University of Leeds
School of Mechanical Engineering
MECH5680M – Advanced Finite Element Analysis Coursework
Dr Marlène Mengoni / Dr Zeike Taylor 2019-2020
Table of Contents
1 Introduction 2 1.1 OverviewofpracticalworkforMECH5680M………………… 2 1.2 Thegoldenrule ……………………………….. 2
2 The research question 3
3 Assessment 5 3.1 Learningobjectives……………………………… 5 3.2 Deliverable………………………………….. 5 3.3 Peerassessment……………………………….. 6
©University of Leeds, 2019 1
Introduction
1 Introduction
This booklet contains information about your coursework.
1.1 Overview of practical work for MECH5680M
Sessions timetabled on Monday afternoons in a cluster should be used to ask questions on all aspects of the module practical work (see module handbook for more information).
Tasks
You are expected to work individually through 5 tasks/tutorials (described in another book- let available on MINERVA). When you have completed each task, you should fill in a short test/questionnaire about it on MINERVA. Completing tasks 1 to 3 will give you the experi- ence in ABAQUS required to complete the coursework.
Coursework
In addition to the lab tasks, you are required to complete a piece of group coursework. Infor- mation about pairs for the coursework is available on MINERVA.
The coursework will be introduced in the lecture of week 2. Student pairs will be allocated at the same time.
Please check the module handbook for information about marking and deadlines.
1.2 The golden rule
It is very easy to make a mistake when you are inputting data into ABAQUS, so you need to check your results. All ABAQUS can do is to fit the numbers you give into a matrix and solve the equations using the options you have chosen, whether they are appropriate or not for the problem you are trying to solve.
You should use your Engineering Analysis skills to be ABAQUS’s brain and check your re- sults are reasonable. In this piece of coursework, you will progressively build up to a com- plex model with simpler cases to start with, to help avoid making mistakes that might not be obvious in the full model.
Rubbish in = rubbish out
2
The research question
2 The research question
A car manufacturer has had faults with their door seals found on their upper range model (Figure 1). In developing their new design, they conducted a Finite Element study that gives them more information about their product and how components can be optimised.
FIGURE 1: The door seal in situ.
In particular, they were interested in knowing how the specific rubber used in combination with their new design affects the “door closing effort” i.e. the force required to shut the door closed, with the seal compressed. They have found that the door closing effort is one of the factors of usability and quality that customers consider.
A rubber has been chosen and material testing in the form of uniaxial tension tests following ASTM D412-16 has been performed (see stress/strain results in Figure 2).
40
35
30
25
20
15
10
5
0
0 50 100 150 200 250 300 350
engineering strains (%)
FIGURE 2: Uniaxial tensile test data for Ethylene Propylene Diene Monomer (EPDM) sponge rubber.
3
engineering stress (MPa)
The research question
The new design for the seal cross-section has been developed and a drawing is shown in
Figure 3.
1 mm 1 mm
door fram
e
rubber
sea
l
y
car frame
x
FIGURE 3: Cross-section of door seal, placed in relation to door and car frames.
The main measures of door closing effort are (1) the force required to attain a sealing effect, and (2) the average distance between the door and the frame at maximum sealing capacity. The sealing effect is measured in percentage of surface area within the rubber seal that is closed.
The technical team asks you to help them justify and analyse their computational choices. The models you will build and solve for this coursework are representations of the new door seal. Uniaxial tensile data is available on MINERVA (data required to create Figure 2). The drawing of the cross-section in relation with the door and body frame (Figure 3) is also on MINERVA; it is suggested you use tools such as WebPlotDigitizer to create coordinates of points to use as support for your abaqus sketch. You are required to use both the test data and data from the drawing.
This coursework builds upon the first three ABAQUS lab tasks. You are required to undertake a more complete finite element analysis and write a structured report. Simplifications you should make are that both frames are rigid, and when a car door closes, it is displaced with respect to the car frame, in the direction of the y-axis as depicted in Figure 3.
Specific objectives and deliverables for this coursework are given below. 4
3 Assessment
3.1 Learning objectives
• Tocompletefiniteelementanalysisofatubularelementunderexternalcompression.
• To understand how finite element work can be written up for clarity and future refer-
ence.
• Tounderstandanddescribethenatureandlimitationsofmodellinghypothesis.
• To conduct and write up a computational parameters (e.g. contact algorithm choice) analysis.
• To use engineering analysis to choose which features should be included in a finite element model.
• To demonstrate understanding of the nature of outputs from finite element analysis and how they can be used.
• Touseengineeringanalysistochoosewhichoutputsshouldbeanalysedinaparticular research question.
3.2 Deliverable
You are asked to prepare a summary of your methods, results and discussion in a scientific journal paper format.
You are advised to
1. initiallycreateaverysimplifiedgeometryoftheproblem(forexampleaperfectringof
similar dimensions rather than a more elaborate sketch);
2. consider a 2D-approach (using either plane stress or plane strain elements as appro- priate for the problem);
3. create intermediate complexity models by including one non-linearity at a time to analyse the effects of computational choices.
Don’t forget that in 2D representations of a 3D problem, forces are forces per unit length.
You are required to report at least
1. an analysis on the effect of computational parameters – which can be done on a sim- plified but representative geometry and/or loading conditions (for example effect of contact algorithm or mesh controls);
2. an analysis on the effect of material model – which can be done on a simplified but representative geometry and/or loading conditions;
Assessment
5
Assessment
3. an engineering analysis of one final geometry (which can be 2D) with material and computational parameters based on the outcome of your two other analyses, and dis- cuss outcomes related to the door closing effort.
For aspects 1 and 2, you need to justify how you chose what parameters/models to analyse and how you decided which are fit for purpose. Your analysis may be based for example on theoretical concepts, convergence issues (i.e. the ability to reach a given sealing effect), or computational costs.
In many cases, the finite element method is not capable of capturing the full physical be- haviour of the real object and approximations will have to be made. You need to include in your paper a discussion on the approximations that were made in order to build a simpli- fied model. In particular, you will need to discuss what aspects of the model are unrealistic, including aspects you are required to follow (use of material test data, rigid frames, closing movement).
For this Assignment, marks will be awarded for making well thought out simplifications. It is not intended that your models are of world class standard, rather that you get to experience the decision making process and the amount of thinking required to design and build an FE model.
The paper you will submit (see assessment section in MINERVA for how to submit it) needs to follow a template available on MINERVA. The template provides both a structure to follow and some guidance on what to report in a finite element study.
You are expected to analyse values of interest in your model and present them in a concise manner.
IMPORTANT NOTE: Due to very large element distortion, you may not be able to design a Finite Element Model that can converge to 100% sealing effect. You should at least run a short mesh analysis to discuss the effect of mesh size on capacity to converge. You do not need to use mesh management methods that would enable convergence at larger contact area.
3.3 Peer assessment
If there are any issues with the contribution of individual team members, you should com- plete a peer assessment document. You will need to complete an assessment sheet per group member and attached these to your submission. It is important you join information from all group members. The sheet is available in MINERVA.
If there are no issues then you do not need to complete the peer assessment!
The peer assessment is primarily a reporting mechanism to start a discussion about how the team work has gone. The ability to make grade adjustment is somewhat restricted.
If there are serious issues, for example one group member has been absent for a substan- tial amount of the semester, please let us know about these sooner rather than later (email: [email protected]).
6
Reviews
There are no reviews yet.