Discrete Event System Based Pyroprocessing

Modeling and Simulation

Hyo Jik Lee, Won Il Ko, Sung Ki Kim, Seong Yeol Choi, Han Soo Lee,

Geun Il Park and In Tae Kim

Department of Nuclear Fuel Cycle System Development, Korea Atomic Energy Research Institute,

989-111 beon-gil Daedoekdaero, Yuseong, Daejeon, 305-353, Republic of Korea

Keywords: Discrete Event System, Pyroprocessing, Material Flow, Material Balance, Operation Model.

Abstract: The pyroprocessing operation-modelling is characterized as complicated batch type operation and tangled

material flow logic, and handling many numbers of chemical elements. Discrete event system modeling was

performed to build an integrated operation model, a simulation of which showed that a dynamic material

flow was implemented. All data related to a dynamic material flow were recorded in database tables, and

used for verification and validation in terms of material balance. Compared to equilibrium material balance,

dynamic mass balance showed that the amount of material transported upstream and downstream in the unit

process satisfied the mass balance equation at every batch operation. This study also showed that a dynamic

material flow, which is a basic framework for an integrated pyroprocessing simulator, was well working.

The integrated model built thus far will be improved in a few years toward an integrated simulator with

safeguards assessment, technical feasibility, and economic feasibility modules.

1 INTRODUCTION

The Korea Atomic Energy Research Institute

(KAERI) has been developing pyroprocessing

technologies, which can reduce the increasing

amount of spent nuclear fuel (SNF) and dramatically

decrease the disposal load, through recycling and

destroying toxic waste such as long-life fission

products in the SNF. Pyroprocessing technology has

not been fully demonstrated in terms of

commercialization and technology maturity. To

navigate the right direction of pyroprocessing

technology development, a demonstration in an

integrated facility is certainly a tangible solution, but

it is too costly and time consuming to construct a

fully integrated facility including all unit processing

and remote handling equipment. Actually, modelling

and simulation enhance an understanding of known

systems, provide qualitative and quantitative insight

and guidance for experimental work, and produce

quantitative results that replace difficult, dangerous,

or expensive experiments (DePaoli, 2011).

Therefore, a technology assessment and

breakthrough by modelling and simulation would be

preferable even in pyroprocessing technology

development. In this study, the main concern is to

build a consolidate framework able to describe the

material flow of an integrated pyroprocessing

facility and to build a model on that. This study is

on-going mid-term research to aim at a multi-

purpose integrated pyroprocessing simulator. As a

basic frame of the simulator, the material flow

modelling and mass balance management were

carefully designed and applied to the simulator.

Mass balancing model was studied about iron ore

terminal example by using mixed discrete and

continuous model (Béchard, 2013). In this study, a

discrete event based system (DES) appropriate to

build a model of a batch type process is applied to

the configuration of the pyroprocessing material

flow. Dynamic in-out material balance in the unit

process is managed in the database whenever events

according to the material flow occur. The progress

on the simulator was verified in terms of rigorous

implementation of operation logic and mass balance.

2 PYROPROCESSING

An integrated pyroprocess is under consideration to

590

Jik Lee H., Il Ko W., Ki Kim S., Yeol Choi S., Soo Lee H., Il Park G. and Tae Kim I..

Discrete Event System Based Pyroprocessing Modeling and Simulation.

DOI: 10.5220/0005010005900596

In Proceedings of the 11th International Conference on Informatics in Control, Automation and Robotics (ICINCO-2014), pages 590-596

ISBN: 978-989-758-039-0

 2014 SCITEPRESS (Science and Technology Publications, Lda.)

process the spent oxide fuel discharged from PWRs

and fabricate metallic fuel containing transuranic

(TRU) elements for a future sodium cooled fast

reactor (SFR) (Yoo et al., 2008). The process

includes head-end process, electrolytic reduction,

electro-refining, electro-winning, and a salt waste

treatment system.

The pyroprocessing also includes many complex

recycling flows. Since it almost consists of batch-

type processes even though some are more like

continuous processes, a discrete event system is

preferred to model it. A lot of effort has been put

into an investigation of principle (Song et al., 2010;

Lee et al., 2011). Since current experimental studies

focus on the unit process technology, and not an

integrated process, it is hard to predict the overall

behaviour and mutual influence. However, modeling

and simulation can make it possible to see

unforeseeable behaviour.

3 MODELING AND SIMULATION

3.1 Operation Model

3.1.1 Discrete Event and Hybrid System

The operation model in the pyroprocessing simulator

is located between the process model and facility

model (Lee et al., 2013). Although the process

model is involved in a pyrochemical reaction within

a batch, the operation model is engaged in states in

border of batch, that is, states driven by the start and

end of the batch operation. The state transition

driven by event in the operation model is what the

DES modeling describes the best. On the other hand,

a time-dependant state in the process model can be

well described in a continuous variable dynamic

system (CVDS). This is why the pyroprocessing

simulator is a hybrid system. In this paper, the

operation model is focused more than in the process

model because the process modelling and hybrid

system modelling was explained in a previous study

(Lee et al., 2013).

3.1.2 Operation Procedures

We do not have as much information on the

pyroprocessing operation as a real existing facility

operation. Therefore, gathering the operation

information is limited. However, baseline operation

procedures for a rational process operation can be

drawn by professionals involved in the process

development. Since pyroprocessing often has

recycling and complex operations, to build an

operation model to meet such requirements is not

that simple.

For example, the oxide reduction process (Karell

and Gourishankar, 2001; Herrmann et al., 2005; Hur

et al., 2008) includes three unit processes such as

electrolytic reduction (P2-1), cathode processing

(P2-2), and LiCl purification (W4-1) as shown in

Figure 1. P2-1 converts two types (pellet and

fragment) of oxide fuels into metal ones by

electrolytic reduction, P2-2 evaporates and recovers

entrained salt in a cathode product carried from P2-1,

and then recycles the recovered salts during the first

campaign (1st through 40th batch) but sends them to

W4-1 after the first campaign (41st batch ~ ). The

recovered salt is added in P2-1 every other batch

(3rd, 5th, 7th, … 39th batch) operation during the

first campaign. The recovered salt is regenerated in

W4-1 to be recycled to P2-1 from the third campaign

(81st batch ~). During the second campaign (41st

batch through 80th batch), P2-1 needs new salt to

compliment insufficient salt corresponding to

entrained salt accompanied by a cathode product,

and P2-2 holds the recovered salt until it reaches an

amount sufficient to feed W4-1. From the third

campaign, P2-1 receives the regenerated salt during

every other batch operation, and new fresh salt can

then be added if it is insufficient.

Figure 1: Material flow diagram for oxide reduction.

The above operation requirement is changed

according to the batch operation number.

Consequently, the material flow direction changes.

To reflect such complex flow change in a model, a

well-designed logic based model should be built.

3.1.3 Operation Modeling

Operation logic in 3.1.1 was implemented in an

ExtendSim

v8. Routing modeling is built based on

DES, as shown in Figure 2, as a feed material for

one batch is considered to be an item. The oxide

reduction model begins with the transport block that

P2‐1

Electrolytic

Reduction

P2‐2

Cathode

Processing

W4‐1

LiCl

Purification

Newsalt

(LiCl,Li

Reduced

metal

withsalt

RecoveredSalt

campaign)

RegeneratedSalt

(from3

campaign)

Reducedmetal

Pellet,

fragment

ConcentratedSalt

DiscreteEventSystemBasedPyroprocessingModelingandSimulation

591

Figure 2: Logic model of feed material receipt in P2-1.

Figure 3: Blocks needed for operation logic.

represents three types of feed material (SNF,

recovered salts, and regenerated salts) receipt. SNF

is always needed for every batch operation.

However, the recovered salts and regenerated salts

can be received or not according to the batch

operation number. Such routing logic is

implemented by an equation block and batch blocks,

as shown in Figure 3.

Equation block includes complex logic

behaviour describing the operation condition, for

example, whether the current batch operation

requires the addition of new, recovered, or

regenerated salt. If an addition is needed, a

corresponding batch quantity in the batch block is

set to TRUE. The pseudo-code for the equation

block is as follows:

// Batch #: 1

if(BatchNum == TRUE)

addNewSalt = TRUE;

addRecycledSalt = FALSE;

addRegenSalt = FALSE;

// Batch #: 3, 5, 7, ... ,39

else if((BatchNum <= 40) &&

(Realmod(BatchNum-1, 2) == 0))

{

addNewSalt = FALSE;

addRecycledSalt = TRUE;

addRegenSalt = FALSE;

}

// Batch #: 2, 4, 6, ... ,40

else if((BatchNum <= 40) &&

(Realmod(BatchNum, 2) == 0))

{

addNewSalt = FALSE;

addRecycledSalt = FALSE;

addRegenSalt = FALSE;

}

// Batch #: 41, 43, 45, ..., 79

else if(BatchNum >= 41 && BatchNum <=

80 && Realmod(BatchNum, 2) == 1)

{

addNewSalt = TRUE;

addRecycledSalt = FALSE;

addRegenSalt = FALSE;

}

// Batch #: 42, 44, 46, ..., 80

else if(BatchNum >= 41 && BatchNum <=

80 && Realmod(BatchNum, 2) == 0)

{

addNewSalt = FALSE;

addRecycledSalt = FALSE;

addRegenSalt = FASLE;

ORF eedInORF eedIn

RecSaltInRecSaltIn

D U

BatchNum

y=f(x)

D U

y=f(x)

{...}



demand

addRegenLiCl

RegenSaltInRegenSaltIn

D U

addRegenLiCl

addNewLi2O

addNewLiCl

addRecycledLiCl

VesselRegLiCl

FeedFormOR

TR U

BB RegLiCl

Stats

Calculate

P1-3-ORf eed

Stats

Calculate

P2-2-toORRecLiCl

r L

Stats

Calculate

W4-1-RegenSalt

r L

G_O pen

demand

r L

BatchNum

ORFeedInORFeedIn

D U

RecSaltInRecSaltIn

D U

RegenSaltInRegenSaltIn

D U

InCNT

y=f(x)

addRegenLiCl

addNewLi2O

addNewLiCl

addRecycledLiCl

addRegenLiCl

Transport block

Equation block

Batch Blocks

ICINCO2014-11thInternationalConferenceonInformaticsinControl,AutomationandRobotics

592

Figure 4: Mass composition calculation of feed (input) and product materials (output) in P2-1.

}

// Batch #: 81, 83, ....

else if(BatchNum >= 81 &&

Realmod(BatchNum, 2) == 1)

{

if(ReqSalt > RegSalt)

addNewSalt = TRUE;

else

addNewSalt = FALST;

addRegenSalt = TRUE;

addRecycledSalt = TRUE;

}

The first inputs of two batch blocks are for the

receipt of pellet/fragment, which starts from the

ORFeedIn connector. The batch blocks merge the

first and second input items into one new item. The

second input item number is controlled by the

equation block to be set to 0 (FALSE) or 1 (TRUE).

In the case of FALSE, the batch block converts only

the first item into one new item.

A receipt of regenerated and recycled salt means

the P2-1 process has a recycling material flow from

another unit process. In this way, complex recycling

is easily incorporated into a routing model through

equation and batch blocks.

3.1.4 Material Flow Management

Material dealt with in the pyroprocessing simulator

consists of 52 elements for SNF and 23 elements for

chemical additives needed for pyroprocessing. A

total of 75 elements are calculated and written in a

record of the database table whenever each event

occurs. To access ExtendSim’s internal database

table, equation blocks for the mass balance

calculation are inserted at an appropriate position in

the model. The first equation block in Figure 4,

which is an item equation block different from the

equation block in Figure 3, calculates the salt

composition in a bath when an item passes through

the equation block after receiving and merging a

pellet/fragment with recovered or regenerated salt,

and the item then passes through pre-process,

process, and post-process sequentially. Because an

electrolytic reduction generates two types of

products such as O

and cathode product, the

material should be separated into two products. An

unbatch block plays a role of an item separation and

each equation block actually calculates each product

composition, as shown in Figure 4.

Figure 5: Blocks for mass calculations and process

operation delay.

Figure 5 shows equation blocks to calculate an in-

bath salt composition after receiving the feed

material, and product composition after process

operation. The codes in the equation blocks for mass

composition calculation directly access the database

tables to write, read, delete, and create data. The

equation blocks in Figures 4 and 5 are item blocks

and are able to be located anywhere in the item flow.

Therefore, a mass composition can be timely

calculated according to the material flow and event

driven modelling for a material flow to be possible.

D F

preprocess

D F

process

D F

postprocess

y=f(x)

NewCampaign?

CathodeDeposit

r L

Pretime

Posttime

addRegenLiCl

addNewLi2O

addRecycledLiCl

addNewLiCl

y=f(x)

Cathode Deposit

y=f(x)

r L

y=f(x)

AddedLiCl

AddedLi2O

ReqLiCl

ReqLi2O

Pretime

Posttime

Proctime

y=f(x)

Cathode Deposit

addRegenLiCl

addNewLi2O

addRecy cledLiCl

addNewLiCl

y=f(x)

AddedLiCl

AddedLi2O

ReqLiCl

ReqLi2O

Product

Reaction

Salt addition

D F

preproces s

D F

process

D F

postprocess

Pretime

PosttimeProctime

DiscreteEventSystemBasedPyroprocessingModelingandSimulation

593

Process operation (electro-chemical reaction) is

simply modelled with three activity blocks so that

each activity (pre-processing, processing and post-

processing) consume each specific time. In this part,

the processing activity block can be replaced with a

process model if it exists.

3.2 Material Balance

3.2.1 Equilibrium Material Balance

The integrated pyroprocessing simulator is designed

on a basis of dynamic material balance. On the other

hand, a flowsheet study (Lee et al., 2012) is based on

equilibrium mass balance at a specific time, for

example, at the end of year. Since equilibrium

material balance simply indicates the accumulation

of material transported through steam at a specific

time, the calculation of equilibrium mass balance

after a long period has an average effect such that

transient changes are diminished over time. An

equilibrium material balance represents the overall

static characteristic, but a dynamic one shows the

exact material balance change at any specific

instance in time. An equilibrium mass balance in

process P2-1 is shown in Table 1. This result is

obtained from the accumulated transported mass via

input and output streams after 200 batch operations,

which correspond to 10 tons/year of annual

throughput, have finished.

Table 1: Equilibrium material balance in P2-1.

Material via Stream type SNF mass (kg)

new salt feed -

pellet/fragment feed 11,331

recovered salt feed 5

regenerated salt feed 6

Input Sum 11,341

cathode product product 9,997

product 1,331

Output Sum 11,328

remaining salt hold-up 13

Since equilibrium mass balance shows accumulated

results over numerous batches, a difference of each

batch is ignored. Process P2-1 has a total of four

inputs and two outputs. Sums of inputs and outputs

are not the same. However, considering that process

P2-1 can hold a small amount of SNF in its bath, the

mass balance is satisfied. We cannot predict from

the equilibrium mass balance any result affected by

the operation procedure described in section 3.1.1.

3.2.2 Dynamic Material Balance

Table 2 represents the mass of inputs and outputs

calculated whenever every process batch operation

is completed. Process P2-1 receives 50 kgHM/batch

from a previous process. The second column in

Table 2 represents the mass of pellet/fragment oxide.

Excluding the oxide weight, it becomes 50

kgHM/batch. The operation procedure in sections

3.1.1 and 3.1.2 indicates that the recovered salt is

added at every other batch during the first campaign,

and the third batch operation expects a receipt of the

recovered salt from P2-2. However, in the third

batch operation, P2-1 does not receive the recovered

salt because P2-2 has not prepared another recovered

salt by then. Such a dedicate behaviour cannot be

estimated in the equilibrium mass balance.

Table 2: Dynamic material balance in P2-1.

batch # fragment/

pellet (kg)

recovered

salt (kg)

regenerated

salt (kg)

remaining

salt (kg)

cathode

product (kg)

(kg)

1 56.67 - - 0.28 49.72 6.67

2 56.67 - - 0.54 49.73 6.67

3 56.59 0.02 - 0.83 49.74 6.59

4 56.67 - - 1.08 49.75 6.67

5 56.59 - - 1.33 49.75 6.59

6 56.67 - - 1.57 49.76 6.67

7 56.59 0.05 - 1.85 49.77 6.59

8 56.67 - - 2.08 49.77 6.67

9 56.59 0.08 - 2.38 49.78 6.59

…

41 56.59 - - 10.25 49.97 6.59

42 56.67 - - 10.27 49.98 6.67

43 56.59 - - 10.30 49.97 6.59

44 56.67 - - 10.32 49.98 6.67

45 56.59 - - 10.35 49.97 6.59

46 56.67 - - 10.37 49.98 6.67

47 56.59 - - 10.40 49.97 6.59

…

81 56.59 - 0.08 11.03 49.99 6.59

82 56.67 - - 11.04 50.00 6.67

83 56.59 - 0.08 11.13 49.99 6.59

84 56.67 - - 11.13 50.00 6.67

85 56.59 - 0.08 11.21 49.99 6.59

86 56.67 - - 11.22 50.00 6.67

87 56.67 - 0.08 11.30 49.99 6.67

…

194 56.67 - - 13.30 50.05 6.67

195 56.67 - 0.11 13.37 50.04 6.67

196 56.67 - - 13.32 50.05 6.67

197 56.67 - 0.12 13.39 50.04 6.67

198 56.67 - - 13.34 50.05 6.67

199 56.67 - 0.12 13.41 50.04 6.67

200 56.67 - - 13.36 50.05 6.67

sum 11,331 5 6 13.36 9,997 1,331

ICINCO2014-11thInternationalConferenceonInformaticsinControl,AutomationandRobotics

594

It shows that regenerated salt is provided at every

other batch operation from the third campaign (81th

batch ~) in the 4th column in Table 2, as expected in

section 3.1.1. Since the remaining salt always exists

in the bath of P2-1, the 5th column in Table 2

indicates the current accumulated state in the bath.

The amount of products increases as the campaign

increases because a small amount of SNF elements

accompanied by regenerated salts is added in P2-1

from the 3rd campaign. A summation of the input

and output materials over 200 batch operations in

Table 2 means accumulated transported mass

through input and output streams. The accumulation

of each batch operation results is exactly the same as

the equilibrium mass balance in Table 1. Compared

to the equilibrium mass balance, the dynamic mass

balance gives a lot of information on the state at a

specific instance in time, i.e., how much material has

been processed, how many products have been

produced, how much of the material has been held

up, and how much material remains in temporary

storage while waiting for the next process.

4 VERIFICATION

AND VALIDATION

4.1 Models & Codes

Logics and models for operation modeling are

verified from a material flow point of view. Many

debugging tools and utility blocks support the

building of an integrity model. Debugging the item

flow and equation are necessary to detect problems

and fix them to accomplish the completeness of a

model. In the model level, a pausesim block is

located anywhere and stops the simulation progress

at a user intended point. In the code level, a set break

point will stop the programming code in the

equation block running at a user intended point. The

integrated pyroprocessing model includes a very

complicated item flow, and thus a debugging

process is needed to guarantee item flow logic. It

also includes many equation blocks to build a

complicated operation procedure and calculate the

mass composition. Therefore, to debug such

equations is also necessary. Whenever a unit process

model is developed, not only is the model itself

debugged, the integrated operation model is also

verified.

Validation is quite difficult to perform at this

moment because there is no existing integrated

facility using SNF. However, lab-scale based

experimental results using simulated fuels, which are

not real SNF but are able to simulate real SNF to a

certain degree, can be validation data. Such results

have already been used in the model to obtain

material composition of the product. Therefore, the

material composition of every product should be

checked at every batch operation to satisfy the

separation factor as obtained in the experiment.

4.2 Material Balance

Once the unit process is modelled, the mass balance

equation must be satisfied for every batch operation

through an investigation into the related database

tables.



,







,













 (2)

where  is the number of inputs,  is the number of

outputs,  is the current number of batch operations,



,

is the k-th input amount of mass transported

through the i-th upstream, 

,

is the k-th output

amount of mass transported through the i-th

downstream, and 



is the hold-up until the k-th

batch.

The bracket in equation (2), which represents the

difference between the current and previous hold-up,

means the contribution by only the current batch

operation because a hold-up inherently involves

accumulation. For example, for the 9th batch

operation in Table 2, the mass balance equation (2)

is satisfied as follows:

Table 3: Mass balance equation in the 9th batch operation.

9-th

inputs

9-th outputs 8-th hold-

9-th hold-up

56.59

0.08

0.00

49.78

6.59

2.08

2.38

56.67 = 56.37 + 0.30

After the current batch (9th batch), an operation

hold-up represents 2.38 kg, but the contribution by

the current batch operation is exactly 0.30 kg. Any

other batch operation satisfies the mass balance

equation as the result of the 9th batch operation. The

above results must also be satisfied even though a

gross mass decomposes into the type and element

level. The proof is skipped in this paper due to

limitations.

DiscreteEventSystemBasedPyroprocessingModelingandSimulation

595

5 CONCLUSIONS

The integrated pyroprocessing features are a batch

type operation, complicated recycling, and tangled

operation logic. An item flow model based on DES

enables a flow control, mass balance calculation, and

basic framework of an integrated pyroprocessing

simulator. Compared to a static or equilibrium

material flow, a model-based dynamic material flow

provides detailed information and thus a careful

analysis of every batch is necessary to confirm the

mass balance results. Verification and validation

regarding the model built thus far has been

performed in terms of the mass balance calculation,

and shows the completeness of the model. However,

the modeling has not been finished but is still under

progress.

To improve the operation model toward a multi-

purpose integrated pyroprocessing simulator, various

modules must be incorporated at a facility level. One

of the issues on a new recycling process such as

pyroprocessing is to guarantee integrated safeguards

in terms of material accountancy and security. Mass

tracking is the most fundamental requirement for a

model to cope with for safeguards assessment. Since

the material flow framework in the current model

can support a perfect mass tracking on an element

basis, a safeguards module is expected to be

developed without difficulty and to be added in an

integrated simulator. Technical feasibility can also

be supported by an integrated simulator to determine

or recommend process operation conditions by

adding an optimization module. Compared to other

reprocessing technologies, economic feasibility must

be tested in a simulation by developing a cost

evaluation module. It is expected that an integrated

pyroprocessing simulator fulfilling the above

described functions by add-on modules will be

released in a few years.

ACKNOWLEDGEMENTS

This work was supported by Nuclear Research and

Development Program of National Research

Foundation of Korea (NRF) funded by Ministry of

Science, ICT and Future Planning (MSIP).

REFERENCES

DePaoli, D., 2011. Modeling and simulation of nuclear

fuel recycling systems, short course of “Introduction to

nuclear chemistry and fuel cycle separations.”

Bechard, V., 2013. Simulation of mixed discrete and

continuous systems: an iron ore terminal example, In

Proceeding of the 2013 Winter Simulation Conference,

1167-1178.

ExtendSim Simulation Software. Imagine That Inc, 2014.

Web. 24 Jun 2014. <http://www.extendsim.com>

Lee, H. J. et al., 2012. Pyroprocessing baseline flowsheet

v4.0, talks in KAERI.

Lee, H. J. et al., 2013. Design for integrated

pyroprocessing plant level simulator, Annals of

Nuclear Energy, 60, 316-328.

Phongikarron, S., Herrmann, S., Simpson, M., 2011.

Diffusion model for electrolytic reduction of uranium

oxides in a molten LiCl-Li2O slat, Nuclear

Technology, 174, 85-93.

Yoo, J. H. et al., 2008. A conceptual study of

pyroprocessing for recovering actinides from spent

oxide fuels. Nucl. Eng. Technol. 40, 581–592.

Song, K. C., Lee, H., Hur, J. M., Kim, J. G., Ahn D. H.,

and Cho, Y. J., 2010. Status of pyroprocessing

technology development in Korea, Nuclear

Engineering Technology, 42(2), 131-144.

Lee, H., Park, G. I., Kang, K. H., Hur, J. M., Kim, J. G.,

Ahn, D. H., Cho, Y. J., and Kim, E. H.,

Pyroprocessing Technology Development at KAERI,

Nucl Eng Tech, 43(4), 317-328.

Karell, E. J., and Gourishankar, K. V., 2001. Separation of

Actinides from LWR Spent Fuel Using Molten Salt

Based Electrochemical Process, Nucl. Tech., 136, 342.

S. D. Herrmann, S. X. Li, and M. F. Simpson, 2005.

Electrolytic Reduction of Spent Oxide Fuel – Bench-

Scale Test Results, Proc. Global 2005, No. 488,

Tsukuba, Japan, October 9-October 13.

J. M. Hur, I. K. Choi, S. H. Cho, S. M. Jeong, C. S. Seo,

2008. Preparation and Melting of Uranium from U3O8

J of Alloys and Compounds, 452, 23.

ICINCO2014-11thInternationalConferenceonInformaticsinControl,AutomationandRobotics

596