Interdependencies of development and manufacturing and its effect on the production ramp-up Abstract Number: 003-0245

Sixteenth Annual Conference of POMS, Chicago, IL April 29th to May 2nd, 2005

Jan Jürging Industrieseminar, University of Mannheim Schloss Südflügel, 68131 Mannheim E-Mail: [email protected] Tel.: +49 – 621 – 181 1750 Fax: +49 – 621 – 181 1579

Abstract Manufactures offer more often new products or varieties, which demands changes in the product or its production processes. According to that the number of production rampups a company has to cope with increases. These production ramp-ups face even more problems in the light of decreasing development times, so that more product and process changes are required during and after the ramp-up. A System Dynamics Model is developed that analyses the interdependencies between the development and manufacturing phase. The model will assume a development process following the simultaneous engineering approach and special attention will be paid to the development of required changes during development and ramp-up. The model shows that a shorter time to market results in a slower ramp-up and thus might result in a longer time to volume depending on how early production is involved in the development process and how it is managed.

Introduction The depth and width of technological knowledge undergoes an exponential growth. The resulting technological progress has a direct link to the competition between companies that are selling products in the same technological area. Schumpeter already described these interdependencies (Schumpeter, 1947): Agile companies try to push technological progress and turn the gained insights into new or improved products, in order to be more competitive and at least for a certain time earn monopoly profits. As soon as the innovation is established at the market other companies will imitate the innovation and counterbalance the competitive advantage. Regaining these advantages the search for new innovations must be commenced again. A lot of empirical surveys show that this spiral of innovation accelerated over the last 20 years and the product life cycle has been shortened (Bullinger, 1990). These implied fundamental changes for the competition: the ability to turn new technologies quickly into sellable products determines the success or failure of a company. This is true for the technological leader as well as the follower, since the follower cannot afford to loose the technological contact. The development speed becomes a competitive factor (Stalk and Hout, 1990, Rosenthal, 1992). There have been a lot of rapid-innovative Japanese companies that have advantaged the spiral of innovation, and a lot of the Japanese concepts were transferred with success to other countries.(Smith and Reinertsen, 1991). The duration from development start until the start of production is continuously reducing. The start of production (SOP) is often regarded as the end of the innovation process, which is a false assumption. The diffusion of an innovation within a company, measured until a stable and satisfactory

-2-

level of production is reached after the introduction of new products or processes, still is unchanged and too long. This time range is termed the time to volume. Time-consuming production ramp-ups have in the circumstances of an increased competition in innovations and shortened life cycles disastrous economic consequences (Bullinger and Wasserloch, 1990,): •

The market cannot be supplied with sufficient new products and the aspired position as the technological pioneer is lost, because of a competitor with shorter ramp-up times.

•

Because of lower cumulated production quantities compared to competitors with shorter ramp-ups, experience curve effects cannot be realized; the cost position becomes worse.

•

Profit contributions lost at the very beginning of a product life cycle because of lower sales cannot be compensated later when the market is in its saturation.

•

At the start of sales the cumulated cost of a development project reaches its maximum and if then the earnings are delayed because of lower production volumes especially smaller companies with a narrow product portfolio will run into liquidity troubles.

•

Releasing products late can result in 1/3 lower life-cycle-earnings. (Hendricks and Singhal, 1997)

Once a company has been trapped the financial resources decrease, which results in lower budgets for new- or variants developments. The lower budget and time pressure to release new or modified products to the market can lead to a longer production ramp up. This cycle, shown in

-3-

Figure 1, is iterated more often and quickly on behaves of steadily shortening life cycles and the rising number of variants. Production ramp up is a critical issue in a products life cycle concerning time and money and becomes an important competitive factor.

production ramp up developement efforts -

-

learning curve

competetive position

life cycle earnings

-

Figure 1: Reinforcing loop in production ramp-up The goals of productions, such as short lead times, low costs and flexibility in the processes, the requirements for its ramp up can be derived: A controlled achievement of the stable production status. The problem is obvious: “Companies can simply not afford any more to design a product, transfer it into production and debug or adapt it during a period of sometimes two years” (Dierdonck, 1990). The transfer from development into production seems crucial from a timely and an economic perspective: the product is short before its market entrance and time slacks exist no longer. Simultaneously with announcing the next product generation at least part of the customers will delay their consumption and wait for the next generation. Because of that sales will decrease and the demand for the product in ramp-up is rising and it has to be released quickly in a sufficient quantity (Inness, 1995). But how are the differences in production ramp-up determined? Already in the 1930th the ramp-up of production processes have been empirically tested and individual and collective learning processes have been identified as a reason (Wright, 1936).

-4-

Differences in ramp-up times cannot only be based on learning curves, especially in an automated production environment. Probably the reasons can be found at the differently handled transfer from development into production. Two aspects have to be considered: •

On the one hand the transfer involves cooperation of organisations from development and from production,

•

on the other is a physical transfer of development results from laboratory environments into series production.

An isolated view of the ramp-up is completely deficient. The question is where are the problems coming from? According to a study “Fast Ramp Up” undertaken in Germany in 2002 four main deficits in production ramp-up have been identified: •

There is insufficient knowledge about the interfunctional project progress

•

By now it is not possible to analyse occurred problems according to their impact on the entire project or their roots

•

Problems or disturbances are recognized only after they occurred

•

Action taken to solve problems are only based on employees experiences

Right know it is not completely possible, according to the surveyed companies, to solve problems or encounter disturbances in advance by a more intense, or sophisticated planning. Methods and tools are demanded that take proactive actions to avoid these problems, but developing these tools requires a deep understanding of the causal relationship occurring within the time to volume. The state of the art in the ramp-up scientific literature shows, that a holistic approach on the interconnected processes involved is not far enough developed. Mostly reactive approaches from the project management are chosen to solve the problem (Fischer and Dangelmaier, 2000, Kuhn, 2002, Bennedetto, 1999).

-5-

The idea that at the start of production (SOP) the buying department has all the parts at the right time and in the right quantity and quality in its place and the producer switches the machines on and full production capacity is reached is at this time too removed from reality. Complexity, dynamics and interdependencies of parallel executed processes, e.g. product development and the build up of manufacturing resources require a time consuming ramp-up management. Securing a goal oriented procedure requires an evaluation of economic connections, an evaluation of the technical complexity of a new product and identifying main reasons for disturbances. Analyzing the ramp up first the economical consequences of quantity losses has to be evaluated. Next the complexity has to be categorized in order to take proper actions in advance to minimize these losses. An overdrawn ramp-up management can over compensate the potential further earnings. The economic losses because of ramp up disruptions can be analyzed on two levels, the business volume and on the cost side. Evaluation on the business volume side considers that especially in the phase of market entrance a unique selling position can be achieved, which is rewarded by the customers with higher prices. On the cost side all costs are taken into account, which differ from an optimal ramp-up. As quantification for the ramp-up speed and the quantity losses during ramp up a ramp-up factor (Fup) can be assembled. It is defined as Fup =

produced pieces max # of produced pieces

. Theoretically a factor of 1 would be possible.

Classifying a new product development project according to its parameters influencing the required time to volume, there can be four factors identified: •

Degree of innovation (FI): on the product side it can be a new, variant or advancement development and on the process side there are the options to manufacture with existing or new production processes.

-6-

•

Complexity (FC): Out of how many parts is the product assembled and how many different manufacturing processes are involved

•

Variant (FV): How many variants are there, and if, how complex are they

•

Manufacturing resources complexity (FMRC): The ramp-up can be done on standard machines or product specific ones.

In the model all product specific factors can reach from 0 to 1 where 1 indicates the highest level of difficulty. Controlling a goal-oriented innovation process it is necessary to define the desired goal. At the end there should be a product that fits in form, function and price to the customer’s demands, but also within the company a production system that is capable of producing quantities in the quality the market is asking for. Therefore the result of the innovation process has three dimensions: •

The product dimension, that fulfills the customers demands towards the product concerning form and function. During this process the product properties are fixed,

•

the process dimension, concerns the companies ability to produce a product in the right quality,

•

and the capacity dimension, where the company provides the required manufacturing resources, which also includes services and products by suppliers.

All dimensions have different goals and perspectives towards the product and the conflicts during the innovation process are preassigned. The functional areas have different priorities in the timely progress of the innovation process.

-7-

Productions goal

viewpoint

Capacityization

developments viewpoint

Current status processrealization Productrealization

Figure 2: Dimensions of the innovation progress

Typically development sees the innovation processes path different than production members. This multidimensional status of the innovation process is a measurement for the series-production readiness and most of the troubles during the ramp-up are linked to unready products (Bungard and Hoffmann, 1995). Measuring this also gives hint when to take what action to bring the NPD-project back on track.

The model and simulation results The System Dynamics model consists out of four modules that are interconnected. All the work is done in the work-model (appendix Figure 6). A new product development project (NPD) exists out of three phases: the product development phase, process development phase, and the production ramp-up phase. Each phase has the generic structure shown in the figure and the progress in each phase determines how much work is available in the dependent phases. Consequently the phases can be simulated in a

-8-

strict consecutive order or, following the nature of simultaneous engineering, can be performed in parallel with different degrees of overlap. Performing tasks in parallel is shortening lead times on the first sight, but also more mistakes are generated because downstream phases begin work upon preliminary information. When already released tasks have to be changed in NPD project it is spoken of engineering change orders (ECO). The development of ECOs is modelled in the ECO module (appendix Figure 7). As already mentioned the project progress controlling is very important and in the series-production readiness a valuable measurement is developed. It is modelled in Figure 9. Production ramp-up as the main focus is modelled as shown in Figure 8. The following Figure 3 shows the simulations result for four different scenarios.

Produced pieces per time unit

SOP at time 150 20 15 10

3

3

5

4 1

184

4

2

34

2

234

34 1

2

1

234123412

1

2

1

2

2

2

4 0 123 1 140 162

4 4

4

3

3

3

1 1

1

206

production : NO_overlap_inno

228 250 Time 1

1

272 1

294 1

316 1

1

338 1

360 1

- production : overlap_innov 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 2 2 2 2 2 2 2 2 production production :: overlap_innov overlap 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 production : overlap production : NO_overlap 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 production : NO_overlap Figure 3: production ramp-up in four scenarios

-9-

The four scenarios differ in the factor of innovation and in the degree of overlap. Both factors have obvious a great effect on the ramp-up curve and especially the innovation factor. Overlapping the phases, which in the first place establishes communication among the phases, improves the ramp-up, but not to a satisfactory level. The improvement is due to early problem solving, which is supported by concurrent engineering and the parallel execution of tasks. Figure 4 shows the development of hidden ECO.

Total #HiddenECOs 200 150 1

100 1 1

50

1

2

2

0

2 1

0

34

3

40

4

3

80

4

2 3

120

4

1

2 34

2341234123412341234123412

160 200 240 Time (Month)

280

320

360

400

1 1 1 1 1 1 1 1 1 1 1 1 1 1 "Total #HiddenECOs" : NO_overlap_inno "Total #HiddenECOs" : NO_overlap_inno 2 2 2 2 2 2 2 "Total #HiddenECOs" : overlap_innov 2 "Total #HiddenECOs" : overlap_innov 2 2 2 2 2 3 3 3 3 3 3 3 3 "Total #HiddenECOs" : overlap "Total #HiddenECOs" : overlap 3 3 3 3 3 3 3 3 "Total #HiddenECOs" : NO_overlap4 4 4 4 4 4 4 4 4 4 4 4 4 4 "Total #HiddenECOs" : NO_overlap

Figure 4: Development of ECOs In all four scenarios not all ECO are discovered at the start of production at time 150. But clearly the overlapped ones build up a lot less undiscovered ECO and production can start more smoothly, because of less disturbances and changes made to the product or production processes. It is important to note that every hidden ECO will be revealed

-10-

as soon production has started and from this holistic viewpoint these ECO can be regarded as another quality factor for the innovation process. These changes are very important, because production depends on two main factors. Production has two main input factors, machines and workers. For the workers’ effectiveness the model incorporates a learning curve, which starts again at a slightly higher level each time an ECO changes the production process. The availability of machine hours depends on their maintenance and the set-up times. The development of these two factors also follows a learning curve. These improvements by overlapping the phases are very noticeable, so that the solution for a short ramp-up seems to be found in a maximum degree of overlap. But surprisingly simulation shows a different systems behaviour.

Production rate 15 11.25 2

3

7.5 3.75 0 12 150

1

2

3 4

1

175

1

2

3

1

2

3

3

1

2

3 12

3

4

234123412341 12341

4

4

4

4

4

200

225

250 275 300 Time (Month)

325

350

375

400

production : concurrency level of 5 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 12 production : concurrency level of 10 2 2 2 2 2 2 2 production : concurrency level of 5 23 23 23 23 23 23 2 production level level of 7 of2130 production: concurrency : concurrency 34 34 34 34 34 34 34 production : concurrency level of 1 4

4

4

4

4

4

4

Figure 5: Fixed innovation factor at different overlapping strategies The optimum is located somewhere in the middle at an overlap level of 5, on a scale where 1 means no overlap and a 10 the maximum overlap possible. This effect comes

-11-

from several causes. On one side congestion effects when all the work is executed in parallel and on the other downstream phases commence their work without with very preliminary information, which results in a lot of rework. Identifying the causes for this behaviour the analysis of the series-production readiness development gives insights. The run number 3 in Figure 5 had the most simultaneous development of the product, process and capacity dimension. In the three dimensional cube from Figure 2 it would be the innovation path on the diagonal. Deviating from this path is penalized in other runs with more ECOs and a longer time to volume and all the economic hazards coming along with that.

-12-

Literature

Bennedetto, C. A. D. (1999) Journal of Product Innovation Management, 16, 557--568. Bullinger, H.-J. (1990) IAO-Studie F&E heute - Industrielle Forschung und Entwicklung in der Bundesrepublik Deutschland, München. Bullinger, H.-J. and Wasserloch, G. (1990) Office Management, 38, 22--29. Bungard, W. and Hoffmann, K. (1995) Innovationsmanangement in der Automobilindustrie, Weinheim. Dierdonck, R. v. (1990) R&D Management, 20, 203--209. Fischer, W. and Dangelmaier, W. (2000) Produkt- und Anlagenoptimierung, Berlin Heidelberg. Hendricks, K. B. and Singhal, V. R. (1997) Management Science, 43, 324--341. Inness, J. (1995) Erfolgreicher Produktwechsel, Landsberg a. L. Kuhn, A. e. a. (2002) Praxiswissen, 54. Rosenthal, S. (1992) Effective product design and development: how to cut lead time and increase customer satisfaction, New York. Schumpeter, J. A. (1947) The Journal of Economic History, 7, 149--159. Smith, P. G. and Reinertsen, D. G. (1991) Developing products in half the time, Van Nostrand Reinhold, New York, NY. Stalk, G. and Hout, T. M. (1990) Competing against Time, New York. Wright, T. P. (1936) Journal of Aeronautical Sciences, 3, 122--128.

-13-

Appendix





RCCoord

Interation



ReceiveChanges

Coordinate

Coord process duration



Tasks to Change Workforce

NomRelTime

ReDoTasks

current Total Workload

RelPackSize

Total #tasks PS RelTime

RCCHT ACHD Approved Tasks

Tasks Compl not Checked ApprT

ChT RCCT



Tasks released RelT

ACTD



Tasks not Completed

TPS

HiddenECOs IPC

FPS %Avail Internal Concur

TTA

CT

Phase Fraction Released



CTAvail intern Fraction Avail Ext

External Process Concurrence



CTAvail

Figure 6: work module

ResolveECOs



ECOsin Change

ECOtoChange

NotChange dECO

ChangeECOs

ReleaseECOs



Undiscoverd ECOs

discoverECOs

Total #HiddenECOs PotECOs

Fv genECOs Fi FC



F MRC Concurrency Factor







Figure 7: Generation of ECOs (Engineering Change Orders)

-14-



MinTime factor input

learning disturbance

disturbance

time for last piece

learning rate Produced pieces

production

SOP

build up cap

Available Machine Hours

break downs









Figure 8: production and ramp-up module

ProductLevel increase Prod

ProgressDiff decrease Prod

increasePro

Process Level

DecreasePro





Process KnowHow

increase Cap



Capacity Level



Figure 9: Series-production readiness module

-15-

Decrease Cap