Work method for 3D modeling in pro/ENGINEER

(1)

Undergraduate Thesis 15 credits

WORK METHOD FOR 3D MODELING

IN PRO/ENGINEER AT

ATLAS COPCO ROCK DRILLS AB

Stefan Kandelid

Bachelor degree in Mechanical Engineering 180 credits

Örebro May 2012

Examinator: Johan Kjellander

ARBETSMETOD FÖR 3D MODELLERING I PRO/ENGINEER

Örebro universitet Örebro University

Institutionen för naturvetenskap och teknik School of Science and Technology 701 82 Örebro SE-701 82 Örebro, Sweden

(2)

ABSTRACT

This is a 15 credits thesis in mechanical engineering performed at the PLM Solutions group at the Rocktec division within Atlas Copco Rock Drills AB, Örebro, during spring 2012.

When designers working with 3D CAD modeling uses different work methods in Pro/ENGINEER (Pro/E) it sometimes results in problems. It is also a problem when designers do not follow the specific work methods defined by Atlas Copco.

The purpose of this thesis was to identify the most common problems with 3D models at Atlas Copco Rock Drills AB (RDE) Örebro related to work methods, for example why models crash, why they cannot be checked in to Pro/Intralink or why there are unstable references.

The objective was to present a work method to avoid one or two of the most severe problems identified at RDE Örebro.

To achieve the objective I started with going through the CAD support call data base, to find out in what areas the organization needed help from the CAD support. The result shows that the engineers request most support in how to use both the modeling and the

drawing modules in Pro/E.

I also came up with a suggestion for redesign of the CAD support call data base system, that could reduce the time needed for this kind of analysis from days to minutes.

Thereafter a number of engineers, all with high skills in Pro/E, were selected for personal interviews. The topic was to identify any lack in defined work methods in Pro/E causing problems that are taking long time to correct. The result shows that the biggest issue for the users is references.

I also performed a benchmarking with two other companies within the Atlas Copco Group looking at their CAD guidelines regarding the issues found during the interviews.

Thirdly, one assembly each from six different departments were selected and sent to PTC for an in depth analysis with their software tool Expert Model Analysis. The goal was to find any systematic issues regarding work methods in Pro/E.

The analysis confirmed what the engineers earlier had brought up as the main issues, namely, references, mass/weight handling and structure in the model tree.

(3)

SAMMANFATTNING

Detta är en 15 hp examensarbete i maskinteknik som utförs på PLM Solutions Group på Rocktec divisionen inom Atlas Copco Rock Drills AB i Örebro, under våren 2012. När konstruktörerna arbetar med 3D CAD modellering använder de olika arbetsmetoder i Pro/ENGINEER (Pro/E) vilket ibland leder till problem. Det är också ett problem när konstruktörerna inte följer de arbetsmetoder som definierats av Atlas Copco.

Syftet med detta examensarbete var att identifiera de vanligaste problemen med 3D-modeller på Atlas Copco Rock Drills AB (RDE) i Örebro relaterat till arbetsmetoder, till exempel varför modellerna kraschar, varför de inte kan checkas in i Pro/Intralink eller varför det finns instabila referenser.

Målet var att presentera en arbetsmetod för att undvika ett eller två av de mest allvarliga problemen identifierade vid RDE Örebro.

För att uppnå målet började jag med att gå igenom CAD-supportens samtalsdatabas, för att ta reda på inom vilka områden användarna behövde hjälp från CAD-support. Resultatet visar att konstruktörerna behöver mest hjälp med hur man använder både modellerings- och ritningsmodulerna i Pro/E.

Jag tog också fram ett förslag till hur CAD-supportens samtalsdatabas kan göras om, som skulle kunna minska tiden som behövs för denna typ av analys från dagar till minuter. Därefter valdes ett antal ingenjörer ut, alla med hög kompetens inom Pro/E, för personliga intervjuer. Syftet var att identifiera eventuella brist i befintliga arbetsmetoder i Pro/E, vilka orsakar problem som tar lång tid att rätta till. Resultatet visar att det största problemet för användarna är referenser.

Jag utförde också en benchmarking med två andra bolag inom Atlas Copco-gruppen genom att jämföra deras CAD rekommendationer kring de problem som kom fram under de tidigare intervjuerna.

Som tredje del i examensarbetet valdes en CAD-modell från vardera sex olika avdelningar och skickades till PTC för en fördjupad analys med deras program Expert Model Analysis (XMA). Målet var att hitta systematiska problem gällande arbetsmetoder i Pro/E.

XMA-analysen bekräftade vad ingenjörerna tidigare hade fört fram som de viktigaste frågorna, nämligen referenser, massa/vikt hantering och struktur i modellen trädet.

(4)

ACKNOWLEDGEMENTS

I want to thank my supervisor at Atlas Copco Rock Drills AB; Maria Åberg, PLM Solutions team leader, and Mona-Lisa Refsbäck, CAD support engineer.

I want to thank You both for all support and guidance that You have given me. Thank you Kim Hedstrand from Atlas Copco Tools AB and Magnus Persson from Dynapac Compaction Equipment AB for sharing information about your company Pro/E work methods.

Also, I want to thank Sören Hilmerby, title in mechanical engineering, my supervisor at Örebro University, for guidance and support.

I would like to thank all the engineers that have participated in interviews and answered my follow up questions. You have all been very helpful by sharing Your valuable knowledge.

Thank You.

______________________________ Stefan Kandelid

(5)

Table of Contents ABBREVIATIONS ... 1 1 INTRODUCTION ... 2 1.1 Background ... 2 1.2 Purpose ... 3 1.3 Objectives ... 3 1.4 Delimitations ... 3

1.5 Thesis divided into three parts ... 3

1.6 ModelCHECK ... 3

2 CAD SUPPORT STATISTICS ... 4

2.1 Method ... 4

2.2 Categories ... 4

2.3 Result ... 5

2.4 Summary and improvements ... 6

3 WORK METHODS IN 3D CAD MODELING ... 8

3.1 Method ... 8

3.2 Interview result ... 8

3.2.1 References ... 9

3.2.2 Training and information ... 11

3.2.3 Skeletons... 12

3.2.4 Mass / weight ... 12

3.2.5 Structure in the model tree ... 13

3.2.6 BOM-lists ... 13

3.2.7 Layer management ... 13

3.2.8 Naming of layers, features and simp.reps ... 14

3.2.9 Coordinate system ... 14

3.3 Benchmarking ... 15

3.3.1 References ... 15

3.3.2 Others ... 16

3.3.3 Guideline user interface ... 16

3.4 Summary and improvements ... 19

4 EXPERT MODEL ANALYSIS ... 20

4.1 Method ... 20

(6)

4.3 Summary ... 26

5 CONCLUSION ... 27

6 REFERENCES ... 28

APPENDIX 1 INTERVIEW SUMMARY ... 29

(7)

1

ABBREVIATIONS

RDE: Atlas Copco Rock Drills AB

CMT: Construction and Mining Technique business area RTE: Rocktec division

SDE: Surface Drilling Equipment division URE: Underground Rock Excavation division

LHD: Loaders Haulers and Dumpers department, part of URE RBM: Raise Boring Machines department, part of URE

TME: Tunneling and Mining Equipment department, part of URE SES: Specialist Engineering Services

CAD: Computer Aided Design 3D CAD: Three dimensional CAD CSYS: Coordinate system

PDM: Product Data Management (DEMO, Intralink, Drawing Archive) PLM: Product Lifecycle Management (Teamcenter)

PTC: The company PTC - Parametric Technology Corporation. Pro/E: 3D CAD program Pro/ENGINEER Wildfire 4 from PTC

DEMO: Atlas Copco developed software for handling structures and Bill-of-Material lists

Mapkey: A short command (macro) in Pro/E

Intralink: Product Data Management (PDM) system from PTC BOM: Bill-of-material

(8)

1 INTRODUCTION

1.1 Background

Atlas Copco Rock Dills AB (RDE)

Construction and Mining Technique (CMT) business area. RDE is divided in three divisions: Rocktec (RTE)

Excavation (URE). The products are used in world.

Rocktec is the development center for CMT. Rocktec has

Rock Drills, Automation and Specialist Engineering Services (SES). SES provides consultancy to all of CMT, and is divided in

mechanics, Measurements Technique laboratory, Patents and PLM Solutions. The thesis work is done within the PLM

Figure 1: Atlas Copco Rock Drills AB; Organization

Today there are problems where the engineers have different

stability in models and assemblies There are already several

are only recommendations and are

The PLM Solutions group provides training, development, maintenance and support for the Computer Aided Design (

software at RDE.

During 2012 and 2013 RDE will replace the PDM systems for a unified Lifecycle Management (

go through an extensive reorgani Instead of being a support owner of both the new P

INTRODUCTION

Atlas Copco Rock Dills AB (RDE) in Örebro is part of the Atlas Copco Group and Mining Technique (CMT) business area. RDE is divided in three

(RTE), Surface Drilling Equipment (SDE) and Underground Rock . The products are used in tunnel and mining industries all over the

Rocktec is the development center for CMT. Rocktec has three departments: Automation and Specialist Engineering Services (SES). SES provides

CMT, and is divided into the specialist groups Applied Measurements Technique, Industrial design, Materials & rock drills

PLM Solutions.

The thesis work is done within the PLM Solutions group, see figure 1.

Rock Drills AB; Organization

there are problems where the different departments and even individual work methods in Pro/E. This leads to problems, both with s and assemblies, costing RDE many man-hours.

several work methods issued by the PLM Solutions group only recommendations and are not always followed by the engineers.

The PLM Solutions group provides training, development, maintenance and support Computer Aided Design (CAD) and Product Data Management (

During 2012 and 2013 RDE will replace the PDM systems for a unified

(PLM) solution. In that context the PLM Solutions group will go through an extensive reorganization that also gives them a different role at RDE. Instead of being a support organization, PLM Solutions group will be the application owner of both the new PLM System and the CAD software.

INTRODUCTION

2 in Örebro is part of the Atlas Copco Group

and Mining Technique (CMT) business area. RDE is divided in three and Underground Rock mining industries all over the

departments: Automation and Specialist Engineering Services (SES). SES provides

Applied , Industrial design, Materials & rock drills

Solutions group, see figure 1.

different departments and even individual This leads to problems, both with

the PLM Solutions group, but they followed by the engineers.

The PLM Solutions group provides training, development, maintenance and support a Management (PDM)

During 2012 and 2013 RDE will replace the PDM systems for a unified Product solution. In that context the PLM Solutions group will

a different role at RDE. , PLM Solutions group will be the application

(9)

INTRODUCTION

3 This gives PLM Solutions group new possibilities to set up mandatory work methods instead of recommended best practices. With that background PLM Solutions group want to investigate what the main problems with 3D modeling in Pro/E are, to see in what areas it is needed to create work methods that will avoid those problems.

1.2 Purpose

The purpose of this thesis is to identify the most common problems with 3D models at RDE Örebro related to work methods, for example why models crash, why they cannot be checked in to Pro/Intralink or why there are unstable references.

It also includes a comparison of work methods at RDE Örebro respectively at other companies within the Atlas Copco Group, to identify any differences.

1.3 Objectives

The objective for this thesis is to present a work method to avoid one or two of the most severe problems identified at RDE Örebro.

1.4 Delimitations

This thesis will be limited to the models and work methods at RDE Örebro and the work performed in Pro/E only.

The number of work methods that will be addressed will be decided after the current state analysis.

1.5 Thesis divided into three parts

The thesis have been divided into the following three parts,

- CAD support call statistics, to give a better picture of what support the different division in the company need, and to get an input to the following work method interviews.

- Work methods in 3D CAD modeling, to identify the most common problems with 3D modeling related to work methods at RDE Örebro.

There will also be a benchmarking performed with two other companies within the Atlas Copco Group looking at their CAD guidelines.

- Expert Model Analysis.

A software analysis tool for 3D models in Pro/E, regarding Geometry Quality, Design Intent Strength and Model Complexity [1].

1.6 ModelCHECK

ModelCHECK is mentioned in various parts of the thesis. This is an integrated application in Pro/E that can be run manually or automatically to control the models. It can be customized to control a variety of parameters and settings in the models and drawings as configured by the user, to meet the company standards.

(10)

CAD SUPPORT STATISTICS

4

2 CAD SUPPORT STATISTICS

2.1 Method

To get a better understanding of what problems there are with the CAD models and work methods I went through the CAD support call data base with all the support cases concerning Pro/E from 2011-04-01 to 2012-04-01.

In the data base overview, each support case is represented by one line of subject description, which then can be opened for further information.

2.2 Categories

When registering a support case, the support engineer can categorize the call with a number of different options. Of all the options available only Functionality and

Training are related to the work in Pro/E. It is unclear what the difference between Functionality and Training is, different support engineers log the cases differently. If

none of those are applicable to the case, N/A can be selected. This is followed by a short explanation, up to 5-6 words, of the case in free text under “Subject”.

Picture 1. A section of the CAD support system, choosing category.

By scanning the registered support cases and based on Mona-Lisa Refsbäck’s, CAD support engineer, experience, all the support cases in the categories mentioned above were divided into the following new categories:

- Archive/PiPlot, mainly issues with printing/plotting drawings. - BOM/balloons, “how to” use the BOM/balloon function. - Drawing, “how to” use the drawing module.

- Import/export, importing/exporting using different file formats. - Mapkeys, issues regarding the mapkey function.

- Mass/weight, “how to” set mass/weight properties. - Modeling, “how to” use the modeling module.

- References, reference issues causing check-in problems to pro/Intralink. - Shrinkwraps, issues with the shrinkwrap function.

(11)

2.3 Result

When the case files were opened, there was a large span of how well t described. Some problems were easy to categori

description were of course more difficult. Then I had only the “Subject” information to use for categorizing. This work took a bit of time since

opened.

The result after compiling the information is shown in table 1 and graph 1.

This shows that the top 5 issue types, Modeling, Drawing, System, Import/export and Archive/PiPlot, represents

represents 61%.

LHD (Loaders Haulers and Dumpers) and TME (Tunnel and Mining Equipment) are departments within the URE division.

Graph 1. CAD support cases regarding

Table 1. CAD support cases regarding Issue type LHD Modeling 66 Drawing 49 System 34 Import/export 10 Archive/PiPlot 26 Reference 16 Symbols 9 Mass/weight 12 Mapkeys 4 Shrinkwraps 2 BOM/balloons 0 Total 228

When the case files were opened, there was a large span of how well t

. Some problems were easy to categorize, but the ones without any further were of course more difficult. Then I had only the “Subject” information to

This work took a bit of time since most of the case files had to be

This shows that the top 5 issue types, Modeling, Drawing, System, Import/export and Archive/PiPlot, represents 80% of the registered cases. The top 3 issue types,

LHD (Loaders Haulers and Dumpers) and TME (Tunnel and Mining Equipment) are departments within the URE division.

Graph 1. CAD support cases regarding Pro/E.

Table 1. CAD support cases regarding Pro/E per department.

LHD TME RTE 29% 26 23% 10 21% 11 21% 18 16% 5 10% 14 15% 23 20% 15 31% 9 4% 7 6% 9 19% 18 11% 13 11% 0 0% 3 7% 9 8% 2 4% 4 4% 6 5% 4 8% 3 5% 1 1% 0 0% 2 2% 7 6% 2 4% 1 1% 1 1% 0 0% 5 0% 3 3% 1 2% 0 114 48 70

5 When the case files were opened, there was a large span of how well the problem was

e, but the ones without any further were of course more difficult. Then I had only the “Subject” information to

ase files had to be

This shows that the top 5 issue types, Modeling, Drawing, System, Import/export and ssue types,

LHD (Loaders Haulers and Dumpers) and TME (Tunnel and Mining Equipment) are

SDE Tot 16% 113 25% 20% 86 19% 13% 81 18% 26% 44 10% 4% 42 9% 6% 31 7% 4% 22 5% 3% 15 3% 1% 14 3% 7% 8 2% 0% 4 1% 460

(12)

The top 2 issue types, Modeling different functions in Pro/E.

The System issue type consists of Software installations, License lost (too few software licenses) and System not responding.

The Import/export issue type is mainly about “how to” import and export

dxf and step file formats to/from Pro/E.

Graph 2. CAD support cases regarding

2.4 Summary and i

The top three issues found in this analysis repres CAD support help desk. The issues are:

- Modeling 25 %, - Drawing 19 %, - System 18 %,

The type of information found

would like to give directed information or training within a specific area to the engineers in a specific department

It is also good for each department support or training.

To make this analysis easier and faster in the future, I would recommend that the CAD support call data base system is slightly modified.

I recommend the use of multiple

engineer to fill in when registering the cases. First you choose category at level 1,

Modeling and Drawing, are both a question of “how to” use the

different functions in Pro/E.

issue type consists of Software installations, License lost (too few software licenses) and System not responding.

issue type is mainly about “how to” import and export file formats to/from Pro/E.

Graph 2. CAD support cases regarding Pro/E, divided per department

Summary and improvements

The top three issues found in this analysis represent 61 % of the registered calls to the CAD support help desk. The issues are:

Modeling 25 %, “how to” use the modeling module. “how to” use the drawing module.

software installation, slow system, license lost. nformation found in this analysis could be very useful if

would like to give directed information or training within a specific area to the department.

department to see in what area they have the biggest

To make this analysis easier and faster in the future, I would recommend that the CAD system is slightly modified.

the use of multiple-choice boxes in at least two levels for engineer to fill in when registering the cases.

First you choose category at level 1, followed by the choice of category for level 2. CAD SUPPORT STATISTICS

6 a question of “how to” use the

issue type consists of Software installations, License lost (too few

issue type is mainly about “how to” import and export models with

epartment.

ent 61 % of the registered calls to the

software installation, slow system, license lost.

in this analysis could be very useful if PLM Solutions would like to give directed information or training within a specific area to the

the biggest need for

To make this analysis easier and faster in the future, I would recommend that the CAD

choice boxes in at least two levels for the support category for level 2.

(13)

7 In table 2, below, I have as an example listed suggestions for dividing the categories in two levels. The categories are based on what I found in my analysis.

For each support case, this should be followed by a Subject description and a free text area, the same way as it is done today.

Table 2. Suggestion for problem categories in two levels.

For each support case, this should be followed by a Subject description and a free text area, the same way as it is done today.

Level 1 Level 2

□ Modeling □ Save/check-in

□ CSYS

□ Piping

□ Simp reps / master reps

□ Layers

□ Welds, Mapkeys □ Others

□ Drawing □ Hidden layers

□ Datum plane □ BOM table □ Mapkeys □ Views □ Preliminary text □ Others □ System □ Installation □ License lost □ Not responding

□ Drawing templates lost □ Others □ Import/export □ dxf □ step □ Add mass □ Set colors □ Others

□ Archive/PiPlot □ Archive problems

□ PiPlot □ Others

□ References □ Circular references

□ External references □ Others

(14)

WORK METHODS IN 3D CAD MODELING

8

3 WORK METHODS IN 3D CAD MODELING

3.1 Method

To learn what problems the users are experiencing due to the lack of defined work methods in Pro/E, a number of skilled engineers were interviewed.

The interviewees were selected together with Mona-Lisa Refsbäck and Maria Åberg based on the engineers amount of time spent with Pro/E per month, general Pro/E skills and knowledge of Atlas Copco work methods.

From the result of the interviews, the top issue shall be compared with the work methods used at the Atlas Copco companies Dynapac Compaction Equipment AB and Atlas Copco Tools AB. Both these companies are using Pro/E today. The idea is to get new input on how to avoid the problems.

3.2 Interview result

15 engineers from different departments were selected for the interviews. LHD: 5

TME: 4 RBM: 2 RTE: 2 SDE: 2

LHD, TME and RBM (Raise Boring Machines) are departments within the URE division.

Most of the issues that were brought up can be summarized in one headline:

Pro/E the Atlas Copco way.

This is all about what work methods that are used at Atlas Copco RDE, and how to spread this information to the end users.

A complete list with all problems brought up during the interviews can be found in appendix 1. To the right on the list, shows the number of engineers commenting each issue.

Below I have listed the most common problems (all commented by more than one engineer) that were brought up during the interviews.

For most of them, there is also a recommendation on what can be done to solve the issue.

Together with my supervisor it was decided that the focus of my work for the rest of this section and the following benchmark, should be on the top issue; References.

(15)

9

3.2.1 References

The top issue was about references, circular-, external- and missing references. • Models with circular references cannot be checked-in to Intralink today, but

there are still some old models in the system that have these problems. • External references are created when a feature of one component in an

assembly reference to a feature of another component.

(This shall not be mixed up with constraints within one assembly, which is OK to use.)

Why are external references dangerous?

If a model has an external reference, it becomes dependent on another model. This means that if you modify a dimension in a model to which another one is dependent, the next time you retrieve the dependent model, it may have changed unexpectedly. [2]

Recommendation:

External references must be used with great caution. To avoid problems with external- and broken references it is generally recommended to use datum features such as planes and axis as much as possible. Do not use edges [4]. If you use the outline or a sketch from one part to model another part, make sure to remove all references back to the first part before saving. [2]

This control is implemented in modelCHECK today and set as a warning. I recommend that it is set as an error, which has to be corrected to be able to check-in the file.

• References in assemblies.

The use of constraints when building assemblies is one of the underlying root causes to the problems with broken references.

When another engineer later on is making changes to a part in an assembly, it might happen that a feature that is used as reference to another part is

modified. Then the reference is easily broken.

I found that the usage of constraints and references when building assemblies vary between the different departments and different engineers.

Some engineers make use of the strength of Pro/E's built in feasibility to use constraints when mating a part in the assembly.

Others, on the other hand, have experienced so many problems with broken references when opening large assemblies in Pro/E, that they prefer to remove all constraints and set the part as fix before saving. This way the assemblies are more stable.

(16)

10 Recommendation:

When choosing references for the constraints, make sure to choose stable references as planes, axis, surfaces or skeletons.

Do not use edges as reference. This can cause unexpected behavior, and is generally not regarded as good modeling practice.

To avoid the continuous problems with broken references it is necessary that all engineers put more effort in finding possible problems when redesigning a model. This can advantageously be done before the redesign starts.

Questions to be asked could be:

- What are the features used for, what references might have been used on this model?

- What other components can be affected when I do this redesign?

When the redesign is finished, go a few levels up in the model structure to find out if any issues have occurred.

I recommend that the whole RDE together define how to work with this issue. How many levels in the model structure shall the engineers have to control for possible effects before implementing a change? How much time has to be spent in trying to find these issues? Define the work method, and make everybody working accordingly.

Before releasing a model, an assembly or a drawing it is recommended to verify that it is OK. This is done by “refreshing” the system. Create a

completely new workspace, check-out and open the model/drawing once again and control that no problems occur with references or other issues. Otherwise there might be information left in your workspace making the model work fine, but when another user creates a new workspace, checks-out and opens the model there will be problems.

The Hydraulics is always set as fix.

See the Piping Guideline at PLM Solutions home page for further information. • Who shall repair false assemblies?

Many of the assemblies that have these problems today are rather old, and it is impossible to tell who is originally responsible for causing them. It was therefore suggested by the interviewees that a special central team was appointed to repair all false assemblies. There are probably so many false models in the system that such a team would be busy for a rather long time.

(17)

11 This is a big interference in the work for many of the engineers on a daily basis.

Recommendation:

If there is no agreement between the divisions to sponsor such a central team, my suggestion is that each division as soon as possible starts repairing the models they are responsible for.

• ModelCHECK

When I opened models during my work with this thesis I often got warnings that already must have been alerted in modelCHECK when the files were saved by the design engineers. This indicates that the engineers are not taking care of the warnings that show up in modelCHECK, resolving the problem. During the interviews I got the same indications, that some engineers don´t care about these warnings, they just ignore them and save the file.

PLM Solutions are considering making it impossible to save a file that still has warnings in modelCHECK, but for the time being, that is only set as a

warning/recommendation. Recommendation:

To get more stable models and drawings, all engineers have to take the warnings in modelCHECK more seriously and resolve the problems causing them.

3.2.2 Training and information

Next on the list was how to get in particular new users to work according to Pro/E the

Atlas Copco way. To some extent, this also applies to users already working in Pro/E

today.

There was also a question of how to spread new information in the best manner from the PLM Solutions group to the Pro/E users.

The introduction to Pro/E and recommended work methods are generally rather short for new users. They are shown by a colleague how the systems work and where to find the guidelines for Pro/E work methods, thereafter they start working on their own. The interviewees were not sure what information and guidelines that actually is available at PLM Solutions home page. If PLM Solutions could update their home page more often and include other important information and news, the interviewees thought that they would read it more frequently.

(18)

12 Recommendation:

My suggestion is that all new employees and consultants get at least a 2 to 3 days training course in Pro/E the Atlas Copco way. PLM Solutions already have a number of training courses covering most of the areas discussed in this thesis, all available for the employees and consultants to participate in.

3.2.3 Skeletons

The interviewees are lacking work method rules whether to use skeletons when modeling or not. Sometimes it could be very useful, making the models more stable. The use of skeleton can minimize the parent/child relationships between parts in an assembly. It can also allow drastic changes to the assembly without requiring a lot of redefinition. If an assembly is built using a skeleton as reference and one part, with other sub-parts assembled to it, is replaced, the assembly will still regenerate fine [3]. Recommendation:

PLM Solutions defines and documents a work method on how to create and use stable skeletons. Each division should define their work method, considering using skeletons.

3.2.4 Mass / weight

Some of the problems with mass/weight that were brought up are: - Locked mass.

Will give the wrong value if the model is changed or copied. - Wrong mass unit.

Pounds instead of kg. - Wrong density.

For example set to 1. - Wrong model density unit.

Pounds/mm3, pounds/inch3.

When mass/weight is not correct it will result in the wrong total mass for the whole model or assembly.

Recommendation:

This control is implemented in modelCHECK today and set as a warning. I

recommend that it is set as an error, which has to be corrected to be able to check-in the file.

(19)

13

3.2.5 Structure in the model tree

The interviewees were lacking a clear recommendation for how to structure the model tree.

- Structure in the model tree?

- Shall screws and other small part be placed at the end of the model tree? Shall they be in a special group?

- Shall groups be used for parts or features? Recommendation:

It is generally considered good modeling practice to place features as rounds and chamfers late in the model tree, it will make the model more stable [4].

This should be included in PLM Solutions guidelines.

3.2.6 BOM-lists

Shall the Bill-Of-Material (BOM) list be presented on the drawing or not? Today this is handled differently in the different divisions.

The governing document is always the BOM-list in DEMO, which over everything else.

Recommendation:

To avoid the risk of making mistakes, with the same information in two different places, PLM Solutions recommend to not having the BOM-lists on the drawings, only in DEMO.

The best would be if RDE could have a united approach to this question, but at least each division should define their own approach.

3.2.7 Layer management

The interviewees were unsure of how to handle layers, what is allowed to do etc. There is a Layer Routine guideline available on PLM Solutions home page. Recommendation:

The guideline could be updated with more information regarding Layer Management, clear instruction of what is OK to do etc.

(20)

14

3.2.8 Naming of layers, features and simp.reps

There is a wide variety of how to name for example layers, features, simp. reps and more. Sometimes it is rather easy to understand what the layers or features contain, sometimes the name is very cryptic and doesn't give a clue.

Since there often are someone else than the original design engineer that is making changes or updates of the models, the work would be done faster if it was more clear what for example different features are used for.

Recommendation:

Use functional naming to make the use of layers, features and simp.reps more clear. This should be included in PLM Solutions best practice document.

3.2.9 Coordinate system

The interviewees thought that the work methods for Coordinate system (CSYS) are not defined. There is no defined work method for how the parts are oriented to the CSYS. It would be easier, and the work would be done faster, when an engineer is assigned to modify a part if they were all oriented the same way.

They would like to have instructions like: - Orient the part this way, X - Y - Z. - Name the planes right - top - front.

- Always put the front to the left (or to the right).

- Model the part as it will be assembled. (Today some parts are modeled upside down.)

There was also one comment that it would be good if the different divisions could orient their rigs/machines the same way.

Recommendation:

It would be good if all divisions at RDE could agree on a common work method regarding CSYS. At least each division should define their own work method to speed up the design work.

(21)

15

3.3 Benchmarking

To get ideas on how to solve the issues brought up in the previous section, a benchmarking was conducted.

Atlas Copco Tools AB (TOOLS) and Dynapac Compaction Equipment AB (DYNAPAC) were contacted, both companies are within the Atlas Copco Group. They were kind enough to let me take part of their guidelines regarding Pro/E. Focus was on how to handle the issues with References.

3.3.1 References

TOOLS

There is a clear note about the risk to brake assembly references when making a redesign.

"If you make major changes to a part there is a big risk that the assembly references will be broken, if that happens there will be problems with not fully placed

components in the assembly." [5]

"When assembling screws, pins e.g. use constraint align to axis, do not use insert. The reason for this is that you will get a stronger relation and decrease the risk to get unplaced components." [5]

DYNAPAC

There are clear information about external references and to use stable references.

"Forbidden to have external references. Exceptional case is cables, lists and hoses. To check external references use ModelCHECK or Global ref viewer." [6]

"The references for a sketch shall be surfaces, axis and planes (not edges). Avoid having rounds and chamfers as references".[6]

Recommendation:

Include the comments from both TOOLS and DYNAPAC in the future Best practice document, "forbidden to have external references" and clearly explain the risk for broken references when making redesigns of a model.

(22)

16

3.3.2 Others

Here I have listed topics from TOOLS and DYNAPACs guidelines that correspond to the issues that were brought up during the interviews at RDE.

They should be taken into account when PLM Solutions create their best practice document.

TOOLS

- Divide up your design into clear structure.

- Which feature is the base feature? Start always with a simple base feature, this

feature will probably not be changed during the parts lifecycle.

- Which dimensions may vary? Think FUNCTION!

- The 3D model is the information carrier. Do not create "drawing" dimensions

on the drawing.

- Build up the sketch with dimensions that you wish to show on the drawing. - Orientation; instruction for how to orient the model in the CSYS.

- Assembly features must be used with great care.

- Set logical names on base features and self-defined datum-planes/curves etc. [5]

DYNAPAC

- Naming standard for layers and features.

- Only internal sketches allowed for features (brake all external references). - No buried features are allowed.

- Suppressed features are not allowed. - Avoid small entities in sketches.

- Do not sketch holes, rounds and chamfers.

- Place rounds and chamfers at the end of the model tree. - Use pattern when its more than one of a feature.

- Use function Parent-Child, or Global references viewer, to check references

before modifying, redefining or deleting features. [6]

3.3.3 Guideline user interface

When looking through the CAD guidelines from TOOLS and DYNAPAC I noticed that they both have a very user friendly web interface. It's easy to get a good overview and the most important information is well presented. For more information on a topic, there are links to more detailed instructions.

TOOLS are using a pdf-file presented in a web interface. There is a short explanation of recommended work method in each area on separate pages, followed by a

(23)

17

Picture 2. Atlas Copco Tools AB, CAD guidelines, Overview.

(24)

18 DYNAPAC are using a web interface which is very easy to overview and use.

On the left hand side there is a list with all main topics. When you click one of them, the information of that topic is shown. The list with topics are always visible to the left, this makes it easy and fast to navigate in the guidelines. See picture below.

Picture 4. Dynapac Compaction Equipment AB, CAD guidelines.

RDE Örebro are also using a web interface with good overview and a topics list to the left hand side. When clicking one of the topics, the list disappears and a complete instruction is opened describing how to use the function.

(25)

19 Recommendation:

The RDE guidelines are very thorough and covers many important topics.

To get a better overview of the information and faster navigation, I recommend to redesign the Pro/E page at PLM Solutions to look more like DYNAPAC guideline page. When clicking on a topic in the list to the left, only show the most important information at first. If more information is wanted, use a clickable link to open the complete user instruction. The topic list shall always be visible.

3.4 Summary and improvements

The result of the interviews shows that the biggest issue, due to the lack of defined work methods in Pro/E, is References.

Problems with external references and broken references are most common. When large assemblies are opened, there are relatively often parts with missing references, which the engineers commonly set as frozen or fix, instead of repairing the model. The recommendation is for PLM Solutions to focus on spreading the knowledge and best practice on how to use stable constraints and references. There is also the possibility to use skeletons more than today, to achieve more stable assemblies. There is also the question on how to handle all the models in the system today that have these problems. This is a problem that many of the engineers are struggling with every day, costing a lot of time and money.

To save time spent on this issue in the long run, and money, I recommend that each department within RDE takes this problem seriously and in a structured way starts repairing their false models.

With basis from the benchmarking with TOOLS and DYNAPAC I would recommend PLM Solutions to put together a list with best practice in Pro/E to show on their home page. The best practice should include how to handle external references, how to get

(26)

EXPERT MODELING ANALYSIS

20

4 EXPERT MODEL ANALYSIS

PTC, the supplier of the Pro/E CAD software, has developed an analysis software called Expert Model Analysis (XMA).

This software allows companies to run an in-depth analysis on their model data to spot opportunities for skills or best practice improvement. Using over 50 measures and objectives, XMA helps uncover potential issues in the 3D designs. This is a great opportunity for the companies as it allows them to detect sub-optimal areas and take corrective action [1].

4.1 Method

3200 parts, assemblies and drawings were selected and sent to PTC for analysis. The analysis was performed by Jorge Moreno at PTC in Gothenburg.

The models chosen for the analysis were from completely different products and from different departments, which are shown in the pictures below.

(27)

21 Picture 7, LHD Power Pack

(28)

22

Picture 9, RBM Picture 10, SDE Feeder arm

(29)

23

4.2 Result

One report for each department was returned from PTC, each on approximately 80 pages. Please contact Maria Åberg at PLM Solutions for further information. From the reports I have chosen 26 measures and objectives for comparison between the different departments. For each measure and department there is one column for

XMA Score and one column with % of the models with the issue in question. All

numbers and issues presented are from the assessment done by the XMA software. A comparison between the departments and these objectives can be found in appendix 2.

Below I have chosen to discuss 10 of these objective more in detail below. These objectives I consider more important or there is a significant difference between the departments.

Use absolute accuracy when appropriate

75-84 % of all models use relative accuracy. The PTC comment to this is:

Using relative accuracy allows regeneration calculations to execute quickly and use minimal memory. While relative accuracy helps performance, it makes assumptions about the size of the geometry in your model that is sometimes invalid. If the

assumptions are invalid, your models will have geom. checks and/or regeneration issues that should be fixed by switching to absolute accuracy.

XMA considers relative accuracy a risk because of the strong correlation between relative accuracy and both geometry and regeneration issues. If you do not have either geometry or regeneration issues, then using relative accuracy may be acceptable [4]. The comment from Jorge Moreno was;

There is a variation between the models from 1 to 5 decimals accuracy.

To make the system work better, his recommendation is that PLM Solutions define what decimal accuracy to be used within RDE and that the models are fixed to this.

Tune accuracy for small geometry.

This objective gives a warning when there are relative small geometries compared to the overall size of the model.

If the accuracy is not set appropriately, geom. checks could make the geometry unreliable in downstream work, as well as cause regeneration issues [4].

What stand out is that RBM and TME have about 22% of their models getting this warning, when LHD Frame only have 2% of their models with the warning. The reason for this difference needs to be further analyzed.

(30)

24

Recalculate mass properties after change (mass properties out of date).

This warning, mass properties out of date, varies from 14 % of the models for LHD Frame to 29 % for LHD Hydraulics.

The mass should be updated automatically if the settings in Pro/E are correct and all models have the same weight units. This indicates that something is wrong and need to be further investigated.

Avoid edge reference

Use planes, axis or surfaces as reference.

Do not use edges as reference. This can cause unexpected behavior, and is generally not regarded as good modeling practice [4].

SDE and RBM got the best result with 7 respectively 8 % of their models with this warning, which is still considered to be too much, while TME was worst with 23 % models with this warning.

Of course the goal must be 0 % edge reference warnings.

The best practice, of not using edge references, must be spread among the users.

Improve feature order.

For the 4 best departments, 3 to 5 % of their models are considered by XMA to have rounds and chamfers too high up in the model tree. One department stands out by getting the score 13 %.

Rounds, chamfers, and drafts should be placed late in the model tree to discourage creating child features of them.

For example, since the internal structure and nature of the round's geometry can vary significantly with small changes, referencing the round's geometry is unreliable if changes are anticipated. Chamfers and drafts are similarly complex features whose resulting geometry can vary significantly with small changes [4].

Use suppressed components appropriately

The use of compressed components is not that common. Two of the departments had 6 % of their models with this issue.

If the suppressed components or features are no longer needed, then they should be removed before you release the model. The presence of suppressed features adds confusion to your models, as well as runs the risk of having the features

unintentionally resumed (or unsuppressed) sometime in the future [4].

Set model density when appropriate

5 of the departments are handling this object well with only 0 to 2 % of their models with wrong density or length units. But one department, LHD Power Pack had 34 % of their models with some density/length issue.

(31)

25 LHD Power Pack has models with inch instead of mm as length unit, and density presented in gram, pounds and “tonne” per mm3.

I recommend that this control in modelCHECK is set as an error, which has to be corrected to be able to check-in the file.

Model-centric compared to drawing-centric definition

In model-centric design, the model is the master for capturing the design's geometry and other key design information. The drawing may still be a record of authority, but it is always associative to its models and will not override any of the design captured in the model. The purpose of the drawing is to describe the design in the model and add detail as necessary [4].

In drawing-centric design, the drawing is the master deliverable and record of authority when describing the design.

Drawing-centric design risks having inaccurate and out-of-date information in the model. While drawings are an important record of the design, it usually not possible for downstream work to be based solely on the drawing. Most manufacturing processes, for example, need to reference a 3D model. Subtle but important changes that are made in the drawing but neglected in the model can go unnoticed, causing downstream work to be invalid [4].

There is a range from 0 % drawing-centric models for SDE to 18 % for RBM, but LHD Frame stands out by having 60 % drawing-centric models.

There is a big need to look through this issue, getting the engineers to understand the importance of the correct work method.

Keep drawing views organized

According to XMA, the following drawing properties are usually unintended: - Out-of-bounds entities (entities that are outside the drawing's sheet/paper) - Overlapping views

- Draft geom not related to a view - Drawing views without a scale - Unused sheets

All drawings in this analysis were considered having this issue. (Note, no drawings from LHD Hydraulic were analyzed.)

This issue has to be addressed further, to learn if it is a real problem or not.

Model Complexity

The Model complexity XMA Score varied between 25 to 34 %.

According to Jorge Moreno at PTC this is good and indicates that there are small models, put together into larger assemblies.

(32)

26 A high, medium, or low complexity score is not in of itself a good or bad thing. A highly complex set of models may be the nature of the product design with which you are involved. Review the findings from XMA to find opportunities to simplify and/or expand your modeling approaches as the case may be [4].

Miscellaneous

When discussing the analysis result with Jorge Moreno at PTC he also gave the following comments.

• Simplified representations (simp reps).

It seems like the engineers have some lack in knowledge about what Simp reps are and how to use them.

He recommended that all engineers are brought up to the same level of knowledge and usage of Simp reps.

• Layer management.

The analysis indicates that there are no real standard that the engineers follow. He recommended that all engineers are brought up to the same level of

knowledge and usage of Layers.

4.3 Summary

I think this was a really good analysis of models from different departments at RDE, and it gave an indication on some areas that can be improved to make the work in Pro/E to run more smooth and stable.

Some areas that confirmed the issues found earlier in this thesis are; how to handle references, especially edge references, the importance of having control of the density and mass units. Also the question of how to structure the work, with for example Skeletons, has been discussed earlier.

This was a good opportunity to evaluate the XMA tool. Now PLM Solutions know what information XMA can give, if they in the future might consider to run XMA on a larger number of files.

PTC’s recommendation is to implement best practice regarding: - Use robust references, surfaces rather than edges

- Remove circular references (Note; no circular references found in the analysis) - Investigate Geom Checks

- Assign material to models, or at least specify a density - Check that the relations are calculating correctly

PTC offers further XMA scans on separate projects or on a larger amount of files. There are also possibilities for help to write best practice documents or customized workshops according to Atlas Copco´s requests.

(33)

CONCLUSION

27

5 CONCLUSION

In the first part of the thesis, Cad Support Statistics, the analysis showed that the most frequent topics when contacting the CAD support are how to use both the modeling and the drawing modules in Pro/E.

This kind of information is useful for PLM Solutions to know in which areas the users in different departments need information and training.

Since the compilation of data from the data base took rather long time, I had to open up most of the cases to be able to categorize them, I came up with a suggestion on how to improve the CAD support call data base system. This would reduce the time needed for this kind of analysis from days to minutes.

Regarding work methods in 3D CAD modeling, the interviews with skilled design engineers in all three divisions, URE, SDE and RTE, showed that the biggest issue for the users is with references. If the problems with external references and broken references were improved, a lot of time, money and frustration would be saved. PLM Solutions together with each division, should focus on improving the best practice among the users, especially handling of references. It is also recommended to give a 2 to 3 days training course in Pro/E the Atlas Copco way for all new users. I would also suggest a compressed training/information on the same topic to all current users.

The benchmark with Atlas Copco Tools AB and Dynapac Compaction Equipment AB regarding CAD guidelines showed that they are more distinct in what is best practice and what is not OK to do, also when it comes to references.

After the interview part of the thesis it were interesting to make an objective comparison, now with the software tool Expert Model Analysis from PTC.

This analysis confirmed what the engineers brought up as the main issues, namely,

(34)

REFERENCES

28

6 REFERENCES

[1] PTC University 2012, XMA Sales material.

[2] Pro/Engineer Wildfire 4.0 ModelCHECK Help Topic Collection. [3] www.ptc.com

[4] PTC Expert Modeling Analysis report for Atlas Copco Rock Drills AB. 2012. [5] CAD guidelines from Atlas Copco Tools AB. 2012.

(35)

APPENDIX 1 INTERVIEW SUMMARY

Interview summary.

Issue Comment Sum

References (external / circular / missing) 11/15

Responsibility for what YOU create See report 3

Use real references/constraints instead of Fix See report 3

Fix coordinates makes the exchange of parts smoother See report 2

Update all assemblies the part is included in See report 2

Update all shrinkwraps that the part is included in See report 1

Appoint someone to clean up faulty models See report 1

Create training course about references Shall be included in basic training Pro/E the Atlas Copco way 1

Pro/E the Atlas Copco way, rules/best practice 11/15

Real training courses for new engineers, not only guidelines Better spreading of information / rules

A few days Atlas Copco cource is recommended for all users.

Update the home page more often. 5

Work with skeletons? To be defined by each department 4

Mass/weight. Wrong density. Locked mass creates problems. The standard part has no real material set, no densit!

Can modelCHECK control density within reasonable limits? Inch? Pounds? Tonne?

The start part do have density set!

4

Structure when modeling, Structure in the model tree,

Group (parts/features) To be defined by each department

3

Show or hide BOM-table on the drawings?

PLM Solutions recommend not spresenting the BOM-list on the

drawing, same info on two places. 4

Layer management. (Show/hide BOM-table. Options.) To be defined by each department 3

Unstructured naming of parts/layers/holes/features/etc) Functional naming is recommended, guideline needed. 3

CSYS not structured. Model like this: X-Y-Z, front-back-left-right,

model as the part is mounted To be defined by each department 2

CSYS not located/rotated the same on different frames/rigs To be defined by each department 1

CSYS on the std-part is not PTC standard -- 1

Shrinkwraps. How to create good skwps, dependent/associated Shall be included in the Large Assemblies trainig course 2

Pattern makes the work faster Shall be included in basic training Pro/E the Atlas Copco way 2

Reengineer the right way. Don´t cheat. Shall be included in basic training Pro/E the Atlas Copco way 2

Use Sheetmetal application when possible. Shall be included in basic training Pro/E the Atlas Copco way 2

Make imported models lighter, make only 1 or 2 solids. Shall be included in basic training Pro/E the Atlas Copco way 1

Use "form" for easier handling of BOM-tables and balloons Shall be included in basic training Pro/E the Atlas Copco way 1

Mate to an axel created in sketch Shall be included in basic training Pro/E the Atlas Copco way 1

Use a checklist to ensure nothing is missed Shall be included in basic training Pro/E the Atlas Copco way 1

Cleaning up the WS often makes the system work better Shall be included in basic training Pro/E the Atlas Copco way 1

Imorted models crash more often -- 1

The hydraulics are often more instable -- 1

Master reps / simp reps Management Shall be included in basic training Pro/E the Atlas Copco way 1

Large assemblies, how to make them "lighter"? 2

Welding symbols not updated (too much/missing text) Contact CAD support if update needed 1

Family tables, parts not correct in outer dimensions (hydraulic

couplings) Contact CAD support if update needed 1

Special macros between Pro/E<>DEMO Not approved due to IT security 1

Structular differences between Pro/E<>DEMO -- 1

How to make the hydraulics visible in shrinkwraps?

Tubes modeled in Cabling are not visible in shrinkwraps. -- 2

Assembly cuts. It would be better to have final processing at

part-level instead, then the model can be used in further analysis. To be defined by each department 1

Create model and drawing at the same time, otherwise dims on the

drawing might be "dead". Shall be included in basic training Pro/E the Atlas Copco way 1

Update the CAD-support info more often, more people would read

it PLM Solutions will see how to improve 1

More clear whats included in the different Pro/E licenses There is a guideline at PLM Solutions home page 1

Give the sub-suppliers access to Intralink Not approved due to IT security 1

Intralink is slow Old IT-equipment. Hopefully faster with Teamcenter 5

Old version of Pro/E Update possible with Teamcenter, not with Intralink 4

License lost Too few licenses. More licenses recently installd. 2

No autimatic sync between the drawing <> workspace -- 1

(36)

L H D H y d ra u lic L H D P o w e rP a c k L H D F ra m e R B M S D E F e e d e r a rm T M E C a b in X M A S c o re % o f th e m o d e ls X M A S c o re % o f th e m o d e ls X M A S c o re % o f th e m o d e ls X M A S c o re % o f th e m o d e ls X M A S c o re % o f t h e m o d e ls X M A S c o re % o f t h e m o d e ls G e o m e tr y Q u a lit y 7 6 % 7 2 % 7 8 % 6 8 % 8 0 % 6 1 % E lim in a te G e o m C h e c ks 8 6 % 2 % 8 2 % 3 % 7 9 % 2 % 8 3 % 4 % 9 7 % 1 % 7 0 % 6 % U s e a b s o lu te a c c u ra c y w h e n a p p ro p ria te 3 0 % 8 4 % 3 1 % 7 8 % 3 0 % 8 3 % 3 0 % 8 4 % 3 1 % 7 5 % 3 0 % 8 1 % T u n e a c c u ra c y fo r s m a ll g e o m e try 7 4 % 1 1 % 6 3 % 1 5 % 9 4 % 2 % 4 7 % 2 3 % 6 8 % 1 3 % 4 7 % 2 2 % F ix s u rfa c e s w ith u n in te n d e d g a p s 1 0 0 % 1 0 0 % 1 0 0 % 1 0 0 % 1 0 0 % 1 0 0 % D e s ig n In te n t S tr e n g th 8 9 % 8 2 % 8 4 % 8 4 % 8 8 % 8 6 % R e s o lv e re tri e va l a n d re g e n e re tio n is s u e s 8 7 % 8 8 % 9 5 % 8 9 % 8 9 % 9 1 % M is s in g re fe re n c e s 7 7 % 2 % 7 2 % 4 % 1 0 0 % 8 4 % 1 % 7 8 % 2 % 9 1 % 1 % C irc u la r r e fe re n c e s 1 0 0 % 0 % 1 0 0 % 0 % 1 0 0 % 1 0 0 % 1 0 0 % 1 0 0 % F re e z in g c o m p o n e n ts 9 8 % 1 % 1 0 0 % 0 % 1 0 0 % 9 3 % 3 % 1 0 0 % 1 0 0 % D e s ig n in te n t u n c le a r 7 9 % 5 % 9 7 % 1 % 8 7 % 3 % 7 4 % 8 % 8 2 % 4 % 7 2 % 1 0 % R e c a lc u la te m a s s p ro p e rti e s a fte r c h a n g e 5 0 % 2 9 % 6 0 % 2 1 % 6 7 % 1 4 % 5 9 % 2 1 % 6 1 % 2 0 % 6 0 % 1 9 % O n ly a s s ig n p a ra m e ta rs o n c e 8 1 % 4 % 8 9 % 2 % 7 3 % 8 % 8 7 % 3 % 9 8 % 0 % 7 7 % 6 % M o d e l f o r r o b u s tn e s s 8 5 % 7 8 % 8 3 % 8 8 % 8 9 % 8 1 % A vo id e d g e re fe re n c e s 6 7 % 1 7 % 4 7 % 1 4 % 6 7 % 1 6 % 8 % 8 % 8 3 % 7 % 5 7 % 2 3 % R e m o ve q u e s tio n a b le fe a tu re s 8 6 % 6 % 8 2 % 8 % 9 5 % 2 % 8 5 % 7 % 8 6 % 6 % 9 4 % 2 % Im p ro ve fe a tu re o rd e r 8 7 % 5 % 8 7 % 5 % 7 3 % 1 3 % 9 2 % 3 % 9 1 % 3 % 7 9 % 9 % K e e p s k e tc h e s s im p le 9 6 % 2 % 8 9 % 4 % 9 8 % 1 % 9 3 % 3 % 9 9 % 0 % 8 9 % 5 % U s e s u p p re s s e d c o m p o n e n ts a p p ro p ria te ly 9 2 % 3 % 8 5 % 6 % 9 0 % 4 % 9 7 % 1 % 8 8 % 5 % 8 5 % 6 % S e t m o d e l d e n s ity w h e n a p p ro p ria te 1 0 0 % 0 % 6 3 % 3 4 % 1 0 0 % 9 7 % 2 % 9 9 % 0 % 1 0 0 % S h a rp e n d ra w in g c la rit y -6 9 % 5 1 % 6 7 % 8 6 % 7 5 % P re fe r m o d e l-c e n tri c to d ra w in g -c e n tr ic d e fin itio n -8 1 % 9 % 4 2 % 6 0 % 6 8 % 1 8 % 1 0 0 % 0 % 8 1 % 9 % K e e p d ra w in g v ie w s o rg a n iz e d -3 0 % 1 0 0 % 3 0 % 1 0 0 % 3 0 % 1 0 0 % 3 0 % 1 0 0 % 3 0 % 1 0 0 % R e m o v e u n u s e d m o d e ls fr o m d ra w in g s -7 1 % 9 % 1 0 0 % 0 % 1 0 0 % 0 % 1 0 0 % 0 % 1 0 0 % 0 % M o d e l c o m p le x ity 3 3 % 2 7 % 3 0 % 3 4 % 2 5 % 2 6 % U n its in c h /m m O K In c h /G ra m /P o u n d s /T o n n e O K O K O K O K