Open Access - Lund University Publications
Contents
1. This procedure moves the load from the dock to a specified storage place in the Silo This procedure takes us through four systems It starts at the docks control points in the path mover system inside the DC and moves on to the conveyor system which is feeding the Silo After the conveyor the load is moving into the buffer which is simulated with an invisible and stationary conveyor system see 5 1 2 The Peterborough DC Highbay Silo From the buffer the load is entering the ASRS system in the Silo and finally the load finds the specific storage place Order List OL When a load has found a storage place in the Silo the load is placed in an Order List Here the load is waiting to be called for Lowbay storage P_SetStoragePlaceLow Sets a storage place for the load according to predefined pattern racks associated to this type of load There are two possible choices which can be set up for the Lowbay storage e The first choice assigns a random place in the rack s and from this place the load finds the first available storage place If no storage place is found the load is send to a Dump place queue e The second choice uses zones The Lowbay area is divided into eight different zones each zone referring to a country Here the algorithm first tries to find a storage place as close as possible to the floor within the zone and after that it searches for empty places in the same tier If this is not possible it changes to the ne2. 203 00 202 00 201 00 200 00 199 00 198 00 197 00 196 00 195 00 194 00 193 00 192 00 1 2 a Number of Zones Figure 19 Number of Zones The analyses of the Silo when divided into zones are shown above in figure 19 The analyse states that the usage of two zones shows the best throughput Though as described for the analyses of the storage mode the difference cannot be statistically determined with a confidence level of 95 39 Simulation of the IKEA Distribution Centre in Peterborough UK 8 Final discussion This chapter will present the authors own findings There will be a minor discussion with the objective to illustrate the main questions and connect the analysis with the authors own findings When looking at the first scenario and the analysis made of the number of cranes it shows that four cranes are insufficient and five to seven cranes manage the demanded throughput We believe the best solution is to use six cranes Five cranes manage the flow but we have not taken maintenance and breakdowns into consideration Seven cranes do not increase the throughput according to the analysis and therefore we find seven cranes unnecessary Hence six cranes are the best solution The second scenario analysing the storage mode and the number of zones in the Silo differences can be shown Though for both cases there cannot be any statistically secured results using the confidence level of 95 When con
3. Handled Flow in the Silo Units Hour Demanded Flow from the Silo Units Hour Figure 17 Silo throughput 37 Simulation of the IKEA Distribution Centre in Peterborough UK The analysis of the Silo in AutoStat Is shown above see Figure 17 The y axis displays the total Silo capacity per hour while the x axis displays the number of Automated Cranes From the staple diagram you can make the conclusion that four SRM s operating in the Silo is insufficient The SRM s cannot deliver the demanded loads We consider five SRM s the capacity is sufficient within the margin Six cranes as well as seven cranes are sufficient within the margin The difference between six cranes and seven cranes considering the throughput are slight 7 2 Analyse Scenario 2 Storage Mode 210 00 209 00 208 00 207 00 206 00 205 00 204 00 203 00 202 00 201 00 200 00 199 00 198 00 197 00 196 00 195 00 194 00 193 00 0 Storage Mode 0 Systematic Storage Mode 1 Random Storage Mode Figure 18 Storage Mode 38 Simulation of the IKEA Distribution Centre in Peterborough UK Figure 18 above shows the difference between the systematic storage mode and the random storage mode Though the analysis is done with the confidence level of 95 and the figure is showing that the difference between the modes cannot be statistically determined The differences are within the interval Number of Zones 205 00 204 00
4. and loads increment blocks automatically when moving through the physical space defined by the block Loads can also claim blocks in process logic as directed by the user Blocks can have any shape 3 4 7 Variables and Counters Data may be stored in an AutoMod model using variables Variables can be used in calculations or can be compared to other variables In addition to integer real and string types variables can also be used to store references to other process system entities such as processes queues resources order lists counters loads and locations Storing these references gives the users more power and flexibility to expand and extend models to match modifications of the actual system Counters have a maximum capacity making them useful for traffic control When a load tries to increment a counter that is at its capacity the load will be delayed until another load decrements the counter Statistics for counters are collected automatically 3 4 8 Functions Subroutines and Source Files Functions and subroutines are used to create modular models This allows the models to be extended more easily Users can define their own functions in the AutoMod and these functions can be called from any where in the AutoMod model Subroutines help eliminate duplication of action statements in the model Source files contain the logic for the model 2 Brooks Automation Inc 2001 4utoMod User s Manual Bountiful 30 1h Ib
5. 7 T tuw v ns Jord a bo Co ama GENRE S r A GN NATAN KA 5 2 2 2 QUANTITATIVESIMOTN OG GA scaza kawad aaa O usq Sua Sas uui AN AA 5 2 2 3 Choice of scientific methodology LLL aR kakak ERK ARR RERA KKR AA 5 ZED NAVI T YZL L E mana tat aches be a S a aa ca da da una Saku sta Susa a EE ua east ym asya u ata Sa 5 2 4 REWABIB ESA NG ku Bahan nck obi Kaha cach cobs u a m E Sua sis ui mutu ys ukyu sa ane 6 2 5 METHOD IOP GOLEEGCTION I Lau R lun Qul Dum a ka h deal ad slet AT a aaraa S uya aiaia 6 LO PIM a s u u NA RAD a i skua yp akak A SKER pa Hand 6 PAN VAIMOIVIOW secs in sanu mu a S ANGARA BB ssd aalas 7 DIA SUIVOY ete isd alas Bel ia Sua SSS Ar A aasma ai sau uds aa 7 29 12 GADITIWINOS u u Sua Du s m ua Su am Jed beta ia 7 LO 2 ECON GNA arterna u aaa aa NAN AA PAANAN ap reo ARA RS 7 2 6 50URCE CRINGISMI 3NA Ka an IE d 7 2 60 13 BabkgrouNA m m ass s a aa tae basa AN AN h a LA a uya uma ita da 7 2 6 2 Criticism of primary dale uuu u aa ANN carro 8 2 6 3 Criticism of secondary data ooooonnocccononococinnnonoccnononn cinco nono carroceria 8 3 THEORETICAL FRAME u 9 3 INTRODUCTION ada 9 3 2 BACKGROUND 22 ida dba 9 3 3 AUTOMOD INTERFAGE usina tias 10 3 4 AUTOMOD S WORLD VIEW aia iaa 10 S41 PROCESSES a a AA AA AT KAN AA AN ng NA 11 342 He A EES ee SRS EA 11 34 REOSQUICOS it Pais akwa amu Sona Tana AE SR REN ia naa Lan asqa pusay Saku Sta Aa 11 SA SALES vis nanahan Ngan Ba
6. in this procedure which starts the simulation either with an empty DC no loads in the racks or with loads in the racks 5 3 1 2 Inbound procedures P_TruckArriving The initiation part starts up eight procedures one procedure for each incoming truck The procedures are waiting but when an incoming truck is initiated one procedure is activated The procedure is run trough and when that is done the procedure is ready for another truck to initiate it again looping The procedure simulates the arrival sequence of a truck from different countries finds a suitable dock according to load types and country P_MoveTruck This procedure moves the trucks to a specified dock in the path mover system This is the visually part of the arriving trucks Eight different trucks can be visualised It creates the number of loads in the truck It also does a capacity check which calculates the maximum number of loads allowed in the truck P_LoadArriving This procedure creates one procedure for each load type The P_LoadArriving procedure can set up to twenty eight load types in each arriving truck The loads are placed outside the truck into the dock zone This procedure sets the frequency attribute The load now stands in front of two storage alternatives in the simulation The storage depends on what kind of type the load belongs to see Figure 12 28 Simulation of the IKEA Distribution Centre in Peterborough UK Alternative storage P Loa
7. nr rr ann kr rn RR ARA Snr Snr rn 18 5 MODELLING U ia 19 S T MODBELEINGIDESCRIPTION i son Naan dde ku Baon BANA 19 9 151 Inbound IOGISHCS u u uu AA NANA PANA BANANA ah 19 O T Ts ATUCIeS ERo80Ssu ouu u nuna maana Aaahhh KANA sl tere a Naan an 19 Dill ATING TICKS u al unuman uspha titan alias 20 S 12 The Peterborough DG iat ma u Q uum Sunan PANANDA Paa NAN PARA FARAN 20 5 1 2 1nbound OCK ii A pana eee ieee 20 Dali 232 NA BEE u m uni a na Ann aaa da aaa dana aag lora 21 5 1 2 3 Lowbay 7 Conventional storage m aaananwwwaannaaawanananawananasaaanananaananasasanannanssnanans 23 5 1 2 4 Highbay S00 suu sua kanaman nn an aa aaa a a 24 5 1 3 QU1b0UNnd 109ISNI GB ma u U au Aa AI BN Pa AN AN PA RER ARA 26 9215321 OULDOUNG COCKS sia isu a ak paha kwan namn Aid canal tetas ces qam aa ainia 26 0 1 3 2 Departure frucgkS usu u nan aula asas uapa huta shua uum aaa 26 5 2 GRAPHIC DESCRIPTION snm ARA Ba AA C asss eina BA asum a yanasa 26 5 3 PROGEDURE DESCRIPTION a adde u A ha A aaa AA 27 DISTRO COS O A LG A aaa 28 5 3 1 1 Model Initiation asinna aa aaae ae eaaa Ea RANA RASK ENAS ASKR DOK ies 28 5 3 1 2 Inbound procedUreS u Lagana an PANA ere rss Ers SHAN enas AABANG 28 5 3 1 3 Outbound procedures arr arr nr rr en kr nr narr rr en nn nn nr RR AR ARR RR RR AR RBK 31 6 EXPERIMENTAL DESIGN u 3
8. of simulation We will run scenarios with different conditions in order to evaluate where bottleneck can be developed and find the best and most optimal solution for an efficient material flow Two important activities Run capacity simulations of the Silo Look for bottlenecks in the material flow Ikea Intranet Thrapston 2002 10 15 Simulation of the IKEA Distribution Centre in Peterborough UK 1 4 Demarcations The Peterborough DC includes a Custom Distribution Centre CDC but we will not take this part into consideration during this report Thirty percentage of the total material flow are picked pallets2 though this influence on the material handling this will not be taken into consideration Further we will not do any economic calculations nor will we consider any organisational restructions We will not do any collection of empirical input data We will have the data and information needed from already existing CAD drawings done by IKEA Further more we will not collect or measure the input data to the simulation our self this is also done by IKEA 1 5 Target group This report is mainly aimed towards the management of IKEA Thrapston in UK Though this report is written by authors who have been studying at least six terms at different scientific faculties in Sweden And while we consider our frame of reference to be theoretical we have also chosen to direct this report to other students and teachers with equivalent k
9. the approach of the examination The strength with the quantitative method is that collected data render possibilities for generalisation 2 2 3 Choice of scientific methodology This report departs from a quantitative point of view Our project is mostly built on figures and numbers therefore we prefer the quantitative method in front of the qualitative method Our result will also be presented numerical which is strengthening the quantitative approach Though we have not totally rejected the qualitative method we have actually chosen a hybrid between the two scientific methodologies Like described above our project departs from the quantitative approach with measurements The translation of the hard facts between the CAD drawings and the graphics in AutoMod are strictly quantitative But when it comes to the translation of the reality or real life scenarios into syntax which means interpretations and simplifications of the reality we have chosen a qualitative approach Interviews have been done which have had a qualitative characteristic 2 3 Validity A result shows high validity when you measure a phenomenon you have planned to measure and nothing else Holme Idar M amp Solvang Magne K 1997 Forskningsmetodik Om kvalitativa och kvantitativa metoder Studentlitteratur Lund Ibid 1 Backman J 1998 Rapporter och uppsatser Studentlitteratur Lund Holme Idar M amp Solvang Magne K 1997 Forskni
10. the load is set down and the domestic truck arrives waiting to be loaded Fork lifters from the DC are loading the domestic trucks When the loading procedure is finished the domestic trucks depart and the simulation system ends 4 2 Layout description The schematic figure below see Figure 2 illustrates the layout of the DC in Peterborough The Silo or the high stacking area is in the middle of the DC On both sides of the Silo there are Lowbay areas or conventional stacking areas Automatic storage cranes SRM s are working inside the Silo Standard fork lifters are operating in the conventional stacking area The inbound docks and the outbound docks are located along the longs sides of the DC Schematic layout of the Peterborough DC O a i i SILO i LI Se Seon Ah P o wa a Lowbay or Conventional storage area E Inbound Outbound docks Figure 2 Schematic top view of the Peterborough DC 17 Simulation of the IKEA Distribution Centre in Peterborough UK 4 3 Facts and figures Peterborough DC e The DC building is 313 excluding the CDC metres long by 186 metres wide e The surface area of the DC is therefore 58 218 square metres e The high bay Silo area of the DC has 24 aisles and 70 000 pallet locations e The external roof height of the high bay Silo of the DC is 23 metres e The external roof height of the low bay conventional are
11. 2 6 1 INPUT DATA PROGNOSIS sa LL uum n in an lis pee NAA een y nr pedal 32 6 2 SCENARIO STRUCTURE t i uu S Tun ae Samana E da IE LAM an Fahad 32 02 SLO SCENARIOS u m op An ANA Ka LANG BONN So aman am R Laupala s Suha ated 33 6 3 1 Scenario fu esau su A in Some ANG aq e et cae Vet taza 33 6 3 1 1 Number of Automated Cranes u Kr RR rn nan 33 6 3 2 SCCM AIO Lit E sls ANAN aquta asa an O a a pad 34 0 32 17 Storage MOE aman aka NAAN adelaida pa Ga idad ibi NAG 34 0 3 2 2 NUMDEOF OF ZONCS Lulun unu dl tt a ede eae aa 35 7 OUTPUT DATA ANALYSIS ii ann n ann RAR RR RAR R RAR ARR a RR RAR RR eed 37 T V ANAEYSESCENARIO Distintas o cede ead das eee eed red ete ead BAN ee 37 2 ANALYSE SCENARIO 22 aan aaa a eee nade need needa a al aaa 38 8 FINAL DISCUSSION u 40 9 FUTUREAPPROACG Ha AANGAT NG NANANA NGANGA AIGANAD 41 LIST OF REFERENCE us s uuu ANAKAN MAA ANA cin 42 APPENDIX aNG NANANA NNNANNNANINANANKAA 43 Simulation of the IKEA Distribution Centre in Peterborough UK 1 Introduction This introductory chapter deals with several fundamental parts of an academic report The chapter begins with a short background Followed by problem setting the objectives of the report demarcations target group and the last part illustrates the structure of this report 1 1 Background Today s companies require higher p
12. Articles Loads In the simulation there are four different load types Each load type represents a group of articles in the IKEA assortment e The first load type represents euro pallet type and is dedicated to the Silo e The second load type represents euro oversize pallet type and is dedicated to the conventional area e The third load type represents Ikea oversize pallet type which also is dedicated to the conventional area e The fourth load type is a special pallet with overhang Especially hard to manage Euro oversize pallet differs from the Ikea oversize pallet in dimension This means that there are certain racks and which in other words means that there are certain areas dedicated for each pallet type in the conventional stacking area see Figure 4 19 Simulation of the IKEA Distribution Centre in Peterborough UK Pallet type location in the Peterborough DC JIE 5 Figure 4 Schematic pallet type layout of the Peterborough DC Each load type has a number of attributes which are set during the simulation For example which origin country the load type is coming from In which dock the load is loaded unloaded The direction of the load is of importance when the fork lifters prioritise the work in their work lists The load types also have attributes for their destination For example what aisle what bay and what tier they are going to 5 1 1 2 Arriving trucks The arriving trucks are simulated in a path mover system
13. Hundred fifty inbound trucks arrives to the DC every week e Hundred fifty outbound trucks arrives to the DC every week We will shortly describe the analyse methods used when analyses were done in AutoStat One factor analyse This model analyse how the model behaves when you change one factor while keeping other factors constant A factor could be for example the number of resources or a machine s processing time What the output analyse will focus on is what happens with the Silo throughput and how it affects when factors are varied 6 3 Silo Scenarios 6 3 1 Scenario 1 6 3 1 1 Number of Automated Cranes We will focus on the Silo and isolate it from the rest of the DC We will consider variations in the number of Automated Cranes This scenario underlies further analysis of the Silo Though isolating the Silo each scenario is taking the whole system into consideration From the arriving trucks the material flows through the DC and finally when the loads are leaving the DC with the departure trucks The scenarios are aimed to attain our problem definition and our objective see 1 2 Problem setting 1 3 Objective Number of Automated Cranes Factors e Vary the number of Automated Cranes SRM S between 4 and 7 Response Silo throughput per hour measured against the demanded output from the Silo Figure 13 Scenario Number of SRM s 33 Simulation of the IKEA Distribution Centre in Peterborough UK 6 3 2 Sce
14. Lund Institute of Technology Lund University Department of Design Science Division of Packaging Logistics Master Thesis Simulation of the IKEA Distribution Centre in Peterborough UK Authors Tutor William Jovalli Mats Johnsson Gregor Nalepski Logistic Manager Darrell Harvey 11 04 2003 Title Department Authors Tutor Key words Problem Purpose Method Objective Conclusion SUMMARY Simulation of the IKEA Distribution Centre in Peterborough UK Department of Design Science Division of Packaging Logistics at Lund Institute of Technology William Jovalli and Gregor Nalepski Mats Johnsson Capacity simulations bottlenecks efficiency AutoMod and effectiveness IKEA want to avoid building a DC with insufficient capacity and with imbalances in the material flow therefore they want to run simulations of the system to see if the material flow is of satisfaction The focus of this report is aimed towards capacity simulations of the Silo the automated high storage area Run different scenarios of the Silo with different conditions in order to evaluate where bottleneck can upraise and aim for optimal solutions with retained efficiency in the material flow The report is built on both quantitative and qualitative studies Information gathering has been done from interviews and CAD drawings in combination with secondary sources Information from secondary sources was gathered from literature as books bro
15. Packaging Logistics at Lund Institute of Technology and IKEA in Thrapston UK The collaboration is to demonstrate attainable benefits with simulations The simulation software used in this project Is called AutoMod We would like to thank our tutor Professor Mats Johnsson and Darrel Harvey Logistic Manager at IKEA in Thrapston for great support and feedback during the project We would also like to thank Dantel Hellstr m PhD candidate at the Division of Packaging Logistics for great guidance within the software AutoMod and finally Adri Kraa Rick Hadley Keith Morris and Miranda Stephens for good insight of the logistics at the DC s in Thrapston and Doncaster Lund 2003 04 11 William Jovalli Gregor Nalepski Table of contents INTRODUCTION cusco dada 1 1 14 BACKGROUND mamintas A A A aa at a ati 1 12 PROBLEM SETTING nrerin araa e a oia ds 1 1 3 OBJECTIVE a a p DADA n ARN LATA a A aa a ADA tb BANNA 1 1 4 DEMARGATIONS Wick ug u L Su a l a aa BANANA 2 1 5 TARGET GROUP us qa u odd m n uu daa a SA KABA alaq tl caba 2 1 6 STRUGTURE OF THE REPORT 4 spa un a dl NLA AKA acacia 3 2 METHODOLOGY cedida nica RR RR RR RR RR RR KKR RR RR RR RAR 4 2 1 SCIENTIRIC APRROAC HID Apan aNd ah E una a kagak Ea e 4 221510 OXT EAV 0A AAE A Daag anna AI A A A dd 4 ZN ZAS MENU diia A 4 2 2 2 Choice of scientific approach u u s O un RQ bo SARA oas aaa aussi iss 4 2 2 SCIENTIFIC METHODOLOGY aa kunuk aaa 5 2 2 4 QUalitative method u
16. a of the DC is 12 metres e The maximum storage capacity of the DC is 98 000 cubic metres e There are 78 loading and unloading docks in the DC The Peterborough DC will be a similar design to the Doncaster DC It will be a combination between conventional storage and crane based Silo A third party operator will run the Peterborough DC It is common to use a third party operator in UK The option of IKEA operating the DC has been considered and rejected This external service provider ESP is using the FIFO rule They will use the ASTRO warehouse management system 4 Ikea Intranet Thrapston 2002 10 15 18 Simulation of the IKEA Distribution Centre in Peterborough UK 5 Modelling This chapter will describe how the real life scenario Is converted into modelling and which systems used in AutoMod The chapter will also deal with the different processes that appear in the simulation This chapter will also describe how the authors have chosen to model the different procedures in the system 5 1 Modelling description A similar schematic system description will be followed as in chapter three The inbound logistics the Peterborough DC and the outbound logistics see Figure 3 Under each part we will more thoroughly describe the system used in AutoMod 2 The Peterborough DC 1 Inbound logistic a 0 3 Outbound logistic Figure 3 Schematic drawing of the system 5 1 1 Inbound logistics 5 1 1 1
17. an identify specific areas in the facility There are several ways to create a layout of the system to be modelled AutoMod comes with a three dimensional graphics editor ACE that allows the user to construct objects from standard graphics primitives 3 7 Run time environments In keeping with AutoMod s interactive features the user has complete control of the model in the run time environment The model can be viewed with the animation on or run with the animation off AutoMod uses concurrent animation the simulation progresses as the animation picture is being updated With animation off the simulation does not render the animation but it performs all simulation calculations The user can suspend the simulation at any instant to review statistics through pop up windows to take resources down to set break points or alarms or to control the view of the animation without constraint 3 7 1 View Control AutoMod provides a comprehensive and flexible method of interacting with a model during model execution If the simulation project is in the experimentation phase where only parameter changes are made and the model needs to be re run several times AutoMod provides the ability to run in batch mode without animation 3 7 2 User Interaction AutoMod provides advanced debugging and trace facilities A model can be single stepped at any time during the animation Also the ability to set breakpoints and alarms allows the user to sus
18. an nng AA na u Siga qata nna Naga ET ER 11 3 4 5 Queuesand Order Dista NANANG NAK AMONG AAKALA 12 34 61 BIOCKS Aa m hZ mus Sua ank aa a EA na NG KANO Pana naan 12 3 47 Variables ANd COUMLCTS ii NN vi Sak sa atus Ona ALAALA 12 3 4 8 Functions Subroutines and Source Files aaa 12 A A O la TF PA nha kha muu kas nG adn NLA Thon itu ka 13 3 4 10 Types and Random Stearns sessirnir ianea EANA A VENERAR RENYA REARDAN AA S RAN NE SRK RNA 13 SAMA RUN CONTO E AA amaka NTA SR RES ltd 13 3 4 12 Business Graphi65 Ju u LA a a ira 13 O XA adm Y aaa usu SS A a tua ua ukasa Sasha S Gaya aasan A ww ma a aaa sQ 13 S 6 GRBAPHIGCSIN D a kana AN ag a aa aa AG haa a caches Ka a Ta Ka aha 14 3 7 RUN TIME ENVIRONMENTS n sr snar sn SR SR SR RASA RR KR KRKA R RKA RR KR RR RR RR RR RR KRA R SAS a aan 14 SAL MISW COM AA 14 31 2 User Interacci n idad a k Shuwa wee dak ani sae u u SYR S uksa 14 3 B AUTOSTAT N u Qu RAANG adn AN kan bn askan SANN h ama Quam Saa aku s ua quan tas 14 4 SYSTEM DESCRIPTION mpa NABABAGABAG 16 AAC SYSTEM SIMULATION a a ar dd one p C a u ANA 16 4 1 1 Inbound IOGISUICS sc akan asa a KA sq s na ua AS AAA GEA 16 4 1 2 Material flows inside the D Os u u u uuu uuu oss s gs und NON AN eaaa Sual aaa NER bene 16 4 1 3 Outbound lOgIStiCS u u uu u nn anida tr s 17 4 2 LAYOUT DESCRIPTION LU naaa toldo 17 4 3 FACTS AND FIGURES PETERBOROUGH DC rr ns sn annan rr
19. as possible the primary data collection is very vital for our project Therefore we spent a lot of effort on this part 13 Holme Idar M amp Solvang Magne K 1997 Forskningsmetodik Om kvalitativa och kvantitativa metoder Studentlitteratur Lund 14 Arvidsson P amp Rosengren K E 1992 Sociologisk metodik Liber Lund S Winter J 1973 Problemformulering unders kning och rapport Liber Arbnor L amp Bjerke B 1994 F retagsekonomisk metodl ra Studentlitteratur Lund 17 Eriksson L T och Wiedersheim Paul F 1991 Att utreda forska och rapportera Liber Malm Simulation of the IKEA Distribution Centre in Peterborough UK 2 5 1 1 Interview To gather the primary data we made interviews with persons who have been working in IKEA for several years with expert knowledge s within the area requested During the interview the tasks within the project group where strictly divided One member of the group put general questions as well as specific questions and the second member had the answers written down Furthermore we used an open time schedule during the interviews which made it easier for us to gather information a character of a more freely held discussion Also this made the respondent more relaxed and not strictly questionnaire focused We found this alternative approach to be very complementary in our search for information regarding the interviews 2 5 1 3 Survey Surveys have been don
20. chures and articles in addition to the Internet Run capacity simulations of the Silo and look for bottlenecks in the material flow The analysis made of the number of cranes states that four cranes are insufficient and five to seven cranes manage the demanded throughput We believe the best solution is to use six cranes Five cranes manage the flow but taken maintenance and breakdowns into consideration six cranes are to prefer Seven cranes do not Increase the throughput according to the analysis and therefore we find seven cranes unnecessary Hence six cranes are the best solution The analysing of the storage mode and the number of zones in the Silo shows differences in the throughput Though for both cases there cannot be any statistically secured results using the confidence level of 95 When considering the variation of the number of zones we consider the zoning could be done in a different way Because of the fact that there are not SRM s in every aisle the cranes have to change aisles Using the zoning from the front to the back the cranes have to move a long distance when changing aisle for the fast moving articles in the front of the Silo Instead we believe the zoning should be done from the left to the right or the way around and the cranes in the fast moving zones should have a crane in every aisle The other zones could share the other cranes Preface This Master Thesis is the result of collaboration between the Division of
21. comprehensive level see Figure 112 Further we will present and illustrate the procedures on a more detailed basis though simplified by following one load throughout the simulation this will show what different choices and algorithms the load undergoes An additional and more detailed and thorough description of the program structure are illustrated in Appendix Program structure Schematic program structure Model Initiation lt 8 procedures are created TruckArriving MoveTruck LoadArriving Outbound procedures MoveOutLoadSilo MoveOutLoadLowbay RetrieveLoads MoveDomesticTruck 11 Outbound procedures are also created waiting to be called upon Figure 11 Schematic Program Structure 27 Simulation of the IKEA Distribution Centre in Peterborough UK 5 3 1 Processes 5 3 1 1 Model Initiation The model initiation is the first part of the program and the starting point of the processes This part is needed to do the processes ready for simulation Initiation function The function initiates the procedures and makes them ready for simulation P_Set In this process all the variables in the simulation are set For examples the number of incoming trucks what frequency number of different load types in each truck and choice of either mixed inbound outbound flow or not P Init This procedure set certain procedures variables to zero and initialise certain procedures For example you can make a choice
22. d by AutoView are also defined in the run control 3 4 12 Business Graphics Graphs in AutoMod are easy to define and they update in real time with the animation Graph types include bar charts pie charts and timelines Any model entity can be attached to a graph including transporters vehicle velocity number of loads on a conveyor section or average utilization of a machine Graphs can be printed or plotted 3 5 AutoView Post processed animation extension can be used to view the animation records created by running an AutoMod model The AutoView product allows the user to predefine all views and time periods in the model animation and then play them back to generate presentation quality animation files AutoView also has single frame capture capability to AVI and MPEG format files This feature allows users to create animations that can be shared with others without the need for any additional software Brooks Automation Inc 2001 AutoMod User s Manual Bountiful 35 Th Ibid 37 mia Ibid 38 Brooks Automation Inc 2001 AutoMod User s Manual Bountiful 13 Simulation of the IKEA Distribution Centre in Peterborough UK 3 6 Graphics in 3D Both dynamic and static objects can be displayed during model execution Dynamic objects represent loads vehicles resources queues and statistics The static layout is the background graphics of the plant such as columns aisle markings and walls Labels c
23. dArriving 1 max 28 Load type 1 Load type 2 3 4 Figure 12 Alternative storage Additional the two alternative storages create two different paths The different paths are illustrated below see Table 1 According to which processes the load is entering the load can end up in either Order Lists Silo OL Silo or Order List OL Lowbay Load type 1 2 3 4 Alternative Silo storage Lowbay storage P i P SetStoragePlaceHigh P SetStoragePlaceLow kinang P MoveLoadToSilo P MoveLoadToLowbay Order List OL Silo OL Lowbay Table 1 Alternative paths Silo storage P SetStoragePlaceHigh Sets a storage place for the load according to predefined pattern racks associated to this type of load There are two possible choices which can be set up for the Silo e The first choice assigns a random place in the rack s and from this place the load finds the first available storage place If no storage place is found the load is send to a Dump place queue e The second choice uses zones The Silo is divided into six different zones Here the algorithm first tries to find a storage place as close as possible to the exits and 29 Simulation of the IKEA Distribution Centre in Peterborough UK secondly it searches for empty places further back in the Silo This alternative minimizes the distance to the exits If no storage place is found the load is send to a Dump place queue P_MoveLoadTosSilo
24. dary data collection This project is unique and therefore there have not been any performed investigations before 2 6 Source criticism 2 6 1 Background We have performed a hybrid examination between a quantitative process and a qualitative process and need to take considerations from this point of view regarding our source of material for information gathering In qualitative examination processes the influence of the researcher cannot be stated merely enough In a quantitative examination process the statistical character of the information would suppress our personal and more subjective 18 Eriksson L T och Wiedersheim Paul F 1991 Att utreda forska och rapportera Liber Malm Simulation of the IKEA Distribution Centre in Peterborough UK views There is also the potential risk of misunderstanding the respondent in some way Also there is a more vital loss in the information gathering process were the respondent subconsciously or not is hiding information from us 2 6 2 Criticism of primary data The people we have been in contact with during our research and examination period have been very forthcoming and interested regarding the interviews Most of the questions were answered without hesitation during the interviews Though taken into consideration is the respondent s position within the company The reason for this is that the respondent does not want to relate bad will towards the company nor himself herself On
25. divided into two zones no illustration picture are shown The loads are placed randomly all over the Silo More frequent loads are placed in the first zone in the front The second zone in the back is used for the least frequent loads 35 Simulation of the IKEA Distribution Centre in Peterborough UK Random storage using 3 zones The Silo is divided into three zones More frequent loads are placed in the first zone lower part of the figure below and thereby closer to the exits The middle zone is used for minor frequent loads and the third and last zone in the back are used for the least frequent loads The loads are placed randomly within the zones Figure 16 Random storage 3 zones 36 Simulation of the IKEA Distribution Centre in Peterborough UK 7 Output data analysis This chapter will present the AutoStat analysis In this analysis chapter our model and scenario connections will be put together There will be a minor discussion under each headline with the objective to illustrate the main questions and to present the authors own findings 7 1 Analyse Scenario 1 Number of Automated Cranes Analysis of Cranes 210 00 200 00 190 00 180 00 170 00 160 00 150 00 140 00 130 00 120 00 110 00 100 00 90 00 80 00 70 00 60 00 50 00 40 00 30 00 20 00 10 00 0 00 4 5 6 7 Number of Cranes l Handled Flow Silo Units Hour I Demand Flow Silo Units Hour
26. e B 1994 F retagsekonomisk metodl ra Studentlitteratur Lund Arvidsson P amp Rosengren K E 1992 Sociologisk metodik Liber Lund Backman J 1998 Rapporter och uppsatser Studentlitteratur Lund Banks J 2000 Getting Started with AutoMod Brooks AutoSimulations Division Brooks Automation Inc 2001 AutoMod User s Manual Bountiful Eriksson L T och Wiedersheim Paul F 1991 Att utreda forska och rapportera Liber Malm Holme Idar M amp Solvang Magne K 1997 Forskningsmetodik Om kvalitativa och kvantitativa metoder Studentlitteratur Lund Thur n T 1991 Vetenskapsteori f r nyb rjare Runa Winter J 1973 Problemformulering unders kning och rapport Liber Electronic sources Ikea Intranet Thrapston 2002 10 15 Brooks AutoSimulations Division www automod com 2002 11 10 Interviews Thrapston Distribution Service e Darrell Harvey Logistic Manager e Adri R Kraa Business Support Manager e Miranda Stephens Logistic Trainee e Nick Ridley DC layout Doncaster Distribution Service e Keith Morris DC Manager Appendix Input data prognosis Dev Type4 Dev Type3 Dev Type2 Dev Deviation Amount per week Type1 Truck Arrive Every h 10 10 10 10 14 14 12 12 22 88 18 90 9 90 12 84 29 87 22 88 29 87 11 88 TYTOON Ox MO o T NYO ONO 152 Total Dev Type4 Dev Type3 Dev Type2 Dev Truck R
27. e in order to establish further confirmation in our quest for correct and relevant information The surveys have been carefully assembled and the aim was twofold we searched for either new information or either to confirm or reject the thesis from our interview process The surveys have been made continuously throughout the project We have sent the surveys to several persons with the aim of receiving correct information This strengthens the report s validity and reliability from the perspective of information gathering see Validity and Reliability From this point we feel very confident that we still reach a high level of validity and reliability in our analytical process of the report 2 5 1 2 CAD drawings The translation between the CAD drawings and AutoMod is of great importance To make the simulation true the measurements need to be correct In order to attain the main goal see Objective we have done a careful examination of the CAD drawings of the DC in Peterborough Our measurements are carefully implemented so we find that the graphic in our model is equal to the CAD drawings To make the model and the simulation as close as possible to the conditions in real life we have been in close contact with our tutor at IKEA this to increase the correctness of the method 2 5 2 Secondary data Secondary data is material that has already been produced in books magazines newspapers and articles etcetera We have minimized the need of secon
28. etrieve Every h Deviation Amount per week Type1 10 10 10 10 10 14 14 14 12 12 12 DA O O ORO O O OO O O O O OO OO O 0 O SS Sf o OT QO NILO OL re Tr Orr NY Ora O DD LO O N OM O LO O CGN c 2 F LO ON ODOT 151 Total JUNOJOTISHXASPeoT d O ISpeO ppnOSAo N d AeQMmoTpeoTINQoaow d ABtgmoTO peo AON d jJUnoDOJISI9JUASPRLOT d MOTOI GIBVIOISIOS d Anoedeppeo40 d 87 BULALINYpeoT d TIW ABUIYI0AIIS A PUDHOJ9ALIY d 8 7 SALIV NI d Z ZUIALLIYADNIL d PS d pda osuu d UOUESI EDIUI por Je goJ Aedd d MITE 99t 43UNDO IAS A ainjonis uleJJBo d uonounj yo Fred wajsAs A88 uonounj YO YIOM WASS ABB amnpaso1d YaIeas ASP WAS S ABE uonounj YO YIIRAS peo Yon wid MUDHOJ ALHIV d X9NIJINSIWO GIAN d A A EW PASA TA A AE VELA BIRTE Pp Re Seg gE eg ea ye I o xopu OTIS TO 19P10 J t 4 4 MUNUOO 0 BEMOT TO PIO 19JUNOJO ISWOJASPLO PURI d oe dsurniood1oS A SpeEOTIAIININ d ZOO ASULJOoIS A Questionnaire CAD DRAWING We have opened the CAD drawing from an AutoCAD viewer Some information where missing Which parts of the DC are we suppose to focus on The DC high bay racking R G 86 DC low bay racking R G 83 DC low bay racking R G 84 The CDC DC low bay racking R G 85 Or all of them At t
29. f the model including system procedures and algorithms Further on the experimental design show chosen scenarios and in data In the output analysis we present the results from AutoStat We started to analyse the parts using a more integrated form of mind not to loose our main objective Further after a thorough analytical performance we went on to our final discussion thoughts and conclusions Finally we discuss the future work and approach of this project Simulation of the IKEA Distribution Centre in Peterborough UK 2 Methodology In this chapter we will describe the scientific approach the scientific method and which method the authors have chosen to work from Classic instrument are presented to facilitate the understanding for the readers who are not familiar with these scientific conceptions and also to verify our choice The authors will show how these instruments have been used to secure the quality and the correctness of the report Further we present how we have gathered the data for the report followed by criticism of the sources 2 1 Scientific approach Science is about to interpret and understand Through science It is possible to create understanding and knowledge about the reality we live in To be able to examine and explain the reality we have to do assumptions about how this reality looks like The assumptions made results in different approaches Today there are two different main approaches positi
30. fication and thereby an increased understanding of the simulation system 4 1 System simulation To simplify the simulation approach we divide the system into three parts The first part is the inbound logistics the second part is the material flows inside the DC and the third part is the outbound logistics see Figure 1 2 Material flows inside the DC 1 Inbound logistic 0 3 Outbound logistic Figure 1 Schematic drawing of the system 4 1 1 Inbound logistics The simulation system begins with the inbound system The inbound logistic begins when trucks arrive to the Distribution Centre DC The frequency of the arriving trucks depends on the given prognosis The trucks drive to an empty inbound dock 4 1 2 Material flows inside the DC At the dock the unloading procedure starts Fork lifters from the DC start to unload the pallets in the truck Depending on what kind of article load type that is unloaded the fork lifters drive the load to a specified pallet location either into the Silo which is an automated high storage area or to the Lowbay which is a conventional storage area see Figure 2 16 Simulation of the IKEA Distribution Centre in Peterborough UK Depending on the specified prognosis or specified demands for the outbound loads the demanded or outbound load is picked up by a fork lifter which drives from the pallet location to an outbound dock 4 1 3 Outbound logistics At the outbound dock
31. from an external data file or attached to a statistical distribution AutoMod has predefined random distributions that can be used to fit most real world random events Loads can have attributes such as colour priority and time in the system Attributes can be accessed and modified in AutoMod s action statements Loads have 3D shapes and dimensions like most other entities in AutoMod 3 4 3 Resources Resources in AutoMod are used to represent machines operators fixtures containers and any other finite capacity objects There are two default categories for a resource s state The first is the working category busy or idle and the second is the availability category up or down During the animation state colours indicate the status of each resource Statistics are automatically collected for every resource in a model Loads use resources for specified processing times Resource cycles can be created and attached to specific resources to indicate when or how often events such as random failures 3 4 4 States The user can define states other than the default states These states might be used to represent conditions such as blocked starved offline etcetera The state of a resource can then be changed using AutoMod s actions so that statistics can be tracked for each state In 2 Brooks Automation Inc 2001 AutoMod User s Manual Bountiful 27 qh Ibid 11 Simulation of the IKEA Distribution Centre in Peterboro
32. he meeting here in Lund you mentioned a distribution 60 20 20 between the different racking systems Did you mean 60 to the DC high bay racking R G 86 20 to the DC low bay racking R G 83 20 to the DC low bay racking R G 84 To be able too draw the DC as well as be able to do a proper simulation we need some dimensions Height of the racks Number of tters storage places up How high m per tier height per storage place Number of bays per rack storage places along the rack Width per bay width per storage place Depth per bay storage place Which docks what are the dimensions of a dock are intended for the incoming flow loads and which ones are intended for the outgoing flow loads As we understand there are cranes SRM s in AutoMod only in the DC high bay racking R G 86 and one crane support 4 aisles Is this accurate Do the other sections use fork lifters trucks We will need to have data of the SRM s and the fork lifters E g velocity pickup time set down time IN DATA For the capacity check SILO Are there any articles that have to be addressed to a specific rack zone or do it not matter Is the arrival fixed or with probability per day What kind of arriving departure procedure How many trucks are incoming day We can set up to 8 different trucks How many trucks are outgoing day We can set up to 11 different trucks How man
33. he random situation is actually a special case of the above described situation when the fork lifters are operating in zones Instead of four zones see Figure 5 the DC is divided into one zone only This means that the assignments are spread all over the DC The work list appears as follow see Figure 7 Work list Randomly No work Park 1 2 1 Inbound docks 2 Silo outbound Work found Work found y Load moves to v Load moves to SILO LOWBAY y Dock Silo oubound Silo oubound 1 Inbound docks 1 Inbound docks Figure 7 Work list ForkLift1 822 5 1 2 3 Lowbay Conventional storage The Lowbay area or conventional storage area is divided into two parts in AutoMod the visual part and the numeric part e Visual part The physically the racks is done as ASRS systems These systems are not active they are static it is just for the eye To be able to show the loads we used a conveyor system with the velocity zero Every station in that system represents one storage place in the Lowbay area The dynamic part is the Loads in the racks e Numeric part The Lowbay area in the numeric part is simulated as a matrix to keep the computer power at minimum 23 Simulation of the IKEA Distribution Centre in Peterborough UK 5 1 2 4 Highbay Silo Inbound flow There are four inbound point to the Silo one point in every corner The four points are inbound stations In a conveyor system which serves the Sil
34. id 12 Simulation of the IKEA Distribution Centre in Peterborough UK 3 4 9 Labels AutoMod provides the ability to use text labels to enhance model understanding Labels may be added to any location in the model s physical space and they can either be static or dynamic during the simulation Labels can rotate when the animation view changes or can be attached to a fixed position on the screen Tables in AutoMod supply the user with the means to collect statistics on any model entity and to classify those statistics for better understanding of their distribution Tables automatically provide average standard deviation maximum and minimum for all values entered in the table 3 4 10 Types and Random Streams Types in AutoMod provide users with a powerful tool for creating among other things lists of entities such as resources or stations and then making complex decisions based on the contents of the lists Random streams may be defined to give each element of randomness independence from other elements The number of different random streams used by AutoMod models is without limit The current version of the software 10 0 supports the Linear Congruential Generator LCG 3 4 11 Run Control Run control in AutoMod allows users to define the warm up and steady state periods for the model Reports can be printed for any run control period Business graph output can be automatically created Post processed animation periods use
35. ifter2 Referring to figure 8 above e ForkLifterl are operating in the upper part of the DC and they are serving dock 46 to 78 They use Park 1 as their parking lot e ForkLifter2 are operating in the lower part of the DC and they are serving dock 1 to 45 They use Park 2 as their parking lot The DC storage can be divided into four different zones as a maximum The fork lifters are operating in zones in their work list to create a more foreseeable material flow in the DC The zones decrease the distances for the fork lifters They also minimize collisions in the DC The fork lifters are ranking their works in order of priority The work list for ForkLifterl appears as follows see Figure 6 21 Simulation of the IKEA Distribution Centre in Peterborough UK Work list Using zones 1 Inbound docks 2 Silo outbound Work found Yo Load moves to O 2 SILO LOWBAY Look for work at No work found Silo outbound ZONE 1 ZONE 2 ZONE 3 ZONE 4 Silo outbound Silo outbound Silo outbound Silo outbound 1 Inbound No work Work found TH Load moves to y Dock 1 Inbound docks Figure 6 Work list for Forklift1 The work list for ForkLift2 is almost identical to the work list for ForkLiftl Though with some minor differences ForkLift2 uses Park 2 as their parking lot and they serve dock 1 to 45 instead 22 Simulation of the IKEA Distribution Centre in Peterborough UK Randomly T
36. in AutoMod The path mover system is build separately The arriving trucks can be varied up to eight different kinds of trucks The variation represents different trucks from different countries Depending on what prognosis we have the arriving frequency will vary Likewise what kind of load types the arriving trucks contains will vary 5 1 2 The Peterborough DC 5 1 2 1 Inbound dock The docks are represented as queues in AutoMod This is the meeting point for the arriving trucks in the outside path mover system and the fork lifters in the inside path mover system The loads are moved from the trucks into the queues by a dispatched fork lifter Well at the queue the load is waiting to be picked up by the other fork lifters 20 Simulation of the IKEA Distribution Centre in Peterborough UK 5 1 2 2 Fork lifters As mentioned above the fork lifters in the DC are simulated in a path mover system in AutoMod It is only one path mover system in the DC This means that we have not discerned the inbound flow from the outbound flow in the DC This is done to be more flexible for changes to minimize the paths and thereby the computer power The fork lifters are working either In demarcated zones see Figure 5 or randomly all over the DC Fork lifters operating in zones Dock 46 78 Dock 1 45 Figure 5 Schematic grouping into zones Using zones The fork lifters are divided into two different types ForkLifterl and ForkL
37. istic New VR Graphics allows creation of visually accurate 3D scenes in their models making them more convincing Additional graphical enhancements include templates for conveyors and path movers that provide realistic graphics for widely used material handling systems Material movement systems include 2 e Path Mover vehicle systems such as fork lifters and AGV S e Conveyors e Automated Storage and Retrieval Systems ASRS e Robots e Bridge Cranes Brooks Automation Inc 2001 AutoMod User s Manual Bountiful 2 Banks J 2000 Getting Started with AutoMod Brooks AutoSimulations Division 2 Brooks Automation Inc 2001 AutoMod User s Manual Bountiful Simulation of the IKEA Distribution Centre in Peterborough UK e Power and Free Chain Conveyors Tanks and Pipes To define material movement systems the user simply defines movement elements such as paths and stations and then inputs the operating parameters such as velocity and acceleration AutoMod then automatically creates the corresponding model logic 3D animation is created automatically as well providing a realistic picture of how a facility will look and operate Model animation can be viewed from any angle or perspective in real time providing visualization capabilities unmatched in other simulation tools AutoMod s key strengths are e Flexibility to model complex systems accurately e Unlimited size of models e High performance simulatio
38. n engine e Bestin class statistical analysis features e Graphic environment for creation of geometry e True to scale import from CAD tools The model execution environment is very interactive allowing the user to stop and start the simulation run without animation in an accelerated time scale and select objects from the animation screen to gather detailed statistics about the system being modelled Statistics can be viewed at any time during a simulation run These model execution features make it easier to verify and validate models of complex systems 3 3 AutoMod interface An AutoMod model consists of one or more systems A system can either be a process system in which flow and control logic are defined or a material movement system Each model must contain one process system and may contain any number of movement systems Loads can move between processes with or without using movement systems All inter arrival and event times can be represented by deterministic values or be derived randomly from one of many statistical distributions AutoMod s interface is window oriented utilizing pop up and pull down menus 3 4 AutoMod s world view Any number of movement systems can be defined in an AutoMod model and a process system connects the movement systems to the logical flow of products In the process system loads articles parts etcetera move between processes locations and compete for resources equipment operators and q
39. nario 2 6 3 2 1 Storage mode According to the scenario described above we will use the optimal number of Automated Cranes for the following scenarios Random mode Using the random method the loads in the Silo are placed with a random algorithm The algorithm finds a location randomly If that specific storage location is vacant the load is placed in that location If the specific location is occupied the algorithm searches for a location next to the previous until a vacant location is found see figure 15 below Systematic mode The systematic algorithm is shown below see Figure 14 The Silo Is divided into two zones According to the frequency the loads with the highest frequency are placed in the front and the loads with minor frequency are placed in the back The algorithm finds the bay nearest the exit still within the zones Then the load is placed from the bottom location to the top location within the bay When the bay is full the algorithm searches in the next bay and so on Figure 14 Systematic storage method 34 Simulation of the IKEA Distribution Centre in Peterborough UK 6 3 2 2 Number of Zones Random storage using 1 zone The Silo is not divided see Figure 15 One zone is the total area of the Silo The loads are placed randomly all over the Silo Twenty thousand loads are the initial value Figure 15 Random storage 1 zone Random storage using 2 zones Like described above though the Silo is
40. ngsmetodik Om kvalitativa och kvantitativa metoder Studentlitteratur Lund 12 Andersen H 1990 Att skriva vetenskapliga rapporter Studentlitteratur Lund Simulation of the IKEA Distribution Centre in Peterborough UK It is of the outmost importance that the material made for the interviews is gathered and structured in a thorough manner In this course of action we aimed for maximum level of preparation before the interviews we were ready and prepared for our aim establishing a high level of validity and reliability see Reliability By preparing in this way and conducting the research process thoroughly we also minimised potential misunderstandings which is of great importance regarding validity and reliability The formation of the questionnaire was a result of this process The questions were of an open nature The reason for this was that open questions are not likely to lead the respondent s answers in any direction thus giving way for better and more reliable validity and reliability The validity indicates the correctness of a method In order to obtain a high level of accuracy information has been gathered thoroughly Surveys has been made and sent to a number of persons relevant to our studies and our objectives of this report 2 4 Reliability Reliability means the degree of authenticity within a measure which is the lack of random errors Another definition is In what degree the results from a measure are n
41. nimation and a flexible simulation language to create realistic simulation models of manufacturing facilities material handling systems and Distribution Centres DC s Through experimenting with a facility in the virtual computer world optimal configurations equipment specification and operational improvement can be determined for implementation in the real world before any change or capital outlay is made Simulation is the imitation of a real world process over time Simulation is an indispensable problem solving methodology for the solution of many real word problems Simulation is used to describe and analyse the behaviour of a system ask what if questions about the real system and aid in the design of real systems Both existing and conceptual systems can be modelled with simulation Whether designing a new system or modifying an existing one engineers want to take the estimation out of finding the best possible solution While there are many analysis methods for designing industrial systems simulation remains the method that provides the highest level of confidence that a system will perform A well written simulation model can be a valuable tool in the design analysis and operation of manufacturing and other complex systems 3 2 Background The main focus of the AutoMod simulation software is on manufacturing and material handling systems With the release of AutoMod version 10 0 in 2001 graphics in AutoMod are more real
42. nowledge However this does not prevent people with interests in the same field to read the report Interview Darrell Harvey Logistic Manager 2002 10 10 Simulation of the IKEA Distribution Centre in Peterborough UK 1 6 Structure of the report Structure scheme of the report 1 Introduction 2 Methodology 3 Theoretical frame 4 System 5 Modelling description 6 Experimental design 7 Output analysis 8 Final discussion Figure 1 Structure scheme of the report In the introductory to our choice of subject we initialised our main course of structure for the report as demonstrated above The introductory part of the report was discussed and was structured as the foundation where we wanted to build the forthcoming report This first chapter also includes the main building blocks for the report together with the methodology chapter Regarding the methodology chapter we aimed for a hybrid approach This chapter was then structured to explain the certain more mandatory parts of the report Important in this chapter is our criticism regarding our research method and our sources of information This demonstrates our awareness regarding possibilities and difficulties when writing an academic report The main purpose with the system description is to give the reader a quick overlook and to simplify the understanding of the model and simulation Regarding the modelling chapter this is a thorough description o
43. o Silo The Silo is simulated with two ASRS systems in AutoMod The Silo is fully automatic which the ASRS systems fully represent see Figure 8 and Figure 9 When the load travel on the conveyor and enters the Silo there are two SRM s working across the aisles and serving a buffer The buffer rack is also placed perpendicular to the other racks in the Silo The buffer Is simulated with an invisible conveyor system From the buffer the load then can enter the real storage The Silo is also divided in one visual part and one numeric part just like the Lowbay see 5 1 2 3 Lowbay Conventional storage Silo 3D side view i A s Load Figure 8 3D view of the Silo 24 Simulation of the IKEA Distribution Centre in Peterborough UK Schematic top view of the Silo O Station in e Station out EZ Eli Conveyor Figure 9 Silo Top view There are six possible attributes which can be set for the Silo This means that two different choices can be set The first choice sets the Silo to one zone all over which means when a load enters the Silo the load can be placed randomly all over the Silo The second choice divides the Silo Into six different zones One attribute is equal to one zone This means that frequent articles loads could be placed as close as possible to the exits and this implies shorter distances and thereby increased efficiency and increased throughput see Figure 10 25 Simulation
44. of the IKEA Distribution Centre in Peterborough UK Side view of the zones in the Silo Figure 10 Silo divided into zones Outbound flow The outbound flow is simulated with a conveyor system The number of conveyors and outbound stations is equal to the number of aisle in the Silo 5 1 3 Outbound logistics 5 1 3 1 Outbound docks The docks are represented as queues in AutoMod This is the meeting point for the fork lifters in the Inside path mover system and the departure trucks in the outside path mover system The fork lifters in the DC dispatch the loads to the queues At a dock one single fork lifter continues to dispatch the loads into the truck 5 1 3 2 Departure trucks The departure trucks are simulated in the same path mover system as the arriving trucks Though there are eleven different departure trucks each represents different stores in UK 5 2 Graphic description The graphic in the different systems in AutoMod represent only a minor part in the simulation Most of the graphics in the model are built and imported from ACE The graphic is unique and customized for the DC The Peterborough DC is mainly built with static graphics The only dynamic graphics are the fork lifters and the SRM s in the DC 26 Simulation of the IKEA Distribution Centre in Peterborough UK 5 3 Procedure description To make it understandable we will start to illustrate a schematic program structure on a basic and
45. ot inflicted by accidental occurrence Previous to the interviews we discussed and considered different kinds of problems that could inflict and affect our attempt in gathering relevant and correct information from the respondents It is of great importance that the respondent correctly understands the questions asked from our side By asking the same question but in different ways and by formulating it differently we can secure the consideration of the respondent s level of understanding the question to some extent we can also verify the answer in this way The method mentioned above has been used throughout the interview process in aim of reaching the highest possible reliability for the report We will have the input data and information needed from already existing CAD drawings done by IKEA The reliability in the input data and measurements are thereby hard to secure However we have been in close contact with IKEA explaining how the program is working and in what form we wanted the input data Hence we feel confident in the reliability of the input data and measurements 2 5 Method of collection There are two main approaches when collecting data The first technique is collecting of primary data and the other one is the collecting of secondary data 2 5 1 Primary data The primary data is the material that the researcher collects him self for the settled purpose with the examination In order to do our simulation as true
46. pend the simulation when a certain event occurs or when a specific clock time is reached AutoMod also provides comprehensive reports The reports can be displayed on request at any time during the simulation 3 8 AutoStat The AutoStat product is the statistical analysis tool in the AutoMod product family AutoStat does extensive output analysis on an AutoMod model including e Confidence intervals e Warm up determination e Sensitivity analysis e Factor response analysis e Design of experiments 14 Simulation of the IKEA Distribution Centre in Peterborough UK e Optimisation using evolution strategies AutoStat also has support for running scenarios across a network of machines Support for multiple machines and CPU s give users the ability to make many more runs of the simulation than was possible before The evolution strategies algorithm used by AutoStat is well suited to finding the optimum solution without getting trapped at a local optimal value 8 Brooks Automation Inc 2001 AutoMod User s Manual Bountiful 15 Simulation of the IKEA Distribution Centre in Peterborough UK 4 System description This chapter is mainly intended to simplify the understanding of the simulation system The chapter starts to describe what approach the authors have chosen to work from where the simulated system will start and where it ends The chapter continues with a layout description to glve a visual simpli
47. re in Peterborough UK 6 Experimental design This chapter describes the input data The chapter also contains different scenarios aimed to attain our problem definition and our objective The first scenario will investigate the optimal number of Automated Cranes for the best throughput in the Silo This scenario underlies further analysis of the Silo The following scenarios will depart from this optimisation and Investigate how different storage mode will affect the throughput 6 1 Input data prognosis For figures of input data view Appendix Input data prognosis The run control of the simulation is set to Run time Six days a week based on a sixth teen hours working pattern In the weekdays and eight hours shift on Saturdays Eighty eight hours total a week 6 2 Scenario structure The best solution are used for Scenario 2 Combining Storage mode and Number of Zones Number of Zones Storage mode 1 3 Zones Random or Systematic Figure 15 Structure of the scenarios 32 Simulation of the IKEA Distribution Centre in Peterborough UK Common assumptions for the scenarios e Both scenarios start with loads in the Silo and the conventional stacking area There are already twenty thousand loads in the DC Randomly spread e The DC utilizes mixed inbound and outbound docks e The security level of the Silo is set to ninety five percent e The fork lifters are operating in zones e FIFO rule is used e
48. roduct quality and faster deliveries This among other things increases the demands on the distribution systems To be able to satisfy these increased demands IKEA needs to improve and develop their distribution systems and storage facilities The IKEA distribution strategy in UK is to build Distribution Centre s DC s at two additional sites over the next two to three years The first of these two are the DC in Peterborough The site is placed in Peterborough in order to take advantage of the planned Customer Distribution Centre CDC that has potential synergies The DC will be similar in design to the Doncaster DC It will be a combination of conventional storage and automated crane based Silo IKEA have a clear construction plan including the physical layout for the implementation but during the building process they want to do a capacity control of the DC 1 2 Problem setting IKEA wants to avoid building a DC with insufficient capacity and with imbalances in the material flow therefore they want to do a simulation of the system to see if the material flow satisfies the requirements and to see if the DC s capacity is sufficient The focus in this report is towards capacity simulations of the Silo the automated high storage area 1 3 Objective Our main goal with this project is to do capacity checks of the Silo and look for bottlenecks in the material flow where maximum efficiency and productivity is of first priority by the case
49. sidering the variation of the number of zones we consider the zoning could be done in a different way Because of the fact that there are not SRM s in every aisle the cranes have to change aisles Using the zoning from the front to the back as described the cranes have to move a long distance when changing aisle for the fast moving articles in the front of the Silo Instead we believe the zoning should be done from the left to the right or the way around and the cranes in the fast moving zones should have a crane in every aisle The other zones could share the other cranes Doing this we believe the time it takes to change aisle could be minimized When looking at the staple diagram all the staples show a slight insufficiency considering the capacity We though believe a longer simulation time would equalise the staples These analyses show only a small part of what can achieved with the simulation Many other things and factors can be varied and calculated upon but this is a project for future studies 40 Simulation of the IKEA Distribution Centre in Peterborough UK 9 Future approach This chapter will present the authors findings of future work The chapter is divided Into two parts which is two different approaches The first part present future work within the capabilities of the model and the second part presents future expansion of the model The model is build to be able to simulate the total Internal material flo
50. steori for nyb rjare Runa 9 Eriksson L T och Wiedersheim Paul F 1991 Att utreda forska och rapportera Liber Malm 7 Thur n T 1991 Vetenskapsteori for nyb rjare Runa Simulation of the IKEA Distribution Centre in Peterborough UK 2 2 Scientific methodology You can discern two different scientific methods The starting point is how you examine soft data or hard data these are called qualitative methods or quantitative methods 2 2 1 Qualitative method The qualitative method is distinguished of a close contact to the science object which leads towards increased understanding of the reality which is studied The strength with the qualitative method is that it renders a possibility for increased understanding in social processes and social consistence A characteristic of the qualitative method is flexibility That is because of the possibility to change the approach during the implementation Another characteristic is that the qualitative method does not use any hard data which is numbers or figures 2 2 2 Quantitative method Studies with the quantitative method should be able to be measured and the result should be able to be presented with numeric data Experiments and surveys are examples of quantitative methods In opposite to the qualitative methods the quantitative methods are inflexible standardized approaches are often used New knowledge uprising during the implementation must not lead to changes in
51. the contrary companies tend to reflect themselves in a more positive manner It is important to be aware of this as researcher and interviewer The input data for this project has been collected by IKEA This is important to point out Concerning the translation from the CAD drawings into our model we aimed for maximum level of equality and we feel confident in the accomplishment Though the absolute best result could be established if there would have been a direct translation between AutoMod and AutoCAD 2 6 3 Criticism of secondary data As described we have minimized the need of secondary data Though scientific articles and literature within the area of our objective dominates the chosen secondary data We did this to secure the correctness and reliability of the collected information We do consider the secondary data we use Is covering the area selected for the report s problem setting in a satisfying way Simulation of the IKEA Distribution Centre in Peterborough UK 3 Theoretical Frame In this chapter we want to establish a frame of mind that will enhance the special aspects which should be taken into consideration when simulating in AutoMod The content in this chapter though is mainly presented as an introductory foundation of simulations using the AutoMod software 3 1 Introduction AutoMod is simulation software for material handling systems It helps design and optimises facilities It combines CAD 3D a
52. ueues The load is the active entity executing action statements that are connected to the processes Typical action statements give users the ability to 2 e Use machines operators Brooks Automation Inc 2001 4utoMod User s Manual Bountiful Ibid Ibid gt Ibid 10 Simulation of the IKEA Distribution Centre in Peterborough UK e Move into queues e Clone new loads e Change load types e Wait on user defined delay lists e Increment decrement counters e Set variable values e Read from data files Send to other processes 3 4 1 Processes The process system is the spine of an AutoMod model providing modelling a wide range of real world problems While material movement is important it is not always critical in manufacturing No value is added to product while it is being moved around a manufacturing facility Machines and people perform value added operations AutoMod s process system includes a simulation language based on action statements Action statements combine the power of a structured language with the ease of use of English like manufacturing oriented syntax AutoMod models are not limited in any way so model logic can be of any size and complexity The power and flexibility of the AutoMod language makes is easy to model almost any real world situation 3 4 2 Loads Loads are the active entities in AutoMod and can be created in many ways The inter arrival rate for loads can be read
53. ugh UK addition state monitors can be defined and used to track states for entities other then resources such as vehicles conveyors or particular areas of a facility 3 4 5 Queues and Order Lists Queues in AutoMod are both graphical and statistical elements and they can have any user defined capacity When a queue has reached its capacity the next load trying to enter that queue must wait until there is space available Queue contents can be shown dynamically in the animation and loads can be stacked in any direction in the queue Loads that are at a queue or process may be sorted or delayed until they are explicitly ordered to leave To determine which action should take place next the loads must place themselves on order lists An order list OL is not a physical entity like a queue but a logical element that provides a way to sort loads that have been delayed for any reason To remove a load from an order list another load must execute an order action Loads can be ordered to move to another process or order list or to simply continue where they left off in their processing Order lists can be sorted by load priority or other load attributes either in ascending or descending order 3 4 6 Blocks Blocks control the number of entities occupying a physical space making them very useful for controlling path mover vehicles Blocks may have any capacity from one to infinity Path mover vehicles fork lifters AGV s etcetera
54. vism and hermeneutic 2 1 1 Positivism Positivism is built upon experiments quantitative measurements and logical discussions The positivism describes and explains which means observation and examination The positivistic consider an objective reality which is depending on the viewer and can be studied without being influenced This means that the positivism have an objective approach 2 1 2 Hermeneutic The hermeneutic studies demand an entire understanding an insight Hermeneutic it self means learning about interpretation Unlike the positivism which has an objective approach the hermeneutic approach is subjective Within the hermeneutic approach our own understanding and our own experience is used to interpret and understand other people s understanding and experience In other words it is our own interpretations of the reality that control how we act 2 2 2 Choice of scientific approach We have chosen a positivistic approach Our project is unique and built upon a concrete problem Further on our measurements are of quantitative characteristic and our project is also built on logical discussions and because of the fact that our conclusions are objective a positivistic stance has been selected 3 Arbnor I amp Bjerke B 1994 Foretagsekonomisk metodliira Studentlitteratur Lund Eriksson L T och Wiedersheim Paul F 1991 Att utreda forska och rapportera Liber Malm Thur n T 1991 Vetenskap
55. w of the Peterborough DC In this report though we have chosen to focus on a minor part of the DC The Silo is isolated and it is only this part that we have chosen to analyse in AutoStat The future approach we would like to see is continued simulations and measurements of other parts of the system We think the following questions of issue and investigations could be of interest Future analysis Run capacity analysis of the dedicated racks in the Lowbay area Is the capacity of the dedicated areas enough Possible bottlenecks Run analysis of the fork lifters What happen when they are working all over the DC What happen when they are demarcated into zones Number of fork lifters needed in the DC Utility What happens with the input output Mixed inbound outbound What happens with the fork lifters pattern Total material flow Future model expansion A future model expansion and approach would be to investigate how picking inflict on the material flow Almost thirty percentages of the total pallets are picked pallets in the material flow This has a big impact on the material flow and we see this as an interesting area to look into Another suggestion of future works and expansion of the model would be to take the Customer Distribution Centre into consideration 41 List of reference Published sources Andersen H 1990 Att skriva vetenskapliga rapporter Studentlitteratur Lund Arbnor I amp Bjerk
56. xt tier above and so on This alternative minimizes the distance to the floor and thereby the docks If no storage place is found the load is send to a Dump place queue P_MoveLoadToLowBay This procedure moves the load from the dock to a specified storage place in the Lowbay area It all takes place in the same path mover system inside the DC Order List OL When a load has found a storage place in the Lowbay area the load is placed in an Order List Here the load is waiting to be called for 30 Simulation of the IKEA Distribution Centre in Peterborough UK 5 3 1 3 Outbound procedures P _MoveOutLoadSilo After the load is being claimed this procedure moves the load out of the Silo The load is moving in the backward order referring to P MoveLoadToSilo though with the slight difference of exiting the Silo through the exit stations in the conveyor system P_MoveOutLoadLowBay This procedure take the load from the control point in the Lowbay area and moves it in the path mover system inside the DC to an outgoing dock control point P_RetrieveLoads Sets the amount of outgoing loads for the domestic trucks depending on truck load type and country origin Finds a suitable and free dock for the truck P_MoveDomesticTruck Moves the truck in the path mover system outside the DC to the dock Here the truck waits to be full according to the order and then it drives away 31 Simulation of the IKEA Distribution Cent
57. y loads are in each truck We can set up to 4 different loads for each different truck both outgoing and incoming fixed number or with deviation for instance 28 or 24 4 For the outgoing trucks we can also set the country origin for the load The number of racks and location of the racks that are dedicated for each country How do the loads enter and leave the silo The CAD drawing and the excel sheet differ from each other concerning the flow of gods in We have trouble to interpret them Is the entering and leaving processes to the silo taking place from two directions PD stands where How long time does it take for a crane to change aisle Referred to the CAD Drawing Excel sheet note The excel sheet Is upside down compare to CAD drawing O OZ gt m gt O Kx Q s 2 N S o E Which of these schematic drawings above are most accurate Maybe you can give an example of the transfer for one load entering process transfer process storage process leaving process Can you explain how the flow of goods to the Silo Is working please What function do the conveyors have Where can a load enter leave the silo
Download Pdf Manuals
Related Search