 1.0 mm with a nomtal diattibution <br />of diatrtetets about the mean). Two increases in the siu of the imperfections were tested, based on increasing tits <br />diameter according to the standard deviation multiplied by a factor appropriate For the confidence levels. Further, <br />the ntunber of defects within any interval enc^^~^, ~~=ino other thmt as even number aI a~w was determined by a <br />random number generating program and the wtnpmations repeated [o derive a die[[ibutioa and rates for each level <br />of ranfidence. As an exarople, an interval with 1.34 surface acres of liner was assumed [o have a 34 pen;en[ <br />probability of 2 imperfa:tions. This procedure results in the development of the various confidence levels plotted <br />on [he graphs included with this report, whirb then ere used w project attticipated flows as a function of the depth <br />of [Ire pregnant solution (where. of course, the area of liner involved is nom-linearly proportional to the pregnant <br />solution hued). Based o¢ these cotttpu[ations, it wan confirmed the[ the current Phase 11 LVSCS flow tarry are <br />higher than the confidence intervals used ty prepare cstit¢atc for the 1993 submission for Amend[¢ert[ Number 6. <br />Part of this diflcrence is anributable ty using the as-built typography which differs from the topography used in <br />developing the Applicmioa a~ construction drawings. OFcoutee, the field data continue to suggest, even az higher <br />pregnant solution heads, thaz the rates remain acceptable based on the pumping rypabilities of the LVSCS. The <br />relationship of [he field data to the 2t 1 teen multi-imperfection configtuation is shown on the graph of daily flow <br />v, pregnant solution hued provided as Atnchtneat 5. Thu cottfiguratioa ernploya 2 t 1 nun defxts az every location <br />and may be referred ty as "2t 1 that + 2t1 rmn." <br />CC&Y a¢d Golder Associate then repeated the computations For other perarrability wafiguncions to identify one <br />or more configurations that would better describe the distribution of the field elan far Phase Ii. The same esdmatiog <br />technique was adjusted to assume teary Jifferettt combinations of imperfection sine and location, l¢itially, the <br />resulting range of probability curves trill did vo[ include all of the held measurements. The field elan suggested <br />a slight decline in the rate of LVSCS flow iactease with increasing pregnam solution height, noel this trend was nol <br />being aaounted foe. In addition to simple variations based on the tease defect diameur, the variable of the !oration <br />of the lowermost imperfection was examined during these compuntions, u was changing the mean defect diameter <br />a[ different elevations. Based on numerous trial computazions, it was decided that the first "defect" might be located <br />soatewhere aMve the standard location for the lowermwst defect (the lowermost median elevation is 9247.5 feet, <br />halfway between the base at 9245 and the top of the firs[ slice, which is 9230 fart). Trill co¢,puta[ioas la:ated [he <br />lowermost Imperfection within the third-lowest interval, at 9757.5 fxt. Hmong the imperfx[ion diameters used <br />for wmpuntion were St2.5 rem, 6t3mm, and 713.5 mm. The results for these are reported here, though one <br />