Solve the five difficult problems the manufacturing business faces with TPiCS
When you hear it say, gsolve the difficult problems with TPiCS,h you might think, gHow could no more than 1 or 2 million yen systems do such a thing?h It could happen. Of course, you canft solve the gdifficult problemsh with it just because you bought it. You have to understand the basic concepts of production control we advocate, and to put them in practice.
We are not talking about empty wishes. We do have gdefinite answersh to the problems of production control. The road is never smooth, but it surely leads to real solutions.
We hope you will understand the gdefinite answersh we have, and we would like you to take a step towards the gsolutions to the problems.h
TPiCS Laboratory, Inc.
Yoshio Ninomiya, President
When I write about only an aspect of gyou can freely correct data,h people unexpectedly evaluate it from the viewpoint of ginternal controlh these days. So I would like to mention in the beginning that there are functions such as gAuthorization Managementh and gData Correction Log Managementh in TPiCS-X. But my explanation in this booklet continues from here on without all of that.
Table of Contents
Although people say manufacturing business in a few words, what is made is different company by company, and the kinds of products and markets are different. Ways of thinking of the management and the skills the employees might have are different. Moreover, as the history in which the factory have come so far is different, ghow to make things,h namely, gproduction control methodh is naturally different.
If you look at it from a different angle, however, in order gto make thingh therefs a narrowly defined production method in which there might be drawings, composition lists and recipes. And therefs a widely defined production method in which there are equipment to make things with and materials to order and to process. These are common things in any manufacturing business.
This is quite easy to understand if compared to gman.h Men are all different but if you think each man has a head, arms, legs, and so on, men are all the same. And you can think of gdiseaseh in the same way. If you eat too much, you have an upset stomach. If you drink sake too much, you ruin your liver. Then, if anybody in the manufacturing business is incapable of sensitively dealing with the needs of customers, he misses a chance. If he responds to them only with inventory, he goes into increasing dead stock in the worst wayc This problem is universal.
But what is really difficult is from here on.
Some live up to 90 years old smoking cigarettes all the way and some die of a lung cancer at 40 years old without smoking a cigarette.
If there are products that have few variations and engineering changes, and of which life cycle is long, the way to run the production with plenty of inventory is one answer.
If you think this way, it brings you back to the beginning. But smoking cigarettes is certainly bad for your health - so they say, - and less inventory is always better. To find general problems in specific problems and to create specific answers from general answers; you will need these aspects.
Now what problems are people in the manufacturing business having to begin with? Of course, they must have lots of problems, but I would like to pick up the ones to be solved easily with the production control system among them, or the ones not to be solved only with it but to be improved with it.
And I would like to highlight the mechanism of the problems by generalizing them as much as possible, and to lead them to specific answers.
Products have become diverse and their life cycle has been shortened since the bubble economy burst. And working speed in the whole society has quickened up due to the rapid progress of IT and we have had to supply for the demand as soon as possible.
For example, it was commonplace to deliver orders in a few months after accepting them in the old days, but people canft wait that long nowadays.
Trying to supply them from stock results in vast amounts of inventory value, and the production for short-term delivery canft keep up with parts and materials.
Or, since orders requiring short-term deliveries are already coming from customers, you would simply say with no significant methodology, gWe just have to do it,h and you are working in a blind effort. Such a situation we call gproblem,h and Ifm going to talk about how to solve it.
(2) Attempting to make the factory floor observable
The word gmanageh basically means to gcreate the state where anybody can quickly understand things.h Then, the essence of gobservableh is, by utilizing gto be able to see,h to improve things and to prevent troubles from happening. Ifll talk about how to realize it using TPiCS-X.
They say, like a half-funny anecdote, in developing countries even now, gOne to two days sounds like common sense.h But, of course, there is no way we can. Ifll talk about how to solve the problem of delay here.
When you think about the problem from a standpoint of ginventory,h ginventory reduction and the production for short-term deliveryh is the contradictory problem in a sense. And when you think about problems of the production control, there are so many gdamned if you do, damned if you don'th type of problems.
The above four problems described here are linked all together to one root. From such a standpoint Ifm going to talk about how to solve these two problems with balancing them.
The life cycles of products are shortened and the shortening of development periods of new products is requested. At the same time it is requested that the production is also anxious to be able to run in top gear. This problem is somewhat different from the above four in nature, and depends largely the gtools.h Ifll talk about how to solve the problem.
(1) Achieving the production for short-term delivery and ready for whatever changes come along with TPiCS-X
¡ What is the production for short-term delivery?
Nobody knows whether end consumers are going to buy the products or not, and when they are going to buy what specific product and how many of it. All the corporate activities and all the productions are influenced by the trend of end consumers (= humans.) When you are distant (for example, research and development) from end consumers, you have trouble seeing this connection between you and them. No research or development must be done for what is free from the possibility of being accepted by the world. Although you manage to predict the behavioral principles of humans that are difficult to predict, and the direction to aspire for accuracy improvement of the prediction is one of the answers, as a different answer it is also important to reflect the fluctuation of demand in the production at the point of time as close to the transition of the demand as possible.
More importance on gaccuracy improvement of predictionh had been placed in the high-growth period before the bubble economy burst compared to gkeeping pace with demand fluctuation.h But as we are aware of our limitations of gprediction,h it seems that the road to keep pace with demand fluctuation has entirely been demanded these days.
Since gkeeping pace with demand fluctuationh is too broad for a theme, however, we call it gthe production for short-term deliveryh in TPiCS, and Ifm going to deal with this problem.
(1) You want to realize the system to ship orders from your customers as soon as possible with as low in inventory as possible.
(2) You are running the production getting pushed around by the daily customer orders because it is difficult to predict the market with such as the sales plan. But you want to improve that somehow.
(3) You are receiving the forecast information from your customers. But since the firm orders close at hand change a great deal, you are being forced to always accept the confoundedness.
(4) You are receiving requests for shortening the deliveries from your customers.
(5) You want to more smoothly respond to urgent orders and orders from sudden visitors.
The gproduction for short-term deliveryh is the one to solve these problems. You should be able to respond to them in a brain match instead of a physical match with guts and toughness. We call the ability to achieve them with as low in inventory as possible the gproduction for short-term delivery.h
As for the production for short-term delivery, many people have great misperception and I guess such a situation of gthough we understand the needs, we canft quite move forward.h
(1) gAdding to our misery that we have a hard time of delays in production even now, therefs no chance to produce more for short-term delivery.h This is a typical misperception about the production for short-term delivery. Now examine more in detail about this problem, sorting out the causes of delays in production.
Ifll start with a store for you to most easily understand. For instance, assume that the capacity of equipment is 900 pieces per day and the production is behind to produce the scheduled quantity of 1,000 per days. Since it doesnft matter that the 1,000 was decided a few months ago as long as the equipment capacity is 900, the delay in production due to this reason has nothing to do with the production for short-term delivery or the shortening of delivery period.
Therefs the answer in the function of TPiCS-X.
One of the basic ideas of f-MRP in TPiCS-X is that gparts and materials are purchased at the right time as required.h The parts that have no other choice to be purchased three months before, even with the production for short-term delivery, are to be purchased three months before (described later for details.)
When you think about the production for short-term delivery, you must gnot make arrangements for required parts after the production is decidedh but gsimulate whether or not to be able to produce from things that have already been arranged.h Of course, since you had better make arrangements for the things that can make it in time even after the decision of the production once it is decided, the more with shorter procurement period you have, the more possibility of being able to you get as a result of the simulation. In order to further increase the possibility, you ought to set up a greserve (buffer) for the fluctuation of the productionh for the parts of which procurement period is long, and increase the setting value.
If you implement the production for short-term delivery only with the parts and the materials that have already been arranged, and when you have no functions to use as the gbufferh and forcibly implement it, that gstirs uph the productions in your suppliers and results in delays in delivery of the parts and the materials.
If you set up a required buffer, however, delays of the parts never increase even when the production for short-term delivery is implemented as TPiCS-X can effectively utilize it.
We are not thinking about gachieving the production for short-term delivery by forcedly producing in a hurry.h As with purchases for parts and materials, therefs no other choice to produce taking three days whenever it takes the process three days. Thus, defects donft necessarily increase even if the production for short-term delivery is implemented.
(2)What is preventing you from the production for short-term delivery most, however, may be a thought of gAre you asking us to produce for shorter-term deliveries though we are so busy even now? That we canft do.h
Letfs think about this story separating factory floor and office.
A story on the factory floor would go: you can understand if you think about it as with the misperception of the (1) That is, therefs no difference between the 1,000 decided three days before and those decided two months before because the workload would be the same. The busyness on the factory floor never changes even for the production for short-term delivery.
Office work wonft go that way as might be expected. The office work that occurs in proportion to production volume wouldnft increase even for the production for short-term delivery as long as the quantity is the same 1,000 pieces. But the gwork that should be done by todayh would increase. That gsense of urgencyh might make you feel ghaving been busy.h Since TPiCS-X is something like a gdedicated machine for the production for short-term delivery,h however, you can achieve it with ease if you use TPiCS-X.
(3)The problem of the leveling also prevents you from the production for short-term delivery.
Letfs think about an extreme case as an example to make it easy to understand. Itfs a case of receiving an order today and producing it today. In this case, the quantity of the customer order would directly be applied to that of the production.
That is, since the fluctuation of customer order volume directly becomes that of production volume, the leveling becomes difficult.
When you wanted the leveling to occur on a weekly basis, you couldnft do the adjustment work for the leveling itself unless the schedule of the following week is at least determined by Friday. And when you thought about an intermediate or first process, you should be doing work for the week after next at the end of the following week. Therefore, the schedule for the first day of the week after next should also have to be decided.
You have a tendency to think that the schedule of the nearest two weeks has to be determined, or that you fail to reflect the needs of your customers.
(4) The last misperception is the case of gmaking the work period shorter is impossible due to the long work from start to finish.h
Although I may not be able to call this case the misperception, you would give up without knowing TPiCS-X, or you end up choosing the road that doesnft lead to an answer.
In the case of producing something that canft be held within the factory (= put into inventory) on the grounds of the problem on the quality control or something once the work has been started, theref is no way out. But if itfs possible, you can solve this problem.
Least of all, in the case of something becoming a different product in the middle of the process depending on the processing method or the embedded part, you can get big benefit out of it.
If you are flipping through the pages of this booklet with a cynical thought of gthat system boy is only saying good things anyway,h nothing will come of it.
I would like you to read these texts in a serious manner thinking, gwe may be able to do the production for short-term delivery.h
In order to think about this problem, you have to think about what should be done to achieve the production for short-term delivery to begin with. Letfs have a run-through on gwhat is needed for the production?h for that. When you open any textbooks of the production control - To be honest with you, I completely forget about it because that was nearly forty years ago, - you always find g5 major elements of the productionh written in the first chapter.
(1) Equipment and place. (2)Humans. (3)Parts and materials. (4)Drawings and specifications. (5)Funds. If my memory is correct, what is written in the textbooks are such as they are. I feel like I might be able to add g(6) Outsourcing suppliers and partnersh to the list. Anyway, these are the 5 major elements indispensable for the production. It is the (3)that is largely concerned with the production for short-term delivery among them. Now letfs move on focusing on the procurement of parts and materials.
We are going to think about what is different in the procurement of parts and materials by comparing the gproduction for short-term deliveryh with the gtraditional production to be 3 to 4 months ahead.h It used to gmake arrangements after the production schedule is determinedh in the past. To put it the other way around, it used to gproduce at the time when arrangements make it in time.h To think about the gproduction,h you canft finish it with only a part missing no matter how many parts are used and no matter simple the part is. Therefore, you have to adjust your production to something that will take the most time to get it among the parts and materials to be used. Thatfs why it was greflected in the production to be 3 to 4 months ahead.h
When you think about the procurement period of parts and materials, what you shouldnft forget about is the power balance. Normally the procurement period is determined by the production schedule and the point of view of the supplier. But if the side to purchase has overwhelming buying capacity (ability to buy a lot, technical capabilities, ability for leadership, etc.,) it can make the supplier side to deliver in 3 or 4 days after the purchase ignoring the schedule on the supplier side.
Let the consideration here focus on the general manufacturing business instead of these gparticular circumstances of big boys.h
In order to purchase parts and materials as requested (as scheduled,) you have to make a purchase considering the timing to be able to get the part (purchase lead time.)
When you make a purchase at the time way before, however, you end up making the purchase before gwhat to make eventuallyh is determined among the production for short-term delivery.
This means that you have to think about gwhat to makeh among the parts and the materials that have already been arranged in order to do the production for short-term delivery. That is, it means that you have to think about the production schedule based on the situations of arrangements.
Of course, the procurement period varies depending on the part and the material. There may be a lot of cases where it will be the same among suppliers in reality, but the part and the material essentially determine it.
When you think about the production schedule on that basis, you have to consider gsomething that can make it in time if making a purchase from nowh and gsomething that canft make it in time any moreh by part and material.
According to the MRP calculations of TPiCS-X, they are performed definitely separating, by part and material, the period where purchase orders have already been released, and the period where purchased goods can make it in time if purchase orders are released from now on. The general production control systems are designed based on the assumption of gthere were schedules for products in the beginning,h and the idea of making a purchase for the parts required for the implementation. You already understand at this point that it will be too difficult for the general production control systems to handle the production for short-term
delivery because of that.
Since we can think of gwhat is already being purchasedh and gwhat is already being ordered for productionh as the gscheduleh here, wefll call these the galready-arranged schedules.h
When you think about the relationship between gparts for a long procurement periodh and the production for short-term delivery, you begin to realize that it is similar to the problem of equipment impacting on the production schedule. Generally speaking, it takes a certain amount of period to build a piece of equipment, and itfs not a matter of today and tomorrow. Then, that leads to an idea of gmaking a production schedule within the equipment capacity that is already fixed.h However, there is something slightly different in purchases of parts and materials from the equipment capacity. Therefs an aspect of gwhere might is master, justice is servanth to some extent in purchases of parts and materials, which offer great flexibility more or less comparing to those of equipment. They may rather be similar to the problem of staff in terms of offering flexibility.
If you have to make a purchase of parts and materials more or less before determining what and how many to produce, you have to make the production schedule considering those galready-arranged schedules.h
To make the production schedule wondering whether the parts get short or not, and whether they make it in time or not, we normally call it a simulation. Yes, an idea of simulation is extremely strong in the MRP calculations of TPiCS-X.
On the contrary, the general production control systems stay at an idea of one-way traffic where grequired quantities are calculated and work order slips are issued.h Allocating inventory and performing the rounding in lot-size to required quantities, and purchasing parts and materials to suppliers and releasing work orders to production, they would eventually say, gKeep up the good job for the rest.h They would shut their eyes ignoring whether you can do it as expected or not. I call them hrunaway production control.h We generally call purchased schedules gorder-released schedules,h and we exclude them from the following MRP calculations. Or, by differentiating the period subject to the MRP calculations from the period not subject to the MRP calculations, that is, the period not subject to schedule changes in the system, we separate them with the concept of gtime fence.h When they say, gSince we have already released it as the system, we no longer include it in the calculations,h you believe at a glace that they are right on the money. Or, you may have believed it as gall too realh until this point. But once you have to implement the production for short-term delivery until the latest moment and come to get into the time fence, you have trouble then.
Even for the most recent schedule we do the calculations on it with the system in a proper way, do the maintenance on it creating the vision for being possible or impossible, and we see all the production activities in action around that production schedule.
This is our idea of the production for short-term delivery.
Therefs a hurdle to overcome, however, in order to achieve this. When you calculate something with the system, you wonder based on what the system is going to calculate. Thatfs masters pre-registered in the system. It uses inventory information, too. And what you shouldnft forget about is the Schedule data, already arranged and finalized.
Now what happens if therefs gsomething that has nowhere to goh mixed up in the Already-arranged Schedule data?
Since the system doesnft know whether that something has gsomewhere to goh or gno where to go,h the system will calculate based on the assumption that deliveries are to be made anyway according to the Already-arranged Schedule data including those schedule data in the calculations. That results in the calculation contradictory to the reality that the current schedules satisfy the stock needs although a new schedule that uses the parts in question for its product comes in. As the parts in question are not going to be delivered, however, they shouldnft have been used. To put it differently, the system canft give you alert information when they shouldnft actually be used.
In order for the MRP calculations to deliver correct simulation results,
(1) Masters to be registered in the system,
(2) Inventory volume in the system,
(3) Schedules themselves recorded in the system,
the above has to be maintained accurately.
Generally the (1) and the (2) are talked about but the (3) is not really. Letfs think about why.
There used to be a manner of operation before in general that the MRP calculations were performed only once a month. Using the case where the MRP calculations are performed on the 1st every month, it will be on the 1st of the following month that the MRP calculations are to be performed next. If the MRP calculations were performed on the 1st of the month of N, for example, the next MRP calculations would be on the 1st of (N+1.) When the 1st of the month of (N+1) comes, all the days from the 1st to the 31st of the month of N are obviously excluded from the re-calculations of MRP since they are all past days. As a result, the schedules entered to the MRP systems were called gMaster Planh or gMaster Schedule,h which was positioned as the gschedule to order parts and materials with.h The schedule to actually produce on a daily basis with was called gExecution Plan,h and people got to think, gthose two are different by definition.h And the idea of gchanging the Master Plan based on the Execution Planh didnft have its need or its function in the system.
Since the general systems are designed as an extension of that idea even now, the schedules of which orders have been released are excluded from the re-calculations. And we call the interval not to re-calculate (possible to re-calculate) for gtime fence,h and explain as follows. gYou are saying to change the schedules for today, tomorrow and the most recent, but we canft. We had trouble up until now because you forced us to do it. Since our system protects you with the time fence, itfs going to clear up the confusion like before.h Whether itfs tomorrow or the day after tomorrow, however, the situation where we have to change the schedules does occur. Although the assumption, to begin with, is that the gMaster Planh and the gExecution Planh are different, he says, gThe system protects the Master Plan with the time fence.h I guess she must not be thrilled at all to hear thatc
Now, we are going to think about what happens if the gMaster Planh and the gExecution Planh are different when using the system?
(A) Even if parts and materials have been delivered according to the order sheet calculated from the Master Plan, nobody knows for sure whether or not there will be a shortage in parts by just looking at the gExecution Plansh for today and tomorrow. Even with the production control system at hand, you have to work overtime every day and verify with a calculator in hand that the parts required for the production for tomorrow and the day after tomorrow will really be sufficient.
(B) You canft tell what will be produced and made today and tomorrow by looking at the production schedule from the system.
(C) Since the system doesnft process the most recent production changes for you, you will have to handle them with gexceptional handlingh if you produce the most recent additions.
However, the TPiCS-X-style logic of gno time fence = what has been order-released is always subject to the re-calculations until its completionh is not a completely ghappy-go-lucky guy.h There are restraints of these: Even the schedule that has been order-released must not lose touch with reality of the situation. The Schedule data in the system must always be adjusted to the gExecution Plan,h which is a feasible plan.
I believe those who are experienced in business of this field will understand both the difficulty and the importance of achieving them.
Something might be worked out as to internal matters if you did your best. But when it comes to information on an external supplier, you couldnft quite get it. I always tell the following story in the training workshops my company holds.
gI was working on the production control in a manufacturing business some 20 years ago. At that time, if I had participated in the training workshop of TPiCS and heard someone say, gIn order to proficiently use TPiCS, I have to do the maintenance on the schedule down to the level of parts and materials to make it constantly feasible,h after having gone back to my office, I would have written the report that reads, gItfs impossible to use TPiCS-X in our company.h
The story continues a little bit. gBut about 2 years ago, someone who had heard this story thought, gThe schedule without the maintenance on it never leads to essential problem resolution for sure.h And he bought TPiCS-X and put the operation to maintain the schedule in a proper fashion in practice. I hear that as a result, he could get very big benefit out of it.h
When we had requests of moving delivery dates forward and additional orders from our customers, previously it took us about 2 weeks to get an answer of being possible or impossible to handle. But it takes us just 1 day since the implementation of TPiCS-X.
As we have diligently been doing the maintenance on the schedule, we have come to understand gwhat will complete and when that will beh or gwhat we should do nowh whoever watches the screens of TPiCS-X.
Delays in delivery to our customers have reduced.
The sales have become as 1.5 times as the last yearfs.
I hear that he has seen such effects as above.
I doubt that the implementation of TPiCS-X had an impact on each one of them including the sales having become 1.5 times, but those are music for my ears.
He has proven for me, gyou can do it if you try hard enoughh and gyou can really make a difference if you try hard enoughh although it appeared to be impossible.
In order to achieve the production for short-term delivery, however, there are two important things for TPiCS-X users to do:
(1) Constantly do the maintenance on the Schedule data in the system in line with the Execution Plan.
In order to make the MRP calculations (simulation) in TPiCS-X correct and meaningful, you have to do the maintenance on all the Schedule data as described above.
(2) Run production lines observing the schedule in the system.
Needless to say, nothing will start without the attitude of gobserving the schedule.h Only in observing the schedule, you can realize the consistent production and achieve the production for short-term delivery.
We call such way of thinking and manner of operation gSchedule Management.h
Letfs not have a misunderstanding.
If you are satisfied with the level of operation equivalent to the general production control systems, you donft need the gSchedule Managementh even using TPiCS-X.
When looking at the screen of the system, it is displaying something different from what is going on now on the factory floor.
When a new order comes in, you canft tell whether or not itfs possible to produce without a calculator.
As the system canft handle the most recent schedule changes, you have to frequently issue additional orders as special processing.
Or, anticipating it, a purchasing guy in charge is purchasing an extra.
Even if parts are delivered up to the order sheet, you always have to do the inventory count and confirm it with an calculator as you donft know whether or not they will cover the production for today and tomorrow.
Thinking about the above, you canft help having too much inventory.
If you donft mind these situations to remain, you donft need the gSchedule Managementh even using TPiCS-X.
While you manage to apply the in-house Execution Plan to the system if you try hard enough, it is very hard to get accurate information from external suppliers in the case of purchased items.
Isnft it on the purchasing side, however, that most of the reasons to make it hard lie?
I have never actually found myself in such a position in my days of the manufacturing business because I was in a different department. But let me think about what I would have done if I had been informed by a supplier, gThe delivery seems to be behind schedule.h
When a gstretching requesth came in from a supplier, the systems in those days had no functions to enter the information to them with. As a result, people had to cover all of such things with handlings outside the system while they were constantly working with it. Since it was the system of the general-purpose machine 20 years ago, and the system operation was mixed with manual labor, which came first in a sense, it must have rather been easy. But that was a troublesome job at any rate. It wasnft a kind of job appreciated by everybody and if I did such a thing in the factory day in and day out, my grades would be getting worse.
They used to be purchasing parts for their delivery in 3 to 4 months back then.
So if I were informed of a stretching request, I would like to handle it as follows:
When a supplier called me right after they received my order sheet, I would say, gDo something about it! Because you still have enough time until the due date.h
On the contrary, when they called me right before the due date, I would say, gWhat the hell were you doing up until now?!h Then, I would hang up right away.
Thinking that all I do would be to hang up right away, no suppliers would call me and say, gIt appears to be late.h
gNo stretching requests or no actual delays in deliveryh will be the best, but gno delays in deliveryh isnft that easy in reality. It would be me who received a cancellation at the last minute. Getting the information as soon and widely as possible, I would have to do something right. If I had known the delay in delivery in advance, I would be able to think of something right. I would be at a loss on the very day to hear, gWe are sorry.h
In order to solve this kind of problem as much as possible, what we have developed is Strategic Due Date Adjustment Option. (Refer to the 5th chapter of gInformation Sharing with Suppliersh on the Strategic Due Date Adjustment Option.)
Think again here about, gWhy canft we achieve the production for short-term delivery?h
We have gone through the gProduction Control Systemh several times until now, but we havenft adopted it, thinking that it wouldnft fit our job.
We canft make the job doing the production control with hand working or Excel.
There may be these cases.
Or, you are already using the production control system,
(1) Using the gsystem adopting the time fence,h We canft adjust the schedule of the system to the Execution Plan.
(2) We canft tell whether or not we cover the production for today or tomorrow with parts already delivered even using the production control system.
(3) The person in charge of materials has to confirm with a calculator.
(4) We are checking the production and delivery statuses for parts within the nearest few weeks in the confirmation work since we canft make it when we find a shortage in parts even if we are making some noise about today and tomorrow.
(5) We donft want the schedule to be changed once we have checked its period delivery status because checking it again and again is painful. Besides we canft make it anyway even if we find a shortage during that time.
Whenever youfve gone a long way to think about the production for short-term delivery, you bump into the problem of the leveling.
Thinking about the know-how of the production like greceiving an order today, producing it today, and shipping it todayh or greceiving an order today with its shipment in 3 days,h you find the production volume linked to the customer order volume. Itfs good to have the customer order volume leveled, but if itfs not leveled, the production volume ends up swelling according to the customer order volume.
Besides, the schedule has to be determined in first for the period you want the leveling of since the leveling job itself cannot be done unless the production schedule is determined.
TPiCS-X solves this contradictory problem of gWe want the leveling, and the production for short-term delivery as wellh with the following two functions:
(1) Function of gmake the production schedule fixing the nearest two weeks temporarilyh using the parameter of gPeriod to Fix.h
Although the schedule in the required period (for example, the nearest two weeks) is fixed in terms of the MRP calculations, it is put in the state where no orders have yet been issued - where the production schedule is temporarily fixed issued - otherwise. The Customer Order data are entered every day and the MRP calculations are performed every day. If no shortage or overage is found, which means they are within a set acceptable range, in the temporarily fixed schedule, the schedule is executed as-is. If a shortage or an overage is found, the production schedule is corrected then as the system tells you so.
There are two kinds of schedules in the MRP calculations of TPiCS-X; the fixed schedule (similar to what is called the independent demand) and the schedule derived from calculations (similar to the dependent demand.) While the schedule of the independent demand can only apply to the finished goods in the general production control systems, it can be calculated having any process and any end part and material fixed in TPiCS-X. Having them fixed once, you can further specify how many nearest days to be fixed for. We solve the problem of the leveling using this function.
(2) Shop Floor Control Option.
The origin of the leveling lies in the following. When a person on the factory floor starts his work according to the work order sheet, he says to himself, gI have tons of work today but I will only have half of it tomorrow.h He is in trouble with such a situation. Thinking about the problem returning to this origin, we can think that all thatfs required is for him to know, gHow much I should finish today in order for me to go home?h
Although the shop floor control option allows him to judge, gWhat should I do now?,h it also allows him to know, gHow much should I finish?h at the same time. Even if the production schedule itself has a quantity of unevenness, it can clearly display up to what work he should finish today or whether or not he has to get to the work for tomorrow on the terminal of the factory floor, assuming the leveling, for example, from Monday to Friday.
(Refer to the 4th chapter of gHow to make the production schedule, having the nearest two weeks temporarily fixedh and the 6th chapter of gInstructions of the factory floor and the levelingh respectively.)
Generally speaking, people say, gIn order to for MRP to correctly be calculated, actual results must correctly be entered. It is the same as with TPiCS-X that the data entered to the system must all be correct, but the schedule management (to accurately maintain the schedule) is more important than the actual results entry in TPiCS-X only in terms of getting the accurate MRP calculation result (simulation result). Because the MRP calculations can be done with some degree of accuracy except for special cases (occurrence of a large amount of defect, etc.) without actual results being entered since the MRP calculations in TPiCS-X are done based on the schedules, as described in g¡ A little bit difficult story,h as far as unfinished schedules are concerned.
In contrast, the system couldnft accurately calculate without the information being applied to it even if a supplier informed me, gThe delivery will be late.h
(2) Attempting to make the factory floor observable with TPiCS-X.
Try to think about what happens if the gSchedule Managementh can be done successfully. The schedule data in the system are all living, the factory floor runs according to the schedule, and the suppliers also come to deliver things according to the schedule.
All will be known to everybody, then, by just looking at the screens of the system like what is going on now on each factory floor and what they can do for tomorrow. So, situations of the factory floors will be grasped by just looking at the screens of PCs without actually going there. This is what I mean gobservable.h
The gessence of observable,h though, is not in simply observing but in making it observable to know what to do next after the observation.
You should lead the result to a campaign or implement appropriate measures in advance like the gchilly hath activity of safety control in factories: watching a little signal (predictor) prevents big troubles from happening.
A little bit crafty programmer could easily make the system which only displays data on the screen.
But itfs hard to make the system display constantly living data on the screen.
And itfs harder to actually utilize the data.
I might say, gto put the soul into the system.h
Meanwhile, the gTPiCS-X Production Control Dashboardh is provided as the gmake it observableh tool on the side of actual results. This tool is to create administrative and analytical data based on the data of TPiCS-X. It calculates and displays the data from TPiCS-X of progress status, inventory value changing situation, inventory turnover days, actual production results (amount,) operating efficiency achievement rate, purchasing cost reduction achievement rate, customer and arrangement amount, yield, master analysis, disused inventory.
The Production Control Dashboard allows us to analyze the actual results based on a variety of angles. For instance, the display switches to the breakdown analysis graph upon click of the circular graph of the progress status. What becomes a bottleneck most in the analysis of the progress status is consideration to the influence due to delays in components and previous processes. In this dashboard such analysis to incorporate delays in previous processes, and of the follow-up status to the moving-forward schedule from the original delivery date is possible. Also, the inventory analysis allows us to analyze by inventory value and the number of days in inventory for each item, digging further down to the most recent purchase date and a purchase approver.
(The Production Control Dashboard can be used for free only for those who use TPiCS-X Version 3.2 and have StiLL-X (Version 1.00.019 or later) attached free of charge at the time of the purchase of TPiCS-X.)
(3) Solving the problem of delay with TPiCS-X.
I think that those who have been engaged in the production control for a long time still remember the gFire accident in the Nihonzaka tunnel,h and the gPiston ring of the Chuetsu Offshore Earthquakeh in recent years. We canft prevent delays in production and parts due to sudden accidents and disasters from happening no matter what we try. I think that this is a kind of problem that should rather be handled with the point of view of gRisk Management.h
In contrast, as for the delays that occur in the course of our daily lives, therefs no other way of resolution to gforesee troubles soon and take measures soon.h
There are two kinds of delays that occur in our daily lives, delays due to external factors and those due to internal factors. It is obvious that poorly tackling them causes delays in shipment to occur and a lot of problems to customers at any rate.
If the factor of delays lies close to the production control operations, TPiCS-X can solve most of them.
(1) Delays due to external factors.
In order to get rid of delays in delivery from suppliers, you have to do a purchase to keep the delivery deadline, getting rid of the cause (unreasonable orders) of delays in delivery the purchasing side makes first.
The purchasing can be done without gstirring uph the production in suppliers in TPiCS-X by gpurchasing parts and materials at the time when they are required,h and utilizing the function of a greserve (buffer) for the fluctuation of the productionh as described in g¡ Misperception about the production for short-term delivery.h
In addition, the gOn Schedule and Delay Listh of TPiCS-X plays an important role in the sense of taking preventive measures against delays. For example, the report containing outstanding orders, of which scheduled delivery date will be forthcoming in the next week, is sent to all suppliers by e-mail with a click of the button every Friday to have them confirm the schedule. If therefs something that has already been delayed then, it will obviously appear on the report as gdelayed.h
No matter how hard you try to execute these things, however, there wonft be a day when there are no delays. And if there is a delay, you have to foresee the risk as soon as possible as descried in terms of the Schedule Management, constantly getting living information such as notification of a stretched delivery date.
When you receive notification of a stretched delivery date from a supplier, you have to know whether or not the delay has impact on the current schedule, and what and how much you should make adjustments to in the first place. If you put off the schedule of products, you look for something to be produced instead since you donft want the production volume reduced for that day. You can find answers to what you can make and how many of them you have enough parts to cover, performing the simulation on those in the system with TPiCS-X. Even if you still have delays to occur by this, you can minimize effect of the occurrence.
Just to be safe, Ifd like to add that you shouldnft necessarily accept notification of a stretched delivery date without any reason as a matter of course. It is very important in the conduct of the production control to gkeep the scheduleh – to build a framework to have your suppliers keep delivery dates.
And yet, as whatfs really impossible canft be helped, we are going to think about the production schedule as a realistic solution among whatfs possible.
(2) Delays due to internal factors.
(A) Delays due to oversight of customer orders beyond your ability.
(B) Delays due to oversight in applying to the schedule and arrangement mistake.
(C) Delays due to not-thorough instructions to the factory floor.
(D) Delays due to design mistake.
(E) Delays due to troubles with equipment, jig, and mold.
(F) Delays due to downturn of attendance ratio.
Citing some of the causes, there is something that can be solved by just what is called gmanaging in the system,h and something that canft. That is something that can be prevented by systemization as to (A,) (B,) and (C.)
And even if delays should occur, you still can minimize effect of them by promptly making an executable recovery schedule as with delays due to external factors.
If the data in the system are improper, however, they wonft help you here, either.
(4) Balancing inventory reduction and the production for short-term delivery with TPiCS-X.
As they say, ginventory comes at the end of all production activities,h there are many factors that influence inventory. No matter how many factors there are, however, you canft get an answer without analyzing them. Sorting out the source of inventory in the first place, wefll think about inventory reduction on that basis.
(1) gDead stockh due to miscalculation and arrangement mistake, or engineering change,
(2) gOperating inventoryh that spontaneously occurs in day-to-day production activities,
(3) Self-defensive inventory that prepares for any changes,
Firstly, the source has been divided into the above three.
In order to reduce gdead stockh it is necessary to gimprove prediction accuracyh or gto eliminate arrangement mistakes.h As the countermeasure we can think of the gcalculations with the multidimensional analysis system,h the gimplementation of double-checking as to confirmation work and important mattersh and such. No matter how eagerly you try such countermeasures, you wonft be able to predict the future 100% or completely eliminate mistakes.
The world has been globalized, and the factors to get the market going have increased but never decreased.
In addition, the quality standards and the safety standards to be required must go up but never go down from now on. Engineering changes because of it will increase more and more, and engineering changes because of cost reduction will also last for years to come.
If gsomething you donft want to happenh like these mistakes and miscalculations never happens again, itfs an important answer as well to minimize the damage when gsomething you donft want to happenh did happen. Thatfs where the production control is put to the test.
And it is necessary to conduct speedy monozukuri in order to achieve this. If you shorten procurement period for parts and materials, and make procured lot sizes small, you can reduce dead stock for parts and materials. Ditto with the production for products, you can reduce dead stock for products by the shortening of manufacturing lead-time and compacted lot size. If you purchase parts for 3 months, therefs a possibility that you could have dead stock worth 3 months when some trouble comes up. But if you purchase them only for a month, the dead stock could remain as much as only a month. In corporate activities gswiftness solves 100 problems.h
gSwiftnessh is also the key to reduce operating inventory.
Since gall that come into the factory and go out from there become inventory,h all you have to do is shorten the period between in and out in order to reduce inventory.
Even if parts and materials have been delivered, you canft start your work until you have everything you need. Even if you start it, processes wonft go off smoothly with frequent defects and reworks. If it takes time for tooling change, the production period until finished goods are shipped out will be long. Unnecessary stagnation makes the production period long. All the work results in the production period and inventory.
Operating inventory is controlled by gdelivery lead-time and manufacturing lead-timeh and glot sizeh in TPiCS-X. Lessening these values can make operating inventory reduced.
In the state of the gSchedule Managementh having been achieved you will find the critical point at which these settings canft be lessened any more while they are being lessened. Thatfs the bottleneck in terms of reducing inventory. You are going to have to improve it.
The general production control systems do not have a function to re-calculate the schedule when the most recent schedule has changed, allocating the schedule for parts and materials that have been arranged. The person in charge of arrangements makes the purchase, adding some to what is called the required quantity with gseat-of-the-pantsh and sometimes gguts.h Needless to say, missing his guess leads to dead stock and the inability to handle changes when less quantity is purchased. Although work relying on the gseat-of-the-pantsh must essentially be the supposedly impermissible gworking methodh in a moment of cool-headed, rational thought, such manner of operation is permitted by any manufacturing business due to the restrictions of the general production control systems. Work is sort of judged by a good or bad ginstincth in a sense here.
Therefs a function of ghandling changes with a setup bufferh in TPiCS-X. It automatically reserves a buffer in the MRP calculations as set up in the master. And when a change - difference between forecast and firm orders, urgent order, etc. - arises, it can build the schedule by utilizing it and absorb the change with the buffer.
Additionally, the buffer can hold the set value constantly and properly with the function of gstandard inventory improvement.h You should reduce unnecessary inventory by turning work relying on the gseat-of-the-pantsh to work managed by the system, and you are going to enable your factory to handle changes.
(5) Handling short-term turnaround of new products, and engineering changes with TPiCS-X.
The cycle of changes in the world has become rapid and the period in which things sell has really shortened. Full production has become necessary as soon as possible not to miss sales opportunities and not to have leftover goods. Drawings for trial production are submitted, arrangements are made, and trial products are made. It is repeated several times, and moves on to high-volume trial production and then to high-volume production. Drawing changes are ceaselessly repeated during that period.
As a matter of course, this problem largely includes the elements other than the production control such as design development and production engineering. I am going to explain how to use TPiCS-X and how to approach to the gshort-term turnaroundh in this kind of problem.
Here are two points to raise:
(1) Function to convert the bill of material information (E/BOM) on the design to the bill of material information (M/BOM) of the production control, including engineering changes.
(2) Manage both the trial and the repetitive productions only with TPiCS-X.
¡ Function to convert the bill of material information (E/BOM) on the design that is kept in CAD and Excel to the bill of material information (M/BOM) of the production control, including engineering changes.
The gCAD Data Conversion Optionh of TPiCS-X coverts the bill of material information for the design to that for the production control. The bill of material information for the design is written to a CSV file, imported by TPiCS-X, and converted to that for the production control. The CSV file can be created from Excel as long as the bill of materials is managed not only in CAD but also in Excel.
In general, the bill of material information for the design is different from that for the production control.
You were to produce personal computers, for example.
I think your designer would treat the power switch as a component of the power circuit.
However, the power switch is mounted on the front panel. When you were to actually produce it, you would purchase it to your partner as the gfront panel assembly,h and they would connect it to the lead wire stretching from the power circuit in the assembly process.
If you follow the way of this production, you position the power switch as a component of the front panel in the bill of materials for the production. You can easily make these changes of the structure using the drag-and-drop functionality on the screen of the CAD Data Conversion Option. After changing to the structure for the production, you are gong to import it to the masters of TPiCS-X (the Bill of Materials and the Item master.) Adding the unit cost, the supplier, the lot size, etc. in addition to that, youfll be done and able to perform the MRP calculations as a new product.
The system development is not so hard if these functions only deal with the conversion of new drawings. When you learn to manage as far as the engineering change information, it becomes hard all of a sudden.
Letfs assume that the power switch changes from SW1 to SW2 in the engineering change. The information of the change on the design is the one of gChange the power switch SW1 as a component in the power circuit to SW2.h But therefs no power switch under the power circuit according to the structure for the production. It appears to be all right if you find the gSW1h and change it to the gSW2h at any rate, but it doesnft go that way in reality.
It is all right in the case that the power switch is used only in one place like that of a personal computer, and that all are to be changed even if it is used more than one place. But if the same part is used in more than one place, you canft identify which part among those used in multiple places has been changed.
The CAD Data Conversion Option of TPiCS-X keeps the information on both structures in one table, and when the structure information on the design changes, you can apply the engineering change to that for the production.
Now, letfs think about how to apply the engineering change to the production.
Since drawings are strongly joined to manufacturing numbers in the discrete production, cancel and additional orders can be issued in TPiCS-X by directly verifying what has already been arranged using the manufacturing number as the key. In the repetitive production, however, changes are applied to the production by specifying the period for the change.
As specification changes other than the above occur everywhere, the system, like any other things, has to be able to process them.
- How to switch to a new part since the time when inventory and what is already arranged are used up (engineering changes for the purpose of running change and cost reduction.)
- How to change the quantity used depending on the period (season) (seasoning with salt in food.)
- How to use a different part only for a limited time period and to turn it back to the original once the period is over (limited-time product.)
- How to use a part of other makerfs when the part from a maker different from the instructions of the drawing is usable in the production (the case that is frequently seen in electronic parts.)
- In the case of producing based on a special specification only this time by customerfs request.
- In the case of the change after the work order issuance and the production instructions.
- In the case of the change after completion.
(Such cases where products are actually to be renovated, and where a master error is found after completion and only the data has to be corrected.)
- This is not an engineering change, but the processing where the structure varies depending on the place to produce at.
(Internal production and external production, and arrangement method gets changed depending on the place to produce at.)
Engineering changes have impact on material arrangements.
The required quantities decrease in parts before change, and those of new parts have to increase.
In the case of producing based on a temporary, special specification other than the engineering change being explained as the premise now, a similar processing is also required.
In the case of a temporary, special specification, you can process it without applying it to the master (bill of material information) in the production control system. When you choose to process it without applying it, you will not need the part in the structure information and will have to make arrangements for a part with the special specification.
When issue instructions are given for the part from the warehouse to the factory floor, the new part has to appear on the issue report.
And the inventory relief processing for materials is influenced at the time of completion.
Furthermore, after they become inventory at the time of completion,
sometimes they have to be treated as something different and sometimes they are treated as the same things as commercial products.
As the explanation would go into too much detail if I were to explain those functions in the system and setting methods, I would like to omit it in this booklet but to say that all can be processed in TPiCS-X.
¡ Manage both the trial and the repetitive productions only with TPiCS-X.
Therefs an aspect where even products for the repetitive production are more manageable at the stage of the trial production with the discrete production-like management.
TPiCS-X has both the functions mainly for the repetitive production and those mainly for the discrete production (g3. Discrete Productionh for detail) in one package. You can manage your production with TPiCS-X as the discrete production (manufacturing number control) at the stage of the trial production, and as the repetitive production (the MRP calculations) at the time of the high-volume production.
When you manage your trial production with the functions for the repetitive production without using the functions for the discrete production, you will have trouble if what was arranged for the trial production is allocated by the MRP calculations for the high-volume production. On the contrary, in the case that the parts are the same as those in old models, there are times where you want the work order sheet to be issued performing the MRP calculations including those for the trial production.
Therefs a function of gexcluded from the allocation by the MRP calculationsh in TPiCS-X, and you can have the parts to be used for the trial production gexcluded from the allocationh or gincluded in the allocationh by it. You can differentiate the stage of the order balances from that after being put into stock with the setting.
If you manage with the same system as for the high-volume production from the stage of the trial production, you can gcheck the masters for the high-volume production while doing the trial production,h and achieve the prompt start-up for the high-volume production.
Itfs a tremendous job to build the proper masters while there are so many changes in the start-up time. One false move and you have a missing part, or you end up arranging an unnecessary part. While the long-term turnaround only generates a small amount of influence from the wrong arrangement, the short-term turnaround will increase the damage.
¡ The essence of difficulty and toughness of the production control lies in gchange and alteration.h
Think about this. If what and how many to make, and the specifications were determined for the past 6 months, the production control would be an easy job to do, wouldnft it?
gChangeh means gchange in quantity and periodh and gspecification change.h
gChange in quantity and periodh can be solved by the concept of the schedule management centered on f-MRP, and TPiCS has its own answer as for gspecification changeh as well.
And, in reality, gchange in quantity and periodh and gspecification changeh are pouring in in a mixed, chaotic state. TPiCS-X can process them as routine.
I would like you to do your job of the production control in a prompt, accurate, and easy manner using TPiCS-X.
gDiscrete productionh in a word really represents a variety of things, but letfs start with pigeonholing the discrete production.
To begin with, Ifm going to think about it in terms of gdegree of discrete.h
- Every time receiving an order from your customer, you start making all the drawings and telling that you have almost nothing to repeatedly use and so is your company with internal inventory. Therefs a gung-ho discrete production such as this.
- Or, although they are different as finished goods in each case, their parts and materials to be used are just about the same. So, therefs also the production with parts and materials in stock to some extent.
- Therefs also a case in which to handle with a combination of units as unitization goes along.
- A case where almost all parts of the products are the same, but only a part of them such as nameplate, ancillary parts, and lead wires changes.
- Furthermore, even a case where the product to be produced is a high-volume production item, but it is managed in a discrete manner (manufacturing number control) as the internal management method.
It is not uncommon that therefs a mixture of something strong and something weak in the degree of discrete in one factory depending on the product.
Let me think about the intensity of parts as another method of classification now.
- In the case of producing a product collecting a lot of pars, and
- producing a product processing one (or a little) material.
That is, whether materials and parts to be used are many or few is an important point that could affect the problem of the production control.
In addition, the method of processing one material can also be broken down:
- In the case of a completely different form and a different processing method in each case, and
- being possible in developing a certain pattern to some extent.
As how to manage is different in these cases, how to use TPiCS-X changes accordingly.
Since I canft explain everything in this booklet covering all cases, Ifm going to give you the explanation focusing on the case where multiple (a number of) parts and materials are used though there isnft a lot of repeatability.
Meanwhile, if you use the function of the discrete production, the gManufacturing Number Control Optionh is indispensable.
On the contrary, if you donft need the MRP calculations with no repeatability at all, you can purchase the system at a moderate price by building your system centered on the gSmall Business Pack,h not the gf-MRP Basic System.h
(1) Achieving the production for short-term delivery and ready for whatever changes come along with TPiCS-X.
Handling short-term turnaround of new products, and engineering changes with TPiCS-X.
As Ifm going to give my explanation based on the production in small quantity here, the gshort-term turnaroundh will be explained in the section (1) together in terms of gthe production for short-term delivery = shortening entire work period.h
¡ To begin with, letfs organize work from customer order to shipping.
(1) Design work ¨ release of drawing,
(2) Production schedule creation, arrangements work, and filling in a form ¨ purchase,
(3) Delivery from a supplier,
(4) Outsourcing operations, internal operations, and assembly completed,
and so on.
(A) Increase the pace of each work (shorten each work period,)
(B) Perform work in parallel.
The above is a theory.
¡ Improvements in an aspect of the work.
We are going to think about (A) Increase the pace of each work first as theory says.
The production control has a direct effect, in a narrow sense, only on (2) Production schedule creation, arrangements work, and filling in a form.
If therefs repeatability, you register master files in the system, and perform material explosion and process explosion using them. If therefs no repeatability, gyou register master files first ch will require more work. If grouping is possible at least, you could have a way of developing a certain pattern out of it, and registering the pattern like a master to create data for arrangements using the master. If even grouping is not possible, you end up directly entering them from the keyboard like a word processor. Operability is all you need in that case. For example, you can get the data, like supplier information, among what is already registered by just entering the supplier code. Selecting it from the pull-down list, and dragging and dropping it sounds even better, and you will be far better off with that alone in doing day-to-day work. However, this function alone leaves you dissatisfied in so much developed IT environment.
And so we developed gCAD Data Conversion Optionh and One Item Production Option.h
TPiCS-X imports the bill of material information for engineering design written to a CSV file. Creating source data for arrangements from the imported data without an operation of master registration, it can directly create the data for order issues, writing the supplier, the unit cost, and the due date to them.
The CSV file is written from the CAD, if it has the bill of material information and can export it to a CSV file, or from Excel, if it manages the bill of materials, and imported by TPiCS-X.
These data of structure information are hierarchized, and if a new product uses an existing unit, just specifying the drawing number of that unit will do, and the data for grandchild components will not be necessary.
In this way, if you have the bill of material information in the CAD or the data of the bill of materials in Excel, arrangement work will improve and speed up dramatically.
Way to use similar data in the past.
When you produced a similar product in the past and if you can use the data at that time, your work will go fast. If you use the data in the past, you usually have a few corrections. For example, the specification your customer has requested this time has a slight change; you must use a substitute item because the old part is discontinued; etc. In the One Item Production Option you can change the past data to meet the specification this time after copying them, and have the order issue data created.
Next, wefll think about (B) Perform work in parallel.
The relationship between design and arrangement is the most important point in thinking about this.
Start your arrangements in the middle of design from wherever possible even if it is not all finished.
But I think that this is something many factories already put in practice.
The biggest bottleneck or something that causes trouble in doing this, however, is engineering change. Isnft it?
Although the production control job is not so difficult individually, it becomes difficult as there are a lot of elements (parts,) and they act disconnected and constantly change.
If you use the CAD Data Conversion Option and the One Item Production Option previously mentioned at this point, you can achieve the following processing:
Upon receipt of engineering change data,
- Unnecessary components in the above source data for arrangements get marked due to the engineering change, and only the unnecessary data are displayed upon click of the [Unnecessary] button.
- New data for arrangements are created for the components that have newly become necessary.
- Displaying the data that have become unnecessary and talking about countermeasures with your suppliers, enter the outcome to each data.
- Release the allocation to the manufacturing numbers among those that canft be cancelled and can be diverted to others, and put them into scrap disposal unless the diversion is possible.
- Issue cancel orders for those that canft be cancelled.
You can easily process engineering changes that occur frequently in a large amount of arrangement data like this.
When hurrying up the design and hastening to make arrangements using the bill of material information (E/BOM) for arrangements, you will face a growing risk in engineering change.
It is very important to accurately and promptly notify your suppliers of the change. You might eventually cause a wrong arrangement and shortage to happen, which could have the opposite effect of the production.
As TPiCS-X can handle engineering changes this much as a system, however, it can solve this problem as well.
¡ Improvements in an aspect of the production.
Manufacturing number for order in advance.
There is a function of gManufacturing Numbers for Advanced Orders (Unofficial Manufacturing Numbers)h in the Manufacturing Number Control Option of TPiCS-X.
You can make arrangements for something, which has some repeatability but takes time in arrangements and production such as components and functional components, with the gManufacturing Numbers for Advanced Ordersh to some extent.
Having made arrangements beforehand with the Manufacturing Numbers for Advanced Orders, those unofficial manufacturing numbers can be assigned automatically when customer orders come in afterwards and assignments of official manufacturing numbers occur.
When thinking about the production for short-term delivery in the discrete production, one of the most important things is whether or not parts and materials are going to be delivered as scheduled.
Needless to say, no shipments can happen if even a part is missing.
So in order to shorten delivery dates, parts have to be delivered as scheduled. And definite follow-ups are necessary for that.
With that in mind, Ifm going to briefly explain the follow-ups next.
Since itfs the follow-ups to prevent delays in delivery, you canft get anything done by making a scene after the fact. Therefs the gOn Schedule and Delay Listh in TPiCS-X. It lists all outstanding orders, for example, that will have become due some time in the following week (depending on the setting.) Emailing it to your suppliers is possible, let alone printing it. If you register the mail addresses of your suppliers, you can send it to all your suppliers with a click of the button off the screen of TPiCS-X. Since it includes delays in delivery as well, itfs going be a sure follow-up.
Or, if you use the Strategic Due Date Adjustment Option, you can e-mail orders and gdue date reply requestsh to your suppliers.
If the suppliers use the terminal program (free) of the Strategic Due Date Adjustment Option, they can automatically receive your e-mail and send back their change request of delivery date with a click of the button off the terminal program. Then, you can automatically receive their change request of delivery date with TPiCS-X and apply the change to the Order Balance data of TPiCS-X again with a click of the button.
When the change is applied to the Order Balance data of TPiCS-X and if anteroposterior processes are reversed due to the delivery date change, you can immediately tell that you are going to need the readjustment of the schedule since it is displayed in red in the Gantt chart.
The readjustment of the schedule can be made by a drag-and-drop of the Gantt chart. Usually a lot of schedules become victim to this kind of readjustment of the schedule. Failing to send a notification of the result of the readjustment is even scary but with the use of the Strategic Due Date Adjustment Option you can send an email of the delivery date change to all of whoever suppliers to be influenced only with some clicks of the button once the readjustment is over. Thus, you can constantly maintain the schedule to build the reliable, lean schedule.
(2) Attempting to make the factory floor observable with TPiCS-X .
As far as making the factory floor observable, essence is the same in the repetitive and discrete productions.
In order to achieve gobservable,h it is necessary to always match the Schedule data in the system with the Execution Plan, and to keep the production as scheduled in the system.
For example, you can see the schedule in the Gantt chart. You can tell how far the manufacturing number has advanced with the Gantt chart. Even if the system has those functions, data are nonsense and so are the contents to be displayed. This is anything but gobservable.h
Next, essence of gobservableh is what to do once itfs seen.
- Like the Shop Floor Control Option, the Gantt chart of TPiCS-X displays , and on it, and allows you to understand what is possible to start now and to relate you to the next action.
- And as you understand the workload status of the job by looking at the workload display of the Gantt chart, you can adjust the schedule before delay in production occurs.
- Or, when stretching requests (delay information) from your suppliers are applied to the Schedule data in TPiCS-X, you can tell a problematic spot immediately and make adjustments to it since delivery dates for parts that are behind schedule against the following processes are displayed in red.
- These schedule adjustments can be done by the drag-and-drop functionality of the Gantt chart. As a result of the adjustments you can automatically issue orders for delivery date change, and directly send data for delivery date change to your suppliers using the Strategic Due Date Adjustment Option, which can prevent you from terribly failing to send the notification to them.
To constantly maintain the Schedule data in the system is as important as to always maintain the master files right.
It is no easy matter but very important to always maintain the data right in the system, to put the soul into the system, and to surely get the most out of the data. (Refer to gobservableh in 2. Repetitive Production for detail as well.)
(3) Solving the problem of delay with TPiCS-X.
As for the problem of delay, the idea is the same as the repetitive production, too.
In order not to make something that shouldnft happen happen, the only way is gto find a sign of delay and destroy the problem beforehandh like the gchilly hath of safety control in factories. (Refer to gSolving the problem of delay with TPiCS-Xh in 2. Repetitive Production for detail as well.)
As for the adjustments to the production schedule for the discrete production, you directly make changes to the schedule with the mouse on the Gantt chart, considering the association of the daily schedule with the previous and following processes and the workload.
(4) Balancing inventory reduction and the production for short-term delivery with TPiCS-X.
We are going to think about inventory reduction by source of inventory even though itfs the discrete production.
Here are those sources:
(1) gDead stockh due to miscalculation, mistakes in arrangements, and engineering changes.
(2) gOperating inventoryh that spontaneously occurs in day-to-day production activities.
¡ Reducing gdead stock.h
Issuing orders as required.
gOur companyfs policy is not to keep inventory because we are implementing the Manufacturing Number Control and we only make as much arrangements as required.h There are many such companies but there are very few companies that hold no inventory after physical inventory at the end of the year.
Without cancel orders from customers, without engineering changes, or without erroneous arrangements, unnecessary inventory must not occur.
As long as gsomething you donft want to happenh never goes away, however, it is an important answer to keep the damage to the minimum when gsomething you donft want to happenh does happen. And thatfs where the production control is put to the test.
In the case of the manufacturing number control, when the drawing comes out, you can reduce the risk of stock in dead storage by not making a purchase until the time has come without purchasing all by manufacturing number right away.
In the case of making a purchase with the Manufacturing Number Control in TPiCS-X, the setting of gOrder Issue Periodh also works. The scheduled order issue date is calculated backward from the delivery date of each part after the manufacturing number explosion. And if itfs not the time to issue orders yet, orders do not get issued. Instead, when the time has come for orders to have to be issued in daily gorder issueh operation, the Order data are automatically populated and the order forms are printed.
Since the method of gissuing orders laterh causes missing orders to happen in manual procedures, people end up issuing orders at a time by any means. Or, if the system is developed based on the manual procedures, it may always be designed to issue orders at a time.
In the case of the discrete production, however, the risk of delay in delivery by suppliers may increase when issuing orders later. Therefore, you must consider the balance between them.
Inventory allocation function by the system.
Setting it aside whether itfs deliberate or unwanted inventory, if therefs inventory in the process of arrangements, allocating it results in the effective utilization (less inventory) and inventory decreases.
When performing arrangement work in manual procedures, there are many things that are gaccidentallyh forgotten. Even if it is processed by the system, there are quite a lot of systems that have no functions of inventory allocation at all in the case of the general Manufacturing Number Control systems. Since TPiCS-X allows you to allocate inventory that has not yet been allocated before to the manufacturing numbers at the time of the next arrangement of the manufacturing numbers, you can reduce inventory.
Reducing mistakes in arrangement work.
It is important to reduce erroneous arrangements as well in order to reduce inventory. I explained it focusing on gmaking it easier and faster processingh in the section of the gproduction for short-term delivery,h but the linkage between the CAD Data Conversion Option and the One Item Production Option of TPiCS-X has another important effect of greducing mistakes in arrangements.h
Needless to say, no erroneous arrangements lead to no inventory caused by them.
¡ Reducing the goperating inventory.h
The operating inventory can be categorized as ginventory that inevitably occurs from coming into the factory and going out from thereh or ginventory that prepares for unexpected orders.h Since greducing inventory that inevitably occursh and greducing delays in production and shippingh have almost the same meaning in the case of the discrete production, I think it is unnecessary to get onto it in terms of ginventory.h
It is necessary, as a matter of course, to gexamine the necessityh or to glessen the lot to purchaseh in order to reduce ginventory that prepares for unexpected orders.h In addition, the gstandardization of the parts to useh also has effect.
If the kind and quantity of inventoried parts increase, you can manage only parts and materials with f-MRP. TPiCS-X allows you to control the production with the manufacturing numbers depending on the item, to process it with f-MRP, and to set up in combination (by mixture) of the above two based on the part and the nature of the unit. And it is also possible for you to allocate something that was arranged in advance with f-MRP to the manufacturing numbers, and to arrange the shortfall with f-MRP if there is any as a result of the manufacturing number explosion. This is an efficient function when the kind and quantity of inventoried parts increase.
4. How to build the production schedule with the nearest two weeks temporarily being fixed
Excerpt from TPiCS report No. 79.
¡ I wrote a similar theme to this in the previous report titled gregarding the instant delivery system and the production for short-term delivery.h
But I received the responses from many people, which says, gtoo difficult to understand.h
I believe that to achieve the production with as little inventory as possible, making gproduction to meet your customersf needsh and gscheduled production,h is a very important issue of ghow fast to make what the customers want,h not the issue of the gsystemh or the gnarrowly-defined production control.hSo Ifm going to explain in a little plainer language this time.
I wrote the following in my previous report. Thinking about the production control, ghow to make the schedule and to maintain ith is very important. But I notice that there are many people who are considering the production control systems but who are not so interested in that point. Although I receive a lot of questions up to the chin regarding gcost,h I find many people, who are considering the production control systems, donft know much about, for example, ghow orders come in from their customers.h
It appears to me that what is expected from the gproduction control systemsh in general is different from what TPiCS-X intends to provide. As our company is a joint-stock corporation and at least a profit-oriented company, all thatfs required is to make and sell what the customers want. But I donft quite feel like doing it because such thoughts as gitfs useless to make such a thingh and gnot a problem resolution at all to use ith are stronger in myself.
As discriminating users could understand me, however, I continue my small-sized, TPiCS-style business with their emotional support. (^J-)
There are words of gsetup finishes 80% of your work.h I like these words so much that I frequently use them. Needless to say, it means gsuccess and failure of your work are determined by the setup.h Replacing gsetuph with gschedule,h we say ggood and bad of the production scheduleh determines good and bad of gproduction.h
However, I wonder why so little importance is put on the schedule generation function. Letfs think about why.
In the case of the general MRP-family systems, human creates the gmaster plan,h enters it to the system, and perform the MRP calculations. In other words, since therefs no function (or very weak if any) to make the schedule of products using the system, or to make the production schedule with the system in the general production control systems, users no longer expect the production control system to do anything for them. Or, assuming that gthe system canft do ith or gthatfs what human should think,h do they expect nothing from the system?
I would like you to find gwhat a way to do with TPiCSh or gseems like I can do what Ifve been thinking ofh in this report.
Since our company is a software package maker, we always think of something, assuming a variety of production situations. When I am aware of the versatility in my explanation, it gets a little difficult to understand. In addition, a story of the production control tends to be long by any means and it ends up with a gbucket shop profitable if the wind blowsh-like story.
Establishment of TPiCS itself started with squeezing what was in my brain out of it first, and ends up in a general-purpose package. It never systemized a userfs management method somewhere into a package. Because of that, you can use the whole system from a variety of standpoints and viewpoints. I guess thatfs one of the factors that makes people feel TPiCS difficult.
¡ To begin with, we are going to think about the relationship between sales plan and production schedule.
There are companies who have what is called gsales planh and who donft have one in the first place. Among those who have a sales plan, there are companies whose plan is good in terms of the degree of accuracy (Case 1,) and whose plan is bad (Case 2.) There are also companies who can approximately figure it out from the past experience even without any sales plans (Case 3,) and who donft know whatfs coming at all (Case 4.)
When you think about a sales plan, you must consider the length of the period in the schedule, the degree of the leveling, and the maintenance cycle.
If you create the production schedule with enough time in the Case 1, you would consult it a great deal. But if you are in the Case 4, if the degree of accuracy of the schedule is bad and not reliable, or if you only have the schedule for a very short period of time if any, you would find it difficult to use the information.
To make the explanation simple, this report only covers the Case 4 with instant delivery for consideration.
Once you understand my explanation and the functions of TPiCS-X, Ifm sure that you will understand the most important concept of the MRP calculations in TPiCS-X, and a concrete way to meet the contradictory request of making gscheduled productionh while making gproduction to meet your customersf needs.h
¡ Ifm going to explain the operating method of the MRP calculations in TPiCS-X based on this use.
The MRP calculations of TPiCS-X allow you to enter customer orders and to perform the MRP calculations on a daily basis while temporarily fixing the production schedule of products. (To tell you the truth, the temporary fix is possible not only on gproductsh but also on parts and semi-finished goods. Itfs a very important function in the MRP calculations of TPiCS-X, but my explanation would be complicated if I wrote it. So Ifm going to skip such explanation as much as possible in this report.) If you have too many customer orders for the temporarily fixed production schedule to meet your customersf needs, or if customer orders stop coming in all of a sudden and inventory goes beyond a setup value, TPiCS-X will let you know that as a report.
Once you notice the shortage and overabundance in inventory, you make changes to the production schedule, considering the leveling and the setup, and perform the MRP calculations again.
Also, the MRP calculations of TPiCS-X allow you to make the schedule with the same quantity as the customer ordersf on the following day after the temporarily fixed period is over. (I guess it is a little difficult to understand here, I will explain it for detail later.) When orders for the current day are issued with the result of the MRP calculations and the final processing is performed, the temporarily fixed period extends one day with the scheduled quantity remaining. Continuing this for a week results in the production schedule before the leveling in the week after next week. After making adjustments to the schedule for the week after next week for the leveling and the setup every Friday, it becomes the final production schedule.
At the same time, you can also review the production schedule for the next week as required.
When you make adjustments to the schedule for the week after next week and if you keep the weekly total for each product unchanged, you can make the schedule that constantly produces just as much quantity as the customer ordersf this week. At the same time as this schedule planning work, you can also adjust the operation capacity. That is, if you had too many orders this week to be able to produce all of them using even the week after next week, you would have to carry it forward to the following week. If you had too few, it may be that you would have to make some extra hot-selling products. These kinds of things are incorporated in the schedule after going through managerial decision-making. Even if you made the schedule with extra quantities for the week after next week in the process this week, the production schedule from which those are subtracted would be shown in the next week. Therefore, therefs no bias in inventory even if careful attention is not paid to inventory.
Once you make the production schedule for products, you have to execute the MRP calculations for the second time. The first MRP calculations only go though the finished goods level and the second ones go through semi-finished goods and parts.
TPiCS-X allows you to execute the calculations as many times as you want as long as orders are not issued.
By running the operation like this, you can achieve the gscheduled productionh with as little inventory as possible, making gproduction to meet your customersf needs.h
¡ Ifm going to explain specific settings and system behavior based on simple goperating assumptionsh here.
(1) The production schedule is updated in a weekly cycle (every Friday.)
(2) There are no forecast or sales plans and the MRP calculations are performed based on customer orders that come in on a daily basis.
(3) The customer orders come in for todayfs and tomorrowfs shipments.
(4) The scheduled production is run for the nearest 2 weeks and the sale is to be reflected in the schedule from the third week on.
(5) The work orders for products are issued for the next dayfs production.
(6) The customer orders for typical products come in 10 pieces per day on an average.
(When you think about this, you can assume 10 pieces are actually shipped out every day in the beginning, and then think over the case where it changes on a daily basis.)
(7) 5 each of the above products exist in inventory in the beginning.
(8) In the meantime, assume there will be no lot-sizing.
(10) In addition, focus on the schedule for products only.
Orders of 10 pieces with no lot-sizing every day mean that 10 pieces must be produced every day.
Then, the figures for production, sales, and inventory will look like shown below:
Receiving orders of 10 pieces for Mondayfs shipment in the evening of the current day (Friday, Day 0,) you expect 10 pieces in the production schedule on Monday in the week after next week (the 11th day) at the MRP calculations. The MRP calculations of TPiCS have a characteristic of making the production schedule to compensate the shortfall on the day right after the fixed period if the scheduled inventory exceeds the standard inventory. Since the scheduled inventory is 95 on the 10th, 10 will come up in the schedule if you set the standard inventory to 105.
The fixed period is set to 11 since 11 days ahead of the current day are to be fixed. The order issue period is set to 1 since orders for tomorrow are to be issued.
With these settings being set, if you issue orders and perform the final processing, the schedule will temporarily be fixed up to the 11th day.
10 pieces are shipped out and 10 pieces of products are completed in parallel with that. The actual results for those are to be entered.
Next, letfs assume that the current day becomes Monday and you have a lot of customer orders for today, which results in 13 pieces.
When the MRP calculations are performed, 13 will come up in the schedule on Tuesday, the 12th day this time.
With this continuing on for one week, the quantities of customer orders received every day from the day 0 to the 4th will appear from the 11th to the 15th as the production schedule. Since daily customer order quantities are unequal, some days have much in the production schedule and some days have a little, which means they are not leveled at all. Thatfs where you have to make adjustments to the schedule on the 5th (Friday,) considering the leveling and the setup.
¡ Ifm going to think about the standard inventory and the physical inventory now.
In the operating method like this, you have to set the standard inventory to a large value. As you understand it right away, however, what actually remains in inventory must be less. If you have customer orders worth 10 pieces every day, this setting should be about 5.
¡ Let me think about the variation of customer orders.
Though you think the standard inventory you set is large enough, this setting unchanged may cause immediate journal writing and changes to the production schedule to happen when big customer orders (more than 15) come in at one time. When considering the variation of quantities of daily customer orders, the standard inventory should be a larger value. On the contrary, if only a small amount of customer orders comes in, the production schedule from the week after next week gets small and is adjusted accordingly.
You might think setting the standard inventory is difficult to a certain extent, but using the gautomated improvement function of the standard inventoryh in TPiCS-X, you can leave it to the system, too.
¡ Ifm going to think about the situation where lot-sizing is applied here.
Suppose the lot-size is 30. That results in the production worth about 3 days at one time. In other word, some quantity appears in the schedule (is to be produced) once every three days. (Suppose initial inventory is 15.)
It is all right that nothing appears in the production schedule on the 11th in the processing for the day 0.
When the 2nd (Tuesday) comes, the scheduled inventory becomes 95 on the last day of the temporarily fixed period. And if you set the standard inventory to 105, 1 lot (30 pieces) will appear in the production schedule. In the case of the lot sizing like this, it is obvious that the quantities will appear in the production schedule in the frequency comparable to the customer ordersf.
¡ Ifm going to think about the actual operation image.
We have thought about just one product so far, but we do have a lot of products in reality. We have more variation of quantities of customer orders, too. And we may also have larger lot-sizes. As a result of that, there are some products that will not be produced for a week, and the schedule on a day will have a large figure. You have to make adjustments to it once a week. Keeping the total value of the week unchanged at that time, you learn to constantly have inventory at the set level at the end of every week. On the other hand, even if you adjust the production schedule to change the total value, it will return to the set level of inventory at the end of the following week as it is re-adjusted in the schedule for the following week as long as the standard inventory is not changed.
The schedule for child parts follows the normal f-MRP calculations based on the required quantities calculated from their parent items as long as it is within the period where the schedule for products is temporarily fixed. Parts that have long legs in procurement are calculated in a similar logic to that for products as explained here.
The f-MRP calculations of TPiCS calculate in exactly the same logic for finished goods, semi-finished goods, and purchased end items and materials.
This is the way to achieve the gproduction to meet your customersf needsh and the gscheduled productionh with as little inventory as possible.
Excerpt from TPiCS report No. 70.
¡ The production for short-term delivery and schedule management.
The production for short-term delivery is intended to have good response to customersf needs and to promptly provide them with commercial products. We used to handle this by keeping inventory for products in olden days. But the life cycle for commercial products has shortened and we have faced a growing risk of inventory for products leading to dead stock. Or, we have started thinking about rather making production in response to customersf needs for a wide range of reasons; for example, variation of commercial products has increased and we have to keep too much inventory if we respond only with inventory for products.
difficult, however, to change the principles of the manufacturing industry
I explain it in the training workshop, too, that gto keep the schedule and to change it are inextricably associated with each other with razor-thin margin.h
If I may use big words, the gessence of managementh is make rules, make them open to the public, and everybody keeps them. If they are bad, change them.
This applies not only to the management of general work but also to the production control.
The grulesh are the gproduction scheduleh in the production control. gMake the schedule, demonstrate it, and keep it. If the schedule needs to be changed, promptly change it and immediately notify whomever concerned of the changed schedule.h Being organized in words, it looks entirely obvious. So I think therefs nobody who disagrees with this thinking.
When people are in the midst of it, however, they tend not to be able to make some sound judgment.
It may be because therefs a problem of gsuitable time,h gdegree,h or gextent.h
The standard of gperfectly suitable timeh varies depending on the circumstances of the world. As the circumstances of the world gradually change in an analog way, it turns out that gthe standard has become outdated before I know it.h
To achieve the production for short-term delivery is, for example, to gchange the schedule for today and tomorrow, and to reflect it in the actual production.h The cycle to make the schedule must be short to that end. And you have to figure out the following in no time at all: gIs the schedule to change going to be feasible or not? If therefs a problem in it, what could that be?h
That is, you need to gmeasure the feasibility using the computer.h
Since the f-MRP calculations of TPiCS not only calculate the required quantities but also perform a function to examine whether or not the current production schedule can handle new situations or what that could be if therefs a problem, and to report it (function to write the problem to the journal,) all you have to do is to enter your customersf needs to the system and to execute the f-MRP calculations in TPiCS.
From this point on is the main issue of this report.
If the schedule in TPiCS has a bunch of malarkey, the calculation result will be meaningless no matter how brilliantly the system does the calculation.
When it come to this story, the discussion always ends in gThatfs why we need to correctly register the masters,h and gInventory in the system has to match the physical inventory, too.h
If no masters or inventory is correct, the system is useless for sure. But that is not enough.
The entire production schedule used in the calculation has to be the same as the schedule you are going to actually implement. For example, what if you only enter the spots to be changed this time and leave the rest of it as it is, which is the schedule made at the beginning of the month. If the calculation is made based on the schedule you are not going to actually implement, the calculation result will be useless. Even if you received a report saying, gWe canft make production because we have shortage in parts,h after the calculation based on the production schedule in which something that is not going to be scheduled for production is included, it would be of no use. On the contrary, even if you were told that you have too many parts in the situation where the schedule based on which you are going to make production is not incorporated in the production schedule, it would be of no use, either.
Normally therefs the production schedule made yesterday, you only enter changes for today to it, and you have the system calculate it. In other words, the portion you didnft enter changes to must always reflect the actual production, too. Since you donft know where to change tomorrow, the entire schedule in the system must always reflect the actual production.
Let me explain it using a little easier example to understand.
For example, suppose somebody received a request from a parts maker who cried, gWe canft make the delivery for December 10 under such and such circumstances. Is it all right if we delay just the half of it by 3 days?h Since there are some purchases to replenish the standard inventory and there are some to lot-size in TPiCS, there are a lot of cases where it doesnft matter that about the half of deliveries is behind schedule. The person who received the call properly looked at the data in TPiCS, confirmed there were no problems at all, and gave the OK. Then he wrote it down in his notebook in a proper fashion, but unfortunately he forgot to apply it to the data in TPiCS.
Afterwards, a product that used the part was added on December 12th, and the child part was required on December 11th.
But the delivery of it was scheduled on December 13th. If this went on, the additional production that should have been added to the schedule was impossible with the data remaining uncorrected in the system | that is, the entire quantity was scheduled to come in on December 10th. So there was no way to give him an alert message. If there was no alert message coming out, there was no way to take the necessary action, assuming that the additional production gwas to be made without any problems.h
If you think about this problem, targeting the schedule a few months ahead | that is, the delivery date is far ahead, | it becomes less important. Even if therefs a problem in the schedule to be shown, it can be handled somehow since therefs enough time on the parts makerfs side. On the other hand, it canft be done means git canft be doneh in the case of the production for short-term delivery.
When I told this story in the workshop we hold every month, a young man asked me, gHearing that story, I feel like itfs contradictory to the story of gkeeping the scheduleh Mr. Ninomiya was telling about a while ago. How can I interpret it?h
That question is too difficult to answer over a short amount of time. So I said, gGood question,h and tried to look for words for the answer. gNever give up the schedule right away as you say, saying, eWe canft do it.f In contrast, knowing that you canft do it, it is no good not to apply it to the schedule, either. The best way to get an answer to this kind of question is to return to the starting point and think it over from scratch. What are we making production for? Thatfs for the customers to buy. It would be useless for them to not buy. We would only be producing industrial waste from factories. We make production to meet the customersf needs. We keep the schedule for it. In case we canft keep the schedule, we change the schedule. Our priority is like the above. To express this in a word, itfs going to be, ewe are going to make customer-oriented production for sure, and we are going to make the schedule for that purpose.fh
I call this gschedule management.h Managing the schedule itself, the schedule is intended to always become one with the actual production.
In the case of the production for short-term delivery, this becomes very important.
I would like to insert disclaimers here to avoid misunderstanding.
If you can get all the parts and materials in a shorter cycle or you can keep as much inventory as possible to handle all changes no matter how fast the production cycle is, the concept of gschedule managementh is not important.
¡ Strategic Due Date Adjustment Option
I assume you were able to understand the concept of this option.
But it is a very hard job to achieve this in a fast cycle.
You receive orders from your customers every day, execute the MRP calculations every day based on them, and make purchases. There are a lot of parts and suppliers involved. A call comes in from a parts maker in the midst of extreme busyness and cries for help, gWe canft make ith and gWe want it delayed.h If you donft have a good command of this, you canft achieve the true production for short-term delivery to survive.
I have developed the Strategic Due Date Adjustment Option in TPiCS-X Version 3.1 this time. It is to be able to simply achieve this job. The Strategic Due Date Adjustment Option consists of the combination of the ghost programh on the purchasing side and the gterminal programh on the receiving side (partners.)
It performs the MRP calculations or the manufacturing number explosion in TPiCS-X to create the Order data for parts and materials. It passes the Order data to the database managed by the host program, and sends them as the gOrder datah or the gDue Date Reply Requesth to mail addresses of the partners via the host program. The terminal program at the partners automatically receives and imports them to the database for the terminal program.
The partners look at the received data to go through whether they can handle them or not, and send the due date replies back to the sender. If the partners also use (purchase) TPiCS-X, they can import them to the Customer Order data and directly execute the MRP calculations.
Going through delivery dates itself is a difficult job to do. With TPiCS-X being used, however, it prints them as the journal from the f-MRP calculations if there are any parts that canft satisfy the situation by even allocating the buffer by part in preparation for the fluctuation of customer orders. Going through a pile of workload and this journal, they send the due date replies (due date change request) back to the sender via the terminal program as required. E-mails with the due date replies are directly sent to the address of the purchasing company side, too.
The purchasing side automatically receives the due date replies sent from multiple partners one e-mail after another, and imports them to the database.
Although they could send back the gReview Requesth after looking at the screen of the host program and writing a comment in the remarks field for gunacceptable requests,h they become unable to decide whether or not they should accept the request when they have a great number of due date replies (change requests) from many partners. TPiCS will give them an answer to this, too. It imports the due date replay data to the database of TPiCS-X at any rate.
Applying these change requests from the partners, they perform the MRP calculations based on day-to-day orders from their customers.
As a result of the calculations, they will be notified as the journal if delivery date on change request is too late. If they could never get some parts, they would have to ask the customer to change the delivery date. Or, they would have to ask unreasonable things of the parts maker again if therefs room for their review. As a result of reviews and negotiations, the schedule for today is finally determined. This returns to the work explained at the beginning: creating the Order data, performing the final processing, and sending e-mails via the host program.
It is gabsolutely impossibleh to perform the MRP calculations with due date change requests from parts makers being incorporated into them in the general MRP calculations. But f-MRP of TPiCS processes them in a quite natural flow. To begin with, since the f-MRP calculation logic assumes the mixture of gsomething where the schedule must not be changedh and gsomething where schedule change is possible,h such processing as this is possible.
On the other hand, as therefs no gmagic calculation logich in the schedule with the manufacturing numbers exploded like in f-MRP, it only relies on visual check. When the host program applies due date changes to the Order Balance data and if before and after processes are reversed because of the changes, you will immediately notice that re-adjustments of the schedule will be necessary since the changes are displayed in red in the Gantt chart.
The schedule re-adjustments can be done using the drag-and-drop functionality in the Gantt chart. Many other schedules get involved in these kinds of schedule re-adjustments. Failing to notify suppliers is scary here, too, but the Strategic Due Date Adjustment Option allows you to send due date changes to all suppliers involved by e-mail with just another some button clicks when the re-adjustments are finished.
When I myself really think about gserious about the production for short-term deliveryh this time, even prepare the Strategic Due Date Adjustment Option, and organize anew a story of the actual system operation, I bewitchingly sing my own praises, re-discovering the magnificence of the logic of f-MRP (flexible-MRP) once again.
(1) A TPiCS-X user who has the Strategic Due Date Adjustment Option installed is called the host user, and a program the host user uses is called the host program.
The system the suppliers use is called the terminal program, and the host user can distribute the terminal program to the suppliers free of charge.
(2) Both sides need to have the dedicated mail addresses.
(3) The [Send] button on the screen of each system triggers the data transmission, and automatically sends the data. As for the receipt of the data, the program on each system automatically receives them from the mail server at the set time interval.
(4) The user at the terminal program, when he buys TPiCS-X, can treat the due date reply request data as the gCustomer Order data,h based on which he can perform the MRP calculations. And if he uses the Strategic Due Date Adjustment Option, he can send/receive due date reply requests and replied due dates to/from their grandchild makers.
(5) When the program on each system receives the data, it automatically sends the data of notification of the receipt to the other independently of userfs operation.
Excerpt from TPiCS report No. 68.
Since I had issued the previous report (No.67,) I tried my very best to think about ghow I can manage to solve this problem.h I finally found out my own answer and immediately incorporated the function into TPiCS-X Version 3.0.
I have been involved in the job of TPiCS for nearly 20 years, but I think that the functional addition this time is another epoch-making job in the history of TPiCS.
I am going to briefly explain the problem itself first that I was trying to solve.
Suppose you are going to use a new production control system instead of TPiCS. To use a production control system means you make purchases for parts and materials, and make production according to the instructions from the system.
In the case of making production according to the instructions from the system, you normally assume the situation where the workload by day shown in the instructions (schedule) coming from it is not leveled. For example, if somebody told you to make 10,000 pieces for today and 100 pieces for tomorrow, you would be working all night today but would have to take a half day off for the afternoon tomorrow. This sounds horrible for those who actually make production in the factory. They would ask you, gWhat are we going to do?h
In order to solve this problem, to gmake the schedule that is leveledh or to gmake the schedule where allocation of the resources such as manpower and equipment capacity is consideredh will be needed.
But it is a very difficult job.
(1) Since therefs the Automatic Leveling Option in TPiCS-X, you can make the schedule that is leveled to some extent. But it is not something perfect.
(2) Therefs a function that coordinates with the scheduler in TPiCS-X. But the operation of the scheduler seems quite difficult in reality.
(3) As the most basic solution, it is the manual labor that makes the production schedule leveled. Since TPiCS-X allows you to move a day to schedule on forward and backward with the drag-and-drop also on the scheduling screen, operability is considerably good. Yet, when the process gets deeper and there are common processes involved, itfs going to be a tremendous work.
(4) Next solution I can think of is gleave it to the factoryh in my previous report. The system is operated with the order issue period a little longer to ghave them adjust it at the discretion of the factory,h but harmful influence by it was a great deal.
(5) The last solution of gmake production with the schedule that is different from the systemfsh is also found in my previous report. The system is only used for arrangements for parts, and the instructions to the factory are made according to another schedule. The case of my previous report was an example of weird results, which were derived from the misuse of the functions of the actual results corresponding to the plan and the actual results corresponding to inventory in TPiCS-X.
At any rate, that was difficult to solve and a down-to-earth bottleneck for running the production control systems.
Are there any ways to solve it? Is there a practical way with a simple logic? Is there a simple solution that doesnft have to score 100 points but that scores 80? As no answers come out by the way anybody can think of, it has to be a completely different idea. When trying my very best to think about it, a hint suddenly springs to my mind.
To begin with, let me explain the points of my solution.
The factory floor is living. It is very difficult to bring momentarily changing situations in the system, and to calculate in a real-time manner and give instructions of the schedule, which is leveled in each process and consistent, despite those situations.
On the contrary, all thatf required is that people at the factory floor know, gWhat and how many should we make today?h at any rate. They donft mean to tell you to decide gevery single movement of chopsticks.h They just want to know, gWhat and how many should we make today? What can we make now? If therefs anything we canft make in todayfs production, why canft we make it? When can we possibly make it, then? What is it we have to make in a hurry?h If you have only to provide the information they want, they make a judgmental decision for themselves as for the rest according to the situations they are in.
Organizing the thoughts this far, the rest is easy.
Therefs the Shop Floor Control Option in TPiCS-X.
To begin with, the Shop Floor Control is to clarify the information of gWhat should we make today? What can we make now? If therefs anything we canft make in todayfs production, why canft we make it? When can we possibly make it, then? What is it we have to make in a hurry?h The only thing the current Shop Floor Control doesnft have is the gfunction to clarify the production level (quota) to be achieved today.h
Now, let me move on to the explanation on how to calculate and display the production level.
When the production schedule, which had not been leveled, has been leveled, it must show work hours of average value. If therefs any order to follow in the production, it is desirable that the scheduled days should just be shifted in that order once itfs been leveled. Considering this average value to be the production level, if you only select the work that reaches the production level and is possible to start in the priority order, they are going to a quota. If you can give the factory these instructions in a real-time fashion, it means you have accomplished the objective.
The average value of how many coming days to take varies depending on the user. Though I find it suitable to be a week cycle in the modern age, there may be some users who say, gWe want a month to be leveled due to the relationship with our union.h
So, we have made the
setting of the leveling period available in the gProduction
Level Calculation Period by
And now, Ifm going to continue with my explanation based on the operation of determining the production level in a week cycle from here on.
(1) Placing the personal computers at pivotal points of the factory, you learn to be able to use the traffic controller of TPiCS-X.
The screen of the traffic controller shows the Manufacturing Order data that have been released for each floor.
(2) You execute the calculation for the average value of the workload for the next week from the current production schedule at the office every Friday.
(3) As much production quantity as the production level is written in the mantissa field of the Work Order data upon click of the [Mantissa] button on the screen of the traffic controller at each floor.
(4) When the previous process is only finished in the middle, or when all necessary parts are not prepared, the quantity possible to produce is written in the mantissa field.
(5) In addition, when you finished your work for the morning and start working in the afternoon to click the [Mantissa] button once again, only the production quantity corresponding to the remaining of the production level is written in the mantissa field, the quantity of the work finished in the morning being subtracted.
(6) The order to fill in the mantissa field follows that of the displays, but you can freely determine it. And the traffic controller also allows you to manipulate the values of the sort keys from the following process | the function to send a message from the following process to the unfinished previous process.)
(7) If therefs anything left unfinished among the work prior to today, it is displayed in red and the work after today is displayed in blue.
(8) Besides, since this function works with the function of gIssue Managementh added in Version 3.0, the quantities possible to start can be calculated based on the quantities of parts issued to the factory floor.
Ifm going to add some more explanation regarding the calculation method of the production level.
(1) The production level is calculated from the current production schedule.
It is not calculated from the workforce enrolled but from the production schedule to be required.
If the manpower of your section is not sufficient, you ask another for the backup. If thatfs not possible, you make changes to the production schedule. That is, you shift the schedule for next week to that for the week after next week. If you canft shift forward any further, the only way is to work overtime or overnight.
(2) On the contrary, if you have a little work, the modern thinking is that gsince therefre no orders, it's a waste of money to make many useless things.h Then, you could go to other sections for backup or do maintenance on machine and equipment. But if you donft think you can in reality, you could make adjustments to the calculated production level to raise it. This is a job somebody must do with positive intent.
(3) The case where the production for this week is already behind schedule.
For example, if therefs delay that cannot be recovered within today as of Friday, you have to prepare a margin to recover next week in the production level.
The calculated working hours for next week are written to the Work Center master, but right next to them are displayed the working hours for this week and the remaining of those for today. The greater the remaining of working hours is, the little greater you have to make the working hours for next week.
If you understand this mechanism, you can consequently tell whether behind or ahead of schedule to occur against the original production schedule that is not leveled. You canft make errors in the leveling or delay and progress zero. The point is, itfs the balance between the permissible limit of delay and progress, and the degree of the leveling.
In addition to that, if you use the Shop Floor Control, you can get answers in completely different directions like never before to such problems as, for example, gwe want to make production in a mass if we use the same mold,h and gsince it is more efficient if it changes from red to white than if it changes from red to yellowch There are so much more functions in the Shop Floor Control, but I must stop here at the explanation of gInstructions to the factory floor and the levelingh as my explanation defocuses.
Above are the answers, which have been achieved by the Shop Floor Control in TPiCS-X Version 3.0, to the gproblems of the instructions to the factory floor and the leveling.h
Excerpt from TPiCS report No. 73.
¡ Ifm going to think about the g5Sfsh which anybody involved in the production control knows about.
I used to make a half-fool of the g5Sfs,h too, when I was young.
gWho doesnft know about organizing things? Adults like you are saying such a thing nowch
I began to understand well, however, how important those fundamental things were, and how difficult to execute them in a proper fashion as I grew old.
You can never achieve the production for short-term delivery if you are saying, gfCause somebody told me to,h and gWhat must we do such a thing for?h
It is difficult to achieve the production for short-term delivery.
Because no ideas to achieve it come out if you greluctantlyh do it.g
When I am told, gTPiCS is difficult,h I say, gIt isnft TPiCS that is difficult but to control the production in a proper fashion is. Since to achieve the production for short-term delivery is more difficult, you find TPiCS particularly difficult.h Since there are a lot of other things to explain in the training workshop held every month, however, I limit my explanation to such a story as, gIn order to achieve the production for short-term delivery, it is important to etruly intend to make customer-oriented thingsf to begin with.h
The g5Sfsh mean always organize things used at the factory floor such as materials, works in progress, jigs and tools, put them right, wipe them, keep them clean, and discipline factory workers so that they can do it of their own will. I had better think about it expanding this coverage to the data in the production control system and entry work.
I have said in the past that the production control is like the health control.
Doing proper exercise, taking well-balanced meals, moderately controlling alcohol and cigarettesc
Thus, everybody knows all about what he must do. But it is difficult to do it every day.
And the difficulty of the production control is like everybody can make a living every day with very little subjective symptom or a resigned tone of gThis is the way it goes,h no matter somewhat high his blood pressure and blood sugar level are.
If you lived in the days when everybody used to be able to do his job at a slow pace, you would have no trouble doing your job even with somewhat high cholesterol level. But when it comes to recent years when things have to be made at a high-speed, you canft do that because itfs like breaking into a run with chronic disease. This is what I say recently.
I have described in my previous reports that it is important to manage the schedule in a proper fashion (schedule management) in order to achieve the production for short-term delivery.
To do the production for short-term delivery means to change the production schedule on a daily basis.
Since that production schedule has to be feasible, it is necessary to simulate the feasibility of new schedules on a daily basis.
In order for the simulation results to be meaningful, the schedule to be simulated has to represent the reality.
All the people involved have to take action, following the production schedule in the system.
All the schedules that change on a daily basis have to be reflected in the Production Schedule data in the system.
It is natural that all the changes on the side of making schedule have to be incorporated into the schedule, but all the information on delays in delivery from parts makers have to be reflected in the Production Schedule data as well. Because the parts involved will be calculated based on the assumption that they are to be delivered according to the old schedule unless the information on delays in delivery is correctly reflected in the schedule data in the system.
People used to make a purchase for parts whose delivery date was 3 or 4 months ahead in earlier times. Even if they discovered a problem with the parts to be delivered after a parts maker received the purchase order, they could solve the problem by taking some sort of action since they had enough time until the delivery date. As the purchasing lead -time for parts gets shorter, however, it will be gtoo lateh when they receive the purchase order.
There must be the increasing case where parts makers ask for the extension of deliver as short-term delivery is accelerated. While those requests are a little, all that's required is to individually handle them on a case-by-case basis. But as they become more common, they have to be able to be processed in a proper fashion as the system.
It is the gStrategic Due Date Adjustment Optionh that has achieved it.
The background of doing it has been clear now. All the tools have been prepared here. But nothing starts unless you do it.
With that having been said, itfs not like gjust do it.h
You must steadily make rules and everybody must keep them.
You must gorganize basic data and put them righth so that everybody can understand them. You must immediately erase the data that have no longer been necessary, and keep the data registered in the server living data all the time.
This is the g5Sfsh of data.
¡ This is the conversation with a customer who visited our company the other day.
the Japanese arm of an American company. They import their main parts from
their headquarters in
He says, gWe import a
To be honest with you, this is a new case for me. He has a concept completely opposite to the concept of my gschedule management,h and he is a customer who appears to be very difficult to implement the concept of my gschedule management.h
I start my explanation with TPiCS-X as usual and end it with the importance of the schedule management, showing the Strategic Due Date Adjustment Option.
I say, gNo matter how
inaccurate their due date replies are, the problem can never be solved at all
forever if you let it be. Since the parts in question donft seem to be so many
according to the appearance of your story, I think it possible to diligently
maintain the schedule. If delivery dates for 10,000 parts were not reliable
because they all kept changing, there would be no way to solve it. But if only
a few parts have the problem, it can be worked out. When you receive due date
replies and change notifications of them from
Then he replies, gNo problems can be solved for sure unless you do what youfve got to do, can they?h
When the schedule keeps changing, there are two methods of thinking corresponding to it.
(1) A method of thinking of gNo need to make a schedule as it changes anyway.h
(2) A method of thinking of gConstantly follow up the schedule that keeps changing, simulate it, and positively keep making the schedule.h
TPiCSf and our method of thinking is (2).
Fast calculation speed and the operability to be able to easily process are needed in order to achieve it.
I think itfs hard for us to develop the system and so are the users to do the job with this method of thinking. They donft do the job in a sluggish fashion but do it in a proper fashion. Thatfs why they need the g5Sfsh for the system.