assembly line balancing and optimization using siemens

Assembly Line Balancing and Optimization Using SIEMENS Tecnomatix Plant Simulation: A Case Study of an Electric Power

Steering Column Assembly Line

by

Zay Yar Myint

A thesis submitted in partial fulfillment of the requirements for the degree of Master of Engineering in

Industrial and Manufacturing Engineering

Examination Committee: Associate Prof. Erik L.J. Bohez (Co-Chairperson) Dr. Huynh Trung Luong (Co-Chairperson) Dr. Than Lin Dr. Mongkol Ekpanyapong

Nationality: Myanmar Previous Degree: Bachelor of Science in Industrial Engineering Previous Degree: Asian Institute of Technology Thailand

Scholarship Donor: AIT Fellowship

Asian Institute of Technology School of Engineering and Technology

Thailand May 2017

ACKNOWLEDGEMENTS

I would like to present special thanks to Associate Professor Erik L.J. Bohez, co-chairman of examination committee for his constant, expert guidance, encouragement and valuable suggestions from the very beginning of the thesis. I would like to express special thanks to the exam committee members, Dr. Huynh Trung Luong, co-chairman of examination committee for his guidance, Dr. Than Lin and Dr. Mongkol Ekpanyapong for their suggestions and encouragements. I would like to express my special gratitue to Mr. Daniel Reinle, Vice President of Siemens Limited, Thailand, who supported me to use the Tecnomatix Plant Simulation Software and allowed me to have very supportive technical discussions with specialists and host factory. I’m also deeply indebted to Mr. Ittipol P. (SLT), Miss Napa Hoonpanich(Premium PLM Co. Ltd) and Mr. Chaowat Kruttin(Premium PLM Co. Ltd) for their kind helps and technical consultations. Moreover, thanks to all my teachers and lab supervisors for their kind teaching, guidance and help throughout my time in AIT. Finally, words are never enough to express my deep gratitude to my family and friends whose love for their inspiration and encouragements to complete my masters program.

ii

ABSTRACT

In A tier 1 automotive steering supplier company from Thailand is one of the leading steering manufacturing company for variety of automobile brands. They produce tens of thousands of steering columns per month for different models of automobile from different brands. The company’s manufacturing facility exists in Samut Prakan district, Bangkok, Thailand. The production process consists of 4 sections, manufacturing, sub-assembly, final assembly and warehouse. Among these, the final assembly line and the warehouse bear the majority of the inventory cost. In order to reduce the tied up inventory cost in the factory, it needs to find out the optimal inventory set up in the warehouse which can also supply the monthly demand without any delivery shortage.

This thesis presents a case study on an assembly line optimization problem in an existing assembly section of a Power Steering Column production line for Honda Fit. The rationale of this project is to build the model of current system, to simulate the entire process of the final assembly line, analyse the resultant inventory and delivery status of current inventory setup and find out an optimised inventory setup for the minimal inventory cost with minimal delivery shortage. The simulation is done by using Tecnomatix Plant Simulation. All the data needed for the line balancing analysis are collected. This data gathered is then simulated in Tecnomatix Plant Simulation.

Keywords: simulation, assembly line, optimization, tecnomatix, plant simulation, siemens

iii

TABLE OF CONTENTS (Cont’l)

iv

CHAPTER TITLE

TITLE PAGE

PAGE

i

ACKNOWLEDGEMENTS ABSTRACT TABLE OF CONTENTS LIST OF FIGURES LIST OF TABLES

ii iii

iv

vi

ix

1 INTRODUCTION 1.1 Background 1.2 Problem Statement 1.3 Objectives 1.4 Scope and Limitation

11122

2 LITERATURE REVIEW 2.1 Assembly Line Balancing 2.2 Terminology Used in Assembly Line Balancing 2.3 Assembly Line Balancing Problems 2.4 Aspects of Line Balancing 2.5 Discrete Event Simulation 2.6 Modeling and Simulation 2.7 Tecnomatix Plant Simulation

333

4 6 7 8

11

3 CASE STUDIES 3.1 Modelling, Simulation and Optimization of Process Planning (Jovisevic, S., Jovisevic, V., &Jokanovic, S., 2009) 3.2 Kanban Simulation Model Production Process Optimization. (Golchev, R., et.al, 2015)

18 18

22

v

CHAPTER TITLE PAGE

4 DESCRIPTION OF SYSTEM UNDER STUDY 4.1 Description of Product Under Study 4.2 Production Process and Stations

26

26 32

5 MODEL DEVELOPMENT AND

OPTIMIZATION

5.1 Basic Model Development 5.2 Worker Operation and Transport Workpieces 5.3 Verification and Validation 5.4 Pull System Configuration

5.5 Implementation of Production Plan

5.6 Optimization Using Genetic Algorithm Wizard

44

4546

5052

55

57

6 CONCLUSION AND RECOMMENDATION 6.1 Conclusion 6.2 Recommendation

REFERENCES

75

75 75

76

LIST OF FIGURES (Cont’l)

vi

FIGURE TITLE PAGEFigure 2.1 An Example of Precedence Graph 4Figure 2.2 Single Model Assembly Line 5Figure 2.3 Mixed Model Assembly Line 5Figure 2.4 Multi Model Assembly Line 5Figure 2.5 Simplified Version of the Modelling Process 9Figure 2.6 Methodology Flow Chart for Simulation Study 11Figure 2.7 Time-Oriented vs. Event Oriented Simulation 12Figure 2.8 Modeling of Process Planning in 2D Environment 13Figure 2.9 Modeling of Process Planning in 3D Environment 13Figure 3.1 Segment from Simulation Model Shown in 2D 19Figure 3.2 Simulation Model Shown in 3D 19Figure 3.3 Simulation Model Located in the Corresponding

Production Plant20

Figure 3.4 Efficiency of Machines before Optimization 20Figure 3.5 Efficiency of Machines after Optimization 21

Figure 3.6 Methodology for Kanban Implementation 23

Figure 3.7 Basic Concept of the Simulation Model 24Figure 3.8 Experimental Results 25Figure 4.1 Electric Power Steering System 27Figure 4.2 Sensor Shaft Assembly E-Lock 28

Figure 4.3 Gear Box 28

Figure 4.4 Cover Assembly 29Figure 4.5 Motor 29Figure 4.6 Electrical Control Unit 30Figure 4.7 Bracket Harness 30Figure 4.8 Harness Assembly 30Figure 4.9 Cover Connector 30Figure 4.10 Bill of Materials in Final E.P.S Assembly 31Figure 4.11 Electric Power Steering Column 32Figure 4.12 Overview of the Electric Power Steering Column

Production Line32

Figure 4.13 Process Flow in Final Assembly Line 36Figure 4.14 Plant Layout of EPS Column Assembly Line 37Figure 4.15 Kanban System of EPS Column Assembly Section 42

Figure 4.16 Kanban Card in EPS Column Final Assembly Line 43

LIST OF FIGURES (Cont’l)

vii

FIGURE TITLE PAGE

Figure 5.1 Modelling Steps Flow Chart 44

Figure 5.2 Station Allocation 45

Figure 5.3 Processing Time Configuration Window 45

Figure 5.4 Resources Menu Bar 46

Figure 5.5 Layout after Workplaces and Footpaths are Added

46

Figure 5.6 Creation Table of Workers in Workerpool 47

Figure 5.7 Workplace Window 47

Figure 5.8 Supported Service Window 48

Figure 5.9 Importer Window 48

Figure 5.10 Exit Window 49

Figure 5.11 Shift Calendar Window 49

Figure 5.12 WorkerPool Window 50

Figure 5.13 The Correlation Inspection Approach 50

Figure 5.14 Summary of One Day Simulation 51

Figure 5.15 Utilisation of Stations 51

Figure 5.16 Performance Chart of Workers 51

Figure 5.17 Kanban Buffer Window 52

Figure 5.18 Kanban SingleProc. Station Window 52

Figure 5.19 Stock Level Transitions in Kanban Buffer 54

Figure 5.20 Initial Method 57

Figure 5.21 Method for Delivery Control 57

Figure 5.22 Genetic Algorithm Wizard Window 58

Figure 5.23 Optimization Parameter Table 58

Figure 5.24 Fitness Parameter Table 59

Figure 5.25 Optimised Results 59

Figure 5.26 Stock Level Graph after Optimization 60

Figure 5.27 Inventory Level on Day1-Day3 62



LIST OF FIGURES

viii

FIGURE TITLE PAGE








Figure 5.37 Inventory Level on Day31 72

Figure 5.38 Utilisation of Stations in Pull System before Optimization

73

Figure 5.39 Performance Chart of Workers in Pull System before Optimization

73

Figure 5.40 Utilisation of Stations in Pull System after Optimization

74

Figure 5.41 Performance Chart of Workers in Pull System after Optimization

74

LIST OF TABLES

ix

TABLE TITLE PAGE

Table 2.1 Verification vs. Validation 8

Table 2.2 Tecnomatix Plant Simulation License Details 15

Table 4.1 Shift Schedule 33

Table 4.2 Description of the Stations 38

Table 4.3 Time Study of the EPS Column Final Assembly Line

39

Table 4.4 Time Study of the EPS Column Final Assembly Line (Cont.)

40

Table 4.5 Monthly Production Plan of EPS Column 43

Table 5.1 Production Plan Table 55

Table 5.2 Delivery Summary Table before Optimization 56

Table 5.3 Delivery Summary Table after Optimization 60

CHAPTER 1 INTRODUCTION

1.1 Background

After visiting a Tier 1 automobile supplier factory, it was found out that the production line is a hybrid push-pull system in which the Manufacturing section is run in push system and the sub-assembly section and final assembly section producing Electric Power Steering Columns are working on the Kanban System(Pull System) to fulfil the fluctuating demand which causes delivery shortages sometimes. At the same time, types of products produced on it are also changing frequently. The company would like to evaluate the outcome of the current inventory setup and find out the best configuration to reduce the cost. Simulation technology is an effective methodology for planning, implementation, and operating complicated technical systems. Many simulation software are available in market just to implement the virtual layout of the assembly area for example, WITNESS and ARENA software. Though, the software used in this project is Tecnomatix Plant Simulation software developed by SIEMENS. Tecnomatix Plant Simulation can be used to improve the effectiveness and profitability of a facility by finding out the solution increasing throughput and utilization of resource and facility. Plant Simulation is also able to find the ways to reduce lead times, needed resources and inventory requirements provided that are all accurate data fed into the simulation program for analysis. Moreover, Plant Simulation is capable of identifying the bottlenecks, WIP reduction, evaluate the effects of capital investments or process alternatives and prevent planning defects as the simulation was implemented virtually before applying to the reality.

1.2 Problem Statement

To achieve an optimal assembly line, it firstly needs to analyze it to know where are the bottlenecks and drawbacks for the current line. And then try to find the alternative configurations to achieve the favourable results. The company would like to find the optimum assembly line model with optimal inventory setup that is efficient and able to fulfil the planned demands economically by using Tecnomatix Plant Simulation.

1

1.3 Objectives

The objectives for this project are to: a) Simulate the existing pull system assembly line b) Determine the optimal minimum stock level and maximum stock level at the warehouse to avoid

• Delivery Shortage and • Over Stock

1.4 Scope and Limitation

Scope This thesis is based on knowledge of the theory and concept of assembly line modelling and optimization. A case study is conducted at a Tier 1 automotive supplier factory. The data of an existing assembly line, producing Electric Power Steering Column Assembly is collected, modelled and simulated using Tecnomatix Plant Simulation software. The research is done with the help of Premium PLM Co. Ltd. in Bangkok, Thailand. Limitation This study emphasizes only on the final assembly line section and warehouse though the entire manufacturing line contains Production, Sub-assembly, Assembly and Warehouse sections. Though they have many different product types, this product only consider a single product, EPSC(Electric Power Steering Column) for Honda Fit.

2

CHAPTER 2 LITERATURE REVIEW

2.1 Assembly Line Balancing

Assembly Line balancing can be defined as the organisation of a production line to obtain an uniform circulation of production between each station. Assembly Line balancing is also a suitable methodology in eliminating bottlenecks by calibrating the takt time of individual station to avoid delays and overload on each station. According to Falkenauer (2000), Assembly Line Balancing, or simply Line Balancing (LB), is the activity of operation and function appointment to each station of an assembly line to get the optimal arrangement. In another aspect, an assembly line can represent a system of a group of workstations along a line that are arranged systematically. Between each stations, materials and workpieces can be transported by means of human worker or other transportation devices. The assembly functions are undertaken in accordance with the precedence constraints of the assembly process.The decision problem of optimally balancing the assembly work among the workstations is pointed out by M.Baskak. et. al (2008) as the assembly line balancing problem.

2.2 Terminology Used in Assembly Line Balancing

According to Pekin N, Azizoglu M (2008), production of an item on a production line needs to breakdown the overall function into a set of elementary tasks. Terms in assembly line balancing methodology are described as follow. Task A task is the smallest, indivisible work element of the total work content. Task time Task time or processing time is the time required to fulfil a task by a particular equipment. Workstation Specific place in which machines and/or human workers are set to perform particular functions. Work content The work content means the total processing time in a workstation.

3

Precedence constraints and Predecessors

The tasks are arranged in priority according to technical requirements which are also known as the precedence relations or precedence constraints. Processing can only be performed in predefined order. The prior tasks which are needed to perform before a particular task are known as the predecessors of that task. The successors of a task are the tasks that cannot be performed before the completion of this task. The precedence relations can be represented graphically as illustrated in Figure 2.1.

Figure 2.1: An example of Precedence Graph

The nodes in the figure stand for the tasks and line between the nodes i and j exists if task i is an immediate predecessor of task j. In the figure task 3 is the successor of task 1 and 2. And it is the predecessor of task 4.

2.3 Assembly Line Balancing Problems

The classical assembly line balancing problem (ALBP) emphisize the job allocation to the workstations. Major objective of the allocation is the minimization of total cost in the mean time it fulfil the demands obeying the limitations like precedence relations between tasks and some particular system constraints ( Pekin N, Azizoglu M, 2008)

Classifications of assembly line systems. Assembly lines can be classified into: • single-model • mixed-model and • multi-model systems according to the number of models that are • present on the line.

4

Single-Model Assembly lines are the assembly lines that are used to produce single type or model of the product. Mass production is usually done to produce specific products with same dimensions and design. In this type of assembly line, operators provide the same performance when a sequence of products pass them at a constant speed. Mixed-Model Assembly lines are the assembly line that are usually used to assemble two or more different models of the same product at the same time. On the line, the products are continuously changing from one model to another. Multi-Model Assembly lines. Many similar products are produced on one or many assembly lines. In order to perform different production processes, re-organizing the line equipment are necessary whenever the products are changed. Hence, the products are assembled in different batches for minimizing set-up arrangement inefficiencies. While enlarging batch sizes reduces set-up costs, inventory costs are increased. (Scholl 1998)

Figure 2.2: Single Model Assembly Line

Figure 2.3: Mixed-Model Assembly Line

Figure 2.4: Multi Model Assembly Line

5

2.4 Aspects of Line Balancing

In present age, assembly lines tends to transform into cellular manufacturing in the aspect of different production. So that, employment of multifunctional machine, tools and human operators is increasing. To get advantages from continuous productions these machine, tools and operators must be employed to the line in the optimal arrangement to achieve the maximum utilisation, minimum number of stations (Agpak and Gokcen, 2005). In the assembly line design phase, the task list along with times required to complete each task and the precedence flow between them are analyzed. Following goals are set when the tasks are categorised into stations according to this analysis. 1. Minimization of the number of workstations for a given cycle time. 2. Minimization of cycle time for a given number of work stations.

Falkenauer (2000) recorded some difficulties to deal with in line balancing tool for applicability in the industry as follow: I. Workstation cannot be eliminated. As individual workstation has its own features, it is sure that these shouldn’t be eliminated except that the workstations were in front or at the end of the line. Otherwise, the elimination can create gaps in assembly line. II. The load needs to be calibrated. A small increment in the maximum lead time may result a significant decrement in load misbalanced. Takt time is normally set by the company’s marketing that sets production target. The cycle time must be under given Takt time. III. Multiple operators. When a workstation requires more than one operator, the workstation’s lead time can be concluded as the simple sum of all assigned operation times in that station. The precedence constraint among the workstations may result idle time between operations that leads to reduce the total efficiency of the line.

Bottleneck

As it was discussed in theory of constraints (TOC) by (Goldratt and Cox, 1986) the idea of manufacturing systems is limited by the performance of bottleneck stations. Mostly, the total production rate of manufacturing systems can improve significantly if the bottleneck machines are systems can improve significantly if the bottleneck machines are organise and supervised properly. But, identifying bottleneck and implementation of bottleneck detection methods are still complicated. Basically, the overloaded workstations in manufacturing line can be defined as bottlenecks. Bottlenecks overwhelm the throughput of all products passes through them. If the production lines are fed with work amount that are more than the capacity of the bottleneck stations, too much work in process (WIP) inventory will be added up.

6

2.5 Discrete Event Simulation

Simulation can be defined as a simplified representing system created according to the understanding of a specialist who aims to find out its possible improvements (TAKO, ROBINSON, 2010). Simulation tends to acquire outcomes that can be applied to the real system(VDI, 1993). By means of simulation, manufacturing system designs can be tested before the purchasing phase providing required data such as number of machines, cost, required space and best location arrangement. Modelling and simulation was progressively used as decision helping tool as it has features that can represent complex systems and analyse the dynamic behaviour of these systems. (BANKS et al., 2010). According to (Sandanayake et al., 2008) discrete event simulation together with production system analysis, intend to improve performance, became more relevant in the last decades. Together with the expansion of the capabilities on computers, discrete event simulation provides better services to specialists in visualizing, analyzing and optimizing complex production processes, within fair period with a equitable investment. Simulation is a representation of a real system implemented in a controlled environment in which its responses can be studied under different scenario costing no physical risks or money(BANKS et al., 2010). The circumstances in the scenario can be analysed regarding quality improvement, machine investment, layout and processing parameter arrangements. Simulation is a popular research tool because of its versatility, flexibility and power of analysis (RYAN; HEAVEY, 2006) and can be applied to investigate any stochastic system (HILLIER; LIEBERMAN, 2010).

7

2.6 Modeling and Simulation

According to Hillier and Lieberman (2010) mathematical simulations are suitable to solve complex problems. According to Leal et al. (2011) a simulation model implementation contains 3 steps as follow: • Conception or Problem Formulation, • Implementation and • Results analysis. In the phase of conception, the researchers state the simulation’s main objectives such as physical layout planning, maximizing the efficiency, defining the suitable buffer space, etc. Then, a conceptual model is developed after all necessary data are collected. The researcher should plan the simulation experiments and then run them iteratively to verify the model. After the conceptual model is verified, the researcher can initialise the implementation phase. In this phase, the conceptual model is converted into a computational model, by organizing all the setups, specifications and methods in detail in accordance with all collected data. Then, the mathematical model must be verified and statistically validated. Verification and validation methodology (Sargent, 2012) is shown in figure 2.5. The main different points in verification and validation of a model can be compared as follow.

Table 2.1: Verification vs. Validation

8

Verification Validation

Does model work right? Is it a right model?Try to detect and remove un-intentional errors

Try to detect discrepancies between model and reality

Debugging Team work, group discussion

Restricted to model Involves real system

Involves only modeler Involves all in project

Develops personal confidence Convey confidence to others

Easy but time consuming Very difficult and subjective

Figure 2.5: Simplified Version of the Modelling Process

Finally, in case the model is affirmed, it can go on analysis phase. At this point, all changes in mathematical model reflect the identical changes and results in the real system.

Simulation Performance and Uses

Functionally, simulation can:

Increase

• Throughput

• Resource Utilization

• Facility Utilization

Determine

• Optimal buffer sizes

• Number of transporters and AGVs

• Number o the workpiece carriers

• Production schedules and sequences

Decrease

• Throughput times

9

• Required resources

• Storage requirements.

Additionally, simulation can also

• Validate the process design in the planning process

• Identify the bottleneck

• Decrease WIP

• Figure out the consequences of capital investments or processes changes

• Optimize control strategies

• Eliminates planning defects

• Protect investments

Simulation can be used to plan a new facility by performing the following tasks:

• Determine and optimize the times and throughput

• Determine the dimensioning

• Determine the limits of performance

• Investigate the influence of failures

• Determine manpower requirements

• Gain Knowledge about the behavior of the facility

• Determine suitable control strategies

• Evaluate different alternatives.

And it also can be used to optimize existing facility by

• Optimizing control strategies

• Optimizing the sequence of orders

• Testing the daily proceedings.

Simulation is also useful to execute a management plan as it can

10

• Provide a template for creating the control strategies

• Test different scenarios during the warm-up phase of the facility

• Train the operators of the machines in the different states of the facility

10

Figure 2.6: Methodology Flow Chart for Simulation Study

2.7 Tecnomatix Plant Simulation

According to Jovisevic, S., Jovisevic, V., & Jokanovic, S. (2009, December), Tecnomatix Plant Simulation is a software system which is designed for modeling, simulation and optimization of manufacturing process planning. Optimization of manufacturing process planning using this software system is based on time-oriented simulation and event-

11

Definition of Problem and Target

Acquire Data

Create Model

Validate Model

Experiment and Analyze Model

Analyze System

Evaluate Results

Recommendation

oriented simulation. Time-oriented simulation considers a wide range of different types of production time. In reality, time elapses continually. When watching a part move along a conveyor system, you will detect no leaps in time. The curve for the distance covered, and the time it takes to cover it, is continuous, it is a straight line. A discrete, event-controlled simulation program on the other hand only considers events (point at a time) that are of importance to the further course of the simulation. That events as example, may be a part arriving a station or departing it moving on to another machine. All the activities in between are of little interest to the simulation as such. It is only crucial that the arrival (In) and the departure (Out) events are correctly specified. When a part enters a material flow object, Plant Simulation calculates the time until it exits that object and enters an exit event into the list of scheduled events of the EventController for this point in time. Thus, the simulation time that the EventController displays, leaps from event to event. This happens as soon as an event is processed.

Figure 2.7: Time-Oriented vs. Event-Oriented Simulation

12

2D and 3D Modeling Environments

Figure 2.8: Modeling of Process Planning in 2D Environment 2D Modelling, shown in Figure 2.3, is used to model complicated optimization problems, mostly related to time distribution in the manufacturing processes, i.e. study of the manufacturing process in terms of time (production times, handling time cycles production, etc.). Modeling in 3D, shown in Figure 2.4, is primarily used for monitoring the distribution of technological systems and devices, which is necessary to spatially arrange in the appropriate production system.

Figure 2.9: Modeling of Process Planning in 3D Environment

13

Modeling in 2D and 3D environments is possible to connect, so when model in 3D environment is created, model in the 2D environment is generated automatically. The course of creating simulation models is carried out as follows: • Generation of the 2D or 3D models of appropriate technological systems(stations),

devices, methods of transport material, inputs, outputs, etc. from the database of mentioned technological units. (Particularly customized items developed in CAD softwares such as NX, Solid Works etc. in *.s3d format can also be loaded and used)

• Development of spatial distribution of stations and their adjustment to the conditions related to real production processes,

• Connecting the appropriate stations in the production line. Thus defined product lines represent the actual product flows, which occur in the appropriate production system,(eg. conveyor between two stations or the worker will move workpieces from one station to another)

• Setting of parameters for each of the selected station, which is a part of appropriate production flows. Data entered in this step should correspond as much as possible to the values of the real production process,

• Defining the appropriate objects, in the form of diagrams, tables, histogram, etc., which have the function of monitoring and presenting the results of simulations of the production process,

• Modeling the production process and its setting in order to create conditions for the process simulation, i.e. testing of the simulation model.

Capabilities of the system Tecnomatix Plant Simulation from the aspect of objects and methods of simulation are reflected through the simulation and modeling of: • Process plan with a number of different strategies of production(Push, Pull, MTS,

MTO,ATO) • Production process using the process planning, • Condition: in malfunction, in work, in pause, • Workers in the work and tasks they perform, • Working shifts systems, • Transportation systems. The table below describes the features available for each

license type in detail:

14

Table 2.2: Tecnomatix Plant Simulation License Details

15

Feature Pro Std App Run.T Res Edu Stud

Limited Functionalities Build simulation models with any of the built-in and with user-defined objects

X X X(+)

- X X X

Build simulation models with objects from libraries and any of the built-in material flow objects (licenses for the libraries are required)

X X X - X (X)* (X)*

Model with length-oriented objects X X X - X X X

Create and modify class objects, especially user-defined objects

X X X - X X X

Create and modify control codes employing the programming language SimTalk

X X - - X X X

Create user-defined dialogs X - - - X X X

Create libraries X - - - X - X

Use the merge mechanism X - X X X X X Import CAD files (DWG, DXF, DGN) as the background image of the Frame

X - X X X X X

Profiler X - X X X X X

Objects and functionalities which are not available for all license types Confidence analysis X - X X X X (X)*

DataFit X - X X X X (X)* Experiment Manager X - X X X X (X)*

Factorial analysis X - X X X X (X)* Neural Networks X - X X X X (X)*

Sequential Sampler X - X X X X (X)* Statistical tools X - - - - - -

Attribute Explorer X - X X X X X Card File X - X X X X X

Queue File X - X X X X X Stack File X - X X X X X

Time Sequence X - X X X X X Flow Control X - X X X X X Optimization objects based on genetic algorithms X - X X X X X

License Types:

Pro: Professional License Std: Standard License App: Application License Run.T: Runtime License Res: Research License Edu: Education License Stu: Student License

X Included - Not available (X)* The object can be loaded, the maximum number of objects might be exceeded the model cannot be saved any more. O Optional product (+) Except for the object Method

Objects Available in Each License Type The following objects are available for each license type:

Built-in Materialflow objects: Connector, EventController, Frame, Interface, Source, Drain, SingleProc, ParallelProc, Assembly, DismantleStation, PickAndPlace robot, Store, PlaceBuffer, Buffer, Sorter, Line, AngularConverter, Converter, Turntable, Turnplate, Track, TwoLaneTrack, and Cycle.

MaterialFlow objects, libraries, and modeled objects: EOM library; Conveyor library, Cross-sliding Car library, High Bay Warehouse library, Kanban Objects library, Portal Crane/Cranes library, and TransferStation.

16

Optimization Wizard X - X X X X (X)*

Generator X - X X X X X HTML Wizard X - X X X X X

Layout Optimizer Wizard X - X X X X (X)* Teamcenter Interface X - - - X - -

Trigger X - X X X X X Variants Generator X - X X X X (X)*

XML Interface X - X X X X X

Resource objects: Workplace, Footpath, WorkerPool, Worker, Exporter, Broker, ShiftCalendar, and LockoutZone.

Information Flow objects: Method, Variable, TableFile, FileInterface, and FileLink.

User Interface objects: Comment, Display, Chart, Button, Checkbox, and Drop-Down List.

Mobile Units: Entity, Container, and Transporter. Tools: Bottleneck Analyzer, LayoutOptimizer, SankeyDiagram, Worker Chart

17

CHAPTER 3CASE STUDIES

The objective of this case study is to enhance the knowledge of properties, features and different application of Tecnomatix Plant Simulation for different purposes.

3.1 Modeling, Simulation and Optimization of Process Planning (Jovisevic, S., Jovisevic, V., & Jokanovic, S., 2009),

Overview This paper is focusing on modeling of the manufacturing and assembly planning for crank shafts using Tecnomatix Plant Simulation in 2D and 3D environment and determining optimal parameters of production.

Optimization PerformanceThis optimization process consists of several steps: • Developing the spatial model and generation of individual processes, which represent

the operations in the production process from the process plan, • Defining the distance between the individual station in order to effectuate the

simulation time which is lost during transportation of parts prior to the following operations,

• Linking individual processes in the flows of materials processing according to designed process plan,

• Defining the required time for individual processes, • Defining the methods and rules of the transition work pieces during processing of

materials, • Defining objects for monitoring and recording the results of simulations • Performing initial process simulation, • Analysis of the results of simulation, • Modification of simulation models and • Performing the final process simulation. Simulation of manufacturing and assembly process plan for crankshafts by the system Tecnomatix Plant Simulation in 2D environment is shown in Fig. 3.1

18

Figure 3.1: Segment from Simulation Model Shown in 2D

After designing satisfactory model of the manufacturing and assembly process plan for crankshafts in the 2D environment, it is transformed into a model of process plan in 3D environment. The purpose of transforming into 3D environment is primarily to determine the spatial layout of machinery and equipment in the production plant.

Figure 3.2: Simulation Model Shown in 3D

By placing the machines and devices in precise defined locations in the production plant in 3D environment, the precise distance between the machines that affect optimization for earlier designed models is determined.

19

Figure 3.3: Simulation Model Located in the Corresponding Production Plant

Simulation model and production plant of the manufacturing and assembly process plan for crankshafts by the system Tecnomatix Plan Simulation in 3D is shown in Figure 3.12 and Figure 3.13. This model is founded on the basis of previous information on the number, layout and positions of the necessary technological systems and equipment. Simulation results Improving the simulation model is performed through initial process simulation and analysis of simulation results. In the model after the analysis of results, bottlenecks of production are identified and accumulation of work in process(WIPs) in each station of the process plan are also observed. Efficiency of machines performing the initial process of simulation, i.e. before performing the optimization process, is unsatisfactory as shown in Figure 3.14.

Figure 3.4 Efficiency of Machines before Optimization

20

After analysis of influential factors, optimization of process plan was performed, where objective is to achieve high efficiency machines and minimum time duration of production cycle, while the limits were defined by the process plan. Resolving the above-mentioned shortcomings, after introduction of buffer zones and increasing the number of machines, it achieved a significantly shorter production cycle time. For example batch of 100 pieces of crankshafts, can save 40% of time in manufacturing process after the simulation. After completion of design and simulation of process plan in the 3D environment, model needs to determine the exact location of machines and devices in the spatial layout of production facilities. Based on this information, manufacturing cycle is reduced by 5% as a result of savings in time in transport parts between machines. Based on the layout and location of machines and devices, it was designed a preliminary solution for production plant with appropriate characteristics (departments, roads, entrances, exits, etc.). Results from simulation of process plan provide cycle times generated for the appropriate batch size and machine utilization of machines in the manufacturing process in percentage. According to optimal designed simulation model, the utilization of machines for the batch size of 100 crankshafts is shown in Figure 3.15.

Figure 3.5 Efficiency of Machine after the Optimization

Modeling the process plan and simulation allow creating models that represent adequate production processes that generates the following benefits: • Improving the productivity of existing production systems,• Reduce investment in planning new production facility and capacity,• Reduce inventory and flow time,• Optimization of production systems dimensions, including backup size,

21

• Reducing investment risk through early proof of production concept,• Maximizes utilization of productive resources,• Improvement in design and layout of production line and machines.

In the example, final result of performing simulations showed data related to the duration of the manufacturing cycle, utilization of resources, as well as the required dimensions of the production plant that meet the set requirements from process plan.

3.2 Kanban Simulation Model for Production Process Optimization (Golchev, R., et.al., 2015).

Overview This article is based on a basic research for possibilities for KANBAN implementation in one metal-working company. In that direction, the basic aim of the article is to present benefits of combined implementation of KANBAN system and methodology for Discrete Event Simulation (DES). After giving the brief theory on KANBAN and DES, the article discusses the options for their integration. The case at the end is focused on one production line in one metal-working company and experiments with the container capacity which directly affects the number of KANBANs. Implementation of Kanban System Methodology One possible methodology for the purpose of systemized and easier implementation of the KANBAN system is presented, consisting of seven fundamental phases as shown in the flow chart in Figure 3.16. This presented methodology seems fairly simple; however, its implementation is a challenge, because in order to be certain that the KANBAN system is well-designed, a variety of stochastic, not to mention expensive, calculations and iterations must be made, (Müller et al., 2012). There are number of possible tools in order to determine the best KANBAN system, but also to experiment with it. One of them is the simulation and its advantage in experimenting with and optimizing performance values. Since there are many variables to experiment with, the simulation shortens the time needed to determine the possible outcomes of the system in different situations, (Hao & Shen, 2008). For the most part, simulations are more than useful in the first three steps of the methodology, especially in the third one, during the actual design of the KANBAN system. This paper is focused on these three steps. As it was stated before, the simulation was used in order to determine the capacity of the container, or the number of KANBAN cards needed in order to achieve a more effective production process. The creation of the simulation model shown below is based on the methodology according to (Banks et al., 2004) a methodology that offers a systemized approach. As a result of the characteristics of the methodology and because its steps are not strictly successive, it allows adjustments to different application.

22

Figure 3.6: Methodology for Kanban Implementation (Gross & McInnis, 2003)

23

Data Collection and Analysis

Kanban Number Calculation

Designation of Kanban

Training

Starting the Kanban System

Is Kanban designed properly?

Maintenance and audit of the System

Improvement of the system

Yes

No

Kanban Simulation Model This section presents the application of the KANBAN system in a simulation model made with the software suite Technomatix Plant Simulation by Siemens. Using Banks’s methodology, each and every problem is defined during the first step. Then, goals are set and a model is conceptualized according to the acquired information. After formulating the model, experiments which in normal circumstances might last for days, months or even years are created in just few minutes. If after the result analysis it is concluded that the experiment data are sufficient and correct, records are prepared and the final step, implementing the solution, can be made. Purpose of the simulation model: Determining optimal KANBAN container capacity in relation to the demand The purpose of this simulation model is through simulation of a number of possible production scenarios with previously determined settings, to get an optimal capacity of a KANBAN container, in relation with the daily needed throughput. This is extremely important because the containers are the ones that when empty, initialize the beginning of the production, and when they are full they stop it. Furthermore, the storage units are an additional burden when it comes to space usage. In order to avoid the main and the biggest problem (to avoid overproduction) it is immensely important to design a precise KANBAN system. Figure 3.17 shows the basic concept of setting the elements in the simulation model, (Robinson, 2004). The information moving direction, as well as the direction of the product can be clearly noticed on the figure, starting with the raw materials and ending in the hands of the customer.

Figure 3.7: Basic Concept of the Simulation Model

Figure 3.7: Basic Concept of the Simulation Model

24

Experiment After the simulation model has been developed, verified and validated, the next step entails conducting experiments and analyzing its outcomes. In Figure 3.18, the results of simulating 13 different cycles are presented. The ordinate lists the daily throughputs done in 3 shifts, 8 hours each, while the abscissa shows the values for which the KANBAN container is limited. First, it can be noticed that by decreasing the container’s capacity, daily throughput does not decrease proportionally. Instead, there are occasions when a container with different capacity has identical daily throughputs. Thus, if the product demand is projected to 750 finished parts by the end of the day, there are two different possibilities for choosing the container capacity. The first possibility is to choose a container with capacity of 75 parts per container, and the other is 50 parts per container. For this specific case, it is better to choose the second option because logically, this type of container has higher number of daily cycles. The higher number of cycles reduces the processing time per container. This makes the system more flexible and resistant to external disruptions including change in demand, defects, change of the product etc. Most importantly, with this type of container, whenever an error occurs, less number of parts will be affected by it. It is the same when the needed daily throughput is 720 finished parts per day. Furthermore, this simulation model can be a perfect basis for future experiments. These can include other important factors, such as: delays, defects, scrap percentage, product changes, workers overload etc.

Figure 3.8: Experimental Results

25

CHAPTER 4 DESCRIPTION OF SYSTEM UNDER STUDY

Premium PLM Bangkok, a Tecnomatix Plant Simulation service provider company which is SIEMENS Limited Thailand’s contracted partner for PLM software helped to get contact with a Tire 1 automotive supplier company for automotive manufacturers in Thailand. According to the customer protection policy, they requested not to mention the name of company from which all the specific data are collected.

Among a large variety of automotive parts produced in the host company, Electric Power Steering Column production line is focused in this study.

4.1 Description of Product Under Study

Electrical Power Steering System

Electric Power Steering for automobile is primarily an energy saving scheme. It is expected to be in widespread use, just to cut down the energy usage in modern vehicles. Further advantages may include enhanced flexibilities in the location of the steering system, the turning of the steering system to cater to the need of the specific automobile or an individual driver and the reliability. The Electrical Power Steering System involves replacing the hydraulic system with an all electric system in which power is delivered to the rack and pinion of the steering mechanism only when required. In this EPS, an electric motor drives the rack and pinion arrangement to steer the vehicle using power from the battery.

Details of EPS designs differ amongst automotive manufacturers. However there are certain components that are intrinsic. Such as:1. Torque sensor2. Electric motor3. Rotational angle sensor4. Controller5. Vehicle speed sensor6. Coupling between motor and steering mechanism

The torque sensor is perhaps the most important component. It measures the effort being applied by the driver to steer the vehicle. The torque sensor output is then used to drive a motor to reduce the effort, while achieving the desired steering. The motor may be located at a number of locations to achieve this. The purpose of the motor controller is essentially to control the torque delivered to the steering mechanism. The vehicle speed must be used to adjust the sensitivity and the performance around the null position of the steering wheel.

The Electric Power Steering Column, the final product of the assembly line under study is part of EPS.

26

Figure 4.1: Electric Power Steering System

Bill of Materials included in the Final Assembly The finished product of the final assembly line which is the Electric Power Steering Column composed of 5 major Sub-assembly groups, - Sensor Shaft Assembly - Column Assembly E-Lock (E-Lock means Electric controlled system) - Bearing Assembly - Bracket Assembly (Tilt) - Cover and - Other Components Some of the components like O-Ring, C-Ring, Spacer Rubbers are directly supplied from store while some other components like Gear Box are Produced according to the production plan of their line. Some special sub-assemblies such as Sensor shaft are produced in sub-assembly line via pull system. The stock level in the buffers on the final assembly line triggers the Kanban system to produce more. Sensor Shaft Assembly Sensor Shaft Assembly is the assembled combination of Sensor Shaft with worm wheel, Ball Bearing, Retention Ring and 2 Bushes. Sensor Shaft is supplied to the final assembly line from the sub-assembly line by pull system while the other components were produced according to their lines’ production plan.

27

Figure 4.2: Sensor Shaft Assembly E-Lock

Column Assembly E-Lock Column Assembly is the main structure of the Electric Power Steering Column that comprises: - Gear Box - Ball Bearing - Worm - 2 Spacer Rubbers - C-Ring - Bolt Flang - Upper Housing - Lower Housing - Column Cover and - Liner Slot.

Figure 4.3: Gear Box

Bearing Assembly Bearing Assembly is composed of two bearing sets with two bushes. Cover Cover contains cover assembly which is produced in the sub-assembly line by pull system and a retention ring.

28

Figure 4.4: Cover Assembly

Bracket Assembly Tilt The bracket assembly is the section to attach the EPS Column to the rest of the system. Bracket assembly contains: - Lever Assembly Tilt - Fixed Cam - Washer - Thrust Bearing - Nylon Nut - Bolt - Bracket Assembly Tilt - Bracket Assembly - Spring and - Retention Ring Other Components Items in other component group are the components that are assembled on the EPS column other than the previous sub-assembly groups such as: - Motor - Electrical Control Unit - Bracket Harness - Harness Assembly - Flang and - Cover Connector

Figure 4.5: Motor 29

Figure 4.6: Electrical Control Unit

Figure 4.7: Bracket Harness

Figure 4.8: Harness Assembly

Figure 4.9: Cover Connector

Final Product of the assembly line is described in Figure 4.11. The complete product breakdown structure including components and subassemblies in the final assembly are described in the Figure 4.10.

30

Figure 4.11: Electric Power Steering Column

4.2 Production Process and Stations

Electrical Power Steering Column Production Line

Electric Power Steering Column production line contains 4 major sections:1. Manufacturing Section 2. Sub-Assembly Section 3. Assembly Section and 4. Warehouse Section Among these, only Manufacturing Section is run in Push system. All the other downstream sections are run in Pull system.

Figure 4.12: Overview of the Electric Power Steering Column Production Line

32

Manufacturing Section

Sub-Assembly Section

Final Assembly

SectionWarehouse

Push System Pull System

Work Flow

Process Flow of Electrical Power Steering Column Assembly Line

This study will be emphasized on the Final Assembly Section in which all the subassemblies and components are assembled as the final product. The line contains 25 processes ran by 17 human workers. According to the collected data in August 2016, production capacity of the line is 721 finished goods per day. The production system is running in 2 normal shifts(Day and Night) and 2 overtime shifts. Although the main flow of the assembly line is sequenced to move the parts from a station to the downstream adjacent station, some parallel stations(GA-4,GA-7,GA0-8,GA-15,FA-1) that are adding supplementary components to the main assembly are configured to transport the processed workpieces to next stations by human workers.

Table 4.1: Shift Schedule

The process flow chart along with the detail description of the activities and added components in each station is described in figure 4.13.The process flow with each process performed by particular worker are described as follow. Worker1 At the beginning of the final assembly process, the Worker1 assembles the Worm End at the Worm End Assembly station (GA-1) with Worm, 2 spacer rubbers, C-Ring and Flang. Then the workpiece is carried by the worker to the Worm End Assembly Fitting station(GA-2) and assembled with Gear Box followed by ball bearing fitting at the Gear Box Bearing Fitting station(GA-3).

Worker2 In parallel, the Worker2 caulks 2 sets of bush and ball bearings at the Bush Caulking station(GA-04). The workpiece at station GA-3 is then taken to the Lower Bracket Assembly station(GA-05) to combine with caulked bushes.

33

Working time Day Night

Normal 6:00-14:00 22:00-6:00

Hr. 6hr.35mins 6hr.35mins

Overtime 14:20 - 17:00 17:50-21:40

Hr. 2hr.40mins 2hr.40mins

Worker3 The Worker3 attaches the sensor column to the workpiece at the Sensor Shaft Assembly station(GA-7).

Worker5 At the same time, the Worker5 is preparing the Cover at Shaft Cover Assembly station(CA-07) combining Cover Assembly and Retention ring and also prepare the worm wheel assembly at the Worm Spacer Caulking station(GA-8) which are then fed to the Worm Assembly station.

Worker4 Worker4 at the Worm Assembly station(GA-9) combines the cover and worm wheel assemblies with the main assembly. Then the assembly is checked whether the fittings and the gears are working properly at the Motoring station(GA-10) Worker6 The sensor shaft is then calibrated to the neutral voltage at the Neutral Voltage Adjust station (GA-11) by Worker6. Seal and No. Label are also attached to the assembly at this station.

Worker7 Worker 7 prepares the housing assembly at the Housing Assembly satation(GA-13) and tests the mechanical movement of the worm gear at Worm Movement Test station(GA-12).

Worker8 The tilt-lever assembly is prepared and combined with the housing assembly from GA-13 in Tilt-PreAssembly station(GA-15) by Worker8. The sub assembly is then carried to Tilt Assembly station(GA-16) and combined with the main assembly by the Worker8.

Worker9 Worker9 tests the function of the main assembly at the Tilt&Telescopic Operation Test Station(GA-17).

Worker11 Worker11 combines the Electrical Control Unit(ECU) and electric Motor at ECU&Motor Assembly station(FA-1). The Worker11 also perform the torque test of the main assembly at the Rotational Torque Test station(FA-3).

34

Worker10 Worker10 attaches the assembly of Motor and ECU to the main assembly at the Motor Assembly station(FA-2).

Worker12 Worker12 also performs the ECU&Motor assembling function at FA-1 station and performs the performance test at the Performance Test station(FA-04).

Worker13 Worker13 inspects the assembly at the Pre-Shipping Inspection station(FA-5) and attaches the Cover Connector at Bunded Table(FA-5.5).

Worker14 Worker14 performs the shipping inspection at (FA-6) and also perform the cover attaching function at FA-5.5.

Worker15 Worker15 performs the overall inspection at the Final Check station.

Worker16,17 Worker16 and Worker17 perform the packing at the Packing station.

Plant Layout of Electrical Power Steering Column Assembly Line

The plant plan of the EPS column final assembly line is described in Figure 4.14. As it can be seen in the figure, all the stations are ran by human workers. And even some of the transportation of the workpieces between stations are also done by human workers. As it was illustrated in the figure, the line includes the inventory stalls for the dedicated items in between the particular stations. To be easier to understand, the description of the station names according to each station’s number are listed in Table 4.2.

35

Table 4.2: Descriptions of the Stations

Time Study of the System

As each and every assembly processes including transportation are done by human workers, the time study of each station has 2 time studies: - Machining Time - Handling and Operation Time Machining Time is the time period that the workpiece is machined while

38

No. Station No. Station Description Done by Worker1 GA-1 WORM END ASSEMBLY 1

2 GA-2 WORM END ASSEMBLY FITTING

1

3 GA-3 GEARBOX BEARING FITTING 14 GA-4 BUSH CAULKING 25 GA-5 LOWER BRACKET ASSEMBLY 26 GA-7 SENSOR SHEFT ASSEMBLY 37 CA-7 SHAFT COVER ASSEMBLY 58 GA-8 WORM SPACER CAULKING 59 GA-9 WORM ASSEMBLY 410 GA-10 MOTORING 411 GA-11 NEUTRAL VOLTAGE ADJUST 612 GA-12 WORM MOVEMENT TEST 713 GA-13 HOUSING ASSEMBLY 714 GA-15 TILT-PRE ASSEMBLY 815 GA-16 TILT ASSEMBLY 8

16 GA-17 TILT&TELESCOPIC OPERATION TEST

9

17 FA-1 ECU AND MOTOR ASSEMBLY 11/1218 FA-2 MOTOR ASSEMBLY 1019 FA-3 ROTATIONAL TORQUE TEST 1120 FA-4 PERFORMANCE TEST 1221 FA-5 PRE-SHIPPING INSPECTION 1322 FA-5.5 BUNDED TABLE 13/1423 FA-6 SHIPPING INSPECTION 1424 QA FINAL CHECK 1525 PACKING PACKING 16/17

the Handling and Operation Time is the combination of transportation, handling and assembling without machining. Time study including the standard time and normal distribution of each station is illustrated in Table 4.3 and 4.4.

Table 4.3: Time Study of the EPS Column Final Assembly Line

39

No. Station Time Study Standard

time (s)Standard Deviation

Lower Bound

Upper Bound

Carried out By Worker

1 GA-1

Machining Time 7 0.79 5.8 8.9

1Handling & Oper:Time 12 0.56 11.4 13.1

2 GA-2

Machining Time 21 0.4767 20.7 22.3


3 GA-3



4 GA-4

Machining Time 20 0.872 18.9 21.9


5 GA-5

Machining Time 0 0 0 0


6 GA-7



7 CA-7



8 GA-8



9 GA-9



Table 4.4: Time Study of the EPS Column Final Assembly Line (Cont.)

40

No. Station Time Study Standard time

Standard Deviation

Lower Bound

Upper Bound

Carried out By

10 GA-10



11 GA-11



12 GA-12



13 GA-13



14 GA-15



15 GA-16



16 GA-17

Machining Time 8 0.7 7 9.7


17 FA-1


11/12Handling & Oper:Time 24 0.37 23.2 24.8

18 FA-2



19 FA-3



41

20 FA-4



21 FA-5



22 FA-5.5


13/14Handling & Oper:Time 15 0.69 14 16.7

23 FA-6



24 QA



25 PACKING


16/17Handling & Oper:Time 48 0.57 47.2 49.1

Kanban System In order to limit accumulation of inventory in WIP and keep the efficient stock level in the warehouse, final assembly line of Electrical Power Steering Column is running in pull system. According to the stock level in the Ware House section after Assembly section, the Kanban cards are circulated to assemble the required number of finished products to fulfil the empty containers. A Kanban box contains 12 identical finished products. Here it can be considered a box Kanban represents 12 identical Kanban cards. For the Electrical Power Steering Column assembly, the product under study, the maximum stock level is 5 boxes(60 pcs) and the reorder stock level is 2 boxes(24 pcs). Once the stock level in the warehouse falls under re-order level(2 boxes), the warehouse sends out the Kanban cards to produce finished products and refill up to the maximum stock level. Similarly, Kanban systems for the sub-assembly lines are also triggered by the stock levels in the buffers on the final line, for example Sensor Shaft Assembly buffer on GA-07 station triggers the sub-assembly line for Sensor Shaft Assembly. But in this paper, these sub-assembly lines are not considered and these buffers are assumed to avoid the shortage.

Figure 4.15: Kanban System of EPS Column Assembly Section

42

Final Assembly LineParts Taken Out By Customer OrderWarehouse

Re-‐order Stock Level>>> 2 Boxes

Maximum Stock Level>>> 5 Boxes

Detached Kanban Card

Kanban card in the EPS Column Final Assembly Line describes the following specifications of the product which will be produced. 1. Product Type 2. Product Picture 3. Production Line to be produced on 4. Product ID 5. Package Bar Code describing the package that the product belongs to 6. Kanban ID 7. ID of Sensor Shaft Assembly used in this product 8. ID of Electrical Control Unit used in this product 9. Quantity of products in this box

Figure 4.16: Kanban Card in EPS Column Final Assembly Line

Production Schedule

From the warehouse, the finished goods are withdrawn to deliver to the customer according to the monthly production plan described in table 4.5. In current condition, according to the varying demand and current system configuration, there are some product shortages on some days and surplus inventory on other days. That became the main issue to be solved by tuning up the stock level.

Table 4.5: Monthly Production Plan of EPS Column

43

ITEM 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 TOTAL

PT22XX0264S Planned 38 0 36 36 36 36 36 0 0 36 36 36 36 36 0 0 36 36 36 36 36 0 0 24 24 0 24 24 24 0 0 698

ACTUAL 36 0 37 36 38 36 36 0 0 23 36 36 36 36 0 0 36 36 48 0 36 36 0 24 24 24 0 24 24 0 0 698

3.

5.7.

1.

2.

4.

8

9.

6.

CHAPTER 5 MODEL DEVELOPMENT AND OPTIMIZATION

The simulation model of the current assembly line is developed according to the collected data. At this point, in addition to the shop-floor data, background production management practices and arrangements that are applied on the production system such as, push or pull system of production, allocation of stores and buffers between the sections etc., will also need to be feed in. The model will be developed in following steps:

Figure 5.1: Modelling Steps Flow Chart

44

Not Validated

Validated

Recommendation

Optimization

Basic Model Development

Worker Configuration

Pull System Configuration

System Performance Analysis

System Calibration

Verification and Validation

5.1 Basic Model Development

Basic model development phase contains the development of the stations with particular settings and material flow routes. First of all single-proc stations are placed according to the plant plan. For the stations that have both operation time and machining time, a pre-station is added in front of the machining station as a separate station to improve the flexibility in modelling. Station allocation in the assembly line plan is shown in figure 5.2.

Figure 5.2: Station Allocation After station allocation, the processing time of each station are entered in each of the station specification. Double click the station to be setup and select Times tab in the station spec window. As the processing time of the stations are in normal distribution, click the drop down icon beside processing time and select Normal. Then key in the mean processing time, standard deviation, lower bound and upper bound in the format shown above the text box. Then click apply and ok buttons. Processing time configuration window of the station GA01_Pre station is shown in figure 5.3. Configure the Operation time and Machining time of all stations according to Table 4.3.

Figure 5.3: Processing Time Configuration Window

45

5.2 Worker Operation and Transport Workpieces

Conceptually, while the Worker works at the machine, he stays on a Workplace at the associated machine. While the Worker is Waiting for a job, he stays in the WorkerPool, which might be the lounge or the staff room of your plant. When the Worker has to perform one or several jobs at a machine, the foreman (Broker ) tells him to do so. The Worker then walks on a FootPath from the WorkerPool to the Workplace of the respective machine and does his job. According to the concept, in order to use the human worker for machine operations, workplace, worker pool, broker and foot path are required to include in the model. These can be inserted from the tab, Resources as shown in figure 5.5.

Figure 5.4: Resources Menu Bar

Add the workstations and footpaths according to the existing plant plan. And then add the worker pool and broker to run.

Figure 5.5: Layout after Workplaces and Footpaths are Added

To declare the number of workers who will be working in the model, configure the workerpool. Open the creation table in Attribute tab. Key in the 17 workers as shown in figure5.7.

46

Figure 5.6: Creation Table of Workers in Workerpool

In the object workplace on Attribute tab, the respective station is described and in the supported service window, key in the StandardService to operate the station. Incase, if the worker is supposed to perform the other services such as servicing, repairing etc. the service has to be declared in this window.

Figure 5.7: Workplace Window 47

Figure 5.8: Supported Service Window

In order to declare the station that the operation at this station will be done by human worker, in the tab of Importer, processing must be activated. Key in the name of Broker. Priority of the station can be set if single worker is working for multiple stations. Importer window of the process CA7-Pre is shown in figure 5.10 for example.

Figure 5.9: Importer Window

To setup the worker to transport the workpieces from one station to another, configure on the Exit tab of the station configuration window of the from-station. Change the strategy to Carry Part Away, select the Broker name. Then define the target station in the MU target text box. Detail setting can be seen in figure 5.11. Repeat the configuration on all the stations that need the workers to transport that are mentioned in section 4.2.2.

48

Figure 5.10: Exit Window

For applying the working shifts to the model, a Shift Calendar is added. Shift details that are discussed in section 4.2.2 are added to the calendar. And then in the worker pool, name of the shift calendar that the workers will work accordingly is added.

Figure 5.11: Shift Calendar Window

49

Figure 5.12: WorkerPool Window

5.3 Verification and Validation

Before the model is configured to the pull system as a part of verification and validation, it is analysed to measure its performance in push system and approach of Correlated Inspection(Law and Kelton, 1991) as shown in figure(5.14) is applied. The system is run for a day and analysed the behaviour of the results. All the configurations such as: - Shift Schedule(Shifts, Overtime, Breaks) - Process Flow - Workers’ Operation Time(Travelling, Handling, Machining) - Workpiece Transportation, etc. are tuned up to get the closest representing model

compared to the real system. After fine tuning, it was found out that the number of throughput is 724 per day while the daily production rate of the real system is 721 per day.

Figure 5.13: The Correlation Inspection Approach

50

Figure 5.14: Summary of One Day Simulation

Utilisation of the stations and the performance chart of each workers after verification and validation are also recorded as shown in Figure 5.15 and Figure 5.16. Here in the worker performance chart, it was described based on 24 hour basis. So that unplanned and paused portions represents the break times of the workers.

Figure 5.15: Utilisation of Stations

Figure 5.16: Performance Chart of Workers 51

5.4 Pull System Configuration

As it was discussed in section 4.2.5, the EPS Column final assembly line works as a pull system. There is a warehouse after the packing station. In the model, after the finished goods are packed into boxes(12 pcs/box) in packing station, these boxes are sent to the kanban buffer. In Kanban buffer, minimum stock level, maximum stock level and initial stock level are defined. Stocks from the Kanban buffer is withdrawn by the Kanban singleproc. station in which the withdrawal rate, in another word, order arrival rate can be defined.

Figure 5.17: Kanban Buffer Window

Figure 5.18: Kanban SingleProc. Station Window 52

Once the model is run, the Kanban buffer will have the initial stock and the Kanban Singleproc will withdraw a box from Kanban buffer. Kanban buffer will withdraw one box in each hour. Once the stock level in Kanban buffer gets to minimum stock level(2 boxes), the Kanban source will produce a Kanban box that contains 12 Kanban cards. The assembly line will start producing the finish goods and the packing station will collect 12 finished goods to pack in a box. Afterwards, the box is forwarded to the Kanban buffer and the stock level will raise. Transitions of the stock levels are shown in figure 5.20. When the stock level gets back to maximum stock level(5 boxes), Kanban source will stop producing Kanban boxes and the assembly line will stop producing finished goods.

53

Figure 5.19: Stock Level Transitions in Kanban Buffer

In this model, the initial stock level is set as 5 boxes.And the withdrawal rate is set as 1 hour constantly. So, In the first stage, once the model is run, a box from the buffer is withdrawn. Therefore stock level at time zero show 4. After an hour as another box is withdrawn again, the stock level drops down to 3. After another hour, as the stock level reaches the minimum level (2 boxes), the Kanban source produces the Kanban boxes to start the assembly operation. After producing a box of finished good, it is transferred to the Kanban buffer and the stock level raised again to 3. The assembly line keeps producing the finished goods till the stock level gets back to the maximum level(5 boxes). As the withdrawal rate is set as 1 hour constant, at each hour, another box of stock will be taken from the Kanban buffer. The model in pull system is verified and validated based on the production rate when it’s in operation. Throughput rate in pieces per hour is considered as the validation factor. The production rate in pull system is observed to be 29 TPH. Here in this point, as the withdrawal rate in real system is working on the production plan discussed in section 4.2.6, the model is modified to withdraw the finished goods according to the production plan on next section.

54

5.5 Implementation of Production Plan

As the final assembly line is run according to the order plan, production plan table by which the finished Electric Steering Columns from the warehouse will be withdrawn and delivered to the customer is implemented in table file as shown in table 5.1. The production plan table contains the planned orders(no. of finished goods) on each day of a month. The finished goods withdrawal time is set to be at 3:00 pm (or) 15:00 for every day in the first column of the delivery summary table as shown in table 5.2.

Table 5.1: Production Plan Table

Once the model is initialized, the initial stock in the warehouse is created at the maximum stock level, min and max values in the warehouse are defined and the piece count in the delivery station is set as zero. First six columns of delivery summary table is created for the planned orders. The delivery status is recorded in the delivery summary table describing: - Delivery Time - Object Name in Model - Number of Orders - Product Name - Attributes - Order Quantity - Delivered Quantity - Delivery Shortage Check. In order to control the delivery, a method is implemented in which the current stock in the warehouse and delivery station are checked whether the order can be fulfilled or not.Till in the warehouse, finished goods are counted in dozen packages. To deliver the finished goods in piece count, the packages are unpacked into pieces in the delivery station after taken out from the warehouse. The method checks the order and the stock level and if the stock is enough for current order, the required number of boxes from the warehouse are requested and unpacked and sum with the remaining quantity at the delivery station. The order is then subtracted from the unpacked quantity. The remaining quantity is kept at the delivery station. If the stock is enough to fulfil the requested order quantity, Delivery Shortage Check is remarked as ‘No’ meaning no delivery shortage in the delivery summary table. If the stock is not enough to fulfil the requested order quantity, only the in hand stocks are delivered. Delivery Shortage Check is remarked as ‘Yes’ meaning there is a delivery shortage and the value for Delivery shortage counter is counted up. When the stocks are available in warehouse, a package is requested and unpacked to keep at the delivery station. Whenever the stock level falls under Minimum Stock level, the assembly line starts production to replenish the warehouse up to the maximum stock level.

55

ITEM 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31TOTAL

PT22XX0264S 38 0 36 36 36 36 36 0 0 36 36 36 36 36 0 0 36 36 36 36 36 0 0 24 24 0 24 24 24 0 0 698

The reorder quantity is defined as: Reorder Quantity = Maximum Stock - Current Stock

When the model is ready to run, the event controller is set to run for a month(31 days). In the early models, the minimum stock level is set to be 2 boxes and the maximum stock levels is set to be 5. In this configuration, the model was observed to have many delivery shortages as presented in table 5.2. The delivery status on each day of the month are described in the delivery summary table as described. As it can be seen in the delivery summary table, 7 delivery shortages are observed in a month.

Table 5.2: Delivery Summary Table before Optimization

56

Figure 5.20: Initial Method

Figure 5.21: Method for Delivery Control

5.6 Optimization Using Genetic Algorithm Wizard In order to optimise the model to minimise the number of delivery shortages, genetic algorithm wizard in Plan Simulation software is used. In the Genetic Algorithm Wizard, it needs to describe the objective, Minimization or Maximization. And to define the number of simulation runs, it needs to describe, - Number of Generation - Size of Generation and - Observations per Generation

57

Figure 5.22: Genetic Algorithm Wizard Window

In the first generation, Plant Simulation evaluates the number of generations which are entered as the Size of generation. In each of the following generations it has to evaluate twice as many individuals. The number of simulation runs, which the GAWizard executes, results from this formula: Number of simulation runs = observations per individual * (generation size + 2 * generation size * (number of generations - 1)) For the case, it is using 50 observations for evaluating an individual and create 5 generations with a frequently used Size of generation of 30, Plant Simulation has to execute 13500 simulation runs (50×(30+2×30×4)). For the optimization parameter, lower bound, upper bound and increment value of the variables in the optimization process are described as shown in figure 5.23.

Figure 5.23: Optimization Parameter Table

For the fitness calculation, the names of the target values are described. When the fitness value is determined by several target values, their weighting factors must be described as well. Here in this case, though the main objective is to minimise the delivery shortages, the levels of minimum and maximum stock levels are also required to keep as low as possible. So, there are three target values in the fitness table with respective weighting factors as shown in figure 5.24.

58

Figure 5.24: Fitness Parameter Table

For the statistical reliability, the number of observations is defined to be 50 for the current case. After setting up all the required configurations for the optimization, the GA wizard is run. After approximately 5 hours, the optimization results are generated as shown in figure 5.26. After the optimization performance the optimum values for both the minimum stock level and maximum stock level are observed to be 3 boxes. The number of delivery shortage in a month is observed to be 1 on the first day. Though the number of product shortage is only 2 pieces, in order to eliminate the delivery shortage absolutely, the maximum stock level is set at 4 boxes. The withdrawal and replenishment statistics of the inventory in the warehouse for a month are plotted in the graph as shown in figure 5.27.

Figure 5.25: Optimised Results

59

Table 5.3:Delivery Summary Table after Optimization

Figure 5.26: Stock Level Graph after Optimization

As the initial stock level is set at maximum stock level, at the beginning, the stock level starts at 4 boxes of finished goods. On the first day, 38 finished goods are delivered and the unboxed 10 pieces are kept in the delivery station. As the inventory level fell down to 0 boxes under minimum stock level(3boxes), the assembly line starts producing.

60

For the first day, reorder quantity is 4 boxes of finished goods. On the second day there is no activity as there is no planned order for that day. On the third day, the planned order, 3 boxes were withdrawn and the stock level fell down under the minimum stock level. The warehouse was replenished again up to 3 boxes. The system goes on in similar behaviour for the following days as it can be seen in figure 5.27. Day by day records of the inventory levels in the warehouse are described in the following figures.

61

Figure 5.27: Inventory Level on Day1-Day3

62


63


64

Figure 5.30: Inventory Level on Day10-Day12 65


66


67


68


69


70


71

Figure 5.37: Inventory Level on Day31

According to the delivery summary table and the daily stock level graphs, it can be seen that the model is optimised for minimum delivery shortage at the minimal inventory in stock to minimise the inventory cost. The inventory cost for each finished product is approximately 10,000 THB. As the optimization can reduce maximum inventory level from 5 boxes to 4 boxes. It saves 1 box of finished good inventory, resulting the reduction of 120,000 THB from tied up inventory cost. Further minimisation for the tied up inventory cost is still possible if the maximum stock level is kept at 3 with a trade-off of letting one inventory shortage of 2 finished good happen. That activity can save one more box of stock costing 120,000 THB. The line performance graphs in pull system before and after the optimization are shown in following figures. It can be seen that the assembling activities of the product under study occupy approximately 1% of the line and human worker capacity. Here the optimization is considered based on monthly basis. Even in case for the quarterly basis or batch by batch basis, the basic model can be used to play with the desired scenario for optimising the most suitable inventory level.

72

Figure 5.38: Utilisation of Stations in Pul System before Optimization

Figure 5.39: Performance Chart of Workers in Pull System before Optimization

73

Figure 5.40: Utilisation of Stations in Pull System after Optimization

Figure 5.41: Performance Chart of Workers in Pull System after Optimization

74

CHAPTER 6

CONCLUSION AND RECOMMENDATION

6.1 Conclusion

The purpose of this study is to analyse the performance measures of the assembly line and find out the optimum inventory setting to minimise the Delivery Shortages at the minimal inventory cost. Over the study during this project, it can be concluded as follow: 1. All the line information of the final assembly line under study are collected from a Tire 1 Power Steering Supplier Company including some historical data and observed data. Samples of the observations were processed to have appropriate input data for simulation model in chapter 4. 2. Tecnomatix Plant Simulation is used to model the assembly line based on the processed data. Initially, it observes the performance of the final assembly line at full capacity (modelled in Push System). In this phase, the model is calibrated to verify and validate with the real system as discussed in section 5.3. 3. The line performance of the assembly line in pull system is observed using Plant Simulation (modelled in Pull System) and validated according to the throughput rate at 29TPH as described in section 5.4. 4. Real Scenario of production according to monthly delivery plan is implemented and observed the weaknesses(Delivery Shortages) in section 5.5. 5. Optimization of the model was performed and an optimised inventory setup to minimise the Delivery Shortages is proposed in section 5.6. 6. Details of the improvements after optimization are discussed at the end of section 5.6

6.2 Recommendation

Some possible future studies on assembly line analysis and improvement are recommended as follow: 1. Sub-assembly lines for the sub assemblies supplied to the final assembly line can be modelled and integrate with the existing final assembly line to observe more realistic behaviour. 2. Warehouse with the inventory management model supplying the outsourced components can be implement and integrate with the existing final assembly line to observe more realistic behaviour.

75

REFERENCES

1. Falkenauer, E. 2000. Line Balancing in the Real World, International Conference on Product Lifecycle Management.

2. Eryuruk, S.H.; Kalaoglu. F; M. Baskak.(2008), “ Assembly Line Balancing in a Clothing Company” Fibres & Textiles in Eastern Europe January Vol. 16, No. 1(66).

3. Pekin N, Azizoglu M (2008) “Bi criteria flexible assembly line design problem with equipment”

4. Agpak, K. and Gökçen, H., 2005. Assembly line balancing: Two resource constrained cases. International Journal of Production Economics, 96, 129–140.

5. Nicosia, G., Pacciarelli, D., & Pacifici, A. (15 april 2002 ). Optimally balancing assembly lines with different workstations. Discrete Applied Mathematics, 118(1-2), 99-113.

6. Goldratt, E.M. and Cox, J. 1986. The Goal - A Process of Ongoing Improvement. Revised Edition. North River Press, Croton-on-Hudson, NY.

7. Jovisevic, S., Jovisevic, V., & Jokanovic, S. (2009, December). Modeling, Simulation and Optimization of Process Planning. Journal of Production Engineering, 12(1), 87-90.

8. TAKO, A. A.; ROBINSON, S. Model development in discrete-event simulation and system dinymics: an empirical study of experts modellers. European Journal of Operational Research, v. 207, n. 2, p. 784-794, 2010.

9. BANKS, J.; CARSON II, J. S.; NELSON, B. L.; NICOL, D. M. Discrete-event simulation. 5th New Jersey: Prentice-Hall, 2010.

10.SANDANAYAKE, Y. G.; ODUOZA, C. F.; PROVERBS, D. G. A systematic modeling and simulation approach for JIT performance optimization. Robotics and Computer- Integrated Manufacturing, v. 24, n. 6, p. 735-743, 2008.

11.HERNANDEZ, M. A. G.; LIBRANTZ, A. F. H. Improvement of the supply chain for the sugar cane exportation process employing discrete event simulation techniques. Acta Scientiarum. Technology, v. 35, n. 4, p. 637-643, 2013.

76

12.RYAN, J.; HEAVEY, C. Process modeling for simulation. Computers in Industry, v. 57, n. 5, p. 437-450, 2006.

13.HILLIER, F. S.; LIEBERMAN, G. J. Introduction to operations research. 9th ed. New York: McGraw-Hill, 2010.

14.LEAL, F.; COSTA, R. F. S.; MONTEVECHI, J. A. B.; ALMEIDA, D. A.; MARINS, F. A. S. A practical guide for operational validation of discrete event simulation models. Pesquisa Operacional, v. 31, n. 1, p. 57-77, 2011.

15.PEREIRA, T. F.; MOTEVECHI, J. A. B.; MIRANDA, R. C.; FRIEND, J. D.Integrating soft systems methodology to aid simulation conceptual modeling. International Transactions in Operational Research, v. 22, n. 2, p. 265-285, 2015.

16.SARGENT, R. G. Verification and validation of simulation models. Journal of Simulation, v. 7, n. 1, p. 12-24, 2012.

17.Golchev, R., Jovanoski, B., Gechevska, V., & Minovski, R. (2015). Kanban simulation model for production process optimization. JEMC Journal of Engineering Management and Competitiveness, 5(2), 55-60. doi:10.5937/jemc1502055g

18.Banks, J., Carson, J., Nelson, B. L., & Nicol, D. (2004). Discrete-Event System Simulation (4th Edition): Prentice Hall.

19.Gross, J. M., & McInnis, K. R. (2003). Kanban Made Simple: Demystifying and Applying Toyota's Legendary Manufacturing Process: AMACOM.

20.Hao, Q., & Shen, W. (2008). Implementing a hybrid simulation model for a Kanban-based material handling system. Robotics and Computer-Integrated Manufacturing, 24(5), 635-646. doi: 10.1016/j.rcim.2007.09.012

21.Lödding, H. (2013). Handbook of Manufacturing Control. Fundamentals, description, configuration. Berlin: Springer.

22.Minovski, R. (2007). Management Information Systems. Skopje: UKIM.

23.Kurkin, O., & Šimon, M. (2011). Optimization of Layout Using Discrete Event Simulation. IBIMABR IBIMA Business Review Journal, 1-10. doi:10.5171/2011.180343

24.Monden, Y. (1998). Toyota Production System: An Integrated Approach to Just-In-Time: Chapman & Hall.

77

25.Müller, E., Tolujew, J., & Kienzle, F. (2012). Push-Kanban – a kanban-based production control concept for job shops. Production Planning and Control: The management of operations(April), 1-13.

26.Ohno, T., & Bodek, N. (1988). Toyota Production System: Beyond Large-Scale Production: Productivity Press.

27.Pisuchpen, R. (2012). Integration of JIT flexible manufacturing, assembly and disassembly using a simulation approach. Assembly Automation, 32(1), 51-61. doi: 10.1108/01445151211198719

28.Robinson, S. (2004). Simulation: The Practice of Model Development and Use (Vol. 67): John Wiley& Sons Ltd.

29.Siemens. Plant Simulation. from http://www.plm.automation.siemens.com/en_us/products/tecnomatix/plant_design/plant_simulation.shtml

30.Sugimori, Y., & Kusunoki, K. (1977). Toyota production system and kanban system materialization of just-in-time and respect-for-human system. … Journal of Production …(April 2013), 37-41.

31.Welgama, P. S., Mills, R. G. J., & Osmond, G. (1995). Use of simulation in the design of a JIT system. International Journal of Operations & Production Management, 15(9), 245-260.

78

http://www.plm.automation.siemens.com/en_us/products/tecnomatix/plant_design/plant_simulation.shtml

Appendix A Init Method

Var tb:table[string,integer,boolean,string,real,time,date,datetime,object,length,weight,speed]

Delivery_Order.Delete KanbanBuffer.Init_Kanban := KanbanBuffer.Max_Kanban Kanban_Chart.MyAnnotation["Value",1] := KanbanBuffer.Min_Kanban Kanban_Chart.MyAnnotation["Value",2] := KanbanBuffer.Max_Kanban Kanban_Chart.Annotations := Kanban_Chart.MyAnnotation Kanban_Chart.YScaleMax := KanbanBuffer.Max_Kanban * 1.25 Kanban_Plotter.YScaleMax := KanbanBuffer.Max_Kanban * 1.25 Kanban_Plotter.XRange := EventController.End for var i := 1 to KanbanBuffer.Init_Kanban .SNSS_MU.Box_PT22XX0264S.Create(KanbanBuffer) next Packing.Flag := 0 Delivery.Part_Unbox := 0 EventController.Date := str_to_datetime(to_str(Order_Plan[2,1]) + "/" +

Order_Plan[1,1] + "/1" + " 00:00:00") for var i := 3 to Order_Plan.XDim for var j := 1 to Order_Plan.YDim if Order_Plan[i,j] /= 0 Delivery_Order[6,Delivery_Order.YDim + 1] :=

Order_Plan[i,j] Delivery_Order[4,Delivery_Order.YDim] := Order_Plan[0,j] Delivery_Order[3,Delivery_Order.YDim] := 1 Delivery_Order[2,Delivery_Order.YDim] :=

str_to_obj(".SNSS_MU.Box_" + Order_Plan[0,j]) Delivery_Order[1,Delivery_Order.YDim] :=

str_to_datetime(to_str(Order_Plan[2,j]) + "/" + Order_Plan[1,j] + "/" + Order_Plan[i,0] + " 15:00:00")

tb.create tb.ColumnWidth := 10 tb[1,1] := "No" tb[2,1] := Delivery_Order.YDim tb[1,2] := "QTY"

tb[2,2] := Order_Plan[i,j] Delivery_Order[5,Delivery_Order.YDim] := tb

79

tb.forget end next next

Delivery Control Method

if @.QTY <= (KanbanBuffer.numMu * 12) + Delivery.Part_Unbox for var i := 1 to Ceil((@.QTY-Delivery.Part_Unbox)/12) KanbanBuffer.Cont.Move Delivery.Cont.Move next Delivery.Part_Unbox := Delivery.Part_Unbox + (Ceil((@.QTY-Delivery.Part_Unbox)/

12) * 12) - @.QTY Delivery_Order["Delivered",@.No] := @.QTY Delivery_Order["DeliveryShortest",@.No] := "No" @.Move if KanbanBuffer.NumMu <= KanbanBuffer.Min_Kanban for var i := 1 to KanbanBuffer.Max_Kanban - KanbanBuffer.NumMu .SNSS_MU.Box_PT22XX0264S.Create(KanbanSource) next end else Delivery_Order["Delivered",@.No] := (KanbanBuffer.numMu * 12) +

Delivery.Part_Unbox for var i := 1 to KanbanBuffer.numMu KanbanBuffer.Cont.Move Delivery.Cont.Move next Delivery_Order["DeliveryShortest",@.No] := "Yes" Delivery_Shortest := Delivery_Shortest + 1 @.Move for var i := 1 to KanbanBuffer.Max_Kanban - KanbanBuffer.NumMu .SNSS_MU.Box_PT22XX0264S.Create(KanbanSource) next end

80

assembly line balancing and optimization using siemens

Documents