CHAPTER 8 NCM SIMULATION

Transcription

1 183 CHAPTER 8 NCM SIMULATION 8.1 INTRODUCTION isolated simulation as a means of trying to model the impact of variability on manufacturing system behavior and to explore various ways of coping with change and uncertainty. Simulation has provided means to support longer term decisions involving resource requirements, equipment needs and sensitivities to a variety of product demand as well as to shorter term decisions such as shop order releases, and shop floor control decisions by Felix and Chan (24). The objective of this chapter is to develop a simulation methodology and to construct simulation models for small to medium companies for helping building of manufacturing model. The important factors to be selected are availability, risk, cost and performance. The simulation tool will be useful in utilizing the resource availabilities of the enterprises, analyzing how the new work order opportunities might change the system workload to determine the time constraints that will be assigned for the new project. For simulation modeling ARENA simulation tool was used and its sketch was prepared by using visual basics. The following subsections contain brief definitions of modeling and simulation. A detailed explanation of the simulation process has been explained in this chapter.

2

3

4 186 explained as, Briefly, steps involved in developing a simulation model, can be i. Identify the problem, ii. iii. iv. Determine the objectives and overall project plan, Collect and process real system data, Formulate and develop a model, v. Validate the model, vi. Select appropriate experimental design, vii. Establish experimental conditions for runs and perform simulation runs, viii. Documentation cum reporting & implementation Although this is a logical ordering of steps in a simulation study, additional steps at various sub-stages may be required before the objectives of a simulation study are achieved Simulation Benefits & Demerits Simulation has many benefits for the users. First of all, it lets users choose correctly among the possible alternatives, provides time compression and expansion according to the type of the simulated event, equips the managers with the tools real system, allows the user to explore possibilities of new policies, operating procedures or methods. With simulation, one can diagnose problems of complex systems that are almost impossible to deal within the real environment, identify constraints that act as a bottleneck for operations, visualize the plan using the animation capabilities of the software used that results in a more presentable design. Simulation is also beneficial to build

5 187 consensus among the members of the decision makers and to prepare for changes by conside support creates training environments for production team, it can also be used to specify requirements for capabilities of equipment and carry out wise investments using all those properties. In accordance with this definition and benefits, simulation has been extensively used as an off-line decision making tool for helping the management with production planning issues such as efficient capacity utilization, sequencing and scheduling and allocation of resources in manufacturing and production. As outlined in the previous section simulation has many benefits and advantages, however despite these advantages, there are things one should considered carefully on carrying out simulation studies. It is a probability that simulation may not be the perfect tool for all types of system analysis. Some researchers underline four main disadvantages of simulation. The first disadvantage is that model building requires special training and it is highly unlikely that models generated by different modelers about the same system will be the same. The second disadvantage is about t most simulation outputs are essentially random variables based on random inputs, it may be hard to determine whether an observation is a result of system interrelationships or randomness. The third disadvantage is that simulation modeling and analysis can be time consuming and expensive especially when enough resource is not allocated for modeling and analysis, resulting in a simulation model and/or analysis that is not sufficient to the task. A final disadvantage is that simulation may be used inappropriately, especially in some cases when an analytical solution is possible or even preferable.

6 Simulation Application One of the largest application areas for simulation modeling is that of manufacturing systems, effectively in the design and analysis of manufacturing systems. The specific issues that simulation is used to address in manufacturing is identified as follows (a) The need for the quantity of equipment and personnel are, i. Number, type, and layout of machines for a particular objective ii. iii. iv. Requirements for transporters, conveyors, and other support equipment (e.g., pallets and fixtures) Location and size of inventory buffers Evaluation of a change in product volume or mix v. Evaluation of the effect of a new piece of equipment on an existing manufacturing system vi. Evaluation of capital investments vii. Labor-requirements planning viii. Number of shifts (b) Performance evaluation i. Throughput analysis ii. iii. Time-in-system analysis Bottleneck analysis (c) Evaluation of operational procedures i. Production scheduling

7 189 ii. iii. Inventory policies Control strategies [e.g., for an automated guided vehicle system (AGVS)] iv. Reliability analysis (e.g., effect of preventive maintenance) v. Quality-control policies As seen from the above discussion, manufacturing and production offers a huge number of issues to deal with Simulation Tools There are several methods to create simulation models on computer. General programming languages such as FORTRAN, Basic, or C/C++ can be used with some routines to be found from the literatures. One of the several commercially available simulation tools can be utilized. These tools can be divided into three basic classes as follows: i. General-purpose simulation languages, ii. iii. Simulation front-ends and Simulation packages. The general-purpose simulation languages require the user to be a proficient programmer as well as a competent simulator. The simulation front-ends are essentially interface programs between the user and the simulation language being used. The most advanced of all, the simulation packages of today utilize constructs and terminology common to the manufacturing community, and offer graphical presentation and animation.

8 19 Information about some major simulation software can be found from the following web addresses is given in Table 8.1, however it should be noted that there are also other software or simulation languages on the market. Table 8.1 Simulation software on the market Name of The Simulation Tool Automod Promodel Arena AweSim Witness Flexsim Extend GoldSim Mast SimCad Web Address for Further Information VISUAL BASIC SIMULATOR Visual Basic Simulator (VBS) for applications is an implementation of microsoft visual basic which is built into all microsoft office applications, some other microsoft applications such as visio and is at least partially implemented in some other applications such as AutoCAD and MSword. It supersedes and expands on the capabilities of earlier applicationbasic, and can be used to control almost all aspects of the host application. Visual Basic for Applications provides a complete integrated development environment that features the same elements familiar to developers using Microsoft Visual Basic, including a project window, a properties window, and debugging tools. VBS also includes support for Microsoft forms, for creating custom dialog

9 191 boxes, and ActiveX controls, for rapidly building user interfaces. Integrated directly into a host application, VBS offers the advantages of fast, in-process performance, tight integration with the host application (code behind documents, cells, and so forth), and the ability to build solutions without the use of additional tools. As its name suggests, VBS is closely related to Visual Basic, but it can normally only run code from within a host application rather than as a standalone application. It can however be used to control one application from another. A visual basic computer simulator with graphical user interface is developed to investigate experimentally the scheduling algorithm. The simulator is modular in design, that is heuristic algorithm can be easily ported to the system. The computing component of the simulator implements a specific heuristic method of scheduling and products an optimal sequence of jobs. The graphical user interface together with the visual model execution engine allows a step wise execution of the visual basic model. The program works with a VB coding file as an input file the SPT heuristic logical code is shown in Appendix II. In the following section an example of a data file for 12 machines and ten processes that were discussed and given below. Each line represents a job and its execution times on different machines. Upon the startup of the tool the main form will display the given data and allow the user to run the method and calculate the schedule. The simulator program will apply the algorithm against the given data. Then it will display the output. The simulator will allow a deterministic execution of the visual basic model. A step by step visualized execution can be performed alternatively an execution can be requested for any given number of units of time. At the end of execution the tool will display the makespan time, and the machines utilization ratios for the calculated schedule. The software development of

10 192 basic 6., Enterprise Edition as a full application development and Ms-access 7. as database engine. The microsoft windows has emerged as the popular graphical user interface environment. Windows provide considerable advantages the features of window are, Since all windows programs have some fundamental look and feel, users no longer expect to spend long period of time in mastering a new program. According to the number of jobs waiting for processing, time taken to process a job varies. A single process cannot be processed, without any interruption. In order to apply VBS for NCM scheduling problem, five forms shown as screenshots from Figure 8.2 to Figure 8.6 are developed in VBS. Out of five forms, two forms are used as data input forms, one is used as calculation form and remaining two forms are used as output data forms. The Figure 8.2 shows the data input form, by clicking the calculation form from package this form will get opened and ask for number of jobs, number of machines and number of times. After entering all above details the process button can be clicked. Figure 8.2 Data input form 1

11 193 Figure 8.3 Data input form 2 Next input form 2 as in Figure 8.3 will opens and ask for the jobs processing times as input. After entering all its processing times and clicking calculation button the SPT algorithm logical calculation will be carried out and calculation form will appear with makespan and idle time results table as in Figure 8.4. Then by clicking the next button output result form 1 as in Figure 8.5 will open. Figure 8.4 SPT algorithm logical calculation form

12 194 Figure 8.5 Output result form 1 Figure 8.5 can display the results about the scheduling outputs like customer demand per day, machine sequence, makespan value, idle time of machine, idle time of product etc. then by clicking the next button output result form 2 as in Figure 8.6 will open and display the results of Nagare cell output like product cycle time, output/cell/shift, TAKT time, number of operators/cell, total parts produced etc. Figure 8.6 Output result form 2

13 Experimentation A sample of one job with 8 machines problem can be taken and the VBS tool is applied. The processing time for the above problem is given in Table 8.2. As the first step the input of number of jobs as 1, number of machines as 8 and the number of times as 1 are entered in form 1 as like in Figure 8.7. Table 8.2 Product A with its processing times Machine m1 m2 m3 m4 m5 m6 m7 m8 Job A The following screen shots demonstrate the work of VB model. One can find a graphical user interface of the VB tool for the heuristic method. From the Figures 8.7 to 8.11 an example of 1 job and 8 machines Figure 8.7 Input for 1X8 problem

14 196 Figure 8.8 Processing time for 1X8 problem Then the processing times from Table 8.2 are entered in the second input form as like in Figure 8.8, by clicking the calculation button the calculation table can be obtained with the details of M1[Ti], M1[To] i.e. machine 1 in time and out time values for all 8 machines as shown in Figure 8.9. then by clicking the next button the first output form will displays the scheduling outputs as like Figure 8.1, i.e. the number of jobs is 1, customer demand per day is 1, machine sequence is m1-m5-m2-m7-m8-m4-m3-m6, the makespan period is 515 minutes, idle time of machine is zero and idle time of product is 4 minutes.

15 197 Figure 8.9 Makespan and idle time determination for 1X8 problem Figure 8.1 Scheduling output for 1X8 problem Then by clicking the next button, the second output form as shown in Figure 8.11 displays the NCM output, i.e. for making 1 numbers of job A the required product cycle time is 39 minutes, output per cell is minutes, TAKT time is 13.5 minutes. Similarly the above steps can be applied for 2 jobs (A and B) 8 machines problem. It s processing times are given in Table 8.3. The VBS forms are shown in Figure 8.12 to 8.16.

16 198 Figure 8.11 NCM output for 1X8 problem Table 8.3 Product A & B with its processing times. Machine m1 m2 m3 m4 m5 m6 m7 m8 Job A Job B Figure 8.12 Input for 2X8 problem

17 199 Figure 8.13 Processing time for 2X8 problem Figure 8.14 Makespan and idle time determination for 2X8 problem

18 2 Figure 8.15 Scheduling output for 2X8 problem Figure 8.16 NCM output for 2X8 problem The final output for 2 jobs and 8 machines problem is shown in Figure It displays the output for making 1 number of job A and B, the required product cycle time will be 19.2 minutes, output per cell will be 142 minutes, TAKT time will be 27 minutes.

19 21 The above procedure can be applied for 1 jobs with 12 machines problem with a total of 1 products in each job with its processing time as given in Table 8.4. Table 8.4 Product A to J with its processing times Machine m1 m2 m3 m4 m5 m6 m7 m8 m9 m1 m11 m12 Job A Job B Job C Job D Job E Job F Job G Job H Job I Job J Figure 8.17 Processing time for 1X12 problem

20 22 Figure 8.18 Makespan and idle time determination for 1X12 problem Figure 8.19 Total ready time, idle time of machine and product determination screen shot. Figure 8.2 screen shot of processing the 1 th product of job J

21 23 Figure 8.21 Scheduling output for 1X12 problem Figure 8.22 NCM output for 1X12 problem

22 An Industrial Case Study The complexity of the scheduling problem has been reduced by decomposing all parts into n k part family and the corresponding machines into m k machines for the k th NCM centre. Thus the first step is forming a number of NCM centres. A program is written to read the input data (k,n and m) for each NCM centres. To demonstrate how the proposed method works, some practical problems from an industrial partner had been used. The unit consists of 5 different components, out of which 31 components are being manufactured using 12 different workstations centres. Each part considered as a job order, which includes many operations where each operation takes a certain processing time. The processing time for each operation in the corresponding workstation is given in Table 8.5. Using the processing times a case study was conducted at NCM centre, to implement the proposed method; the close to optimum machine sequence generated by VBS is shown in Table 8.6 Table 8.5 Processing times of 31 different jobs by 12 machines Machine m1 m2 m3 m4 m5 m6 m7 m8 m9 m1 m11 m12 Job A Job B Job C Job D Job E Job F Job G Job H Job I Job J Job K Job L Job M Job N Job O Job P Job Q Job R Job S

23 25 Table 8.5 (Continued) Job T Job U Job V Job W Job X Job Y Job Z Job AA Job AB Job AC Job AD Job AE Results and Graph To demonstrate how effective the proposed VBS as a stochastic, the above case study practical problem from an industrial partner was used. The processing time for each operation in the corresponding workstation are taken. Using that processing time, for the given case study of NCM centers, the close to optimum machine sequence were generated through the implementation of the VB simulator. The simulator output results are calculated and summarized in Table 8.6. By keeping the customer demand per day as 1, a variety of jobs of range 1 job to 31 jobs were processed through the simulator in a same 12 machines workstation the processing results of machine sequence for arranging the machines in a U shaped product layout has been calculated, then the minimal makespan is observed. Similarly the performance parameters of idle time of machine, idle time of product, product cycle time, output per cell, TAKT time are all calculated and given in Table 8.6. From the table it is observed that by grouping the parts into a part family the makespan can be reduced, product cycle time can be reduced and idle time of machine shows zero time.

24 Table 8.6 Simulator output results for 31 different jobs processed in 12 machines Job CDPD Machine Schedule Makespan Idle Time of machine Idle Time of Product Product Cycle Time Outpu t /cell TAKT Time No. of operator s/ cell Total parts produced `

25 Table 8.6 (Continued) Job CDPD Machine Schedule Makespan Idle Time of machine Idle Time of Product Product Cycle Time Outpu t /cell

26 28 In addition a comparative chart is prepared for makespan and product cycle time which is shown in the Figure It says while product group is increased the makespan value and product cycle time value gets reduced. And the Figure 8.24 shows the comparative charts of all jobs between the idle time and the TAKT time. It was observed that the idle time of machine is zero for all 31 jobs and the idle time of product & TAKT time gets increased by increasing the product group. Time in minutes Number of jobs Makespan Product Cycle Time Figure 8.23 Average makespan and cycle time 3 25 Time in minutes Idle Time of machine Idle Time of Product TAKT Time -1 Number of jobs Figure 8.24 Average idle time and TAKT time

27 ARENA SIMULATION The ARENA modeling system from Systems Modeling Corporation is a flexible and powerful tool that allows analysts to create animated simulation models that accurately represent virtually any system. ARENA employs an object-oriented design for entirely graphical model development. Simulation analysts place graphical objects, called modules, on a layout in order to define system components such as machines, operators, and material handling devices. ARENA is built on the SIMAN simulation language. After creating a simulation model graphically, ARENA automatically generates the underlying SIMAN model used to perform simulation runs. ARENA has many unique properties which are, ARENA has a natural and consistent modeling methodology due to its flowchart style model building regardless of detail or complexity. Even the flowcharts of systems created by Microsoft Visio can be imported and used directly. It is extendable and customizable, which results in a re-creatable, reusable and distributable templates tailored to specific applications. The scalable architecture of ARENA provides a modeling medium that is easy enough to suit the needs of the beginner, and powerful enough to satisfy the demands of the most advanced users. This makes it a perfect tool for continuously improving modeling studies as the other advantage of ARENA is that it is open to interaction with many applications such as Microsoft Access and Excel with its built-in spreadsheet data interface. Arena Packaging is a simulation system for the performance analysis of high-speed, high-volume manufacturing systems Template Overview Arena packaging is one of a family of application solution templates (ASTs) built on the Arena simulation system. It is designed specifically for performing accurate and efficient simulations of high-speed,

28 21 high-volume manufacturing systems, where the processing rates take place at hundreds, even thousands, of entities per minute. The Packaging template enables users to build and run simulation models of high-speed processing lines quickly and easily, and to analyze the results that these models produce ARENA Tools and Features ARENA has three main tools they are, Input Analyzer - can be used to process and classify the obtained data for input data analysis. Appropriate probability distributions can be obtained for being used in the models. Output Analyzer - made the user carry out statistical analysis on the results obtained. Process Analyzer - helps to examine the selected outcomes of several different alternatives dependent on selected controls on the system. The most attractive feature of a simulation study is the animation that accompanies the model. Most people are interested in watching animated actions and graphs rather than straight numbers and texts. ARENA has a powerful animation tool to help the user to pass his/her ideas, studies and results to the audience easily. ARENA animations can be run concurrently with the executing simulation model. For any manufacturing environment, the processing and analyzing can be done from simulation by following five easy steps with Arena: Step 1 Create a basic model. Step 2 Refine the model.

29 211 Step 3 Simulate the model. Step 4 Analyze simulation results Step 5 Select the best alternative Simulation Concepts (i) Entities and Attributes In every simulation model, entities represent the objects moving through the system. Each entity has its own characteristics, refer to as attributes. It can define as many attributes as need for the entities in this system. Each individual entity in the system has its own values of these attributes; these may be assigned at the various processes it encounters. (ii) Queues The primary purpose of a queue is to provide a waiting space for entities whose movement through the model has been suspended due to the system status (e.g., a busy resource). Queues are passive in nature; entities enter the queue and are removed from it based upon the change in state of the system element associated with the queue. There are two types of queues used in Arena. Individual queues Internal queues (ii) Resources Resources are stationary elements of a system that can be allocated to entities. They have a specified capacity (at any point in time) and a set of

30 212 states (e.g., busy, idle, inactive, or failed) that they transition between during a simulation run. Resources may be used to represent people, machines, or even space in a storage area. Resource terminologies Seizes, Releases, Unit, Schedule, Downtimes, Failures Resources are depicted in the animation by a stationary set of pictures representing the states of the resource (idle, busy, etc.) The default pictures can be customized to better represent the resources in this system from more information on animating resources. (iv) Modeling environment The Arena modeling environment will open with a new model window as shown in Figure To model the process it would be work with three main regions of the application window. The Project Bar hosts panels with the primary types of objects that will work

31 213 Figure 8.25 Arena model window with Basic processes (v) Basic Process panel the process. Contain the modeling shapes, called modules, that uses to define (vi) Reports panel simulation runs. Contains the reports that are available for displaying results of (vii) Navigate panel Allows to display different views of the model, including navigating through hierarchical sub models and displaying a model thumbnail. In the model window, there are two main regions. The flowchart view will contain all the model graphics, including the process flowchart, animation, and other drawing elements. The lower, spreadsheet view displays model data, such as times, costs, and other parameters.

32 214 (viii) Exhibit Task Create module, w Create module, from the Basic Process panel. This is the starting point for the flow of entities through the model. 1. Drag the Create module from the Basic Process panel into the some data to support the simulation. a more meaningful description as well as Figure 8.26 Process flow chart model (ix) Process flowchart Build a flowchart the word itself flowchart suggests two of the main concepts behind modeling and simulation i.e. a chart refer to as a process map or a model that describes a flow. Flow refers to as entities that will move through the process steps in the model which is shown in Figure (x) Process module Next in our flowchart is a Process module that represents the Review Application step.

33 Be sure that the Create module is selected so that Arena will automatically connect the Process to the Create module. 2. Drag a Process module from the Basic Process panel into the model window, placing it to the right of the Create. Arena will automatically connect the two modules. As with the Create, (xi) Decide module After the Process, we have a Decide module which determines whether the mortgage application is complete the Object > Auto- Connect menu), be sure that the Process module is selected so that the Decide module will be connected to it. 2. Drag a Decide module to the right of the Process module. If the mortgage application has a complete set of information, it will leave the Decide module from the right side of the diamond shape, representing the True condition. Incomplete applications (False result to the Decide test) will leave via the bottom connection. (xii) Dispose module Dispose module that represents accepted applications, connecting to the True (right) output from the Decide shape. applications.

34 Select the Decide shape so that our first Dispose will be connected automatically. 2. Drag a Dispose module to the right of the Decide module. Arena will connect it to the primary (True) exit point of the -and-drop sequence.) 3. To add the second Dispose module, once again select the Decide module, so that Arena will automatically connect its False exit point to the new Dispose module, and drag another Dispose module below and to the right of the Decide module. 4. Drag and drop another Dispose module, placing it below and to the right of the Decide shape, completing the process flowchart Module Creation In Arena, modules are the flowchart and data objects that define the process to be simulated. All information required to simulate a process is stored in modules. Those are placed in the model window to describe the process. In the basic process panel, these are the first eight shapes used to construct the flow chart: Create: The start of process flow. Entities enter the simulation here. Dispose: The end of process flow. Entities are removed from the simulation here. Process: An activity, usually performed by one or more resources and requiring some time to complete.

35 217 Decide: A branch in process flow. Only one branch is taken. Batch: Collect a number of entities before they can continue processing. Separate: Duplicate entities for concurrent or parallel processing, or separating a previously established batch of entities. Assign: Change the value of some parameter (during the simulation), such as Record: Collect a statistic, such as an entity count or cycle time. Simulation settings are defined in the Run > Setup > Replication Parameters dialog box. There is also a set of data modules for defining the characteristics of various process elements, such as resources and queues Definition of Model Data A basic flowchart can be drawn for one mortgage application process, to define the data associated with the modules, including the name of the module and information that will be used to simulate the process. (i) Create module First the Create module named as Initiate Mortgage Application. Its data will include the type of entity to be created for a mortgage Application. 1. Double-click the Create module to open its property dialog box. 2. In the Name field, type Initiate Mortgage Application. 3. For the Entity Type, name our entities by typing Application.

36 Type 2 in the Value field of the time between arrivals section. 5. Click OK to close the dialog box. Entities are the items, documents, parts, produced, or otherwise acted on by the designer process. Manufacturing models typically have some kind of part running through the process, whether it can be a raw material, a subcomponent, or finished product. (ii) Process module the system being modeled. The application will be reviewed for completeness this will take some amount of time, holding the entity at this point in the flowchart for a delay and requiring a resource module also as Review Application. The designer should specify the minimum time in which the work could be done, the most likely value for the time delay, and the maximum duration of the process. During the simulation run, each time an entity enters the process. For some Review Application process, a minimum time of 1 hour, most likely value of 1.75 hours, and a maximum of 3 hours can be assigned to a resource, to perform the process. 1. Double-click the Process module to open its property dialog box. 2. In the Name field, type Review Application. 3. To define a resource to perform this process, pull down the Action list and select Seize Delay Release.

37 219 Arriving entities will wait their turn for the resource to be available. When its turn comes, the entity will seize the resource, delay for the process time, and then release the resource to do other work. 4. A list of resources will appear in the center of the dialog box. To add a resource for this process, click Add. 5. In the Resource Name field of the Resource dialog box, type Mortgage Review Clerk. 6. Click OK to close the Resource dialog box. 7. Define the process delay parameters in the Minimum, Value (Most Likely), and Maximum fields as 1, 1.75, and 3. (Note that the default delay type is Triangular and the default time units are in hours.) 8. Click OK default values for the other Process module properties. Feel free to explore their purposes through online help or the Modeling Concepts and Resources models in the SMARTs library. (iii) Decide module After the mortgage application has been reviewed, it should be determined whether to accept or return the application. In Arena, whenever an entity selects among branches in the process logic, taking just one of the alternatives, a Decide module is used. For the mortgage application process to determine the outcome of the decision, with 88% of applications accepted as complete.

38 22 1. Double-click the Decide module to open its property dialog box. 2. In the Name field, type Complete? 3. For the Percent True field, type 88 to define the percent of depart through the exit point at the right of the Decide module). 4. Click OK to close the dialog box. (iv) Dispose module For a simple process of reviewing mortgage applications, remove the mortgage applications from the model, terminating the process by a Dispose module. Because there are two possible outcomes of the mortgage application process-applications can be accepted or returned as shown in Figure Double-click the first Dispose module (connected to the True condition branch of the Decide module) to open its property dialog box, and in the Name field, type Accepted. 2. Click OK to close the dialog box. 3. Double-click the other Dispose module to open its property dialog box. In the Name field, type Returned.

39 221 Figure 8.27 Decide module window (v) Resource module Along with the flowchart, define the parameters associated with other elements of the model, such as resources, entities, queues, etc. For the mortgage process, simulation results will report the cost associated with performing the process. To provide the parameters to the model, enter them in the Resources spreadsheet as in Figure In the Basic Process panel, click the Resource icon to display the Resources spreadsheet. 2. Because we defined the Mortgage Review Clerk as the resource in the Review Application process, Arena has automatically added a resource with this name in the Resources spreadsheet. Click in the Busy/Hour cell and define the cost rate when the clerk is busy by typing 12. Click in the Idle/Hour cell and assign the idle cost rate by typing 12.

40 222 Figure 8.28 Resource module spread sheet. (vi) Prepare for the simulation To make the model ready for simulation, one should specify the general project information and the duration of the simulation run. Just by testing the first-cut model, to perform a short, 2-day run. 1. Open the Project Parameters dialog box by using the Run > Setup menu item and clicking the Project Parameters tab. In the Project Title field, type Mortgage Review Analysis; then leave the Statistics Collection check boxes as the defaults, with Entities, Queues, Resources, and Processes also check the costing box. 2. Next, click the Replication Parameters tab within the same Run Setup dialog box. In the Replication Length field, type 2; and in the Time Units field directly to the right of Replication Length, select days from the drop-down list. Click OK to close the dialog box. Save the simulation model by click the Save on the standard toolbar or select the File > Save menu item. of the model definition, including the flowchart, other graphics drawn, and the module data entered. By perform a

41 223 simulation run; the results are stored in a database using the same name as the model file. (vii) Simulate the process With these few short steps, the mortgage application model contains all of the information needed to run the simulation. Start the simulation run by clicking the Go button or clicking the Run > Go menu item. Arena first will check to determine whether a valid model is defined, then will launch the simulation. As the simulation progresses, one can see small entity pictures resembling pages moving among the flowchart shapes as like in Figure Also, a variety of variables change values as entities are created and processed, as illustrated in Figure If the animation is moving too fast, it can be slow down by adjusting the animation scale factor. Open the Run Setup dialog box via the Run > Speed > Animation Speed Factor menu item and enter a smaller value (e.g.,.5) for the scale factor; or Use the less-than (<) key during the run to decrease the scale factor by 2%. Be sure that the model window is active not the Navigate panel or > and < < repeatedly is an easy way to fine tune the animation speed. The greater-than (>) key speeds up animation by 2%. Use the slider bar in the main toolbar. Move the slider to the left to slow down the animation; move the slider to the right to speed up the animation. To pause the simulation, click the Pause button or press the Esc key. With the automatic flowchart animation, one can see how many entities

42 224 have been created and currently in the Review Application process, have left each branch of our Decide module, and have left the model at each of terminating Dispose modules. These variables can be helpful in verifying the model. Figure 8.29 Simulation process One can step through the simulation one event at a time i.e. pause the simulation. Each time by stepping the simulation, an entity is moved movement, (viii) Simulation reports After watching some of the animated flowchart, one can quickly run to the end of the simulation to view reports as in Figure 8.3. Pause the simulation and then click the Fast Forward button to run the simulation without updating the animation. At the end of the run, Arena will display the default report in a report window, as shown below. On the left side of each report window a tree listing the types of information is available in the report. The project name is listed at the top of

43 225 the tree, followed by an entry for each category of data. This report summarizes the results across all replications. Figure 8.3 Simulation model report By clicking on the entries inside the category sections, one can view various types of results from the simulation run. Each report will be displayed in its own window. After viewed the reports end the Arena run session. (ix) Enhancing the visualization process After completing the basic steps for analyzing the simulation application process, one can return to the model and embellish the graphical animation to gain further insight into the process dynamics. Animation will be of great benefit in enticing others in the organization to be interested in process improvement. It has to enhance the visualization components to the

44 226 model. So first a Review Clerk working at a desk, either busy or idle to gain a better sense of how many applications are waiting in the Review Application process over time. Secondly a dynamic plot of the work-in-process (WIP) simulation variable to be added. Now the Arena model will appear as in Figure 8.31 after adding two those two components. Figure 8.31 Review clerk and WIP plot (x) Review clerk animation During the simulation run, the Review Clerk resource can be in one of two states. If no application entity is in-process, then the resource is idle. A picture of a person sitting at a desk to depict idleness can be used. When an entity seizes the resource, busy, for this case the picture will show the person reviewing a document. (xi) WIP plot animation The second animation enhancement is a plot of how many applications are under review as the simulation progresses. It will give us a sense of the dynamics of the workload, which can vary quite a bit when the

45 227 random nature of processes is incorporated into a simulated model as in Figure (xii) Rerun the simulation To make the animation more interesting and valuable, the simulation can be rerun again. Because without changing any of the process parameters the simulation has to provide the same results. By starting the picture change from idle (sitting at the desk) to busy (reading a document) and back again, as application entities move through the Review Application process. The plot as in Figure 8.31 shows some significant peaks in the number of applications that are under review, caused by the combination of the variation in the time between arrivals of applications and the time to process applications. 8.5 NCM SIMULATION Simulation tools are very familiar and widely used for processing the manufacturing systems due to some most important reasons and advantages like i) Realistic models are possible; ii) Options and alternative designs may be considered without direct system experimentation. iii) A computer simulation models directly addresses the performance measures. iv) Non-existent systems may be modeled. v) Visual output helps and assists the end-user in model development and validation; vi) Manufacturing cell sizing, queue sizes, and others design parameters can be done. Simulation models can provide increased comprehension and improved insight into the performance of a manufacturing system. The construction of simulation model forces the modeler to ask the above questions before modeling. Analysis of the numerical results of the simulation

46 228 runs can be used to identify true performance indicators for the system such as total time in the system for a part, work-in-process inventory, and machine utilization for making etc. Most of simulation studies have indicated the importance of workload balancing and machine utilization in determining the advantage of production. The proposed simulation methodology which is described in this subsection is working by the logic shown in Figure The logic is by providing the number of jobs as input, getting processed in NCM cell and released as output by finished product. Figure 8.32 Simulation logic diagram. Figure 8.33 Modeling of NCM system model.

47 Model construction & performance measures The Arena simulation language is used to develop the simulation model for NCM cell. Figure 8.33 shows a modeling of NCM system by Arena software. The construction of simulation model has some assumptions that machine capacities are enough to process all forecasted demand with some considerations incorporated in the model used are, i) All processing times acquired were deterministic ii) Each transfer movement of a job will have durations that are exponentially distributed iii) Queue capacities for processing machines are set at 1 products iv) Jobs can be removed in batches from queues for processing according to the SPT rule v) When jobs are arrived to the queues serving processing machines, they will be placed at the back of the queue according to the scheduling order. The Figure 8.34 shows the zoomed view of NCM model, the main arrivals, cells, and jobs departures. For instance, the module refer as ensures the entry of the job batches within the system as input. Then each batch is assigned a set of attributes such as job type and SPT sequence routing, via module. As the part proceeds through the cell, different attributes record the time delays associated with material handling, processing, machine transfer etc. The module is referred as allows parts to the corresponding cell type. Each part families are assigned to the corresponding cell via modules Then, it is ready to send parts on its way to the transfers an entity to a specified machine station, or the next station in the station visitation sequence defined for the entity. The machine transfer time is ente for each part sequence as a route time. Now that have the arriving batches of parts being routed according to their assigned part sequences, a

48 23 part arrives to the cell, queues for a machine, is processed by the machine, and sent to its next step in the part sequence. J O B A M 6 M 7 J O B B J O B C J O B D M 8 M 5 J O B E J O B F J O B G M 4 M 9 J O B H J O B I J O B J C OM P L E T E Tr u e A C C E P T E D J O B K Fa ls e M 3 M 1 J O B L R E T U R N J O B M J O B N J O B O M 2 M 1 1 J O B P J O B Q J O B R M 1 2 M 1 J O B S J O B T Figure 8.34 NCM Simulation Model by Arena All cells can be modeled by a set of machines, which each one is modeled using the module sequence. Finally, the batches leave the system

49 231 through the module refer as or as finished parts. Here four performance measures like time per entity, time per process, time per resources and time per queue were employed to evaluate the effect of simulation time in NCM system about VA time, NVA time, Wait time, Transfer time etc Simulation Results and Analysis Attention to be focused on the routing and on the sufficient machines capacity in each manufacturing cell. The purpose of procedure consists about evaluation of movements using the initial model which permits the part transfers for all jobs. Before proceeding simulation run, the steady state be established. A plot realized by the depicts the transient behavior of the simulation model after start-up from the "empty and idle" state Figure 8.35 WIP plot By running the simulation a plot as shown in Figure 8.35 is created which explains how the jobs are processing. The plot consists of some SIM expressions M1 WIP, M2 WIP, up to M12 WIP with time range. That plot reports the warm up period as 6293 minutes for several run length.

50 232 Several simulation runs were made for the initial system configuration, each run for total demand. The results of these simulation runs are realized with the help of the simulator ARENA. The result of these runs is shown in Table 8.1 to Table 8.4. Table 8.7 Process summary table

51 Table 8.8 Entities summary table 233

52 Table 8.9 Resources summary table 234

53 235 Table 8.1 Queues summary table Some sample problems of up to 2 jobs and 12 machines job resources, queues, process and entities are shown in appendix section. Though Table 8.1 to 8.4 gives the experiment performance measures for 2 jobs processed in 12 machines, the average number of batches waiting in machine

54 236 queue, number of jobs sized, VA time, other queue time and the machine utilization are shown too. The results indicate that M11 is a bottleneck machine with accumulated waiting time of minutes. It requires priority for machine rescheduling the jobs. But from the Table 8.2 it is observed that NVA time and transfer time was zero. So it decides the effective utilization of machine i.e. idle time of machine is zero. It is observed that in this problem there are generous benefits gained from employing a mixed transfer batch Conclusion A methodology-based simulation for evaluating the NCM scheduling system used to test the industrial case study up to 2 jobs conducted on 12 machining centers to analyze the optimization scheduling parameters was carried out. ARENA simulation software version 11. is used to model this problem and study about results for different performance measures like NVA time, VA time, wait time, queue length and machine utilization. The model is built based on one of the optimum flow shop sequence obtained from SPT heuristic and run for more number of replications. The results are analyzed and modified based on their utilization and queue lengths. Also, the results show that some resources are excessively used and lead to slow throughput. This may drastically reduce the number of parts produced out of the system and increase the average WIP. This causes bottlenecks in the system can be solved by modifications through increasing the machine capacity in the NCM. Necessary changes can be made and simulation results with statistical analysis will enhance the production manager to view in depth all scenarios of the operations and resource limitations and optimization with complete solution.

55 SUMMARY This chapter clearly explains about a development of simulation methodology to construct the simulation models for small to medium companies in helping the building of manufacturing models. Some important factors like availability, risk, cost and performance should be considered during processing. The simulation tools can be used for utilizing the resource availabilities of the enterprises, analyzing how the new work order opportunities might change the system workload to determine the time constraints that will be assigned for the new project etc. The sub-sections contain brief definitions of modeling and simulation through ARENA. A detailed explanation of the simulation process has been explained in this chapter. Further from NCM simulation section one can understand the use of simulation technique for helping the decision maker to have all details about resource analysis and can make scientific decision.