Regulatory Guide 5.26

Selection of Material Balance Areas and Item Control Areas

ML13350A206
Person / Time
Issue date:	06/30/1974
From:	US Atomic Energy Commission (AEC)
To:
References
RG-5.026
Download: ML13350A206 (7)

U.S. ATOMIC ENERGY COMMISSIONREGULATORYDIRECTORATE OF REGULATORY STANDARDSJune 1974GUIDEREGULATORY GUIDE 6.26SELECTION OFMATERIAL BALANCE AREAS AND ITEM CONTROL AREASA. INTRODUCTIONProposed (38 FR 26735) Section 70.58,"Fundamental Nuclear Materal Controls," of 10 CFRPart 70, "Special Nuclear Material." would requirecertain licensees authorized to possess more than oneeffective kilogram of special nuclear material to establishMaterial Balance Areas (MBAs) or Item Control Areas(ICAs) for the physical and administrative control ofnuclear materials. This section would require that:1. Each MBA be an identifiable physical area such thatthe quantity of nuclear material being moved into or outof the MBA can be measured.2. A sufficient number of MBAs be established so thatnuclear material losses, thefts, or diversions can belocalized and the mechanisms identified.3. The custody of all nuclear material within an MBAbe thevresponsibility of a single individual.4. ICft be established according to the same criteria asMBAs except that control into and out of such areaswould be by item identity and count for previouslydetermined special nuclear material quantities.This guide describes bases acceptable to theRegulatory staff for the selection of material balanceareas and item control areas.B. DISCUSSIONThe division of a nuclear plant into material balanceareas and item control areas can provide improvedmaterial control and accounting as follows:1. A loss or theft of material or of an item or items canbe identified as having occurred in a particular part ofthe plant so that the investigation can be more effectiveand the loss or th.-ft mechanism more easily identifiedand corrected or counteracted.2. The assignment of responsibility to a singledesignated individual for the control of the material orthe items in each area could provide more vigilant andeffective control in each area and thus in the total plant.3. The capability for detecting the loss or theft ofmaterial may be improved by taking smaller materialbalances.!Number of MBAs and ICAsThe number of MBAs and ICAs established at aplant will depend on considerations that are specific tothe individual plants. Such consideratiors will have abearing on the definition of the word "sufficient" in thePart 70 requirement that the number of MBAs and ICAsbe sufficient to localize losses or thefts. It is not thenumber of MBAs or ICAs per se that will be sufficient tolocalize losses but the division of the plant into MBAsand ICAs using bases for such division which will permitidentification and location of losses. Among the mostsignificant considerations for establishing MBAs aredetection capability, physical boundaries, and theorganizational structure to provide administrativecontrol in each area. Other factors which may pertaininclude material types, processes and process layout, andfunctional locations such as laboratories, shipping andreceiving areas, or storage areas.Each of these factors will affect the selection ofMBAs and ICAs and the effectiveness of such selecti- nto control material and items and to identify losseswithin an area. For example, if an MBA is selected toconsist of a building in which there are two processesusing different material types (such as two differentenrichments of uranium), there may be some difficultyin identifying to which enrichment a MUF should beapplied. If each process (probably in separate rooms inthe building) is established as an MBA, MUFs for eachprocess could be identified, and losses or thefts fromUSAEC REGULATORY GUIES Copi of published guides may be obtalned bv f owW Indlatina the londeafred to the US. Atomic Eergy Commlsson, Washington, D.C. Ragulatory Guides we issued to deasolbe and make "table to the PUNlc Atte- Ion: DlrWcto of Raguleory Standards. Comments end uggestlons formethods acoeptable to the AEC Regulatory staff of Iniple10ntlng specific parts of ImWove t in theU Uldea we encouraged and should be sent to t- Sacmrearythe regulatiom, to delineae tedclnques used by the staff In of the Commlislon. US. Atomic Energy Commisslon. Washington, D.C. 20545,minsuigar @Pecific probiwa or postulated accidesnt, or to provide qulde" to Ation: Chief, Public ProAednp Staff.alimn. Regulatory Guido am not subaltum for repulatioms end wrfoetnwkh shem I not reqvurad Methods and solutkio doffa!rnt from tho m out In The guils w Issued In the following ton broad divisiom:tdo qds wIII be amcsable If they Povide a boi fOr the findines Ntqulta tothe ofCorntina of epermit or Neon" by the ComgHIsion. 1. Po rs AestwIT 6. ProTd i2. ffesemrch and To"t Reectota 7. Transportation3. Fuesw and Mattei 8. Occupational HealthPublised win be rrAed peatodiaffy, s appropriate, to eoomua odo 4. Emolton,, nteY l atidSiing 9. Anttrut Rasniwcomnn nd oreflct new Informatin to Ot exOiemene. S. Maaislh and Plent Protection 10. Gensral

.each process could be evaluated and investigated asneedid. In this case,-the process and the material typeprovided a definition of the MBA. It would not benecessary for different types of material to be used inthe two processes for them to be established as separateMBAs. Two parallel processes using the same type ofmaterial might be separate MBAs as shown in Cases IIand V in Appendix A. Division also might be madewithin a process to establish MBAs that would improvedetection capability for separate parts of the process.It may be possible to make the conversion step of afuel fabrication process a separate MBA with a measuredbalance around it. The remainder of the process steps(the fabrication steps, pelletizing, sphere formation,alloying, and any other) could constitute another MBAup to the point where the nuclear material is sealed in afuel pin, rod, etc. After sealing, the material could betreated as an identifiable item and sent to another areafor storage or for further fabrication such as welding,assembly, or testing. Transfer of the items from theMBA would be based on the material quantities asmeasured when the items were loaded.If the linal fabrication area or storage area receivesfuel from more than one loading MBA or is in a separatebuilding on the plant site, it would be designated as anICA using item identity and the measured quantititesfrom the loading MBAs for control.It also may he that the conversion step of theprocess is not administratively separated from the rest ofthe process so that it could not be considered a separateMBA. This would not preclude a measured balancearound that step if the produic from the step weremeasurable before it went into the subsequent step ofthe process. With proper control of the material toassure that all is measured once and only once as itmoves from process step to process step, measuredinternal material balances can be taken around processsegments whose inputs and outputs are measurable eventhough separate MBAs may not be established.Detection CapabilityThe basic objectives of material balance accountingfor special nuclear material are to detect the occurrenceof missing material whether it be lost or stolen, andconversely to provide assurance with a stated degree ofconfidence that if any material is missing it is less than athreshold quantity. A prime indicator for attaining theseobjectives is Material Unaccounted For (MUF). The basefor evaluation of a MUF value is the Limits of Error ofthe Material Unaccounted For (LEMUF). If a MUF valueis within the LEMUF value, it can be stated with aspecified probability that the MUJF is due touncertainties of the measurement system. The validity ofthis statement depends on a number of factors, a majorone of which is the validity of the LEMUF itself. TheLEMUF provides the limits which define the thresholdquantity for a detectable loss or theft. A LEMUF thathas been inflated, either intentionally or inadvertenify,can mask a loss or theft by indcating that a MUF is notstatistically significant, i.e., the MUF is the result only ofthe measurement error of the sstem, when in fact theMUF includes a significant loss or theft. Theramifications of the evaluation of MUF bnd thegeneration of data for MUF and I ZMUF are the subjectsof other regulatory guides. it is sufficient for thepurpose of this guide to know that the combination of aproperly generated MUF and LEMUF provides a lossdetection mechanism.In general, the detection c.,pability of MUF andLEMUF varies directly with the quality of the materialbalance measurements and inversely with the quantity ofmaterial in a given balance. In this context, detectioncapability means the threshold quantity of material thatthe system can detect as being missing with some statedprobability. This capability is represented by a LEMUFvalue stated in terms of quantity, e.g., grams orkilograms. Thtis detection capability based on a measuredmaterial balance is associated with MBAs rather thanICAs, since ICAs are controlled on an item basis. In anICA either all items are accounted for or they are not. Ifthey are not, one or more missing items are indicated,and an investigation is required.The selection of MBAs can affect detectioncapability by lowering the quantity of material in amaterial balance, thereby lowering the absolute LEMUF,since with less material there could be a smaller LEMUFand a greater sensitivity. This assumes that only thequantity of material is changed and not measurementquality.Examples showing the effect of this quantity changeusing this assumption are presented in Appendix A ofthis guide. The examples obviously are sinplifiedgreatly. In real situations there would be complicatingfactors such as discard streams, scrap removals fromMBAs, recycle that might cross MBA boundaries, oruneven distribution of inventory or throughput betweenMBAs, in addition to changes in measurement quality.Each of these could affect the selection of MBAboundaries.Physical BoundariesThe physical boundaries of MBAs and ICAs are notspecified in the proposed regulations except that theymust be "identifiable physical areas." The boundaries.zould be no more than lines painted on the floor aroundcertaiyi parts of the process. However, if MIBA or ICAboundaries do not minimize the possibility ofintermixing of materials or items from different areas,either intentionally or inadvertently, the balance of suchan area or the item control for such an area couldbecome meaningless, and the location of a loss or theftof material or items might not be identifiable. Further.5.26-2 hwith boundaries that do not provide physical separationof materials It is more difficult to discharge the custodialresponsibility for a given area. It is too easy for materialto be moved without the proper documentation andappropriate transfer of custodial responsibility in suchcases. Areas bound by walls, such as separate buildingsor rooms within a building, or by grids, such as a storagecrib or a room divider, are well defined and the materialsand items can be kept within the areas more easily.The critical factor is not the physical boundary, butthe identification of an area which can beadministratively controlled as a separate area aroundwhich either measured material balance control or itemcontrol can be maintained. This control would be relatedto the three aspects of improved material conteol andaccounting noted in the beginning of the Discussionsection of this guide, i.e., loss location, responsibilityassignment, and detection capability. The boundariesselected will depend on combinations of considerationsof these three items.Item Control Areas (ICAs)ICAs are differentiated from MBAs to simplify andimprove the control and accountability of identifiableitems. Control into and out of ICAs is required to be byitem identity and count and previously determinedspecial nuclear material quantities. This excludes itemsthat do not have an identity that will differentiate themfrom other similar items, e.g., loose fuel pellets orunsealed, unlabeled containers of SNM. Such itemscould be substituted for other similar items of differentSNM content or the SNM content changed so thatcontrol of the material would not be maintained.Loaded and sealed fuel rods or tamper-safed sealedcontainers of SNM that have been numbered or in someway uniquely identified provide assurance that thequantity of contained SNM remains as previouslymeasured. ICAs for the handling and storage of suchitems provide control without the need for makingadditional measurements for material balances. Storageareas for finished fuel rods or assemblies, processintermediates, or irradiated fuel assemblies could beICAs. Shipping and receiving areas could be consideredICAs if item integrity is maintained in those areas.C. REGULATORY POSITIONA variety of factors that are specific for individualplants and processes pertain to the establishment ofMBAs and ICAs. The effectiveness of the MBAs andICAs in enhancing nuclear material control should beevaluated for each situation. The factors presentedbelow should be considered in the selection andestablishment of MBAs and ICAs.Physical BoundariesPhysical boundaries of MBAs and ICAs should beestablished so that control of the material moving into,out of, and within the area can be maintained to theextent that material assigned to a given area is keptseparate from material assigned to any other area. Theboundaries of the MBAs must be established so that thequantity of material moving into or out of an area canbe represented by a measured value. The boundaries ofICAs must be established so that items moving into orout of an area can be controlled by identity, count, anda previously measured valid special nuclear materialcontent.Detection CapabilityMaterial flows and inventories and the quality of themeasurement of such flows and inventories should begiven primary consideration in establishing materialbalance areas. Model material balances similar to thoseof Appendix A should be prepared to evaluate theeffects of the selection of various MBAs. Such modelbalances should include all of the material flow,inventory, and measurement factors thai will affect thebalance. Such factors would include recycle, discards,scrap inventory, random and systematic error effects,common measurements and their covariant effect, andchanges in measurement or inventory quality as a resultof division of flows or inventories.Material balance areas should provide the maximumpracticable detection capability consistent with otherfactors such as physical boundaries or process operationand layout. To improve detection capability,consideration should be given to changes in such thingsas process layout or process operations, physicalboundaries, measurement techniques, and inventorytechniques. Consideration also should be given toestablishing procedures for material balances aroundprocess segments internal to MBAs.Number of MBAs and ICAsThe number of MBAs and ICAs established in as-ecific plant should be based on considerations ofdetection capability and the physical and functionalaspects of the plant and material that would assist inidentifying and localizing material losses or thefts.Different material should be processed in separateMBAs.The establishment of separate processes as separateMBAs should be considered. Although detectioncapability may not thereby be improved, theidentification and location of losses or thefts would be.Even when separate processes are not Maintained asseparate MBAs, separate material balances should betaken around each process to identify and locate lossesand possibly to enhance detection capability.Functional areas such as laboratories, receiving andshipping areas, and warehouses or storage vaults should5.26-3 bwith boundaries that do not provide physical separationof materials it is more difficult to discharge the custodialresponsibility for a given area. It is too easy for materialto be moved without the proper documentation andappropriate transfer of custodial responsibility in suchcases. Areas bound by walls, such as separate buildingsor rooms within a building, or by grids, such as a storagecrib or a room divider, are well defined and the materialsand items can be kept within the areas more easily.The critical factor is not the physical boundary, butthe identification of an area which can beadministratively controlled as a separate area aroundwhich either measured material balance control or itemcontrol can be maintained. This control would be relatedto the three aspects of improved material contiol andaccounting noted in the beginning of the Discussionsection of this guide, i.e., loss location, responsibilityassignment, and detection capability. The boundariesselected will depend on combinations of considerationsof these three items.Item Control Areas (ICAs)ICAs are differentiated from MBAs to simplify andimprove the control and accountability of identifiableitems. Control into and out of ICAs is required to be byitem identity and count and previously determinedspecial nuclear material quantities. This excludes itemsthat do not have an identity that will differentiate themfrom other similar items, e.g., loose fuel pellets orunsealed, unlabeled containers of SNM. Such itemscould be substituted for other similar items of differentSNM content or the SNM content changed so thatcontrol of the material would not be maintained.Loaded and sealed fuel rods or tamper-safed sealedcontainers of SNM that have been numbered or in someway uniquely identified provide assurance that thequantity of contained SNM remains as previouslymeasured. ICAs for the handling and storage of suchitems provide control without the need for makingadditional measurements for material balances. Storageareas for finished fuel rods or assemblies, processintermediates, or irradiated fuel assemblies could beICAs. Shipping and receiving areas could be consideredICAs if item integrity is maintained in those areas.C. REGULATORY POSITIONA variety of factors that are specific for individualplants and processes pertain to the establishment ofMBAs and ICAs. The effectiveness of the MBAs andICAs in enhancing nuclear material control should beevaluated for each situation. The factors presentedbelow should be considered in the selection andestablishment of MBAs and ICAs.Physical BoundariesPhysical boundaries of MBAs and ICAs should beestablished so that control of the material moving into,out of, and within the area can be maintained to theextent that material assigned to a given area is keptseparate from material assigned to any other area. Theboundaries of the MBAs must be established so that thequantity of material moving into or out of an area canbe represented by a measured value. The boundaries ofICAs must be established so that items moving into orout of an area can be controlled by identity, count, anda previously measured valid special nuclear materialcontent.Detection CapabilityMaterial flows and inventories and the quality of themeasurement of such flows and inventories should begiven primary consideration in establishing materialbalance areas. Model material balances similar to thoseof Appendix A should be prepared to evaluate theeffects of the selection of various MBAs. Such modelbalances should include all of the material flow,invpntory, and measurement factors thai will affect thebalance. Such factors would include recycle, discards,scrap inventory, random and systematic error effects,common measurements and their covariant effect, andchanges in measurement or inventory quality as a resultof division of flows or inventories.Material balance areas should provide the maximumpracticable detection capability consistent with otherfactors such as physical boundaries or process operationand layout. To improve detection capability,consideration should be given to changes in such thingsas process layout or process operations, physicalboundaries, measurement techniques, and inventorytechniques. Consideration also should be given toestablishing procedures for material balances aroundprocess segments internal to MBAs.Number of MBAs and 1CasThe number of MBAs and ICAs established in asi ecific plant should be based on considerations ofdetection capability and the physical and functionalaspects of the plant and material that would assist inidentifying and localizing material losses or thefts.Different material should be processed in separate"MBAs.The establishment of separate processes as separateMBAs should be considered. Although detectioncapability may not thereby be improved, theidentification and location of losses or thefts would be.Even when separate processes are not tnaintained asseparate MBAs, separate material balances should betaken around each process to identify and locate lossesand possibly to enhance detection capability.Functional areas such as laboratories, receiving andshipping areas, and warehouses or storage vaults should5.26-3 be separate MBAs or ICAs. Receiving and shipping areasmay be established as ICAs provided the material is notprocessed or subdivided and is identifiable by item andin a scaled, tamper.safed condition. Warehouses andstorage vaults should be considered ICAs since allmaterial in storage should be identifiable by item and ina sealed, tamper-safed condition.Item Control AreasAreas designated as ICAs should contain only itemsthat are identified to differentiate them from othersimilar items and are in a sealed tamper-safed conditionthat assures the integrity of prior measurements. Suchitems as loose fuel pellets or unsealed, unlabeledcontainers of SNM do not have identities that willdifferentiate them from other similar items and aretherefore not acceptable for control in ICAs.I5.26-4Ill APPENDIX AEFFECT OF MBA SELECTION ON LEMUF AND DETECTION.CAPABILITYTo show the effect of MBA selection on theLEMUF and the detection capability, several examplesare presented. The examples are given for a simplifiedplant consisting of two conversion lines and twofabrication lines. The plant may be represented by thefollowing diagram:-C, C2F, F2where:C1 & C2 = Conversion lines I and 2F, & F2 = Fabrication lines 1 and 2The MBAs used in the example will be:Total Plant -All lines in one MBAParallel MBAs -MBA I = C1 + F,-MBA 2 =C2 + F2Series MBAs -MBA I = C, + C2-MBA2=F, + F2The examples will consider these configurations forboth inventory-dominated and throLllhput-dominatedprocesses. The following parameters are common to allexamples:1. Throughput is in 2-kg batches (Cases I, I1, and III) or2"-kg batches (Cases IV, V, and VI) each of which ismeasured to +/-0.25% (+/-5 grams and +/-50 grams,respectively).2. Fbr simplification it is assumed that there are nodiscards and that there is 100% yield in the form ofproduct batches equal in size to the input batches andmeasured to +/-0.25%.3. The inventory interval is two months.4. Beginning and ending inventories are the same sizebut do not contain any common items or material.5. The total plant inventory is measured to +/-0.2% anddistributed so that when one-half is measured in a singleMBA, it is measured to about +/-0.28%.6. For simplification, only random errors have beenconsidered. In a real situation both systematic andrandom errors would need to be considered.7. For simplification it has been assumed that there are.no common measurements contributing covarianceeffects. In real situations such covariance effects wouldneed to be considered.Case I-Inventory-Dominated Process, Total Plant MBABeginning and Ending Inventories each:250 kg+/- 500 gInput and Output each:30 batches @ 2 kg +/- 5 g = 60 kg +/- 27.4 gLEMUF = 2(27.4)2 + 2(500)2 =+/-708 gThe single total plant MBA detection capability istherefore +/-708 grams.Case Il-Inventory-Dominated Process, Parallel MBAs.For each MBA:Beginning and Ending Inventories each:125 kg+/- 354g /Input and Output each:15 batches @ 2 kg +/- 5 g = 30 kg+/- 19.5 gLEMUF = -,2(9 9.5) + 2(354)2 = +/-501The detection capability has been improved from708 grams for the single total plant MBA to 501 gramsfor each MBA. That is, a loss or theft of 501 grams ineither MBA would have the same probability of beingdetected as a loss of 708 grams in the single total plantMBA.The total plant LEMUF for the two parallel MBAswould'be +/-501 2 = +/-708 grams, the same as thesingle total plant MBA LEMUF. This is because noadditional measurements were made, none of themeasurements were improved by dividing the plant intotwo MBAs, and there were no common transfersbetween the MBAs.Case III-Inventory-Dominated Process, Series MBAs.For each MBA:Beginning and Ending Inventories each:125 kg +/-354 gInput and Output each:30 batches @ 2 kg +/- 5 g = 60 kg +/- 27.4 gLEMUF = -/2(27.4)2 + 2(354)2 = 502 gThe detection capability for Case III is essentiallythe same as for the individual parallel MBAs (Case 11).This would be expected because the inventorydominates and it is divided in half in each case. The totalplant LEMUF does not change, even though there havebeen additional measurements made, i.e., for the transferbetween MBAs. This transfer measurement is assumed tobe the same for both MBAs. That is, the output5.26-5 measurement of MBA I is the input measurement ofMBA 2. When the uncertainties of the two MBAs arecombined to obtain the total plant MBA uncertainty,this transfer measurement is common and drops out ofthe equation for tile total plant.TFlhe assumpticn in this case was that the transfermeasurement is as good as the input and productmeasurement s. To thie extent that this is not true, theindividual MBA LEMUF is increased and the detectioncapability decreased. This effect becomes morepronounced as the absolute uncertainty of the transfermeasurement increases. For example, if the uncertaintyof the transfer measurement were thie same as that of theinventory, i.e., 60 kg +/- 354 grams (3% instead of thc.previously used 0.25%) the LEMUF of the individualMI BAs would be +/-614 grams. There would still be soneadvantage in dividing the plant into the series MBAs butnot as much as when the transfers 1.:1ween MBAs couldbe measured with a precision approaching that of theinput and produrt measurements.It can he seen froin Cases 1. II, and III that striking abalance around portions of the inventory will increaselhe detection capability tor each portion, but not for thetotal plant.In Case I, if anl actual loss of 708 grams hadoccurred, it would be expected that the MUF wouldexceed the LEMLUF of +/-708 grams part of the time. Theprobability of the exceeding tile LEMUF in thiscase could he calculated. When the MUF exceeds theLEMUF, an alarm is sounded and the high MUF isinvestigated as occurring somewhere in the total plant.In Cases If and II the balance is taken aroundsmaller areas so that the detection capability is improvedto 502 grams for each area. If a loss or theft of 708grams were to occur in either area, it would have ahigher probability of detection since the LEMUF is only+/-501 grams. In addition, if such a loss did occur, thearea in which it occurred would be shown by the highMUF in that MBA so that the investigation could beconfined to the smaller area. In order for a person tosteal 708 grams of material with the same probability ofsuccess. i.e., being undetected, as in a single total plantMBA, portions of the material would have to beremoved frmm two different MBAs or over a longerperiod of time in the same MBA. This would expose thethief to an increased probability of detection by theplhysical protection surveillance and alarm systems.If a person wvere to steal 501 grams from each MBAof Case II of Ill the detection capability would be thesame for each NIBA as for theft of the 708 grams fromthe single total plant MBA. The total quantity stolen,however, would he so large that the total theft wouldhave a higher probability of detection upon calculationof the balance for the entire plant. In-the example, thecombined LEMUF for the two MIBAs would be +/-')08grams but the MUF (i.e., material stoien) would le 1002gram!, and probably would trigger an investigAtiin. Th7loeatl'or of the loss within thp p!"'lnt in this case n:ay nol'Wbe known because the MUF of the individual MBAs maynot Lave exceeded the LEMUF.Case IV--Throughput-Doniinated Process. Total PlantMBABeginning arid Ending Inventory each:50 kg +/- 100 gInput and Outptut each:30 batches 61 29 kg +/- 591 g = W00 kg +/- 27- iiLEMUF = vF2(274)2 + 2(l00)` :+/-413 gCase V --Throughput-Dominated Process, Parallel M, AsFor each MBA:Beginning and Ending Inventories each:25kg +/- 71 gInput and Output each:15 batchesV20kg +/- 50g 300 kg +/- 194 gLEMUF = 4 2(71)2 2'2 gThe individual MBA detection capability has beenimproved from 412 grams to 292_rams. The total plantLEMUF will not change (9-_2,v2 = +/-413) hecause noadditional measurements were made nor were anvAdltimprovements made in the measurement of any of thmbalance components.Case VI-Throughput-Dominated Process, Series MBAsFor each MBA:Beginning and Ending Inventories each:25 kg +/- 71 gInput and Output each:30 batches QV20 kg +/- 50 g = 00 kg +/- 274 gLEMUF = x/2(274)2 + 2(71.)- = a400 gThere has been little gain in the delvction capabiliyover a total plant MBA because t0he t.hroughpul is liesame for each of the two ser~es MBAs as t:or a siigl, totalplant MBA. The little gain that is realize"\ is due to thegain obtained by dividing the inventory in half. In ad'di-(ion, if the transfer measurement between MGAs in CaseVI is not as good as the input and produc; measurementsthere may be a loss of detection capability. For example,if the precision of the transfer measurement for eachbatch is +/-0.5% instead of +/-0.25%,, the uncertainty of thistotal transfer measurement becomes 600 kg +/- 547 gramsand the LEMUF for each MBA becomes +/-780 grams.This is a poorer detection capability than the 412 gramsfor the single total plant MBA. The effect of this transfermeasurement is cuore pronounced here than in Case ll1where the inventory dominiated. a5.26.6