Lean manufacturing 3

[Saintor]

I just read a book about it from Womack & Jones. I have the impression that these guys think that continous flow is always the good for all and that lot mfg is always the bad. I understand the other concepts, (added value, value stream). But this one (continuous flow) has a limited application in my book. Right?

They show a few examples with Pratt&Whitney, and others.

If I have an immediate demand for a subcomponent, I have to do it at least the economical batch quantity? And store the "excess" assuming it will consumed in a reasonable time. Right? They don't think so.

It suggest that the rate should be fully harmonized with the final product. But on a workstation, you might have to manufacture several types of components. You can not lowerindefintely the speed the rate of producing, because of costs and you need the resource for something else!

In the end, is it only to build a minimal lot size to match as possible the demand (which I am all about)? I don't understand. Or we don't live on the same planet.

We have turnovers from 3 to 6, and they claim resultats 15+ which is re-producing every 3-4 weeks. In a company I support, we start a production on subcomponents to supply 2-3 mo, except major ones for 5 weeks. Of course, we would be glad to save interests costs and inventory, but whenever you start a production, you have to make sure to minimize all soprts of fee (set-up, transport, order processing).

I read the book, but I am still mystified about this concept. Could anybody shed some light on this subject of universal "flow production"?

 

[MikeHalloran]

I haven't read a lot of books about it, but I've participated in startup of a JIT line. Two, actually, because the first product was a huge success.

In JIT, and I think in 'flow', you don't build 'lots' of anything. A 'lot' implies that you are going to store stuff and then retrieve it; you don't do either.

A JIT line is built, and planned, and documented in excruciating detail, to build finished goods at a fixed rate. The planning includes ways to adjust the rate up or down a little to deal with demand fluctuations, but the basic design assumes a steady fixed demand, say one unit every 20 minutes. Throughout the working day, every 20 minutes on average, a finished unit becomes tested, packed and ready for pickup. And at roughly the same time, every operator on the line finishes a work unit and passes it on to the next operator. And suppliers deliver raw materials and outsourced components. They are free to build ahead in lots, but the inventory costs and management are their problem; they are expected to deliver their product at an agreed rate.

The really hard part, i.e. the part that's mostly a mystery to me, is 'balancing the line', allocating the atomic operations to specific workstations so that everyone works pretty much the same cycle time. Once that was done, each workstation was outfitted to optimize the productivity of the operator for the operations at that particular workstation.

It sounds like you are trying to, um, 'cheap out', and use one workstation for several disjoint sets of operations. I haven't seen that done, and I'm not convinced that you can do it well, or at all, except for minor product variations.

I.e., if you are setting up a hypothetical workstation to make either 15 volt veeblefetzers or 22 volt veeblefetzers to meet domestic and international market needs, that's possible, but if you are setting up a single workstation to make both veeblefetzers and snorkertzers, you are doing it wrong.

And if you are accumulating subcomponents in numbers greater than what it takes to build the next assembly in time, today, you are doing it wrong.

Maybe not wrong for you, but wrong in the context of flow, where the inventory of subcomponents is always and continuously very nearly zero.



Mike Halloran
Pembroke Pines, FL, USA

 

[BillPSU]

The ulitmate goal of lean manufacturing is to achieve a lot quantity of one. The company I worked for was a design to order company with many standard "specials" but also some specialty designed parts.
Parts were laser cut by machine/weldment processed through the secondary operations as a kit and delivered to the welding area, tacked and welded into a subweldment, machined, and tacked and welded into the final weldment and finally hung on a powder coat line. This process was applied to all components of the machine. The throughput time of the complete machine was about two weeks which also included fabrication of fiberglass booms. There were exceptions to the lot quantity of one such as small brackets and other components. In practice the orders did allow some batching but sometimes the specials dictated production of single components.

 

[mechvhstech]

In principle, lean is not to be viewed as a tool, it is more like a tough manufacturing management. Everything is inter-related. Last year we achieved 65% reduction in lot size and 80 % reduction in set up times across all product types… All this was possible only because we identified the bottleneck to be huge lot sizes due in our High Variety Low Volume set up. We have also started moving our machines around in cells so that there is minimum movement and handling. Though we do machine some parts in batches for processes such as heat treatment /powder coating, we employ a supermarket system for such components downstream.
To achieve good product flow line balancing is very important (heijunka in lean terminology). Lean implementation is difficult if some aspects are ignored, for example to ensure cellular manufacturing works, TPM is very important, for TPM to work, there should be a very high level of training and cross training.
I have been a lean researcher/ practitioner for 6 years now and can say with sufficient impunity that lean implementation in every corner of the organization is essential to out perform competition.
Also lean does not necessarily mean copying exactly what Toyota did, one has to develop his own version of lean by understanding what actually adds value and what does not!

-vhs

 

[CompositeGeek]

Okay.....Where do I begin my rant?????

1. Lot size is not the issue.
2. Schedule is not the issue.
3. Inventory is not the issue
4. Workstations and cells are not the issue.

COST is the Issue!!!!

Anything else and your smokin' sumthin.

Okay that said, let us define cost. Is it the cost of inventory? Is it set up cost? Is it Recurring cost of fabrication only? Do I absorb overhead and SG&A into my cost models?

I say all of these questions are spurious red herrings that lead to nothing and signify even less.

Cost is how much money you spend to make parts. Don't make it more complicated. What was the direct labor and direct material costs to fabricate. Do not get caught up in cost accounting; just measure the actual period expenses. Only value material at raw material cost. Only value labor based on payroll expenditure. DO NOT ADD VALUE beyond these two numbers.

More Discussion:
Lean in a Job Shop is addressing vastly different issues than in a continuos assembly line sequence. What I hate most about Lean is the thinking that everything lean touches is important. If that was the case the every company would do lean because they would all see immediate benefits to the bottom line. Somehow that is not the case. Somehow much effort in lean does not realize the savings expected. Why would that be?

If the methods are based in science then the results should be repeatable, don't you think?

Back to Cost:
The reason Lean fails is that the method of evaluating cost in lean is trapped by old cost accounting methods and practices. The underlying assumptions in cost accounting lead to erroneous conclusions about the "Actual Cost" to build a part, which in turns leads to crazy things like single piece flow when you KNOW that you have to process in batches or lots.

Read "The Goal" (Eli Goldratt is the author), find your bottleneck and apply lean tools to that. Use the accounting methods defined by the Goal for measures of operating expense, inventory, and throughput. Then tell me if lean make sense.

In My World, "the throughput world", single piece flow does work on non-bottlenecks, but if you are creating inefficiency by doing single piece flow on a bottleneck resource then you are losing more money than you can imagine. (Again Read the Goal for how the costs really add up under these circumstances.)

I maintain that any company that actually saw benefits from lean applications in their processes because they:
a. were not a job shop
b. hit the bottleneck process due to intuition and luck
c. should take a long hard look at their accounting methods -: cause they jest may be lyin' to themselves.
d. all of the above

Finally, (and then my rant will be over but thanks for staying to this point) I am not critical of the tools in lean, I am critical of the underlying assumptions that drive lean decisions. Assumptions like "single peice flow is always a good thing". I do not buy that because it is not real. There will always be circumstances where your set up time is longer than your run time. When that happens you must run a lot or a batch of parts or your costs will be higher than the guy who does do T.O.C. (Theory of Constraints) but not lean.

ps. Toyota, Lean Gurus also subscribe to the teachings in the Goal, but most lean proponents forget that when they are describing all the wonderful benefits of lean.

Thanks for letting me rant. This is a hot button issue for me if you can't tell.







Composites and Airplanes - what was I thinking?

There are gremlins in the autoclave!

 

[orlorob]

Hey CompositeGeek, or others -

What systems (pen & paper / software) do you use to track you direct labor in a job shop?

thx
-Rob

 

[CompositeGeek]

orlorob,
I have used 4 different MRP systems over the years. None of which was a good choice for various reasons.

the issue in tracking actuals per job is actually backwards to my thinking.

I just look at total sales(thru put) vs. total payroll(one element ogf operating expense) per period (by week, month, day). I am always amazed at the level of variation seen in such metrics but people don't even know it is happening because they do not look at it that way. Their focus is on an individual job not the aggregate (a risk inherent in all cost accounting methods).

Look into DBR by Thoughtware if you are interested in Theory of Constraints management software.

Hope this helps.

Composites and Airplanes - what was I thinking?

There are gremlins in the autoclave!

 

[koeskoes]

orlorob,

our shop uses time cards and the employee would write down the time they took to complete a particular job (braking, punching, sawing etc). The time card has to add up to 9 hours(9 working hours excluding break). Its not accurate because it depends on the honesty of the worker and does not take into account toilet break, fatigue etc.

Anybody has a good idea on how to improve on this system?

Koeskoes

 

[Ashereng]

Have you tried doing the job yourself and timed yourself? This is usually a revelation for me.

Or, work with your worker for a whole week. Note the high and low times, in addition to the average time - now ADD 20% to every number (he is working harder than normal since you are right there, and he can't keep it up all the time).

 

[rorschach]

A few months back, In Mfg Magazine published an article that was in essence trying to say that lean manufacturing would be the savior to those who's supply chains were disrupted by Katrina/Rita. Now, I think the lean manufacturing concept is great where it can be applied to advantage, but this article was snake oil. The problem was that the supply chains were messed up and JIT just went out the window. If you need a steady supply if widgets to make your product and the manufacturer of your widget just had is business destroyed, and the alternate manufacturer is now working 3 shifts to try to make a dent in the backlog, your fundamental problem is not having ENOUGH inventory. Like all management philosophies, when all you have is a hammer, everything looks like a nail. People start trying to apply inappropriate methodologies to solve problems because they don't know any better and they have succumbed to the kool aid sold by the management consultants. I just can't believe somebody at the magazine didn't read the article and say WTF?

 

[MikeHalloran]

The USA's current air transportation system provides a pertinent example of how well a lean system deals with perturbations.



Mike Halloran
Pembroke Pines, FL, USA

 

[kf6bbl]

I have found that you cannot apply lean to all manufacturing businesses. Just as ISO is not for everyone, neither is lean.
I think too many people are attraced to lean because most of the principals are obvious. It's sort of like those commercials for matresses that ask, "Do you wake up tired in the morning?" Most people do.
The most effective tool I have found is Volume Adjusted Costing.

http://www.strategosinc.com/volume_adjusted_costing.htm

This is merely a tool to calculate costs, not a manufacturing system. It cuts through the BS to a basic answer that can't be denied. When compared to activity based costing, it is a lot simpler.

-Bill
CE Designer Forum

http://www.cedesigner.com

 

[jpete]

I have nothing to add, I just wanted to tell CompositeGeek that you are my new hero. I work in a jobshop where they try to apply every new management philosophy that comes down the pike. The end result is a useless amalgam that makes money in spite of itself.

 

[CWoodster]

Gremlins in the autocalves - mnot in mine I can assure you.

I agree the TOC route is the way to go if your looking at the whole facitity and bigger picture etc. You cannot improve everything so target your bottleneck. easiest way to find this is the thing with a massive pile of inventory stuck in front of it and a supervisor with no hair.

Lean tools work really well when applied to the bottleneck because it translates directly into extra capacity.

I wonder wether anyone else has spent hours trying to explain that shortening the length of processes does not 'save' money just as everyone sitting down for a cup of tea does not 'cost more' money.

Only sales inventory and operating expense can affect the bottom line. (for info try the goldratt website

http://www.goldratt.com)

Take it Easy
CW

 

[brookside]

The main problem with ISO and Lean ( all the rest of them for that matter)is in the implementation. There is an order to everything and it begins with identifying who you are as a producer. The second strategic question is "Who do we want to be"? and the third "How do we get there"? The implementation of any quality initiative must be backed up by senior management in order for it to have life and continue as a living document. Without measurable goals that are in perspective you will invariably fail in your planning.