Analysis of the neutron albedo’s influence on TIP deviations in Forsmark 1

(1)

TVE 16005 April

Examensarbete 15 hp April 2016

Analysis of the neutron albedo’s influence on TIP deviations in Forsmark 1

Moa Skan

(2)

Teknisk- naturvetenskaplig fakultet UTH-enheten

Besöksadress:

Ångströmlaboratoriet Lägerhyddsvägen 1 Hus 4, Plan 0

Postadress:

Box 536 751 21 Uppsala

Telefon:

018 – 471 30 03

Telefax:

018 – 471 30 00

Hemsida:

http://www.teknat.uu.se/student

Abstract

Analysis of the neutron albedo’s influence on TIP deviations in Forsmark 1

Moa Skan

In the operation of a nuclear reactor, detailed calculations are performed to predict the power distribution in the core and to ensure the maintenance of a number of safety margins. Repeatedly during a power cycle, the predictions are verified in so-called TIP measurements, and the consequences of possible deviations are investigated.

During the latest power cycles until today, the Forsmark 1 reactor has experienced large deviations in TIP measurements at the edge of the reactor core that have caused reasons to question the limits of the thermal margins needed to ensure fuel safety.

These deviations have especially been present in the first half of cycle 34, which is why only data from this period is investigated in this report. It is expected that the

deviations between predictions and measurements occur due to approximate calculations of the neutron albedo, i.e. the model for how neutrons leaving the core's periphery are reflected by materials surrounding the core.

The objective of this work was to investigate whether the albedo model in the nuclear reactor core calculation programme POLCA7 may be adjusted to reduce deviations. If so, an additional objective would be to give details on the identified adjustments and, if possible, propose a better albedo model that in the future could be implemented on-line at Forsmark 1.

Two analyses are presented, power deviation analysis and CPR (Critical PowerRatio) margin analysis. The presented investigations are based on existing albedo model commands ("cards") in the POLCA7 code, which have been used to introduce alternative albedo models in the calculations. Comparisons between updated and original albedo model cards were done with pre-written analysis scripts in MATLAB that calculated and plotted power deviation and thermal margins.

This approach resulted in one set of proposed albedo parameters with respect to deviations in the power distribution and another set of proposed albedo parameters with respect to CPR margin. The two sets are similar, but not identical. More research in the subject is necessary since the two proposed parameter sets do not coincide and since there still is some deviation between measurements and predictions when applying these adjustments.

In conclusion, the implemented albedo models give improved agreement between simulated and measured data. The results indicate that better predictions for future cycles may be obtained if these model improvements are introduced into future calculations. However, additional studies are required in order to draw conclusions that apply to the general case.

Handledare: Daniel Larsson

(3)

Acknowledgements

Thanks to everybody at the physics department at Forsmark that helped me with everything from understanding the subject to showing me around at the Forsmark power plant. A special thank you goes to my supervisor, Daniel Larsson, for being patient and for answering all my questions during my time at Forsmark.

I would also like to say a special thank you to my subject reviewer, Staffan Jacobsson Sv¨ard, for helping me straighten my thoughts and inspire me to write a better report.

Finally, I would like to say thank you to all of my friends and family for putting up with me and my attempts at explaining my fascination to the subject of this thesis.

(4)

1 Introduction

1.1 Nuclear Power

A large part of the produced electricity in Sweden comes from nuclear power. Sweden currently has three active NPPs (Nuclear Power Plant) of which one is the Forsmark NPP. The electricity produced in NPPs plays an important role in the total energy supply in Sweden as it makes up about 30% of it, according to Energy in Sweden 2015 [1].

The power plant in Forsmark consists of three BWRs (Boiling Water-Reactors) where unit 1 and 2 are of the same design and unit 3 is a bit larger. Unit 1 and 2 are permitted to operate at a thermal power of 2928 MW respectively 3253 MW, Unit 3 at 3300 MW. The reactors were comissioned between 1980 and 1985, which makes Forsmark the newest NPP in the country.

1.2 Energy production from nuclear fission

The energy production that takes place in a nuclear reactor is based upon a self-sustaining chain reaction of fissons, the physical process of splitting atomic nuclei. There exist forces inside the nucleus of all atoms and when these break in a heavy nucleus so that it fissions, energy and 2-3 neutrons are released. The emitted neutrons have a range of different kinetic energies, which is important to know since it affects for example what elements are able to absorb the neutron. The fission process can be induced by making a heavy atom absorb a neutron, as shown in Fig (1).

In a BWR, energy is released from the fission process of the fuel atoms and transferred into water surrounding the fuel rods, thus heating it and causing it to evaporate. The evaporated steam drives a turbine that is connected to a generator that transforms the energy from kinetic to electric.

Figure 1: The Uranium nucleus fissions when it absorbs a neutron, causing a release of fission products, energy and neutrons. Figure from [2]

(6)

In order to maintain the fission chain reaction and thus keep the energy production running, an important criterion must be fulfilled - at least one of the neutrons released in a fission must spawn a new fission reaction.

The multiplication factor k describes whether the power in the reactor core is stable or not

k<1 decreasing power level k=1 stable power level k>1 increasing power level

where a nuclear reactor at a stable state is equivalent to a stable fission rate. However, the multiplication factor and thus the power level inside the reactor are affected by a multitude of parameters that may vary. The relationship between the power level and the affecting factors is complex and accordingly monitoring of the power distribution in the core is necessary.

1.3 Power in the core

The Forsmark reactors are operated in one-year cycles, where ahead of each cycle the oldest fuel assemblies are exchanged with fresh ones. The power distribution during the whole cycle is predicted using core simulators, in the case of Forsmark, the POLCA7 and CASMO codes are used. More information on the core calculations is given in section 3.

The power level inside a nuclear reactor core is monitored using stationary in-core detectors, LPRM (Local Power Range Monitor). Figure (2) shows how the LPRMs are placed throughout the core.

With regular intervals, the LPRMs are calibrated using TIP (Traversing In-core Probe) detectors.

The calibration can be done automatically or manually and results in adjustments of calibration factors in POLCA7/CASMO.

The TIP detectors also give more detailed information from the core than the LPRMs, which is used to validate and update the POLCA7 predictions. More descriptions on how measurements are made inside a nuclear reactor core are presented in section 4.

(7)

Figure 2: An example reactor core illustrated schematically from above. Stationary in-core LPRM detectors are marked in red respectively in yellow. The fuel bundles are placed in groups of four around control rods, marked as black in the figure. Forsmark 1 is designed to contain 676 fuel assemblies and 161 control rods.

Figure from [3].

Large deviations between measurements and predictions in the periphery of the Forsmark 1 reactor (F1) have been observed, especially during BOC and the first half of cycle 34 (c34). This behaviour leads to decreased thermal margins (safety margins during operation of the NPP), which in turn may lead to an override of the designed upper temperature limit. This affects the operation of the NPP since it then has to be designed with larger thermal margins. Larger thermal margins makes it difficult to load the core and the fuel costs may rise.

1.4 Scope of this work

The purpose of this work was to investigate the TIP measurement deviation as a result of the current POLCA7/CASMO albedo (neutron reflection) model. Altering the albedo model is an easy way to study the effect of the neutron leakage on the TIP measurement, since it governs the calculated neutron leakage of the core. This way, the power distribution at the margin of the reactor core can be altered in the calculations, which may reduce future UPDAT/POLCA7 deviations.

Through variation of the albedo model the power distribution has, as well as one of the thermal margins, been observed and analysed in this work. The results are presented in section 6 and may

(8)

give an idea of whether a poor albedo model in POLCA results in the large deviations encountered between TIP measurements and predictions in Forsmark 1. Studies of how the albedo model affects the power distribution may generate knowledge that in the future can contribute enhanced prediction capabilities, and in the longer run to a more effective core design and thus a more optimised use of fuel in F1 and F2. All calculations are made on F1 in c34.

2 Reactor and fuel design

2.1 The Boiling Water Reactor

There are several different models of reactors all over the world, and all of the units in Forsmark are so called BWRs (Boiling Water-Reactor). The main characteristic of the BWR is that the steam used to drive the turbine is evaporated reactor water. The pressure is usually around 7 Mpa and the temperature 286^◦C. Unlike the BWR, the PWR (Pressurized Water Reactor) has a secondary system where a heat exchanger produces the steam used in the turbine. The reactor tank also has a higher pressure in a PWR (around 14.5 Mpa) than in a BWR.

A reactor has an operation cycle of 6-24 months, where in Sweden it is common to use annual cycles.

The operation cycle consists of one outage period and one operational period. During the outage the reactor is shut down and loaded with fresh fuel. Other important maintenance projects take place during this period as well, and this usually takes around a couple of weeks. The operational period is the time when electric energy is produced.

A reactor consists of several systems that work together in order to produce electric energy. Be- cause of the complexity of a BWR, only the most essential details will be mentioned in this report, besides the most relevant theory for this project.

2.2 BWR fuel and core design

The fuel used in a NPP contains uranium and plutonium that fissions inside the reactor core. In Fors- mark, the fresh fuel that is loaded at BOC (Beginning Of Cycle) is made of uraniumoxide enriched to 3-4% with respect to the fissile isotope ²³⁵U (the naturally occurring fraction of ²³⁵U is only 0.7%).

Figure (3) shows how the uraniumoxide, shaped like pellets, are placed in assemblies that later are loaded into the reactor core.

Another fissile nucleus, ²³⁹P u is produced inside the reactor through the absorption of neutrons in

238U , followed by beta decays. At BOC the²³⁹P u makes up 20-30% of all of the fissions while at the EOC (End Of Cycle) it makes up about 30-40%.

In the fission process, splitting of heavy atoms results in the production of lighter elements and the release of a large amount of energy because of the breaking of intranuclear bonds is used in order to heat the reactor water, and further to drive the turbine resulting in power production. When designing the core of the nuclear reactor, it is of large importance to know the physical limits of it.

The design should be made using thermal margins that act like safety parameters during operational

(9)

Figure 3: Uranium dioxide pellets in rods are placed in assemblies that further are placed inside the reactor core as fuel. Figure from [4].

periods. CPR (Critical Power Ratio) is one of the thermal margins relevant to this work and it is usually one of the limiting factors of the operation of Forsmark 1 and 2. The CPR margin describes the dryout-margin of the reactor fuel, i.e. how much the power in a fuel rod may be increased before it risks over-heating. The CPR margin is heavily dependent on the power distribution and the amount of water in the core. Since it is dependent on the power distribution it is important to base the calculation of CPR margins on a good prediction model. Diverse approximations may give deviations in respect to each other. However, if the CPR margins conform for various approximations, the reliability of the calculations is strengthened.

3 Core calculations

In order to calculate different properties of the reactor, the core simulator programme package POLCA7/CASMO is used in Forsmark. The calculations made in POLCA7/CASMO are used to analyse for example

power distribution, xenon transients and track depletion of fuel bundles and control rods. Using POLCA7/CASMO the design of the core is made, a task for which this code package is suitable since the properties of the core can be predicted for an entire operation cycle, until fuel discharge is reached.

POLCA7 is also used in core design to verify the thermal margins and other design criteria, like maximum discharge, shutdown margin etc. Calculations using POLCA7 are also made online during operation in order to verify that the margins are maintained in every operational point.

CASMO is a fuel modelling code that is used together with POLCA7 to calculate physical properties of assembly cross sections throughout the reactor core. As it is today, the CASMO model of the albedo is most likely inadequate since it models the reflector (the material outside the periphery of the core, which may cause neutrons escaping the core to be reflected back into the core again) over-simplified as two rows of water. The reflector model is strongly related to the albedo.

(10)

Water Fuel assemblies

Figure 4: The difference in amount of water between the fuel assemblies and the reactor core periphery, portrayed in a small example core.

In reality the reflector is more complicated than just two rows of water and it varies at different places in the core. Firstly, the reactor tank is cylindrical while the fuel is quadratic, so the distance between the fuel periphery to the reactor tank varies. See Figure (4). Secondly, the reflector is not only made of water, but the core is surrounded by e.g. steel in the moderator tank. It is comprehen- sible that the model is in need of some adjustments because of the differences in the actual and the modelled reflectors.

4 Measurements using LPRM and TIP

In order to monitor the power in the nuclear reactor core, two measurement systems are used. The LPRM detectors consists of a number of stationary power detectors placed throughout the core, as shown in Fig. (2). In F1 and F2 there are 36 channels that each have four axially placed probes.

The LPRMs give continuous information about the powerlevel inside the core. Since the LPRMs are receiving a considerable amount of neutron radiation dose, they age and need to be calibrated every fourth week. Accordingly, mobile in-core TIP detectors are used in order to calibrate the LPRM detectors and to check the validity of the on-line POLCA7/CASMO core calculations.The TIPs are inserted into the core in tubes next to the LPRM detectors.

Since one of the main tasks of the TIP system is to calibrate the LPRMs, all of the TIP channels are placed side by side of the LPRMs. A TIP-measurement is done by having two TIP-detectors

(11)

move axially through the TIP channels. The two detectors are then calibrated with respect to each other in order to give a reliable measured value of the power at each axial location.

Figure 5: Cross sections of a LPRM detector (left) and a TIP detector (right). Both detector types have a central, gas-filled chamber in which ionising radiation causes electric discharges that form the basis for measurement. Figure from [5]

A TIP-detector measures the local power indirectly by measuring the gamma flux at a point in the core. The gamma flux is proportional to the neutron flux that is, in turn, proportional to the power.

Earlier a neutron sensitive TIP detector was used, but the gamma detector was proven to be more sensitive for a larger volume of the core and therefore prioritised. A gamma sensitive TIP-detector is a cylindrical, gas-filled chamber that uses an applied difference in energy potential to create electric signals when being subject to irradiation, see Figure (5) . Electrons can be emitted from the metallic cover of the detector when gamma photons interact with it. Some of these electrons ionize the gas and this creates a current.

In order to make adjustments to the POLCA7/CASMO-calculated power distribution in the core, a correction factor TIPCOR with a value for each TIP detector is calculated. The TIPCOR factor uses data from both POLCA and the latest TIP measurement. TIPCOR is then used to calculate the adjust-

(12)

ment of the base power distribution, which is applied as a coefficient to the LPRM measurement. The latest TIPCOR correction factor is applied automatically to the LPRM when in on-line operation.

However, automatic TIPCOR correction has not been used in this study since the albedo adjustments in the analysis have been done manually.

5 Analysis

Three sets of TIP measurements that were performed during the first half of c34 have been analysed in detail in this work. The dates for the TIP measurements were 140519, 140813 and 141230. In the analyses, the core has been simulated using POLCA7/CASMO, while varying the albedo model and comparing to the results of the processed TIP data. The goal of the analysis has been to minimise the deviation between modelled and measured data, and thus to identify where the simulated albedo is in need of an improved model. Ultimately, the goal is to identify the best possible albedo model within the POLCA7/CASMO framework.

5.1 Albedo parameters varied

In order to enable adjustments of the reactor-core calculations, POLCA7/CASMO uses a set of commands or geometry model cards, that can be accessed by the user. These cards contain information about the physical properties of the core. The albedo model in POLCA7/CASMO has parameter components that can be adjusted individually. The card components which have been varied in this work are presented in short below.

• ASID - information about the radial albedo. The ASID parameters describe the behaviour of neutrons at a) a plane surface, b) inner corner and c) outer corner as shown in Fig (6).

• ABOT - information about the lower axial albedo. The ABOT parameters describe the behaviour of the neutrons at the bottom of the core.

• ATOP - Information about the upper axial albedo. The ATOP parameters describe the behaviour of the neutrons at the top of the core.

Each card has a number of parameters that describe the leakage, scattering and reflection of neutrons at high respectively low kinetic energies. In this study, individual parameter components were varied in order to investigate how they affected the power in the adjusted simulation. ASID, ABOT and ATOP were all varied this way, since they describe the neutron reflection at different places of the core’s periphery.

5.2 Analysis procedure

Two types of analyses have been performed in this work; power deviation analysis and CPR deviation analysis. Both analyses are based on a difference in power between an unaltered processed

(13)

a b c

Figure 6: Concepted illustration of how the set of ASID albedo parameters model the neutron reflection.

Portrayed are the fast radial albedo at a) a plane surface, b) an outer corner, c) an inner corner. Figure from [6].

measurement and an analysis with altered albedo parameters. As described above, three TIP measurements from c34 have been processed through POLCA7/CASMO from dates 140519, 140813, 141230.

The method workflow in both analyses was carried out in the following order:

• Form an original hypothesis of how to set the albedo parameters, based on past results and knowledge (after consultation with supervisor).

• Alter all parameters with equal percentage in order to see the general trend of the changes inferred. In section 6, this type of setting is noted by ”(all)” and by what percentage it is altered.

• Alter each parameter individually to study the result’s dependence on each.

• Depending on what results that have been obtained up to this point, an optimised parameter adjustment is formed. In section 6, the optimised setting is noted by what parameter is altered, e.g ”(ASID11)” and by what percentage it is altered.

It should be noted that the explicit radial reflector in POLCA7/CASMO was disabled in this study. Nor- mally it is added to make the neutron leakage more resistant to changes. However, here the purpose was to infer changes to the modelled leakage, and the explicit radial reflector was therefore disabled.

(14)

5.2.1 Power deviation analysis

The processed TIP measurements are here called BASE files, and the simulations of the power distribution using updated albedo adjustments are called UPDAT. One should note that the TIP measurement data is one of the inputs to UPDAT, implying that a copy of the measurements is also processed in UPDAT. Figure (7) shows the power deviation analysis flowchart.

Analysis flowchart - Power deviation TIP measurement

Copy of original TIP measurement

UPDAT: Simulation of reactor core’s power distribution with altered albedo parameters

Adjustment of albedo parameters

BASE: processing of TIP data to get the power distribution

Comparison of power in BASE and UPDAT

Albedo variation Direct processing of measurement data

Figure 7: Power deviation analysis flowchart.

In the analyses, ASID, ABOT and ATOP were altered in the UPDAT simulation and while doing that, observations could be made of how much a change in the parameter set of respective card affected the radial power. Through comparison with the BASE file, a power deviation could be calculated, giving a measure of to which extent UPDAT resembles the processed original measurement. Its minimisation is therefore used in this study as a way to optimise the albedo adjustment.

A MATLAB script was used to plot the results. An example of the output of this script is presented in Figure (8). The power deviations between the BASE and the UPDAT file were observed in matters of overall standard deviation, radial standard deviation and absolute value of the sum of radial deviations. The importance of each output item can be described accordingly to the following:

• Overall standard deviation is used in order to see how the core is overall affected by the parame-

(15)

ter adjustment. A change in the albedo may e.g. decrease the power deviation in the periphery, but increase it in another area of the core. The overall standard deviation is given in the MAT- LAB output, as presented in the uppermost right in Figure (8)

• Radial standard deviation is used to see how the core is affected radially by the albedo adjustment. Studying how the radial standard deviation changes is important since it is primarily through observation of the radial power distribution that conclusions about the albedo in the periphery can be made. The radial standard deviation can be seen at the uppermost right in Figure (8), below the overall standard deviation.

• Absolute value of the sum of radial deviations is used as a complement to the radial standard deviation. As seen in Fig (8) the power deviation in the TIP channels is plotted in a core map for every radial node. The sum of the absolute value of all radial power deviation components has been recorded since it may give complementary information of the deviation magnitude with respect to albedo parameter adjustments. Here, the summary was done manually, which enabled additional apprehension of what areas in the core that were affected the most by each parameter change.

(16)

Figure 8: The output of the power deviation analysis, plotted in MATLAB. The figure shows overall, radial (and axial, not used in this study) standard deviations, control rod map(upper left), radial power deviation in each TIP channel(upper right), axial power deviation(lower left) and axial power distribution(lower right).

5.2.2 CPR margin analysis

The CPR margin analysis was carried out in a similar manner as the power deviation analysis. See Figure (9) for a CPR margin deviation analysis flowchart. Just like in section 5.2.1, a BASE file and an UPDAT simulation was made. The power distributions of the BASE and the UPDAT files was used because of the CPR dependence on the reactor core power. Using a pre-written script in MATLAB, written for analysis of thermal margins, the CPR margins of both the UPDAT and the BASE file were calculated separately. The script also plots the value of the CPR margin at the node where the margin is the smallest, together with the node coordinates. The smallest CPR margin together with

(17)

its coordinate was calculated for both the BASE and the UPDAT file.

Analysis flowchart - CPR deviation TIP measurement

Copy of original TIP measurement

UPDAT: Simulation of reactor core’s power distribution with altered albedo parameters

BASE: processing of TIP data to get the power distribution

Calculation and comparison of CPR margins

Albedo variation Direct processing of measurement data

Adjustment of albedo parameters

Figure 9: CPR margin deviation analysis flowchart.

The reason why the CPR margin deviation is used to analyse the alternative albedo models in POLCA7/CASMO is because it provides an easy way to keep track of the change between the BASE and the UPDAT data. If the CPR margin has the same smallest value in the same radial node then it is, as previously stated, safe to assume that the power distributions in the BASE and the UPDAT files correspond to each other, which, in turn, indicates that the albedo model applied in UPDAT is adequate.

6 Results

The standard deviations of the power distribution and the CPR margin of the POLCA7/CASMO model file with unaltered albedo parameters is shown in Table (1). These values are used as reference when

(18)

the results of the analyses with alternative albedo models are presented in order to see how a parameter adjustment affect the results.

Table 1: Standard deviations of the power distribution and CPR margin generated from original albedo parameters.

TIP measurement Overall stdev Rad stdev CPR-marg

140519 4.1% 2.8% 4.955%

140813 3.9% 2.7% 5.021%

141230 3.6% 2.3% 5.055%

6.1 Deviations between calculated and measured power

The analysis was carried out as described in section 5.2.1.

A small deviation indicates a good albedo model, which then was of interest to investigate further.

Only some selected results are displayed in Tables (2)-(5), where the results with the smallest deviations are included.

Table 2: Overall standard power deviation of a selection of albedo parameter adjustments.

Adjusted albedo 140519 140813 141230

No adjustment (from Table 1) 4.1 3.9 3.6

ASID +10% (all) 3.8 3.5 3.4

ASID +20% (all) 3.8 3.3 3.3

ASID +30% (all) 3.9 4.1 3.9

ASID +20% (all)

ABOT +5% (all) 3.3 3.1 3.0

ASID +20% (ASP11, ASY11, ASID22), +25% (ASID 21, ASI11)

ABOT +5% (all) 3.3 3.3 3.0

ASID +20% (ASY11, ASID22), +30% (ASID 21), +40% (ASI11)

ABOT +20% (all) 3.3 3.1 3.0

ABOT +30% (ABOT11) 3.3 3.1 3.0

ASID +20% (ASY11, ASID22), +30% (ASID 21), +40% (ASI11) ABOT +20% (all)

ATOP +10% (all)

3.1 3.0 2.9

Table 2 shows that the result with the lowest overall power standard deviation is the ASID +20%

(19)

Table 3: Radial standard power deviation of a selection of albedo parameter adjustments.

Adjusted albedo 140519 140813 141230

No adjustment (from Table 1) 2.8 2.7 2.3

ASID +10% (all) 2.4 2.2 2.0

ASID +20% (all) 2.4 1.8 1.8

ASID +30% (all) 2.6 2.5 2.7

ASID +20% (all)

ABOT +5% (all) 2.0 1.8 1.8

ABOT +5% (all) 2.0 2.0 1.9

ABOT +20% (all) 2.0 1.8 1.9

ABOT +30% (ABOT11) 2.0 1.8 1.9

ASID +20% (ASY11, ASID22), +30% (ASID 21), +40% (ASI11) ABOT +20% (all)

ATOP +10% (all)

2.0 1.8 1.9

(ASY11, ASID22), +30% (ASID 21), +40% (ASI11), ABOT +20% (all), ATOP +10% (all) parameter adjustment. Note that the overall standard deviation exhibits the largest decrease by 1.0% relative to Table 1.

Table 3 shows that the result with the lowest radial standard deviation is the ASID +20% (all), ABOT +5% (all) parameter adjustment. This result is also the lowest for both the overall and the radial standard deviations. Note that the largest decrease here is 0.9% relative to Table 1.

These results shows us a smaller standard deviation relative to the original albedo model, which indicates a better albedo model than the original one. This decrease of standard deviation may not be exceedingly large, but it may show some important improvements for a full cycle simulation. The reader may gain some additional understanding of the result by regarding Table 4, where the absolute value of the total sum of the deviations inside the core is shown.

In comparison to the original albedo model, the largest decrease in the total sum presented in Table 4 is 65 (arbitrary units), and it is obtained in the ASID +20% (all), ABOT +5% (all) parameter adjustment, which is the same as the parameter adjustment with the largest decrease in Table 3. This is no surprise since both Table 3 and Table 4 shows the radial standard deviation.

(20)

Table 4: Sum of absolute values of the power deviation in all radial nodes (arbitrary units), for a selection of albedo parameter adjustments.

Adjusted albedo 140519 140813 141230 Total sum

No adjustment 79 77 63 219

ASID +10% (all) 64 62 51 177

ASID +20% (all) 64 50 57 171

ASID +30% (all) 69 78 88 235

ASID +20% (all)

ABOT +5% (all) 47 51 56 154

ASID +20% (ASP11, ASY11, ASID22), +25% (ASID 21, ASI11).

ABOT +5% (all).

57 51 56 164

ASID +20% (ASY11, ASID22), +30% (ASID 21),

+40% (ASI11).

ABOT +20% (all).

57 51 55 163

+40% (ASI11).

ABOT +30% (ABOT11).

57 51 55 163

+40% (ASI11).

ABOT +20% (all).

ATOP +10% (all).

55 51 55 161

6.2 CPR margin deviations

In the CPR margin deviation analysis, the same parameter adjustments as for the power deviation analysis was made. Only some selected parameter adjustments are shown in Table 5.

(21)

Table 5: CPR margin deviation of a selection of albedo parameter adjustments.

Adjusted albedo 140519 140813 141230

No adjustment from Table 1 4.955% 5.021% 5.055%

ASID +10% (all) 4.335% 4.351% 4.393%

ASID +20% (all) 4.335% 3.55% 3.501%

ASID +30% (all) -0.129% 2.341% -0.726%

ASID +20% (all)

ABOT +5% (all) 3.430% 3.535% 3.563%

ABOT +5% (all) 3.283% 3.396% 3.368%

Table 5 shows the lowest value for the parameter adjustmentASID +20% (ASP11, ASY11, ASID22), +25% (ASID 21, ASI11), ABOT +5% (all). The results are presented as the percentage of the difference between the absolute values of the CPR margins of UPDAT and the POLCA7 calculated original model, kCP RU P DATk − kCP R_{P OLCA}k. In Table 5 it can be seen that one of the results has a negative percentage. This parameter adjustment is rejected because of this - a negative CPR margin difference percentage means that the CPR margin is larger for the POLCA7 model than for the UPDAT model.

This might entail risks of overriding the design margins of the reactor, since the reactor core’s safety margins are designed from observations of a simulated core. Thus, kCP R_{U P DAT}k > kCP R_{P OLCA}k is desirable.

From these analyses, 3 resulting parameter adjustment sets have been identified with the least deviation from the measured result, which has been used as a reference. These results are fairly similar, but have some small differences.

The resulting model adjustment obtained from the overall standard power deviation analysis have in the sides of the reactor tank an increased reflection of 20% of neutrons with lower energies and high energy neutrons in the outer corners of the fuel elements. The scattering neutron reflection has been increased by 30% and the high energy neutrons in the inner corners of the fuel elements are modelled with an increased reflection of 40%. The reflection of all neutrons in the bottom of the reactor tank has been increased by 20% and in the top of the tank by 10%.

The resulting model from the radial standard power deviation analysis has for all neutrons in the sides of the reactor tank increased the reflection by 20%, and in the bottom of the tank by 5%.

The resulting model from the CPR margin deviation analysis has for the sides increased the reflection of neutrons with lower energies and higher energy neutrons in plane surface and inner corner of the fuel elements by 20%. Reflection of scattering neutrons and high energy neutrons in the inner corner of the fuel elements is increased by 25%. Reflection of all neutrons in the bottom of the reactor tank is increased by 5%.

(22)

These changes in the different suggested models may be physically plausible, since the current model is very simplified with only two rows of water. Since all of the three resulting models have many similarities, we can assume that there is some lack in the current albedo model and that a improvement of the model will give simulations of the core that are closer to reality.

7 Discussion and conclusions

7.1 Power deviation

Table (4) displays the total sum of the absolute values of the deviation between modelled and measured power in each radial node. As shown, the parameter adjustment ASID +20% (all), ABOT +5% (all) gives the lowest total sum, which indicates that this albedo parameter adjustment offers the lowest power deviation of the different parameter adjustments analysed. This albedo adjustment also gives the lowest radial standard power deviation relative to the other models and it is identified here as the recommended adjustment in terms of power deviation.

7.2 CPR margin deviation

Considering the CPR margin analysis, two factors are sought: that the updated simulation and the original BASE file give the same CPR margin, and that the node in which the CPR margin is smallest coincides for both the updated simulation and the original processing. These factors indicate good agreement between the power distributions of the BASE file (calculated data) and the UPDAT file (modelled data with parameters from measurements).

Regarding Table (5) it can be seen that the CPR margin for parameter adjustment ASID +20%

(ASP11, ASY11, ASID22), +25% (ASID 21, ASI11), ABOT +5% (all) gives the smallest deviation for all three TIP measurement dates as well as for the total discharge simulation. However, the nodes at where the margin is the smallest for the BASE respectively the UPDAT file do not coincide. Ac- cordingly, there is still a need for further analyses with respect to CPR margins.

7.3 Concluding remarks

Two albedo parameter adjustment sets with the smallest deviation for each analysis have been derived. These parameters have some things in common - that the ASID card, which describes the radial albedo should increase its parameters by at least 20%, and that the ABOT card, which describes the albedo in the bottom of the reactor tank, should do so with about 5%. For the CPR margin deviation analysis, a 5% increase in the ATOP card proved to give a smaller deviation as well.

When observing the similarities and differences of the two albedo adjustments one may conclude that the albedo model in CASMO is inadequate for these specific measurement dates of c34. The results imply that two rows of water do not provide a sufficient reflector model when it comes to radial neutron reflection. It is clear that an improvement in the simulated albedo is needed not only in the

(23)

sides but also in the bottom of the reactor tank, and, as indicated for the CPR margin deviation, in the top of the tank as well. A more carefully designed model that includes realistic materials, which reflect more neutrons in certain areas may be needed.

To confirm that the albedo parameters need adjustment, further studies are needed where these adjustments are applied on full cycles, for both for c34 and for other cycles. Firstly, these analyses have only been made on TIP measurements made on single days during the first half of c34, and it is important to see how the albedo parameter adjustments affect a full cycle to ensure that all thermal margins are correctly modelled and within limits during the entire operational period.

Secondly, one should note that the limits of the CPR margin are individually determined for every new cycle because they are dependent on how the reactor core is designed. A critical CPR margin value in one cycle may be within the margin limits in another. CPR margin limit fluctuation should be kept in mind when regarding other cycles than just c34, making it somewhat difficult to draw solid conclusions.

In conclusion, two albedo parameter adjustments meeting the requirements of a lower deviation between the BASE and UPDAT files have been derived from two analyses. Further research is needed in order to find and validate a new albedo model in POLCA7/CASMO.

References

[1] Energimyndigheten, Energil¨aget 2015.

[2] McMaster Nuclear Reactor,

Schematic representation of an induced nuclear fission event, viewed 2016-03-12.

https://mnr.mcmaster.ca/index.php/about/how-does-it-work/introduction [3] KSU, 2004, Kompendium Reaktorfysik H1

[4] United States Nuclear Regulatory Commission,

September 2015, Nuclear Fuel Assembly, viewed 2016-02-12. http://www.nrc.gov/reading- rm/doc-collections/fact-sheets/storage-spent-fuel.html

[5] Wimpee, O’neil, Ross, Chu,

23 november 1994, viewed 2016-02-13. http://www.google.com/patents/EP0403223B1?cl=en [6] Emma Isberg, private communication,

22 january 2016.

Analysis of the neutron albedo’s influence on TIP deviations in Forsmark 1

Examensarbete 15 hp April 2016

Analysis of the neutron albedo’s influence on TIP deviations in Forsmark 1

Moa Skan

Abstract

Analysis of the neutron albedo’s influence on TIP deviations in Forsmark 1

Contents

1 Introduction

2 Reactor and fuel design

3 Core calculations

4 Measurements using LPRM and TIP

5 Analysis

a b c

6 Results

7 Discussion and conclusions

References