1

Designing a Lean Value Stream

Overview The steps that you will take to create a lean value stream are

1. Select a product family (done earlier)

2. Identify the value stream owner (done earlier)

3. Map the current value stream (done earlier)

4. Determine the rate of production required to meet customer demand, “takt time”

5. Develop continuous flow wherever possible

6. Use Supermarkets and kanbans to control production where continuous flow does not extend upstream

7. Send the customer schedule to only one production process, the “pacemaker process”

8. Level the production mix

9. Level the production volume

10. Develop the ability to make every part every day (or faster) upstream of the pacemaker process.

Continuous Flow The goal of lean is to develop a value stream where value flows continuously to customers with no waste. This means no waiting, no inventory, no defects, no unnecessary motion; none of the 7 types of waste discussed in earlier lessons. Ideally a customer would say “I want 10 widgets delivered at 10AM each day of the workweek” and that would trigger the value stream to create 10 units from scratch and deliver them at 10AM every day. I.e., the needed supplies would be delivered to each process in the value stream at exactly the right time to be worked on, the work would be done perfectly, and the widget would be moved a minimal distance until it was finally delivered to the customer.

Henry Ford’s assembly line is an example of a value stream that could have been an ideal lean value stream. (It wasn’t because cars were identical and made to long-term forecasts, but that’s another story.) What if every Model T that came off of the assembly line was driven away by a customer who had ordered the car? If the production line value stream was triggered by the customer’s order, we would have something very close to the ideal of continuous flow.

Continuous flow is possible when product can be produced and moved to the next process one-unit-at-a-time, without delay and without errors or scrap. When it is possible to do this many benefits result, such as lower costs, instant signals of quality problems, lower costs, no work-in-process inventory, etc. When flow is not continuous it will create the opposite results. In other words, discontinuous flow will result in higher costs, delayed signals of quality problems, work-in-process inventory accumulations, etc..

2

Reasons for Discontinuous Flow Discontinuous flow has many possible causes. Batch production or work scheduling separate processes using long-term forecasts is a primary cause. Other causes include

• Long physical distances between processes and process steps.

• Quality problems due to either sporadic special causes or chronic problems caused by process or product design flaws.

• Poor machine or equipment maintenance.

• Unpredictable customer demand.

• Acts of God causing unpredictable interruptions to supplies.

• Long changeover times.

• Long setup times.

• Supplier issues: unreliable deliveries, long lead times, business failures, multiple suppliers for the same item, quality problems, etc.

Some of these causes can be addressed immediately by the design of the value stream. This will involve making the value stream as close as possible to continuous flow at the outset. The remaining causes are addressed long term by kaizen events, which are short-duration projects that use the DMAIC framework to make quick improvements. Lean Six Sigma is used to tackle especially stubborn problems identified by leadership or during kaizen events. The importance of an organizational culture that promotes the ideal of zero waste cannot be overemphasized.

Steps toward continuous flow

Takt Time To achieve our ideal of continuous flow we need to know how many items our customers want and when they want them. This demand rate is known as takt time. We begin by calculating this value using the equation

available time per daytakt time=customer demand per day

(1.1)

This value is used to synchronize the entire value stream. For example, if a disability claims office works from 9am-5pm with 1 hour for lunch and two 10 minute breaks then it has 400 minutes of time available to work each day. If it receives an average of 50 claims per day then the takt time is 400/50=8 minutes per claim, i.e., the office should complete one claim every 8 minutes. If it completes more claims than this then it is probably overstaffed (wasted resources.) If it completes fewer then customers will have to wait too long to learn the disposition of their claims (wait time waste.)

If a factory produced widgets in the same 9-to-5 schedule with the same 1 hour lunch and two 10 minute breaks, then it would need to know the average customer demand for widgets. If the demand was 50 widgets per day, then the factory should complete one widget every 8 minutes. Faster production would result in accumulated finished goods

3

inventory. Slower production would mean customers would not get the widgets they need when they need them. Takt time is shown in data boxes for processes in the future state value stream map.

Supermarkets and Kanbans When continuous flow is not possible (and it usually is not possible to get perfect continuous flow,) you will use “supermarkets” and “kanbans.” You may recall that in Lean Six Sigma the term supermarkets refers to small, controlled amounts of inventory that are used to schedule production at upstream processes. Kanbans are cards (or other devices, but we’ll use the term cards for all kanbans) used to control production and movement of materials and parts. There are several types of Kanbans as shown in Table 1. Table 1-Types of Kanbans

Symbol Type of Kanban Comment

Withdrawal kanban. Kanban that instructs the material handler to get parts from the supermarket to the process that needs them.

Production kanban Kanbans that tells a process how many of an item to produce and gives it permission to do so.

Batched kanbans A group of kanbans to be picked up at one time

Signal kanban Signals when a reorder point is reached and another batch needs to be produced. Used when the supplying process produces in batches and requires a changeover to produce the required item.

Kanban post The place where kanban cards are collected for pickup

4

Pacemaker Process Takt time is used to schedule the “pacemaker process” in the value stream. This process is used to trigger production by other processes in the value stream. Because of its importance, care should be taken when selecting the pacemaker process. The pacemaker process should be the continuous flow process that is the furthest downstream in the value stream. Since we want pacemaker processes to schedule the entire value stream, there should not be any supermarkets downstream from the pacemaker process. Figure 1 shows a simple example of a pacemaker process for a typical value stream.

Figure 1-Pacemaker Process

If value streams are producing customized, made-to-order items, then the pacemaker process may need to be further upstream. However, the “no supermarkets downstream” rule still applies. Figure 2 Shows an example of this situation. Figure 2-Pacemaker Placement for Custom Production

Production Mix The production mix is set at the pacemaker process. The basic idea is to produce a mix of products that, over a relatively short time interval, matches the mix of products demanded by customers. The ideal flow process is single-item flow, so the ideal product mix would

5

be build-to-order. However, for a variety of reasons, this ideal is seldom possible.1 The primary purpose of creating an even production mix is to eliminate the need for batching. When the pacemaker process produces in batches, the resources it consumes are also used in batches. This creates batching in all upstream processes that feed the pacemaker process, and upstream inventories tend to grow larger as you move further upstream.

A level production mix also makes it easier to respond to different customer requirements with a short lead time and little inventory. The downside is that at assembly there will be a larger number of changeovers and the need to keep components for different assemblies near at hand. The icon for leveling is shown in Figure 3. On the value stream map the icon is shown in the information flow, see Figure 4. Figure 3-Load Leveling Symbol

Figure 4-Example of Using Load Leveling Icon

Production Volume After leveling the production mix, the next step is to level the production volume. This is achieved at the pacemaker process by releasing and withdrawing small and consistent amounts of product in pace with takt time. By doing this we keep the entire value stream responsive to the demands of the customer. Batch-and-queue systems, in contrast, produce according to schedules at each separate process in the value stream, causing all

1 Services are an exception. Produce-to-order is relatively common in the services sector. It would be pointless to treat an emergency patient before they experience the emergency!

6

of the problems discussed earlier. In addition to the chronic problems, the situation is even worse when customer demand changes suddenly and the company is left with large inventories and no demand for them. This has lead to catastrophe that has spelled the end of more than one company.

Consistent, level production rates are a primary goal of Lean Six Sigma. It is only possible when quality is high because poor quality disrupts production flow and is usually compensated for by maintaining large stocks of "just in case" inventory. When small amounts of production are released, say an hour’s worth or less, poor quality will bring things to a halt. This is actually a good thing because it is likely that the cause of the quality problem is still identifiable. The time delay created by inventories is a huge barrier to quality improvement because the cause of quality problems has long since vanished, only to reappear again and again. With this arrangement the causes of quality problems are merely items of historical interest rather than something that can be realistically addressed.

Of course, production is removed from the pacemaker process at the same rate that production orders are issued. This “paced withdrawal” keeps the pacemaker process free from inventory buildup.

The consistent increment of work is referred to as the pitch. The pitch increment is calculated by considering both takt time and the quantity held by a container (pack size.) The formula used is

takt time pack size = time it takes to produce a full containerpitch = × (1.2)

For example, if your takt time is 1 minute and the container size is 10 units then pitch = 10 minutes (1 minute x 10 units = 10 minutes.) This is called the “takt image.” In practice we would, every 10 minutes, tell the pacemaker process to make 10 units and take away one finished pitch quantity from the process. The pitch quantity is the unit by which the production is scheduled for a given product family. Just as we don’t batch parts, with Lean we also don’t batch production schedules.

The pitch time is the time frame used to monitor value stream performance. This is in stark contrast to the usual management time frame, which is often days, weeks, or even months. Just as the delay created by inventories makes it almost impossible to solve quality problems, the delay created by long management time frames does the same with production problems.

Load-leveling Box (heijunka box) One mechanism used to accomplish load leveling is the load-leveling box, Figure 5. Think of a load-leveling box as a data table where each row in the table represents one product in the product family, each column represents a pitch time period, and each cell in the table is a place where kanbans are stored, see Table 2. The bins in the box each hold one or more kanbans. By staggering the kanbans in the load-leveling box at the pitch interval we automatically create a consistently level production mix.

7

Figure 5-Load-Leveling Box (Heijunka Box)2

Table 2-Table Explaining Heijunka Box

Pitch Time →

Product ↓ 9:00 9:10 9:20 9:30 9:40 9:50

A Kanban Kanban

B Kanban Kanban

C Kanban Kanban

There is no separate icon for leveling the production pace, we use the same symbol as for leveling the production mix. Actually, it is not only the icon that is the same, leveling the mix and the pace of production are one and the same since we must level both simultaneously to have lean production.

Every Part Every Day If items (parts, services, transactions, etc.) are produced repeatedly, we’d like to have a system where upstream processes could create the quantity of each item needed by customers immediately. As a step towards this goal we begin with a target of producing every item needed, in the quantity needed, at least every day. Once this is accomplished we can begin the continuous improvement journey and reduce the “EPE” interval from a day to an hour or eventually a pitch interval.

The primary factor that determines the ability of an upstream fabrication process to produce multiple items is the time it takes to changeover from one item to another. As

2 http://en.wikipedia.org/wiki/Heijunka_box. In this illustration, each time period builds an A and two Bs along with a mix of Cs, Ds and Es. What is clear from the box, from the simple repeating patterns of kanbans in each row, is that the production is smooth of each of these products.

8

this time decreases, the number of parts that can be produced in a day increases. In fact, changeover time can be used to determine batch sizes. Calculate the number of hours you have available in a workday, then subtract the number of hours required to produce the daily requirement. The time left over is the time available for changeovers.3

Future State Value Stream Map The principles presented above serve as guidelines for creating a lean value stream. The tools and icons will be used to draw a picture of the new value stream you plan to build. In Lean Six Sigma we are usually not asked to create a brand-new value stream from scratch. Instead we start with the current value stream and apply lean principles to evaluate and improve it. Figure 6 shows a current state value stream map from an earlier lesson.4 We'll use this example to illustrate the process of creating a future state value stream map.

Figure 6-Current State Value Stream Map

3 Actually, this is also done for batch-and-queue processes. But before Lean Six Sigma has been applied the changeover times are frequently so long that huge batch sizes are required to amortize the changeover costs. 4 You may wish to review the lesson 'Establish Baseline 11'

9

Takt Time Obviously, the current state is a far cry from continuous flow, so there's lots of room for improvement. According to the principles presented above, the first step toward improvement is to calculate takt time.5 The customer demand is 16740 uprights per month and the company works one shift per day, so takt time is calculated as follows:

Available work time (minutes): 480-20-30=430 minutes/day=25800 seconds/day Customer demand: 16740/20=837 uprights per day Takt time: 25800/837=30.8 seconds/upright

In other words, Peerless needs to produce one upright every 31 seconds to meet customer demand within its available work time. This does not allow for equipment breakdowns, quality problems, changeovers, or any other events that might occur. Peerless might decide to produce at a faster rate than takt, but takt is determined by customer demand and can't be changed by Peerless. Takt time is shown in data boxes on the future state value stream map.

Where should finished uprights go? We've started with the customer's rate of demand, now we will work our way upstream. The next question we address is whether the finished uprights should be sent directly to customers each day, or whether we need supermarkets. The ideal of continuous flow would favor direct shipping, but the case study in the appendix includes the note “Customer orders are erratic,” suggesting that demand varies quite a bit. Thus, in this case, we’ll need supermarkets to help us smooth the impact of varying demand on the value stream.

Peerless will use the customer’s monthly forecast to determine the production capacity they’ll need in the upcoming period (one day, in this case) and they will adjust staffing accordingly. Actual production will be scheduled based on kanbans coming upstream from the finished goods supermarket. The future state value stream map shows the configuration in Figure 7.

5 Detailed information for this value stream is given in the appendix.

10

Figure 7-Build To Supermarket

If the decision had been to build directly to shipping, the value stream map would look like Figure 8. Figure 8-Build to Ship

Uprights, both short and long, are packed 6 to a box. We will use this quantity as the kanban size. In other words, each box of 6 uprights in the staging area has one production kanban on it. 6 is called the “kanban size.” As the shipping department loads the truck it removes the kanbans from the boxes and sends them back to the paint department. This tells the paint department that the uprights have been removed from the supermarket and need to be replenished.

Continuous Flow The operator balance chart for this value stream is shown in Figure 9. The times are shown for a single operator at each process. Stamping and painting cycle faster than takt. Stamping produces an upright’s worth of side plates in 20 seconds and it takes 15 seconds to paint an upright. A single weld station with one welder takes 45 seconds per upright, longer than takt. It also takes 45 seconds for one inspector to inspect an upright. Thus the weld and inspection stations currently use two operators each, for a total of four workers. We can make an immediate improvement by combining the jobs of welder and inspector. If we add the cycle times of welding and inspection we see that the combined time is 90 seconds, precisely three times takt. Thus we could hypothetically utilize three

11

operators rather than four.6 However, downtime and changeover time mean that we’ll still need four operators for the time being. For now, however, we’ll draw our future state map to show continuous flow between welding and inspection. In a later lesson we’ll discuss the design of a work cell to accommodate this. Figure 9-Takt Time versus Cycle Time

20

45 45

15

0

5

10

15

20

25

30

35

40

45

Stamp Weld Inspect Paint

Cycle Time

Takt Time

With the current layout it isn’t feasible to implement continuous flow that includes the paint department, which is housed in a physically separate facility. It doesn’t make sense to run a single part to another building. Figure 10 shows our progress so far.

6 This will require cross-training so that the three workers are skilled in both welding and inspection.

12

Figure 10-First View of Lean Value Stream

Realizing the other potential gains will require that a number of improvements be made, including:

• Improve the combined welding/inspection operations by improving reliability, reducing changeover time and reducing cycle times. The initial target is to be able to produce to takt time with only three operators.

• Eliminate the defects that make inspection necessary. We could then eliminate the inspection activity. This would be a dual savings because not only would it save labor, it would eliminate scrap and rework. We could then get by with two welders, eventually maybe just one welder.

Kaizen and Lean Six Sigma and other continuous improvement methods can be used to accomplish the above. We discuss how this is shown on the future state value stream map below.

Placement of Other Pull Systems Our value stream needs another pull system. The stamping press can’t be used solely for the production of plates for the uprights. While this is less than the continuous flow ideal, it can’t be helped (at least, not yet.) Until we can figure out how to create the ideal, we need to create a supermarket between the stamping press and the downstream process that uses its output. Then we can use withdrawals from that supermarket to control the stamping press’s production of short and long upright parts.

Scenario A As with the full value stream, we will use customer information to design the supermarket pull system for the press. The “customer” for the press is the weld/inspect work cell. The cell requires enough parts to build 81 short and 756 long uprights per day. The parts for these uprights should be placed in containers that make it easy for the

13

welders to remove and assemble. Since each upright has two side plates and two end plates, perhaps each layer of the container will have these four components arranged in the best order for placement on the welding jig. Or maybe the containers will drop into feeders that automatically offer the parts one by one to the welders in the correct order. The containers need to be small so that Peerless can keep both long and short upright stampings in the work cell at all times. A target size for the containers might be one hour’s total production, say 10 uprights. An hour’s production would require 1 container of short and 9 containers of long upright parts. Each container will have a withdrawal kanban on it. When a welder/inspector begins taking parts from a container its withdrawal kanban will be given to the material handler, who will then go to the stamping press supermarket and withdraw another container of these parts.

The withdrawal kanban is used to initiate movement of parts. The production kanban initiates production of new parts. Peerless will attach a production kanban to each container of 10 uprights in the supermarket. When the material handler removes a container from the supermarket a kanban will be sent to the stamping press. This tells the press operator to produce the needed short or long upright components, place them in a container, and move it to a specific place in the stamping supermarket. The location is known as the market address. Stamping now gets its schedule from the next process in the value stream, not from a central scheduling authority.

A problem with this particular pull system is that the press takes 1-hour to changeover from short to long uprights. If we schedule runs of an hour’s demand of upright components, then we are spending 1-hour setting up the press for a run of only 3 ⅓ minutes (short uprights) or 10 minutes (long uprights.) Changeover time is something that needs to be improved in the long term, but for now we need to design a system that makes sense for the current changeover time.

Scenario B Obviously we need to produce greater quantities at the stamping press between changeovers. In keeping with the lean philosophy of every part every day, we will set the stamping press’s target batch size to one day’s demand: 81 short and 756 long uprights. To allow some buffer for quality problems and downtime we will keep 1.5 days of parts in the stamping press supermarket. We do this grudgingly and we will use Lean Six Sigma, Kaizen, and other methods to improve this situation. For now, it’s the best we can do.

Peerless needs to use a signal kanban to schedule the stamping press. The signal kanban is a triangle which lets the press know that it should changeover and produce another day’s worth of short or long upright plates. The triangle is taken from the supermarket to the press whenever the number of bins remaining drop below a certain point. The stamping press is still scheduled by their customer rather than by central scheduling. We now have the value stream shown in Figure 11.

14

Figure 11-Value Stream Map with Signal Kanban

Upright Blanks Although the supplier of the blanks used by the stamping press isn’t a lean manufacturer (yet,) Peerless will still use kanbans and supermarkets to control production and inventory at receiving. Withdrawal kanbans will be attached to the containers of blanks. As the containers are used the kanbans will be sent to production control, who will use them to order blanks based on the usage of blanks. Production control can order blanks on a daily basis and place the kanban cards in the kanban slots in the receiving area for the day that the blanks are supposed to arrive. When the shipment comes in the kanbans will be attached to the containers of blanks in the shipment. If the expected arrival date comes and goes and the kanban are still there, it provides a visual alert that the supplier is having problems.

Peerless is currently sending the supplier faxed orders weekly. If possible deliveries should be scheduled to arrive daily. Since there are 5 days in a week, this would reduce inventory by 4

5 , or 80%. We’ll keep 1.5 days in the stamping. Figure 12 shows where

we are now.

15

Figure 12-Value Stream Map Including Flow to Supplier

This value stream is far from the continuous flow ideal, but it is a definite improvement over the original batch-and-queue value stream. This system is much closer to achieving what Taichi Ohno calls “autonomation,” or automation with a human touch. Autonomation is a type of automation where production automatically responds to the external environment and to internal problems. With this value stream production is based on short-term customer demand. Parts move quickly from one process to the next (in the case of welding and inspection it moves immediately.) Thus, quality problems will be detected by the “customer process” soon after they were produced by the upstream process. The causes of these problems are much more likely to be discovered than if they went undetected for a week or so, as they would with the original value stream. Also, if there is a production problem at, say, welding, then downstream production will not receive product from them and will not overproduce. Of course, upstream processes will be starved for kanbans and will quickly stop. Lean welcomes the pain signals as an incentive to find and eliminate the causes of problems. Batch-and-queue uses inventory as an anesthetic to dull the pain while failing to treat the disease that is causing it.

Scheduling and Load-Leveling Next we must decide which downstream process to schedule as the pacemaker process. Since there mustn’t be any downstream supermarkets from the pacemaker process, the only real choice for the Peerless uprights product family is to send the schedule to shipping. In a typical day 81 short uprights and 756 long uprights are shipped in packages of six, commonly called “6-packs.” However, it wouldn’t do to simply have all trucks loaded with the day’s forecast of 13 or 14 6-packs of short and 116 6-packs of long

16

uprights and all of the kanbans pulled and sent in a batch to the paint department. This would be a batch-and-queue operation that would not provide any kind of takt image to the upstream processes. In fact, if the different processes received the entire day’s worth of kanbans, chances are they would sort them into short and long piles and batch produce to reduce changeover time. This would, of course, cause the usual set of problems that are caused by batch-and-queue production.

What we need is a system for pulling the kanbans that will result in a mix of short and long uprights being produced. At Peerless we will accomplish this by having Production Control use a heijunka box as a load-leveling device. The heijunka box will be placed in the staging area and it will show a mix of long and short uprights in the sequence. If we use the 6-pack as our container size the pitch interval is 6 x 31 seconds ≈ 3 minutes, but this is too short a time to use for our increment for releasing work. As it turns out, the pack size of six is used because uprights are quite heavy and cumbersome and this is the largest pack size that one or two people can lift without special equipment. In the finished goods area uprights are loaded on pallets of 10. We’ll use this “pack size” to determine our pitch interval, which gives a pitch of 30 minutes.

What this means is that every 30 minutes Peerless will release one batch of kanbans of work. Each column in the heijunka box represents a 30 minute pitch increment. The two rows are designated for short and long kanbans. Every half-hour the material handler brings the next kanban batch to the paint department, and moves the newly painted uprights to the finished goods area from the paint process supermarket. Table 3 illustrates the heijunka box arrangement. For example, at 8:00AM the material handler will move kanbans to the paint department telling them to paint 8x6=48 long uprights and 1x6=6 short uprights. The paint department will remove the needed unpainted uprights from their upstream supermarket and paint them. Table 3-Heijunka Box Kanban Counts

The fact that the value stream includes processes in separate buildings presents a problem. The distance between the buildings also makes it impractical to deliver the kanbans physically to the weld+inspect process. The temptation will be for the material handlers to send the kanbans to the weld+inspect process on a daily basis, since that’s how often the welded assemblies are delivered to the paint process. Of course, this will completely destroy the takt image. Peerless will need to use some means other than physically carrying kanbans to the weld+inspect work cell, such as a computer or fax machine. This should be done at the pitch increment so the takt image will be transmitted to the upstream process. From the lean perspective it would be best if the entire value stream were located in the same facility, as close to one another as possible. This objective should be studied by Peerless.

We now have our complete future state value stream map, see Figure 13.

17

Figure 13-Complete Future State Value Stream Map

This value stream shows substantial improvement compared to the original batch-and-queue value stream. The lead time has been cut from 36 days to 5 days, and value added time has been reduced from 105 seconds to 63 seconds. Of course, there is still room for improvement. Here are some candidates for improvement:

Improve uptime at stamping

Reduce changeover time at stamping

Reduce total work time at weld+inspect to 84 seconds or less

Improve uptime at weld+inspect

Reduce changeover time at weld+inspect

Reduce changeover time at the paint process

Relocate the value stream so that painting is in the same building

These are all candidates for Kaizen bursts. This is shown on the future state value stream map, see Figure 14.

18

Figure 14-Future State Map with Kaizen Bursts

19

Appendix–Data for Current State Process

Peerless Shelving Data Set Peerless Shelving Company produces a full line of shelving for grocery and department stores. This case concerns one product family, the shelving uprights, which the company calls “uprights.” The uprights come in two lengths, regular and short, and in four different colors. The components are shipped to contractors who install the shelving.

Production Processes Peerless’s processes for this product family involves stamping the side plates and the front and rear plates. After stamping, the plates are welded together to form the upright assembly. Assemblies are then moved by truck to the paint shop where they are painted. Parts are warehoused until needed on an order, then they are shipped to customers.

• Switching the presses from one length to another requires a 1 hour changeover.

• Steel blanks are supplied by Omaha Steel Company each Monday morning.

Customer Requirements Customer orders are erratic. The monthly averages are:

• 16740 uprights per month o 1620 short o 15120 long

There are two common order sizes. Small stores: 54 short, 108 long. Large stores: 198 long.

Shipments leave the warehouse daily, 8am-5pm, Monday-Friday. An average of 3 large store shipments/day, 1½ small store shipments/day.

Work Time • 20 days/month • One shift in all production departments • Eight hour shift. Overtime if needed • Two 10-minute breaks/shift; manual processes

stop during breaks • Unpaid lunch

Peerless Production Control • Receives marketing’s 90/60/30 day forecasts

and enters them into MRP system • Issues Peerless 6-week forecast to Omaha Steel

via MRP • Obtains steel blanks by weekly faxed orders to

Omaha Steel • Receives daily delivery orders from sales

department • Generates MRP-based weekly department

schedule based on customer orders, WIP and finished goods inventory, quality reports, downtime reports

• Issues daily expedite lists to area supervisors • Issues weekly build schedules to stamping,

welding, and painting processes • Issues daily shipping schedule to shipping

department

Process Information All processing occurs in the following order. All pieces go through all processes.

1. Stamping. The presses are used to producing upright side plates and end plates and other parts.

o One 50-ton press, manually loaded & unloaded. (another press is available.)

o Cycle: 5-seconds (12 plates/minute) o Each upright–short or long–uses two side plates

and two end plates. Thus, it takes 20 seconds to produce the plates for one upright.

o Changeover time: one hour o Machine reliability 90% o Observed inventory

4000 short and 3000 long blanks before stamping

2000 long side plates after stamping 2000 long end plates after stamping 800 short side plates after stamping 400 short end plates after stamping

2. Weld Station 1 (dedicated) o Manual process with 1 operator

20

o Cycle time 45 seconds o Changeover (jig change) 15 minutes o Reliability 95% o Observed inventory

500 short 800 long

2. Weld Station 2 (available) o Manual process with 1 operator o Cycle time 45 seconds o Changeover (jig change) 15 minutes o Reliability 90% o Observed inventory

600 short 300 long

2. Weld Station 3 (available) o Manual process with 1 operator o Cycle time 45 seconds o Changeover (jig change) 15 minutes o Reliability 90% o Observed inventory

400 short 400 long

Weld Inspection All welded assemblies are inspected before being added to inventory to be sent to the paint shop.

o Visual, dye penetrant, magnetic particle inspection

Manual process Cycle time 45 seconds

o Observed inventory (backlog) 1200 short 800 long

Deliver to Paint Shop There are daily deliveries to the painting facility by truck. Each load delivers the number of uprights needed for the following day’s scheduled paint run. The paint shop is located 3 miles from the fabrication shop.

o Observed inventory o 405 short o 7560 long

Paint Line o Automated process with one person to load

parts and another to unload parts.

o Cycle time 15 seconds (rate at which finished parts exit the paint process)

o Changeover (color change): 30 minutes o Reliability 98% o Observed inventory awaiting paint

o 100 short o 400 long

Shipping Department Orders are shipped to contractors who install the systems on a daily basis. Number of trucks vary.