Determination of key drivers regarding to costs of technical products.
Dvorak, Josef ; Krotak, Stanislav
Abstract: For Design Engineers and Management it is necessary to
decide which variant of a designed technical product should be utilized.
For development of technical products it is thus also necessary to
determine which parameters--"key drivers" of the product have
the highest impact on e.g. product costs. In this paper we would like to
present our approach how to determine crucial drivers of designed
technical products on example--pulley for circular belt--by using
statistic function--correlation coefficient
Key words: costs, correlation, driver, relations, product,
prediction, parameter, dependability
1. INTRODUCTION
Knowledge Management is a key condition of success for Design
Enginnering and Management. Each designer should notify that each his
proposal of dimension or even radius causes costs. So it is advatageous
to know which drivers have the most important impact on costs of the
designed product (Hundal, 1997). As it was mentioned the aim of our
approach is to find drivers of technical product regarding to the costs.
Statistic function called correlation coefient is being used for those
drivers indentification (Hosnedl & Nemec, 2002).
2. THEORETICAL BACKGROUND
Correlation refers to any of a broad class of statistical
relationships involving dependence. Correlation means mutual relation
between two processes or values. If one of them changes the second one
changes and the other way around. If correlation between two processes
or parameters is found out, it is probable that they depend to each
other but it does not mean that one driver of them is cause and that
second one is consequence. But correlation itsef is not able to decide
it.
Relation between two characteristics or quantities (x, y) could be
positive if (approximately) y = kx or negative (y = -kx). A value of
correlation coeficient r = -1 means quite indirect dependence
(anticorrelation). In other words as much as values of one
characteristic increase so much values of second characteristic reduce.
If value of correlation coeficient r equals +1 (Fig. 1) means direct
dependence.
For example if diameter increases then for example cost increases.
If value of correlation coeficint r equals +1 (Fig. 1) the relationship
between two characteristics is linear.
[FIGURE 1 OMITTED]
3. CASE STUDY
To prove that our hypothesis that key drivers can be determined by
using statistic function correlation several technical products were
considered (Hosnedl et al., 2009). For this case study pulley for
circular belts (Fig. 2) were chosen. Technical data about pulley
obtained from catalogue were used (Misumi, 2011). Each parameter from
product specification was tested and the relationships were expressed by
graphs. Software MS Excel for investigation was used.
[FIGURE 2 OMITTED]
4. RESULTS
Results of correlation between costs and individual parameters are
depicted in a form of graphs (Fig. 3-Fig. 11).
[FIGURE 3 OMITTED]
[FIGURE 4 OMITTED]
[FIGURE 5 OMITTED]
[FIGURE 6 OMITTED]
[FIGURE 7 OMITTED]
[FIGURE 8 OMITTED]
[FIGURE 9 OMITTED]
[FIGURE 10 OMITTED]
[FIGURE 11 OMITTED]
By using correlation coefient function we received values of
coefficient r for individual parameters (Tab. 1).
5. CONCLUSION
Our aim is to predict the costs og designed product accurately a s
possible. We obtained the most important property drivers regarding to
costs from the case study. The most important drivers are: Pitch Diametr
PD, Diameter D and diameter [D.sub.1]. The rest of other investigated
parameters behave with positive dependability too but their influence on
costs of pulley is not strong as influence of parameters PD, D and
[D.sub.1].
6. ACKNOWLEDGEMENTS
This paper includes results from Project SGS-2010-049 Complex
support of design engineering of technical products to improve their
properties and competitiveness subsidised by the Czech Ministry of
Education. We want to thank to Prof. Ing. Stanislav Hosnedl, CSc. for
his highly-valued advises.
7. REFERENCES
Hundal, M. S." Systematical Mechanical Designing: A Cost and
Management Perspective. New York: ASME Press, 1997, ISBN 0-7918-0042-3
Hosnedl, S., Nemec, L.: Estimation of a Product Property based on
Similarity of its Parameters, 13. Symposium Design for X, Meerakamm, H.
(Ed.), Neukirchen: Univ. Erlangen-Nurnberg, 10.-11.10. 2002. s 77-84,
ISBN 3-9808539-0-X
Hosnedl, S., Sova, L., Drexler, T., Vachova, J.: Algoritmy pro
predikci hodnoty vybrane charakteristiky vlastnosti technickeho vzorku
na bazi "Case based reasoning". Plzen: ZCU FST KKS, December
2009, 44 s
*** (2011) www.pedf.cuni.cz/kpsp/skalouda/korelace.doc--Correlation--Teaching Kit, Accessed on: 2011-08-10
*** (2011) http://cz.misumi-ec.com/eu--MISUMI catalogue, Accessed
on: 2011-08-10
Tab. 1 Correlation coefficient r of individual parameters
PD R D P dh d D1 T t
0.95 0.72 0.69 0.62 0.59 0.95 0.95 0.55 0.57