FI9800097

&STUKS T U K - Y T O - T R 1 4 4 May 1998

Evaluation of the doseassessment models forroutine radioactive releasesto the environment

Jukka RossiVTT Energy

S T U K • S A T E I L Y T U R V A K E S K U S • STR A L S A KERH E T S C E NTR ALE NR A D I A T I O N A N D N U C L E A R S A F E T Y A U T H O R I T Y

The conclusions presented in the STUK report series are those of the authorsand do not necessarily represent the official position of STUK.

ISBN 951-712-259-4ISSN 0785-9325

Oy Edita Ab, Helsinki 1998

S T U K - Y T O - T R 1 4 4

ROSSI, Jukka (VTT Energy). Evaluation of the dose assessment models for routine radioactive releasesto the environment. STUK-YTO-TR 144. Helsinki 1998. 21 pp.

ISBN 951-712-259-4ISSN 0785-9325

Keywords: nuclear reactor operation, routine radioactive releases, dose assessment models

ABSTRACT

During normal operation of NPP the routine release of radioactive aerosols and particles out of theplant can be assumed to take place as a consequence of large repair and maintenance works, whereasthe gaseous releases are distributed more smoothly over the whole operation cycle. The purpose of thiscontracted work was to evaluate the needs of development concerning the dose calculation models forroutine releases and application of the models for exceptional release situations at the power plantsoperated by Imatran Voima Ltd and Teollisuuden Voima Ltd.

First, the differences of the calculation models concerning input data, models themselves and outputare considered. Subsequently some single features like importance of nuclides in exposure pathwaysdue to change of the release composition, dose calculation for children and importance of time period ofparticle releases are considered.

The existing dose calculation model used by the radiation safety authorities is aimed at a tool forchecking the results from calculations of doses arising from routine releases performed by the powercompanies. Therefore characteristics of an independent, foreign model and its suitability for safetyauthorities for dose calculations of releases in normal operation is also assessed. The needs of improve-ments in the existing calculation models and characteristics of a comprehensive model for safetyauthorities are discussed as well.

S T U K - Y T O - T R 1 4 4

CONTENTS

ABSTRACT 3

1 INTRODUCTION 5

2 ASSESSMENT OF DOSE CALCULATION MODELS FOR ROUTINE RELEASES 62.1 Models evaluated 6

3 ASSESSMENT OF THE CALCULATION MODELS OF THE POWER COMPANIES 83.1 Atmospheric releases 8

3.1.1 Dispersion data 83.1.2 Dispersion models and definition of concentration in the environment 83.1.3 Dose calculation models 9

3.2 Discharges to sea 93.2.1 Calculation of the concentration 93.2.2 Dose calculation 10

3.3 Input data for the models: measurements, derived data and environmental data 103.4 Factors affecting the calculated results 103.5 Adequacy of dose calculation 123.6 Adequacy of documents of power companies concerning demands in YVL-guides 13

4 VALTO PROGRAM OF STUK 14

5 ASSESSMENT OF AN INDEPENDENT MODEL 165.1 Brief description of the atmospheric and marine models 165.2 Performed comparative dose calculations 175.3 Assessment 17

6 DEFINITION OF THE MODEL FOR SAFETY AUTHORITIES APPLICABLE FORCALCULATION OF NORMAL RELEASES 19

7 CONCLUSIONS 20

REFERENCES 21

STUK-YTO-TR 1 44

1 INTRODUCTION

In nuclear power plants radiation and radioactivesubstances appear especially in the reactor fuel,coolant and corrosion products dissolved in thecoolant. If the tightness of fuel pins is impaired,fission products may leak into the coolant. Gaseo-us substances, which are chemically inactive andvolatile like noble gases, are most easily released.Also soluble, solid substances in the temperatureof the coolant like e.g. caesium can leak. Neutronsproduced in fission reactions can activate stablenuclides of the coolant resulting in nuclides suchas Cr-51, Mn-54, Co-58, Co-60, Fe-59, Ag-llOmand Sb-124.

Purification systems remove impurities fromthe loop with the result that impurities are accu-

mulated in the storage tanks. Part of the radionu-clides is carried along with the coolant flow anddeposited on the surfaces of the loop. If flowconditions and water chemistry are changed —such as in the case of the reactor shutdown andstart-up — the radionuclides deposited on the oxi-dation layers of the loop can partly become water-borne again. The release of aerosols and particlesout of the plant can be assumed to take place as aconsequence of large repair and maintenanceworks, whereas the gaseous releases are distrib-uted more smoothly over the whole operationcycle. Use of waste processing systems contributesto the amount of water-soluble releases.

S T U K - Y T O - T R 144

2 ASSESSMENT OF DOSECALCULATION MODELS FORROUTINE RELEASES

The purpose of this contracted work is to evaluatethe needs of development concerning the dose cal-culation models for routine releases and applicati-on of the models for exceptional release situationsat the power plants operated by Imatran VoimaLtd and Teollisuuden Voima Ltd. This kind of spe-cial needs could be formed if the composition andamount of the release is remarkably changed. Thisdoesn't mean real accident cases since dose calcu-lations for those situations can not be performedemploying the models used in the range of normaloperation conditions.

First, the differences of the calculation modelsconcerning input data, models themselves andoutput are considered. Subsequently some singlefeatures like importance of nuclides in exposurepathways due to change of the release composi-tion, dose calculation for children and importanceof time period of particle releases are considered.The needs of improvements in the existing dosemodels are discussed as well.

The dose calculation model used by the radia-tion safety authorities is aimed at a tool forchecking the results from calculations of dosesarising from routine releases performed by thepower companies. Therefore characteristics of anindependent, foreign model and its suitability forsafety authorities for dose calculations of releasesin normal operation is also assessed in this work.

2.1 Models evaluated

The model evaluations in this work are based onthe following computer codes developed for consi-deration of radiation exposures caused by normaloperation effluents:

Imatran Voima Ltd (IVO)• DOSES92, discharges to the atmosphere and

in the sea,

Teollisuuden Voima Ltd (TVO)• ZEUS, KRONIC and PAMEDO, discharges to

the atmosphere• MARI, discharges in the sea,

On the other hand comparisons are made betweenthe following models:

Radiation and Nuclear Safety Authority (STUK)• VALTO (Checking program of the results

provided by power companies), discharges tothe atmosphere and in the sea,

National Radiological Protection Board (NRPB)• PC CREAM, discharges to the atmosphere and

in the sea.

The models used by the power companies dateoriginally from the 1980's, but along the years se-veral required modifications have been made inthe models based on increased knowledge. Themodel of NRPB has partially been funded by theEuropean Commission and it had been completedin 1997. The use of these models is largely basedon the data measured at the power plant site.Releases to the atmosphere and into the sea areeither measured or calculated based on the samp-les. Atmospheric dispersion conditions are measu-red continuously at the site meteorological tower.However the measured rain intensities are notused in the models of the power companies. Desc-

S T U K - Y T O - T R 144

ription of the models supplied by IVO, unlike theTVO documentation, is delivered in a consistentform. The sites of IVO and TVO are located atHastholmen (near the town of Loviisa and therefo-re it is called the Loviisa NPP) on the coast of theGulf of Finland and at Olkiluoto on the coast ofGulf of Bothnia, respectively.

VALTO model exploits as input information thedischarge data and precalculated dilution factorsprovided by the power companies. VALTO is in-tended to be employed for checking the magnitudeof the results supplied by power companies butnot actually to be an independent computing toolfor dose calculations.

STUK-YTO-TR 1 44

3 ASSESSMENT OF THE CALCULATIONMODELS OF THE POWER COMPANIES

3.1 Atmospheric releases

3.1.1 Dispersion data

Dispersion calculation for atmospheric releases isbased on the dilution factors calculated from thedispersion condition data measured continuouslyat the meteorological tower of the power plantsite. Classification of the dilution factors is basedon the appearance of wind direction in dispersionsectors and on frequency in stability and windspeed classes. An alternative approach would beto calculate dose in each hourly dispersion conditi-on, but so far the experiences show this method tobe too slow.

The parameters characterising prevailing dis-persion conditions are measured continuously atthe meteorological tower of the site and resultsare recorded in one hour's intervals to the files.The measured data consist of three key parame-ters: dispersion direction, wind speed and derivedstability category, each of which indicates theprevailing dispersion condition during one-hourperiod. Based on these original data the completeddilution factor data file is compiled and submittedto STUK corresponding each quarter of the year.Possible deficient hourly data can not be detectedamong the dilution factor data.

3.1.2 Dispersion models and definitionof concentration in theenvironment

The dilution factors of DOSES92 model of IVO arecalculated with an ordinary Gaussian dispersionmodel, in which the horizontal concentration isdivided smoothly over the dispersion sector andthe vertical dispersion is calculated employing theaz parameter which in turn depends on the stabili-ty. Atmospheric releases are assumed to take pla-

ce either through the ventilation stack or from theeffective altitude, which is half of the averageheight of the buildings. In the latter case the en-hanced mixing effect of reactor building on diluti-on is taken into account. Dispersion conditions areclassified into seven stability and wind speed clas-ses, dispersion directions are divided into twelvedispersion sectors. Dilution factors are calculatedseparately for the whole year and for the pastureseason.

Depletion of the radioactive plume is consid-ered through radioactive decay and daughter nu-clide build-up, as well as through dry and wetdeposition. Assessment of wet deposition is basedon duration of rain and washout coefficient. Theon-site measurements of rain intensity are notused in the atmospheric dispersion calculations.Instead of that rain duration is estimated to be 16percent based on the measurements at the Helsin-ki-Vantaa airport. Rain is assumed to be uniformand similar in each stability class and the value of104 s-1 is used for the washout coefficient corre-sponding to rainfall at a rate of 1 mmh1 . Due toNordic conditions there is very seldom rain occur-rence in winter (it is snowing) and therefore onlytwo thirds of the total wet deposition is taken intoaccount. On the ground surface the decrease ofthe radionuclide concentration through radioac-tive decay, weathering and migration into deepersoil is taken into account.

Both direct deposition on the plants and trans-fer via root uptake are taken into account inestimating the radionuclide concentrations in theplants used as foodstuffs. Deposition on the plantsis not subtracted from the deposition on theground. Weathering of the vegetation aboveground surface due to rain and wind as well asremoval of radionuclides from the root zone areconsidered. The feed of cattle is assumed to in-clude grass and forage of the pasture and stored

STUK-YTO-TR 1 44

fodder, in which case the storage time is takeninto account as an average value. The predefinedfraction of the radionuclides is transferred intomilk.

ZEUS is the program used for initial sorting ofthe measured data at the meteorological tower. InKRONIC program dispersion is calculated withthe Gaussian model, in which the horizontal con-centration is divided smoothly over the dispersionsector. The Finnish Meteorological Institute hasdeveloped the vertical dispersion parameters. Ra-dioactive decay is calculated during dispersion butdaughter nuclide build-up is not considered. Thedry deposition value of 0.015 m/s is used for alldepositing nuclides. Use of this value for drydeposition implies that the importance of wetdeposition can be assessed to be so small, that it isnot necessary to take it into account separately.On the other hand depletion of the plume iscalculated through radioactive decay, but dry orwet deposition is not taken into account. Radioac-tive decay is calculated in the ground but weather-ing is not included.

Concentration in the foodstuffs is calculatedpartially in the same way as in DOSES92, eventhough there are differences in parameter values.Deposition value used for the pasturing seasondoes not directly correspond to the value of thepasturing season, migration in root zone is notcalculated and integration time of root uptake isdefined differently.

3.1.3 Dose calculation models

External dose from the plume is calculated for thewhole body in DOSES92 with the semi-infinitecloud model. Shielding factors are not considered.Internal dose is calculated through inhalation andingestion pathways to different organs. Also thethyroid dose of child is calculated. Ingestion doseis calculated separately for each foodstuff.

KRONIC program calculates external dosefrom the cloud also with the semi-infinite cloudmodel. Other exposure pathways are calculatedwith PAMEDO program. In the case of externaldose from fallout the roughness of the surface andshielding provided by buildings are taken intoaccount. In the ingestion exposure pathways thedelay between production and consumption areconsidered. Doses to children are not considered.

The ingestion doses caused by C-14 and H-3are calculated both in DOSES92 and in PAMEDOmodels based on the air concentration. In DOS-ES92 concerning C-14 it is assumed that thefraction of radioactive carbon in the body is thesame as that of carbon in the air. Concerning H-3it is assumed that the portion of H-3 in hydrogeningested is the same as that portion in the air atthis location. In PAMEDO the similar calculationapproach is used.

3.2 Discharges to sea

3.2.1 Calculation of the concentration

Assessment of dispersion in the marine environ-ment is calculated employing the compartmentmodelling approach. In DOSES92 two reservoirsare modelled, the smaller one represents the coveof Hastholmsfjarden, in which the discharge is re-leased and the larger one is the Gulf of Finland.The water movement between the two reservoirscreates two-way dispersion flow and in both com-partments the radioactive substance is assumedto be uniformly dispersed in the whole volume.Depletion of radionuclides is assumed to occurthrough radioactive decay and sedimentation.Concentration in fish is calculated on the basis ofa nuclide specific concentration factor. Depositionon the shore surrounding the considered surfacewater body is based on the nuclide specific deposi-tion velocity that is proportional to the value me-asured locally for cobalt.

MARI program includes also two sea-compart-ments, the first one representing local area andthe second one the Gulf of Bothnia. When com-pared with DOSES92 model the transfer process-es are dealt with employing a more comprehen-sive approach which renders a possibility to inputmore parameters which further have an essentialeffect on the results. Those parameters are ob-tained based on the local measurements. In mod-elling particular attention is paid to the transferof radionuclides to the sediment on the shore,because the radioactive material accumulated onthe shore causes remarkable dose. In principle thetransfer takes place from water to the shorethrough material in the water which is disposed tosedimentation. Besides transfer to bottom sedi-ment and to shore and fishes, radionuclides are

S T U K - Y T O - T R 144

transferred to algae, which are gradually accumu-lated on the shore as well. Depletion is caused byradioactive decay and migration into the seabed.

3.2.2 Dose calculation

In DOSES92 and MARI external dose is caused byswimming, boating and stay on the shore. Dose isdetermined as the product of concentration anddose factor. Internal dose is caused by ingestion ofseafish.

3.3 Input data for the models:measurements, derived dataand environmental data

The parameter values and other input data usedin the models originate in literature, measure-ments or other values derived from those. As Tab-le I indicates there exist some differences in thevalues of the az parameter, which is used in theequation of the dilution factor.

Dispersion data is measured at meteorologicaltower of the power plant site; release data arebased on the continuous measurements in theventilation stack. There is a continuous samplingsystem for aerosols and iodine but a periodicalsystem for noble gases and H-3 at the Olkiluotosite. The discharge estimates of C-14 are based onmeasurements during the previous campaign butthe releases are assumed to be proportional to theprevailing energy load factor. At the Loviisa powerplant the aerosol, iodine and noble gas samplesare analysed every week and furthermore the C-14 and H-3 samples each month. Releases into thesea are determined from the samples of the dis-charge tank involved. Releases are reported toSTUK every three months.

If dispersion data measured at the altitude of100 meters at Olkiluoto are missing, the data arereplaced by the values measured at the altitude of60 meters. Stability classification is based ondifferent approaches at Olkiluoto and at Loviisa.In the former case stability is calculated fromwind direction variation (correlation with disper-sion) at the altitude of 100 meters but in the lattercase from the measured temperature difference(correlation with atmospheric stability) betweenthe altitudes of 145 and 45 meters. Imperfect

Table 1. Some values of the sz parameters for stackreleases in the stability classes of A and F used inDOSES92 and KRONIC programs.

Stability

DOSES92 ojm]

KRONIC ojm]

A

347

787

33

25

10000

A F

1565

2000

94

63

dispersion data is not used for the weather datadistributions; in other words the release is dividedequally into the time period including properdispersion data.

The parameterisation in the models used bythe power companies is to some degree different.In DOSES92 the dry deposition velocity 0.005 m/sis used for other substances than noble gases andiodine (for iodine 0.01 m/s) and the washout coeffi-cient 0.0001 1/s is used. In KRONIC the drydeposition velocity 0.015 m/s is used and no valueis used for wet deposition. Table II shows someother parameters that differ from each other.

Doses are calculated for the so-called criticalgroup and for the population living inside thedistance of 100 kilometres from the power plantassuming that locally produced foodstuffs are alsoconsumed locally. IVO takes into account also thesummer inhabitants within 5 kilometres from Lo-viisa Nuclear Power Plant with a weighting factor0.25. The dilution factors for the critical groups inOlkiluoto and in Loviisa don't differ essentially.

3.4 Factors affecting thecalculated results

Annual dilution factor data record can be estima-ted to include stochastic differences; the influenceof those on the results remains relatively minor.This can be seen by comparing the variations ofthe dilution factors for the critical groups at Olki-luoto and Loviisa in 1995-1996. The difference isa factor of two at the most.

Dose assessments are based on the releasesreported, not on the measurements carried out inthe environment. If the doses of the critical groupsat Olkiluoto and Loviisa are compared, one cannotice that the external doses due to the releasefrom the ventilation stack are equal, but external

10

S T U K - Y T O - T R 1 4 4

Table II. Some parameters of the DOSES92 and KRONIC / PAMEDO IMARI models.Parameter

Inhalation rate [m3/a]

Ingested green and root vegetables, grain [kg/a]

Feed of cattle [kg/day]

Effective density of ground [kg/m2]

Sedimentation velocity in local watercourse [kg/m2a]

OOSES92

8400

185

40

215

1.5

KRONIC/PAMEDO/MARI

7972

205

50

240

2.0

Table HI. Individual [Sv] and collective dose [manSv] in the IVO I TVO IVALTO models.

Individual dose

Stack releases

External

Internal

Effluents to sea

External

Internal

Population dose

Stack releases

External

Internal

Effluents to sea

External

Internal

TVO/VALTO

2.3-10-8/2.6-10-«

1.2-10-76.9-1CH

2.3-10-76.5-10-7

1.5-10-75.0-10-7

4.8-10-4/1.4-10-5

2.3-10-3/7.6-10-5

1.8-10-5/7.9-10-4

2.5-10-4/4.6-10-4

1995

IVO/VALTO

1.9-10-s/3.7-10-«

1.5-10-72.3-10-8

2.3-10-9/3.6-10-9

7.3-10-V9.5-10-9

2.0-10-4/1.1-10-4

2.0-10^/1.4-1CH>

3.5-10-72.4-10-6

2.2-10-5/4.5-10-6

TVO/VALTO

1.7-1(H>/5.7-10-«

1.5-10-71.0-10-*

1.7-10-74.7-10-7

1.0-10-73.3-10-7

2.6-10^/7.5-10-4

2.2-10-3/1.1-1CH

1.3-10-5/5.7-10^

1.7-10^/3.1-10^

1996

IVO/VALTO

2.5-

1.7-

1.6-

8.4-

1.3-

1.2-

3.6-

2.7-

10-«/1.1-10-«

10-71.1-10-8

10-V2.1-10-9

10-9/1.1-10-8

10-4/2.1-10-4

10-3/2.4-10^

10-71.4-10-6

10-5/5.3-10-«

dose arising from stay on the shore is hundredfoldat Olkiluoto compared to Loviisa (Table III). Thecorresponding internal doses are equal at bothsites in case of the releases from the stack, but therelease into the sea causes again hundredfolddose at the Olkiluoto site compared to the Loviisasite. The same trend is seen also in the collectivedoses, releases into the sea cause remarkablylarger dose at Olkiluoto than at Loviisa.

Releases of particles are assessed for the mostpart to take place during maintenance workswhereas gaseous effluents are released smoothlydistributed around the year. At Olkiluoto themaintenance and refuelling revisions occurredduring May and June in 1995-1996. The releasesreported from the second quarter of the yearsindicate exactly about the largest releases of theparticles through the ventilation stack, but thereleases of noble gases and releases into the sea

are more smoothly divided around the year. AtLoviisa the revisions were carried out during thethird quarter of the year in 1995 and mainlyduring the third quarter of the year in 1996 butalso during October concerning the unit Lo2. Thereleases reported indicate that the releases ofparticles in 1995 are divided mostly to the thirdand fourth quarter of the year because then alsolarger maintenance and repair works were carriedout. The dominating releases of the particles in1996 took place during the third quarter of theyear. The releases of gaseous radionuclides aredistributed smoothly around the year.

The dose calculation models for normal opera-tion base on the assumption of uniform releaserate around the year. The dose models includefeatures which base on this assumption and there-fore the use of these models in the case of a shortduration release (e.g. during one week due to

11

STUK-YTO-TR 1 44

revision work) can result in incorrect results. Thiskind of short period has to be calculated separate-ly if e.g. during pasturing season iodine would bereleased as a consequence of maintenance work.Because refuelling and maintenance works oftentake place during summer season, the existingexposure calculations could be focused by moder-ate modifications to the models as a consequenceof the prevailing reporting principle in which thereleases and dilution factors are supplied everythird month. Nuclide specific release data recordsindicate essential differences between quarters ofthe year. An important result is ingestion dosefrom summer season. As it depends largely on thetime of deposition it would be important to calcu-late as precisely as possible.

During the last years releases at both powerplant sites have composed of about twenty nu-clides, which are fission and activation products.Concerning the most significant nuclides the re-lease of C-14 depends on the load factor of theplant as a result of which the release of C-14 willincrease when the power level is increased. Theprinciples of dose calculations concerning C-14 areadequately known. The release of cobalt-60, whichoriginates from construction materials includingcobalt, depends on the efficiency of the purifica-tion system. Due to ageing of the plants corrosionproducts tend to accumulate inside power plantsand thus limitation of releases of those productsrequires maintenance of effective purification sys-tems. In addition leaking fuel pins increase fissionproducts in the process.

In particular ingestion exposure pathways in-clude nuclide specific data. If also other nuclidesthan those normally used are calculated, the nu-clide specific transfer and enrichment factorsshould be evaluated.

3.5 Adequacy of dose calculation

In a local scale the behaviour of H-3 and C-14cannot be dealt with using traditional dispersionmodels due to biological behaviour of those nucli-des in the environment. Therefore evaluation ofthese nuclides should be done by using models,which include the specific nature of the nuclides.

Because of their long half-lives Kr-85,1-129, C-14 and H-3 can cause large collective doses due to

global dispersion although individual doses arelow. In this case a local dispersion model is notenough but some simplified assumptions can beused to facilitate handling of global dispersion.Kr-85 can be assumed to be evenly distributed inthe troposphere of the Northern Hemisphere fromwhere it flows gradually to the Southern Hemi-sphere. Three other radionuclides are dispersedbetween atmosphere and biosphere and the be-haviour of these dispersion mechanisms can belargely simplified in the calculation of global dose.

Resuspension models have not been consideredbecause the significance of this phenomenon isregarded as insignificant compared to other expo-sure pathways. The inhalation dose caused byresuspension is orders of magnitudes less thanthe dose inhaled directly from the cloud.

Doses are generally calculated for adults only.Through internal exposure children, especiallyinfants, can be exposed to larger individual dosesthan adults can. Because the amount of infantsamong population is small, about 0.5 %, thisfeature has significance only in the critical groupbut not in the collective dose. Besides environ-mental concentrations, factors having an effect onthe individual doses include: nuclide specific in-gestion dose factors and consumption rates foringestion pathways, nuclide specific inhalationdose factors and inhalation rate for the inhalationpathway. The exposure pathway dependent dosefactors of one-year-old child are at the most anorder of magnitude larger than those of adults,even though commonly the values are only with afactor of 2 to 5 larger. Because the inhalation rateof an adult is six times as much as that of one yearold child, the inhalation dose of an one year oldchild cannot be much more than that of an adult.In total one year old child ingests food about onlyone fourth of that consumed by an adult, but dairyproduces even more than an adult does. If milk isassumed to be from cow, the ingestion dose frommilk can be larger for a child than for an adult ifiodine is the dominating component. If iodine isalso dominating in inhalation exposure similarconclusion can be drawn also for that exposurepathway. But if there is also exposure caused forexample by caesium, the order of significance canbe changed since the dose factors of caesium arelarger for adults than for infants.

12

S T U K - Y T O - T R 1 4 4

3.6 Adequacy of documents ofpower companies concerningdemands in YVL-guides

The reporting done by power companies concer-ning the dose calculations for the releases due tonormal operation shall be carried out according tothe principles presented in the YVL-guides. Theimplementation of the recently revised guides isallowed to be realised with a short delay. The gui-des concerning assessment of dispersion of relea-ses and dose calculations have been published in

revised form in January 1997. Based on the evalu-ation of the documents available, it can in generalbe estimated that if some phenomenon is not in-cluded in analyses, the potential effect can in mostcases be dealt with employing a conservative ap-proach. However some points which are not consi-dered can be found. For instance concerning at-mospheric dispersion: management of missing me-asurement data (YVL 7.3 chapter 4) and plumereflection from the ground, effect of inversion andthermal layers on concentrations are not clarified(YVL 7.3 chapter 2.1).

13

STUK-YTO-TR 1 44

4 VALTO PROGRAM OF STUK

The atmospheric dispersion part of VALTO doesn'thave an independent dispersion calculation modu-le, but it exploits directly the dilution factor re-cords provided by the power companies. Dry depo-sition is calculated, but the depletion of the plumeduring dispersion via radioactive decay, dry andwet deposition is not considered. Build-up ofdaughter nuclides is not considered. Radioactivedecay and other possible decay mechanisms aretaken into account via an additional decay term onthe ground concentration calculation.

Concentrations in the foodstuffs are calculateddirectly using transfer factors. The extra decayterm representing other removal processes can beadded to radioactive decay term. Ingestion dosecalculation is based on a straightforward approachemploying transfer coefficients, where no delaybefore consumption is assumed. The integrationtimes extend to infinity. Only cow milk and greenvegetables are considered as ingestion exposurepathways.

The calculation of doses due to releases intothe sea is carried out by assuming one compart-ment, into which the release is mixed and is thendiluted due to radioactive decay and water move-ment out of the compartment. The two exposurepathways considered include external exposure ofthe stay on the shore and ingestion of seafish.Concentration in fish is calculated employing atransfer factor from water to fish and contamina-tion of shore with a transfer factor from water toshore, which is calculated through the transfervelocity of the sedimentative substance from wa-ter to shore and based on the concentration ofsolid substance in water and a nuclide specificdistribution coefficient.

In Table III the results obtained by the powercompanies are compared with the results calculat-ed by VALTO. In the case of individual doses due

to stack release VALTO indicates equal or anorder of magnitude minor values than power com-panies but in the case of waterborne releasesVALTO shows only some larger values than themodels of the power companies. Differences ininternal doses are larger than in external doses.In the case of collective doses differences due tostack releases are similar as for individual doses.In the case of waterborne releases VALTO givesvalues even order of magnitude larger than powercompanies.

VALTO is a rather simplified estimation pro-gram intended for the review of the dose calcula-tions done by power companies due to normalreleases. With VALTO it is possible to roughlycalculate the magnitude of doses, but as Table IIIindicates the result obtained by VALTO can beeither smaller or larger than the result obtainedby power companies. It should be noticed as wellthat VALTO doesn't take into account C-14 andnuclides with short half-lives e.g. noble gases.

When the most significant nuclides in 1996(1995 values in parenthesis) are compared, it canbe observed that at Olkiluoto Co-60 causes about41 % (48) and C-14 about 34 % (23) of the totaldose, whereas at Loviisa the corresponding valuesare 7 % (not found) and 79 % (68) respectively. Therelease of C-14 is a little larger at Olkiluoto thanat Loviisa. At Olkiluoto the release of Co-60 intothe sea was 2 to 3 orders of magnitude larger thanto the atmosphere. But at Loviisa the release ofCo-60 into the sea, which is only a little smallerthan the release to the atmosphere, is 2 to 3orders of magnitude smaller than at Olkiluoto.Co-60 is an activation product, which is partlyaccumulated from the purification system of reac-tor water into the tanks, from where it is dis-charged into the sea deliberately. In 1995 and1996 discharges of the stored effluents into the

14

STUK-YTO-TR 1 44

sea were not carried out at Hastholmen. There is discharges are released into the cove of Hasthol-a significant difference between Olkiluoto and men and the coolant inflow occurs from the Gulf ofHastholmen sites concerning the water move- Finland.ments in the sea environment. In Olkiluoto the Definition of the most important nuclides indischarge of wastewater occurs with the outflow of VALTO is hampered because nuclide C-14 is notcoolant water into the open Gulf of Bothnia and considered. However, e.g. in the case of Olkiluoto,the inflow of the coolant water occurs from the where Co-60 causes almost 50 % of the totalsmaller cove. In Hastholmen the inflow and out- individual dose, Co-60 is the most significantflow take place, however, in a reverse manner as nuclide of individual dose also in VALTO.compared to the situation at Olkiluoto. Waste

15

S T U K - Y T O - T R 144

5 ASSESSMENT OF AN INDEPENDENTMODEL

Within the framework of this contracted work to STUKa European dose calculationmodel—PC CREAM—was also assessed. The development of this model has been fundedpartly by the EU Commission and carried out by NRPB in UK. Also IPSN-CEA inFrance and CSN, CEDEX and CIEMAT in Spain participated in the development work.The software system is intended to be used in the area of Europe and it is supported bythe extensive environment database EUROGRID, which contains population and agricul-tural production distributions. Grids are ranged to the 62nd latitude but the grid size ishere rather coarse and therefore it can hardly be employed in the Finnish environment.In marine part La. the Mediterranean and the Baltic Sea are modelled. With this softwa-re it is possible to calculate the health effects specified in the ICRP-60. PC CREAM is asuite of six programs that fully exploit graphical user interfaces.

5.1 Brief description of theatmospheric and marinemodels

In the atmospheric dispersion model a Gaussianapproach is used. It is possible to select the Pas-quill/Smith/Hosker model, in which the az para-meter is modified due to surface roughness. Thesecond possibility is to select the so-called Dourymodel, in which az parameter is determined bydispersion time dependent diffusion efficiency andfurther by a coefficient depending on the dispersi-on time. Recommended values are given for stabi-lity dependent boundary layer heights. Calculati-on of dilution factors includes an implicit assump-tion of wind speeds and rain is considered only inC and D stability classes.

Depletion of the plume is considered via radio-active decay, build-up of daughter nuclides, anddry and wet deposition. It is also possible tocalculate plume rise with Briggs-model and build-ing wake effects. External dose from the plume iscalculated with the finite plume model by numeri-cal integration for photons and with the semi-infinite plume model for electrons. Inhalation doseis calculated also from resuspended material. Nu-trition dose pathways are based on the dynamicFARMLAND model. Doses due to H-3 and C-14releases are calculated with separate models.Foodstuffs include milk, eggs, meat, liver, green

vegetables, grain, potatoes and fruits. As animalscow, sheep, goat, pig and chicken are considered,even though the metabolism of animal and pas-ture are considered dynamically only for cow andsheep. Transfer of radionuclides is modelled bymigration into soil, from soil to vegetables andfrom vegetables to feed stuffs of animals andfurther to animals.

The marine model contains a smaller reservoir,into which release is assumed to occur and alarger reservoir surrounding the smaller one inthe compartment modelling approach. Watermovement and transfer of sedimented material isoccurring between the compartments. Inside thesmaller subvolume the distribution of radioactivematerial between solid and liquid phase and re-moval to sediment are considered. Transfer to fishand sediment are calculated on the basis of dilut-ed amount. Exposure pathways that can be con-sidered include i.a. pathways caused by the mate-rial deposited on the shore from seaspray: inhala-tion, resuspension, ingestion and external dose.Seafish is considered as direct exposure. Concen-tration on the shoreline is calculated from theconcentration of the upper sediment of the localcompartment exploiting some measurement re-sults. Inadvertent nutrition and inhalation of sandon the shore are considered as well. Transfer tothe sediment on the shore is calculated in adifferent manner compared with the models of the

16

S T U K - Y T O - T R 1 4 4

Table IV. Inter-comparison calculations from the Olkiluoto releases in 1995.

Exposure pathway

Cloudshine

Inhalation

Groundshine

Meat

Agricultural products

Milk+dairy produces

Green vegetables

Fish

Stay on the shore

Boating

Individual dose [Sv]WO/PC CREAM

2.1-10-74.0-10-9

3.1-10-9/4.6-10-9

2.1-10-73.5-10-11

1.5-10-71.0-10-9

3.5-10-79.7-10"8

5.6-10-8/4.5-10-8

9.5-10-76.0-10-9

1.5-10-72.9-10-7

2.3-10-7/3.3-10-«

8.6-10-^AO-IO-8

Collective dose [manSv]TVO/PC CREAM

4.6-10^/4.2-10^

5.2-10-^/2.2-10^

1.7-1076.0-10"6

7.3-10-74.1-10"4

9.0-10^/3.0-10"2

1.1-10-75.6-10-3

1.9-10^/1.7-10^

2.5-10^/1.8-1CH (total)

1.8-10-7—

1.3-10"7/—

Finnish power companies. Individual doses arecalculated and outputted separately for each nu-clide in age groups: infant, 10-year-old child andadult.

5.2 Performed comparative dosecalculations

In comparison calculations the Olkiluoto site andrelevant data were used. Annual hourly dispersi-on data is used to provide the dilution factor file inthe necessary form for PC CREAM. Dispersiondata are divided into sectors of 30 degrees basedon frequencies of wind directions and stabilityclasses. In addition, rain frequencies in C and Dclasses are given. Wind speed is fixed for theheight of 10 meters. The wind speeds of the realrelease altitude are calculated by the model usingpower law equation. In PC CREAM the windspeeds are at the highest values in stability clas-ses C and D but measurements of Olkiluoto indi-cate that highest wind speeds appear in the moststable stability classes. This feature is apparentlycaused by the stability classification approachused by TVO and it is not found in Loviisa.

The PC CREAM calculations were done withthe release data of 1995. Table IV depicts resultsof individual and collective doses. Input parame-ters were defined as much equal as possible withthe values used in the TVO models. The marinemodel of PC CREAM includes parameters thatwere left in their initial values due to lack ofbetter information.

The collective dose from the discharges into thesea is outputted only as total dose, which is here

shown as a dose from fish. In this case thepopulation exposed or the amount of ingestion arenot explicitly expressed.

5.3 Assessment

When applicability of PC CREAM for use in STUKis assessed a general impression is that the soft-ware could supplement and improve calculationcapabilities to some degree. Calculations could bestarted from the hourly dispersion condition datato calculate dilution factors. Plume depletionthrough radioactive decay and build-up of daugh-ter nuclides, dry and wet deposition would be ta-ken into account. Parameters of the dose calculati-on models could be adjusted to some degree. Usercould for example build an advanced exposure pat-hway for soil-grass-cow-milk.

Doses and concentrations in the output filesare given for the receptor points, which are en-tered via the data entry screens of PC CREAM. Itshould be recognised that sector 1 begins 15degrees east of north. If we are interested in dosesin sector 3 we need our receptor point to bebetween 75 and 105 degrees relative to the refer-ence stack.

Some exposure pathways e.g. resuspension arepeculiar to windy, open areas in southern Europe.Modelling of discharges into the sea and corre-sponding dose models are done with little differ-ent approaches compared to the practice in Fin-land. About 40 parametrisised compartments areavailable in the modelling. Exposure pathwayscaused by seaspray have an emphasised impor-tance, although these pathways are mostly impor-

17

S T U K - Y T O - T R 1 4 4

tant only on the shores of large open seas. The eters is needed once. Afterwards the site-specificparameters affecting population doses remain in- environmental parameters are fixed, calculationsvisible to the model user in marine exposures, can be completed without changes in those param-Release duration of less than one year cannot be eters. Then only hourly dispersion data andcalculated. amounts of releases as well as possibly some input

PC CREAM seems to be applicable above all to parameters representing the operation of theresearch work. If it is used for the checking of the plant are changed. In practice the site-specificresults calculated by power companies (as e.g. parameters can be saved for reuse to the filesVALTO), introduction of customised input param- formed by the user interfaces of PC CREAM.

18

STUK-YTO-TR144

6 DEFINITION OF THE MODEL FOR SAFETYAUTHORITIES APPLICABLE FORCALCULATION OF NORMAL RELEASES

Requirements and specifications of the methods tobe used for assessment of doses for the populationin the environment of a nuclear power plant areintroduced in the chapters 2, 3.1 and 4 of the Gui-de YVL 7.2 and in the chapters 2, 3.1 and 4 of theGuide YVL 7.3. Should the accuracy and quality ofVALTO be improved, the particular phenomena indose calculation models should be taken into ac-count as well as possible. Availability of the hourlydispersion data is also necessary if consequencesof releases discharged during the certain time pe-riod are assessed (e.g. revisions, maintenance andrepair works). The VALTO software should beimproved in the following points:• processing of hourly atmospheric dispersion

data provided by the power companies,• definition of comprehensive dispersion model

for calculation of dilution factors,• calculation of plume depletion due to

radioactive decay and build-up of daughternuclides, dry and wet deposition,

• calculation of the amount of deposition via dryand wet deposition,

• calculation of radioactive decay and build-up ofdaughter nuclides in soil,

• supplement exposure pathways,• modelling of ingestion dose pathways in more

detail.Calculation for the critical group should be

enabled employing the locally existing parame-ters. In the case of population dose it should bepossible to consider the whole area of Finland aswell as the whole globe concerning the nuclides H-3, C-14,1-129 and Kr-85.

The model should be flexible such that it couldbe easily supplemented and expanded in the fu-ture. The user interface should be user friendlye.g. of the Windows type. An advantage of a self-made model is that both the good and bad proper-ties are known that cannot be said about modelsof foreign origin.

19

STUK-YTO-TR 144

7 CONCLUSIONS

This assessment indicates that there are differen-ces and deficiencies in the dose calculation met-hods intended for releases due to normal operati-on. However, those features are not estimated toconstitute especially remarkable effect on the re-sults of calculations. The existing models can alsobe used for calculations of exceptional releases oflong duration, if the nuclides are modelled and thevalues of the nuclide specific parameters are avai-lable. Because atmospheric hourly dispersion dataand discharge data in one week's periods are avai-lable, the input data for short-term releases exists,but the models investigated here are not plannedfor this kind of short-term evaluations. In particu-lar short duration releases during pasturing sea-son would need more comprehensive modelling ofingestion exposure pathways. On the other handthere exist other models used for the radiologicalimpact assessments in the case of accidents. The-se models could be exploited for assessing off-siteconsequences of short-term routine releases as

well.Based on the advanced knowledge and existing

models it would be possible to construct for safetyauthorities a comprehensive model stemmingfrom the guidelines of STUK. Flexibility of theuser interface and extent of input parameterscould be customised according to the require-ments of end users. As there appear to be differ-ences both in modelling and parameters, the im-portance of differences could be evaluated later inthe inter-comparison calculations when the newdiversified model of safety authorities is complet-ed. Then the same input data would be used andthe model dependent parameters would be as-signed to represent similar values in participatingprograms. As a preliminary level the hourly at-mospheric dispersion data, release of nuclides andinternal static and adjustable parameters shouldbe predefined. Inter-comparison should be carriedout by using extensive intermediate nuclide spe-cific output formats.

20

S T U K - Y T O - T R 1 4 4

REFERENCES

Paivi Makinen, Loviisa 1, Loviisa 2, Environmen-tal radiation safety report for the year 1996.Imatran Voima Ltd, 10.4.1997. (In Finnish)

Paivi Makinen, Loviisa 1, Loviisa 2, Environmen-tal radiation safety report for the year 1995.Imatran Voima Ltd, 10.4.1996. (In Finnish)

Paivi Makinen, Calculation model for offsite dos-es—DOSES 92, DY1-G710-0025. Imatran VoimaLtd, 12.3.1992. (In Finnish)

Eero Schultz, Tapio Vahamaa, Annual environ-mental radiation control report for the year 1996,O-TR-M-50/97. Teollisuuden Voima Ltd, 10.4.1997.(In Finnish)

Eero Schultz, Tapio Vahamaa, Annual environ-mental radiation control report for the year 1995,O-TR-M-34/96. Teollisuuden Voima Ltd, 9.4.1996.(In Finnish)

Eero Schultz, TVO I/II—environmental dose cal-culation models, O-1/6/73. Teollisuuden VoimaLtd, 6.3.1992. (In Finnish)

Lauri Pbllanen, Calculation of the environmentaldoses at the Olkiluoto site, 3-KS-M-127/3. Teolli-suuden Voima Ltd, 26.4.1984. (In Finnish)

Lauri Pollanen, Calculation of environmental dos-es from liquid radioactive effluents with the MARIprogram, O-VS-M-27/83/2. Teollisuuden VoimaLtd, 26.1.1984. (In Finnish)

Sauli Pusa, Predicted radiation doses for the years1984—1989 from the releases at the Loviisa andOlkiluoto nuclear power plants obtained by theVALTO calculation model. STUK-B-VALO, Radia-tion and Nuclear Safety Authority (STUK) 1991.(In Finnish)

Guide YVL 7.2, Assessment of the collective dosein the environment of the nuclear power plant.Radiation and Nuclear Safety Authority (STUK)1997. (In Finnish)

Guide YVL 7.3, Assessment of the dispersion ofradioactive releases from nuclear power plant.Radiation and Nuclear Safety Authority (STUK)1997. (In Finnish)

Guide YVL 7.8, Environmental radiation safetyreporting of nuclear power plant. Radiation andNuclear Safety Authority (STUK) 1995. (In Finn-ish)

Radiation Protection 72, Methodology for assess-ing the radiological consequences of routine re-leases of radionuclides to the environment, EUR15760 EN. European Commission, DG XIII, 1995.

PC CREAM, EUR 17791 EN (NRPB-SR296) Na-tional Radiological Protection Board, 1997.

21