on-screen real-time cost estimating

19
This article was downloaded by: [University of Auckland Library] On: 16 November 2014, At: 14:02 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK International Journal of Production Research Publication details, including instructions for authors and subscription information: http://www.tandfonline.com/loi/tprs20 On-screen real-time cost estimating L. B. Newnes a , A. R. Mileham a & H. Hosseini-Nasab a a Department of Mechanical Engineering , University of Bath , Claverton Down, Bath BA2 7AY, UK Published online: 20 Feb 2007. To cite this article: L. B. Newnes , A. R. Mileham & H. Hosseini-Nasab (2007) On-screen real- time cost estimating, International Journal of Production Research, 45:7, 1577-1594, DOI: 10.1080/00207540600942391 To link to this article: http://dx.doi.org/10.1080/00207540600942391 PLEASE SCROLL DOWN FOR ARTICLE Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and should be independently verified with primary sources of information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of the Content. This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://www.tandfonline.com/page/terms- and-conditions

Upload: h

Post on 17-Mar-2017

216 views

Category:

Documents


3 download

TRANSCRIPT

Page 1: On-screen real-time cost estimating

This article was downloaded by: [University of Auckland Library]On: 16 November 2014, At: 14:02Publisher: Taylor & FrancisInforma Ltd Registered in England and Wales Registered Number: 1072954 Registeredoffice: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

International Journal of ProductionResearchPublication details, including instructions for authors andsubscription information:http://www.tandfonline.com/loi/tprs20

On-screen real-time cost estimatingL. B. Newnes a , A. R. Mileham a & H. Hosseini-Nasab aa Department of Mechanical Engineering , University of Bath ,Claverton Down, Bath BA2 7AY, UKPublished online: 20 Feb 2007.

To cite this article: L. B. Newnes , A. R. Mileham & H. Hosseini-Nasab (2007) On-screen real-time cost estimating, International Journal of Production Research, 45:7, 1577-1594, DOI:10.1080/00207540600942391

To link to this article: http://dx.doi.org/10.1080/00207540600942391

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the“Content”) contained in the publications on our platform. However, Taylor & Francis,our agents, and our licensors make no representations or warranties whatsoever as tothe accuracy, completeness, or suitability for any purpose of the Content. Any opinionsand views expressed in this publication are the opinions and views of the authors,and are not the views of or endorsed by Taylor & Francis. The accuracy of the Contentshould not be relied upon and should be independently verified with primary sourcesof information. Taylor and Francis shall not be liable for any losses, actions, claims,proceedings, demands, costs, expenses, damages, and other liabilities whatsoeveror howsoever caused arising directly or indirectly in connection with, in relation to orarising out of the use of the Content.

This article may be used for research, teaching, and private study purposes. Anysubstantial or systematic reproduction, redistribution, reselling, loan, sub-licensing,systematic supply, or distribution in any form to anyone is expressly forbidden. Terms &Conditions of access and use can be found at http://www.tandfonline.com/page/terms-and-conditions

Page 2: On-screen real-time cost estimating

International Journal of Production Research,Vol. 45, No. 7, 1 April 2007, 1577–1594

On-screen real-time cost estimating

L. B. NEWNES*, A. R. MILEHAM andH. HOSSEINI-NASAB

Department of Mechanical Engineering, University of Bath,

Claverton Down, Bath BA2 7AY, UK

(Revision received July 2006)

In the early phase of design the product detail is usually limited, offering generaldetails such as overall size, basic shape and estimated volume. However, it is atthis stage that a large proportion of the avoidable costs are typically created,normally between 50 and 70% of the final cost. Due to this the ability of designersto estimate product costs at this stage with more accuracy provides competitiveadvantage and can avoid some of these in-built costs. This paper describes amethodology of an on-line automated computer system for injection mouldingsthat enables more accurate cost estimating. The approach adopted enablesdesigners to make informed choices during the design process, not only in termsof functionality, but cost implications in terms of design, materials and, in thecase of injection mouldings, complexity. The proposed approach uses a CADpackage, in this case IDEAS solid modeling, in which a special set of injectionmoulding features has been configured for product development. In the proposedapproach the designer builds up the product stage by stage (sub-part by sub-part)and, at each addition to the design, parametrics are used to estimate costs using asmall number of parameters derived from specific component features and otherconceptual design data. This is the Injection Moulding Cost estimating Program.This process will be described in detail within this paper, illustrating the methodof product build-up and how avoidable cost increases are identified easily duringthis process. The approach described provides a concurrent on-screen costestimate as the product is developed that has been validated to within 20% of theactual cost employing the processing parameters likely to be used.

Keywords: Cost estimation; On-screen; Product design

1. Introduction

Designers are mainly concerned with functionality, aesthetics, costs, quality andmanufacturability. At the early or concept stage of design the majority of the designeffort is aimed at meeting the required functionality of the product and thespecification. During this initial design phase the cost estimating requirementsimprove with the information available. In other words, moving from a designrequirement, to the specification and then a concept solution. So in the concept stagethe information available for cost estimating is poor and as the volume of informationavailable increases, the accuracy of the cost estimate improves (Corbett 1986).

*Corresponding author. Email: [email protected]

International Journal of Production Research

ISSN 0020–7543 print/ISSN 1366–588X online � 2007 Taylor & Francis

http://www.tandf.co.uk/journals

DOI: 10.1080/00207540600942391

Dow

nloa

ded

by [

Uni

vers

ity o

f A

uckl

and

Lib

rary

] at

14:

02 1

6 N

ovem

ber

2014

Page 3: On-screen real-time cost estimating

Due to this limited information and low accuracy, cost estimating during the conceptdesign stage tends to be minimized and of a low priority. Paradoxically, it is also at theconceptual design stage that the majority of the unavoidable costs are locked into theproduct and up to 50–70% of the avoidable costs are also generated (Clarke andLorenzoni 1979, Mileham 1993, Currie et al. 1992, 1993).

Once a product reaches the detailed design stage and the design is found to betoo expensive, the course of action many companies take is to try to reducemanufacturing costs and/or use different materials, etc., rather than to redesign outthe avoidable costs. This requires all the effort for cost saving being concentrated ononly 30% of the cost and limits the opportunity. This tends to lead to themanufacturing stages being squeezed and can lead to a reduction in product quality,which, in the long term, increases cost.

The costing information available at the conceptual design stage usually takes theform of historic product costs. The usefulness of this information depends on thedegree of similarity between the old and new products and the time gap betweenthe designs. If no similar historic costs are available, then either an ‘experiencedguesstimate’ can be made or, as is more likely, the cost estimate is deferred until moreinformation is made available, towards the design finalization stage.

With the current drive within companies for concurrent engineering and designingof products in parallel and across a global environment there is a need to design right-first-time in all factors of the product, including cost. To ensure this is effective andthat designers adopt/use appropriate tools there is a need to provide the designer witha simple, accurate method of estimating product costs, on-screen, as the designproceeds during the conceptual stage of design. This approach will assist the designerto make informed choices on design alternatives, incremental improvements and theprocesses/materials to be used. It can also assist in the integration of the supply chainwith the selection with information on materials, processes and complexity reflectingsupplier competencies for the particular design/product.

Cost estimating methods have been investigated by many researchers rangingfrom estimating designers’ time (Bashir and Thomson 2001a, b) as well asinvestigating the most appropriate methodology to use. Chin and Wong (1996)describe a knowledge-based cost estimating approach that was used in a domesticappliance company to assist mould cost estimation at the concept design stage thatused a decision table approach. Decision tables were also used by Qiang et al. (1991)in a package that advised on polymer material selection and Design for Manufacture(DFM) in mould design. Several researchers (Shirley 1998, Jahan-Shahi et al. 1999,Musilek et al. 2000) have used a fuzzy logic approach for cost estimating with goodeffect. Feature-based cost estimating systems for castings have been developed byboth Poli et al. (1990) and Bidanda et al. (1998), the latter incorporating a high levelof intelligence. Researchers have also used a neural networks approach to provide aself-learning environment for cost estimating (Zhang and Fuh 1998). Others havetaken an expert systems approach for early stage cost estimating, for exampleNagarajan et al. (1996).

Other examples of costing research include work by Curran et al. (2002), wherethe use of cost modelling within the conceptual phase of aerospace usescategorization methods such as manufacturing processes, weight of material andvolume. The use of key attributes is one area that is relevant across many sectors ofcost prediction, particularly where these can be validated using historical and current

1578 L. B. Newnes et al.

Dow

nloa

ded

by [

Uni

vers

ity o

f A

uckl

and

Lib

rary

] at

14:

02 1

6 N

ovem

ber

2014

Page 4: On-screen real-time cost estimating

product information. There has also been much research into the issue of top-down(Pugh 1992, 2004) approaches to cost modelling versus bottom-up approaches. Workby Roy et al. (2001) also illustrates that the use of Cost Estimating Relationships(CER) should also focus on design time/thinking time as well as the geometry andspecific CER of the product and uses qualitative and quantitative costing inareas such as software. In addition to this and using a parametric approach,Scanlan (2002) has investigated statistical techniques and parametrics for aircraftdesign at the concept stage.

In line with this demand there is a range of commercial packages on the market,including SEER-DFM, Golorath and PRICE-H. These provide useful tools for costestimating and use historical data to assist in the decision-making process.In particular, they provide accurate costing data for frequent/similar products atthe detailed design level. They include the hardware costs but provide little detail onthe non-recurring costs and through-life costs. In general, these tools are effective forrepeated products that build on a base product model with incremental design changeswhere much of the detail is available. Although they still offer advantages, for thisparticular research there is a need to establish a set of rules and requirements to fullyunderstand the complexity involved with such products in terms of unit productioncosts, through-life costs and non-recurring costs and how these can be represented insuch a way that the rules could be incorporated within a suitable environment. Inparticular, an environment that enables the designer to use cost as one of the designoptimization criteria in parallel with functionality and aesthetics should be developed.

For this paper a detailed description of how to use the cost estimatingmethodology proposed within this research is described. It builds on the generaldescription of the work described by Mileham et al. (2005) and Hosseini-Nasab(2003). Here the authors describe how their proposed costing system had beendeveloped in order to link parametric cost estimating equations to a materialsprocessing database and a design package. These three elements were linked by anexpert computer program that converted the product concept information extractedfrom a solid model and the various material parameters extracted from thematerials database into an on-screen, concurrent cost estimate. Figure 1 shows

Injectionmoulding

costestimationprogram(IMCEP)

Parametricequations

Materials processing database

Design information

Concurrentcost

estimate

Figure 1. Diagram of the IMCEP costing system.

On-screen real-time cost estimating 1579

Dow

nloa

ded

by [

Uni

vers

ity o

f A

uckl

and

Lib

rary

] at

14:

02 1

6 N

ovem

ber

2014

Page 5: On-screen real-time cost estimating

diagrammatically the methodology that was used. The basic methodology uses anapproach where the product is built up, sub-part by sub-part. This paper illustrateshow this operates by a detailed build up of a product.

2. Example of the injection moulding cost estimating methodology

To illustrate the use of the proposed cost estimating methodology the followingsections will show how a product is built up within a CAD system. For this particularmethodology the product needs to be designed using basic features. In this particularcase a total of 17 features (sub-part shapes) were identified from studying currentinjection moulding components. These shapes consist of, for example, a cylinder,hollow cylinder, rectangular plate, etc. For the proposed product shown in figure 2 itcan be constructed using four basic features, namely the rectangular plate, thecylinder, the hollow cylinder, and the wedge. In this example the rectangular plate isused 12 times, the cylinder seven times, the hollow cylinder six times and the wedgetwice. From this a total of 27 developed sub-parts are required to build up the part,as shown in figure 3.

During the build up of this part the user is given a real-time on-screen costestimate. This enables the designer to ascertain what impact a particular sub-part ofthe product will have on the cost. To enable this cost estimate to be undertakenvarious types of information are required.

3. Information required for the cost estimate methodology

The cost estimation methodology needs seven key areas of information to estimatethe product cost. However, most of this information can be obtained as the productis being developed during the design stage and any calculations that are required are

Figure 2. The final injection moulded product.

1580 L. B. Newnes et al.

Dow

nloa

ded

by [

Uni

vers

ity o

f A

uckl

and

Lib

rary

] at

14:

02 1

6 N

ovem

ber

2014

Page 6: On-screen real-time cost estimating

done automatically. Each of the steps below is discussed in relation to the sampleproduct and allows a cost estimation to be made progressively as the design isdeveloped and displayed on-screen for use by the designer.

1. Production volume.2. Projected area for the product.3. Material specification and price per kg.4. Product volume and weight.5. Maximum wall thickness of the product.6. Cycle time, which includes the dry cycle time, the injection time, the cooling

time and any additional time required to produce a tight tolerance, a glossfinish, undercuts, internal threads and/or cross holes.

7. Number of impressions.

For the research described in this paper it is assumed that the following decisionshave already been made with respect to the product and entered into the Injection

1

2

3

45

6

9

8

15

16

17

19

21

22

23

24

25

26

27

18

7

10

13

11

Component shapes 1-27required to build up theproduct within the CADsystem

12

Figure 3. The sub-parts of the injection moulded product required to build up the product.

On-screen real-time cost estimating 1581

Dow

nloa

ded

by [

Uni

vers

ity o

f A

uckl

and

Lib

rary

] at

14:

02 1

6 N

ovem

ber

2014

Page 7: On-screen real-time cost estimating

Moulding Cost Estimating Program (IMCEP) (see figure 1):

. material: acrylonitrile butadiene styrene (ABS);

. production volume: 20 000/year.

From the information given and the details in 1 to 7 the cost estimate can be achievedby the following calculations.

3.1 Projected area

To develop the product it is necessary that one of the 27 sub-parts being consideredbe used as the first sub-part. Others are joined progressively. As a general rule andfor this test product, sub-part 1 (figure 4) was chosen as the first feature, because it isthe largest sub-part. The feature is selected from the Ideas menu, and dimensioned.The dimensions and volume of this sub-part are then transferred from the Ideasdesign software into IMCEP using the properties icon ‘Ixx’ and the information icon‘i’. IMCEP calculates the projected area of the sub-part and as a default this isassumed to be derived from the largest of the sub-part’s surfaces. Therefore, for thissub-part:

projected area ¼ greater of ðL�WÞ or ðL�DÞ or ðW�DÞ

¼ greater of ð43� 63Þ or ð43� 2Þ or ð63� 2Þ

¼ 43� 63

¼ 2709mm2:

Often, up to 20% is added to account for runners, etc.

3.2 Material information

As the ABS is assumed to be the material for the product, the following informationcan be obtained from the material database which is integrated with IMCEP:

. material price: £10.00/kg;

. density: 0.00104 g/mm3;

Figure 4. The first sub-part of the injection moulded product.

1582 L. B. Newnes et al.

Dow

nloa

ded

by [

Uni

vers

ity o

f A

uckl

and

Lib

rary

] at

14:

02 1

6 N

ovem

ber

2014

Page 8: On-screen real-time cost estimating

. scrap rate: 5%;

. melt temperature: 238�C;

. mould temperature: 45�C;

. de-mould temperature: 94�C.

3.3 Volume and weight calculation

The volume of the sub-part can be transferred directly from Ideas into IMCEP sothat the sub-part’s weight can be calculated:

sub-part volume ¼ 5418mm3:

Therefore,

weight ¼ 5418 � 0:00104 ¼ 5:635 g:

3.4 Maximum wall thickness calculation

In this particular research to calculate the maximum wall thickness, as a general rulethe maximum wall thickness is the minimum of the XYZ dimensions. Therefore, forthe first sub-part it can be calculated as follows:

sub-part wall thickness¼ minðlength, width, extrude distanceÞ,

sub-part wall thickness¼ minð63, 43, 2Þ,

product wall thickness¼ 2mm:

3.5 Cycle time calculation

The cycle time for injection moulded products can be calculated using the followingequation:

cycle time ¼ dry cycle timeþ injection timeþ cooling timeþ complexity additions:

To calculate the overall cycle time, each of the individual times can be calculatedbased on the component’s weight, type of machine, thickness, etc. The calculation ofeach of these is described in the following sections.

3.5.1 Dry cycle time calculation. The dry cycle time is mainly dependent on thecomponent weight and based on the information supplied by the manufactures of theinjection moulding machines to be used (Currie 1996). To estimate the dry cycletimes the weight offers a clear guide and using this for a machine that would typicallybe used for such a product the times can be found. These are summarized in table 1for this particular study.

3.5.2 Injection time. The injection time (i.e. time to inject the material into themould) can also be calculated using the weight of the product and the machinedata. The machine data in this case are the injection moulding machine plastisizing

On-screen real-time cost estimating 1583

Dow

nloa

ded

by [

Uni

vers

ity o

f A

uckl

and

Lib

rary

] at

14:

02 1

6 N

ovem

ber

2014

Page 9: On-screen real-time cost estimating

rate data. Previous research by Currie (1996) showed that, for this type of productand machine, only four weight ranges were required for calculating the dry cycletime. These results are shown in table 2.

In this research the proposed cost estimating system focuses on injectionmoulding components. These weight ranges, dry cycle times and injection times havebeen built into IMCEP to enable these times to be identified automatically. Hence inthe case of sub-part 1, the rectangular plate, the weight is 5.3602 g, hence the drycycle time will equal 0.5 s and the injection time 1 s.

3.5.3 Cooling time. As well as the product volume and weight the productcooling time is another of the major cost factors for injection moulded products.There are several different formulae for the calculation of the cooling times forinjection moulded products (Bernhardt 1960, Fenner 1979, Linder 1985). Themajority of these are based upon the general principle that cooling time increasesin relationship to almost the square of the maximum wall thickness. This showsthe importance of wall thickness to the cost of injection moulded products.Equation (1) was suggested by Linder (1985) for the calculation of the coolingtime and for this research was selected as being the most appropriate due to itsease of application:

cooling time ðsÞ ¼t2

2�X� ln

4�ðTm � TwÞ

ðTe � TwÞ

� �, ð1Þ

where Tm is the melt temperature (�C), Tw is the mould temperature (�C), Te isthe de-mould temperature (�C), X is the thermal diffusivity (mm2/s) and T is themaximum wall thickness (mm).

All of the material properties required for this calculation are held in thematerial database. The cost estimating system calculates the maximum wallthickness automatically and then using this and the information fromthe database the IMCEP also calculates the cooling time automatically.

Table 1. Dry cycle time based on weight.

Product weight (g) Dry cycle time (s)

0–60 0.561–200 0.75201–300 1.25301–500 2.30501–1000 3.30

Table 2. Injection time calculation based on weight.

Product weight (g) Injection time (s)

0–60 1.061–200 2.0201–500 2.5501–1000 5.0

1584 L. B. Newnes et al.

Dow

nloa

ded

by [

Uni

vers

ity o

f A

uckl

and

Lib

rary

] at

14:

02 1

6 N

ovem

ber

2014

Page 10: On-screen real-time cost estimating

For sub-part 1 this equals

cooling time ðsÞ ¼2 � 2

2 � 0:128 � �

� �ln

�=4ð238� 45Þ

94� 45

� �¼ 5:62 s:

3.5.4 Complexity. In addition to the geometric features of the product, tighttolerances, a gloss finish, the presence of undercuts, internal threads and/or cross-holes within an injection moulded product, are the main factors for increasing themoulding time or complexity. For example, for a tolerance of less than 0.05mm, alow de-moulding temperature should be used. Two de-mould temperatures thereforeneed to be contained within the material database, the upper de-mould temperatureand the lower de-mould temperature. If a tolerance of less than 0.05mm is specified,then the lower de-mould temperature should be used for calculating the cycle time,otherwise the upper de-mould temperature should be used. For this research, allcomplexity additions are entered manually and then the IMCEP program adds theseinto the overall cost estimate as a complexity factor. This factor has been calculatedbased on the basic findings from Currie (1996). These have also been enhanced toprovide greater detail to take account of certain product requirements. For example,to achieve a gloss finish a higher melt temperature of 150�C typically needs to beused (Belofsky 1995). Undercuts, internal threads and cross-holes usually increasethe mould opening time as extra pins, etc. have to be extracted and therefore thesehave an effect on the dry cycle time. To compensate for such complexity factors thefollowing rules were used.

(a) If either a cross-hole or an undercut is present, the cycle time is increased by 3 s.(b) If an internal thread is present, the cycle time is increased by 4 s.(c) The total increase in the dry cycle time has been limited to 4 s as the pin or

female screw extraction happens concurrently within the mould.

So using the first sub-part, the rectangular plate, the complexity additions arecalculated as follows:

. tolerance less than 0.05¼No;

. gloss finish¼No;

. undercuts¼No;

. internal threads¼No;

. cross-hole¼No.

Therefore, for the first sub-part the complexity additions equal zero and the de-mould temperature remains at 94. Hence, to calculate the overall cycle time thevarious elements discussed above need to be aggregated, i.e.

. dry cycle time: 0.5 s;

. injection time: 1.0 s;

. cooling time: 5.62 s;

. complexity additions: 0.0 s.

Cycle time ¼ dry cycle timeþ injection timeþ cooling time

þ complexity additions,

cycle time ðsÞ ¼ 0:5þ 1:0þ 5:62þ 0 ¼ 7:12 s:

On-screen real-time cost estimating 1585

Dow

nloa

ded

by [

Uni

vers

ity o

f A

uckl

and

Lib

rary

] at

14:

02 1

6 N

ovem

ber

2014

Page 11: On-screen real-time cost estimating

This covers the main material and product details required to calculate the cost

estimate for the product. For injection moulding the number of impressions in the

mould also influences the cost.

3.6 Number of impressions in the mould

There are many empirical equations for the calculation of the number of impressions,

but the most popular and simple is recommended by Dym (1979), namely

number of impressions ¼yearly volume

5�

1

200� ½1=cycle time ðhÞ�

¼production volume� cycle time ðhÞ

1000: ð2Þ

Therefore, for the first sub-part

number of impressions ¼20 000� 7:11670

3 600 000¼ 0:039 � 1:

4. Cost estimation by building the product using sub-parts

As all the necessary information is now available the sub-part cost can be estimated

using cost estimating equations from this research using some of the key parameters

to give the cost estimating relationship. Four levels of estimates can be used, namely

EstCost1(weight) is a rough estimate based on one parameter only, the weight. This

can be refined by providing parameter values for EstCost2(production volume),

EstCost3(cycle time) and EstCost4(machine hourly rate). For the first sub-part this

becomes

EstCost4 ð0�30 gÞ ¼ EstCost3, ð3Þ

EstCost3 ð0�60 gÞ ¼ EstCost2 � 0:83410þC

N � 0:002061366�0:01335, ð4Þ

EstCost2 ð> 9000Þ ¼ EstCost1, ð5Þ

EstCost1 ð0�10 gÞ ¼ ðW � Pð1þ SÞÞ þ ð�0:00172Wþ 0:04217Þ, ð6Þ

where W is the component weight (g), P is the material price (£/g), S is the scrap rate

(%), V is the production volume, C is the cycle time (s), and N is the number of

impressions.Therefore,

estimated cost ¼ Estcost4 ¼ £0:077725:

The information for the first sub-part is displayed on-screen in a separate window

next to the Ideas solid model (figure 5) and is summarized in table 3. The estimated

cost is given to six decimal places because the effect of small incremental changes in

1586 L. B. Newnes et al.

Dow

nloa

ded

by [

Uni

vers

ity o

f A

uckl

and

Lib

rary

] at

14:

02 1

6 N

ovem

ber

2014

Page 12: On-screen real-time cost estimating

the product volume and wall thickness can be seen on the estimated cost during theconstruction of the product.

The process of product model development is then continued using sub-part 1and deducting the six flat cylinders (sub-parts 2 to 5) from it as shown in figure 6 andsummarized in table 4.

Each of the sub-parts numbered 8 to 27 can be a candidate for product modeldevelopment. Sub-part 8 (figure 7) was chosen as the next part to join with sub-part 1and the process of cost estimation was repeated. The information summarized intable 5 was found for sub-part 8.

To develop the product, sub-parts 1 to 7 and 8 need to be joined, as shown infigure 8. This then requires that the product volume, weight, maximum thickness,cycle time and complexity need to be recalculated. This joining is considered as afurther aspect of complexity which can affect the cycle time calculation. These effectsare again calculated automatically.

Figure 5. The assigned window for the cost estimation process.

Table 3. Information for the first step of the product cost estimation process.

StepSub-part(flat plate)

Volume(mm) Material

Weight(g)

Maximum wallthickness (mm)

Cycletime (s)

Aggregatecost (£)

1 1 5418.00 ABS 5.635 2 7.1167 0.077725

On-screen real-time cost estimating 1587

Dow

nloa

ded

by [

Uni

vers

ity o

f A

uckl

and

Lib

rary

] at

14:

02 1

6 N

ovem

ber

2014

Page 13: On-screen real-time cost estimating

Figure 6. The first sub-part after deducting six flat cylinders.

Figure 7. The next sub-part of the injection moulded product (hollow cylinder).

Table 5. Information for the next sub-part of the product.

StepSub-part

(hollow cylinder)Volume(mm) Material

Weight(g)

Maximum wallthickness (mm)

Cycletime (s)

Aggregatecost (£)

3 8 77.754 ABS 0.0809 1 2.9042 0.028396

Table 4. Information for the first step after deducting six flat cylinders.

StepSub-part(flat plate)

Volume(mm) Material

Weight(g)

Maximum wallthickness (mm)

Cycletime (s)

Aggregatecost (£)

2 1 5204.35 ABS 5.4123 2 7.1167 0.076114

1588 L. B. Newnes et al.

Dow

nloa

ded

by [

Uni

vers

ity o

f A

uckl

and

Lib

rary

] at

14:

02 1

6 N

ovem

ber

2014

Page 14: On-screen real-time cost estimating

The maximum thickness of the joined section is calculated using

R ¼ðr1 þ r2 þ 2R1Þ

2þ 4R2 r1 þ r2 þ 2R2 þ 2½ðr1 þ R2Þðr2 þ R2Þ�

1=2� �

4fr1 þ r2 þ 2R2 þ 2½ðr1 þ R2Þðr2 þ R2Þ�1=2

g: ð7Þ

The maximum wall thickness is 2R (mm) and it is assumed that, from figure 9,

R1¼ 1mm, R2¼ 0.5.

Fillet radius ¼ r1 ¼ r2 ¼ 0:2mm,

R ¼ð0:2þ 0:2þ 2 � 0:5Þ2 þ 4 � 1ð0:2þ 0:2þ 2 � 1þ 2 � ðð0:2þ 0:5Þ � ð0:2þ 1ÞÞÞ1=2Þ

4 � ð0:2þ 0:2þ 2 � 1þ 2 � ðð0:2þ 0:5 � ð0:2þ 1ÞÞ

¼ R ¼ 1:116mm,

maximum wall thickness ¼ 2R ¼ 2:23mm:

The results of the joining process are shown in table 6.

Figure 8. The joining of sub-part 1 and sub-part 8.

Figure 9. The effect of joining sub-parts 1 and 8 on the maximum wall thickness.

On-screen real-time cost estimating 1589

Dow

nloa

ded

by [

Uni

vers

ity o

f A

uckl

and

Lib

rary

] at

14:

02 1

6 N

ovem

ber

2014

Page 15: On-screen real-time cost estimating

Sub-part 9, the second hollow cylinder, was chosen as the next candidate to be

added for the product development using the general rule to select the next sub-part

with the greatest volume. This adds volume to the product but no more complexity.

It also did not increase the cycle time, as the increase in thickness caused by joining

sub-parts 1 and 8 is exactly the same as joining sub-parts 1 and 9. To develop the

product features, other sub-parts are then selected and added. The effects of each

sub-part on the product weight, wall thickness, cycle time and product cost are

shown in table 7.The cost information is made available in two different ways to the designer

on-screen as the product model is developed, namely ‘edit’ as shown in figure 10 and

‘browse’ as shown in figure 11.This information can also be displayed graphically on-screen after saving in files,

as shown in figure 12. The main benefit of the graph is that it helps designers to see

which sub-parts have a significant effect on the product cost. This then enables the

designer to make an informed decision and edit the component if required, ensuring

the design for ‘X’ also includes cost.

Table 7. Information for the product model development and the estimated cost.

Jointstep Sub-part

Volume(mm) Material

Weight(g)

Maximum wallthickness (mm)

Cycletime (s)

Sub-partcost (£)

3 10 5437.605 ABS 5.6551 2.2315 7.8407 0.0807864 11 5515.372 ABS 5.7359 2.2315 7.8407 0.0813725 12 5632.341 ABS 5.8576 2.2315 7.8407 0.0822516 13 5722.478 ABS 5.9514 2.281 8.8058 0.0835507 14 5812.478 ABS 6.0449 2.281 8.8058 0.0842098 15 5847.478 ABS 6.0814 2.281 8.8058 0.0845029 16 5882.478 ABS 6.1178 2.281 8.8058 0.08479510 17 5909.967 ABS 6.1467 2.500 10.2761 0.08804611 18 5937.456 ABS 6.1750 2.500 10.2761 0.08819212 19 5958.662 ABS 6.1970 2.500 10.2761 0.08841213 20 5969.162 ABS 6.2079 4.000 23.9668 0.11670714 21 5979.662 ABS 6.2188 4.000 23.9668 0.11678015 22 6020.162 ABS 6.2610 4.000 23.9668 0.11707316 23 6038.912 ABS 6.2805 4.000 23.9668 0.11721917 24 6057.662 ABS 6.3000 4.000 23.9668 0.11736618 25 6072.662 ABS 6.3156 4.000 23.9668 0.11751219 26 6121.662 ABS 6.3665 4.000 23.9668 0.11787820 27 6170.662 ABS 6.4175 4.000 23.9668 0.118245

Table 6. Information for joining sub-parts 1 and 7 together.

Jointstep Sub-part

Volume(mm) Material

Weight(g)

Maximum wallthickness (mm)

Cycletime (s)

Sub-partcost (£)

1 8 5282.107 ABS 5.4934 2.233 7.8407 0.079541

1590 L. B. Newnes et al.

Dow

nloa

ded

by [

Uni

vers

ity o

f A

uckl

and

Lib

rary

] at

14:

02 1

6 N

ovem

ber

2014

Page 16: On-screen real-time cost estimating

Figure 10. The Edit method of depicting the information—here the user edits the material,in this particular case by selecting ABS, as shown at the bottom left of the figure.

Figure 11. The ‘browse’ method for depicting the information—here the user browsesthrough the information, as shown at the bottom left of the figure.

On-screen real-time cost estimating 1591

Dow

nloa

ded

by [

Uni

vers

ity o

f A

uckl

and

Lib

rary

] at

14:

02 1

6 N

ovem

ber

2014

Page 17: On-screen real-time cost estimating

5. Conclusions

This paper has described a methodology for estimating the cost of injection mouldedproducts at the conceptual stage of the design process. This methodology has beendemonstrated through a computer program IMCEP. This is used to demonstratehow the cost of an injection moulded ABS component can be estimated accurately,using only the limited amount of information that is available at the conceptual stageof design. The program provides an on-line concurrent cost estimate as thecomponent is developed. It uses a special set of injection moulding features withinthe IDEAS CAD package, parametric cost equations and a set of rules that convertinformation available at the concept design stage into parameters that drive the costequations. It utilizes a methodology that is considered to be applicable tomanufacturing processes in general and is the first module of a conceptual costcomparison system. The methodology and build-up approach to part design hasbeen illustrated by the use of an exemplar part.

Acknowledgement

This work was carried out as part of the Engineering Physical Science ResearchCouncils Engineering Innovative Manufacturing Research Centre at the Universityof Bath.

Figure 12. Graph of the estimated cumulative product cost versus the product sub-part.

1592 L. B. Newnes et al.

Dow

nloa

ded

by [

Uni

vers

ity o

f A

uckl

and

Lib

rary

] at

14:

02 1

6 N

ovem

ber

2014

Page 18: On-screen real-time cost estimating

References

Bashir, H.A. and Thomson, V., Models for estimating design effort and time. Des. Stud.,2001a, 22, 141–155.

Bashir, H.A. and Thomson, V., An analogy based model for estimating design effort. Des.Stud., 2001b, 22, 157–167.

Belofsky, H., Plastics: Product Des. and Process Engineering, 1995 (Hanser/GardnerPublication Inc.: Cincinnati, OH).

Bernhardt, E.C., Processing of Thermoplastic Material, 1960 (Reinhold: New York).Bidanda, B., Kadidal, M. and Billo, R.E., Development of an intelligent castability and cost

estimation system. Int. J. Prod. Res., 1998, 36, 547–568.Chin, K.S. and Wong, T.N., Developing a knowledge-based injection mould cost estimation

system by decision tables. Int. J. Adv. Mfg, 1996, 11, 353–364.Clarke, F.D. and Lorenzoni, A.B., Applied Cost Engineering, 1979 (MDI).Corbett, J., Des. for economic manufacture. Ann. CIRP, 1986, 35, 93.Curran, R., Kundu, A.K., Raghunathan, S. and Eakin, D., Costing tools for decision

making within integrated aerospace design. J. Concurr. Engng Res., 2002, 9,327–338.

Currie, G.C., A parametric approach to cost estimating at the concept design stage. MPhil,University of Bath, 1996.

Currie, G.C., Mileham, A.R., Miles, A.W. and Bradford, D.T., Conceptual cost informationas an aid to the designer. In International Operations—Crossing Borders inManufacturing and Service, edited by Hollier and Boaden, pp. 119–124, 1992 (North-Holland: Amsterdam).

Currie, G.C., Mileham, A.R., Miles, A.W. and Bradford, D.T., A parametric approachto cost estimating at the conceptual stage of design. J. Engng Des., 1993, 4,117–125.

Dym, J.B., Injection Moulds and Moulding, 1979 (Van Nostrand Reinhold: New York).Fenner, R.T., Principle of Polymer Processing, 1979 (Macmillan: London).Hosseini-Nasab, H., The early cost estimation of injection moulded components. PhD thesis,

University of Bath, 2003.Jahan-Shahi, H., Shayan, E. and Masood, S., Cost estimation in flat plate processing using

fuzzy sets. Comput. Ind. Engng, 1999, 37, 485–488.Linder, I.E., Pre-calculation of the cooling time of injection moulding of simple design BASF.

Technical Report, Plastics Technology Centre, 1985.Mileham, A.R., Currie, G.C., Miles, A.W. and Bradford, D.T., Computer based cost

estimating for conceptual design. International Forum on Des. for Manufacture andAssembly, 1993 (Boothroyd and Dewhurst Inc.: Newport, RI).

Mileham, A.R., Newnes, L.B. and Hoesni-Nasab, H., Automatic on screen cost estimation atthe early design stage, in 12th International Conference on Concurrent Engineering,Fort Worth, USA, 25–29 July 2005, 587–592.

Musilek, P., Pedrycz, W., Succi, G. and Rerormar, M., Software cost estimation with fuzzymodels. Appl. Comput. Rev., 2000, 8, 24–29.

Nagarajan, K., Santos, D.L. and Srihari, K., A computer aided cost estimation system forBGA/DCA technology. Comput. Ind. Engng, 1996, 31, 119–122.

Poli, C., Sunderland, E. and Fredette, L., Trimming the cost of die castings. Mach. Des., 1990,5, 99–102.

Pugh, P., Working top-down: cost estimating before development begins. J Aerosp. Engng,1992, 206, 143–151.

Pugh, P.G., Concept costing for defence projects: the problem and its solution. Defence PeaceEcon., 2004, 15, 39–57.

Qiang, H., Ying, S. and Ruan, X., Knowledge-based system for problem diagnosis ofinjection moulding, in 11th International Conference on Production Research, Hefei,China, August 1991.

Roy, R., Kelvesjo, S., Forsberg, S. and Rush, C., Quantitative and qualitative cost estimatingfor engineering design. J. Engng Des., 2001, 12, 147–162.

On-screen real-time cost estimating 1593

Dow

nloa

ded

by [

Uni

vers

ity o

f A

uckl

and

Lib

rary

] at

14:

02 1

6 N

ovem

ber

2014

Page 19: On-screen real-time cost estimating

Scanlan, J., Hill, T., Marsh, R., Brit, C., Dunkley, M. and Cleeveys, P., Cost modelling foraircraft design optimization. J. Eng. Des., 2002, 13, 261–269.

Shirley, J., Fuzzy logic cost estimation method for high production volume components.Thesis, West Virginia University, 1998.

Zhang, Y.F. and Fuh, J.Y., A neural network approach for early estimation of packingproducts. Comput. Ind. Engng, 1998, 34, 433–450.

1594 L. B. Newnes et al.

Dow

nloa

ded

by [

Uni

vers

ity o

f A

uckl

and

Lib

rary

] at

14:

02 1

6 N

ovem

ber

2014