As I mentioned in my last post, I will be presenting a Tutorial at the upcoming SAMPE 2013 convention in Long Beach, CA.
Composites Design and Analysis is a 2-part tutorial that will span the first day of the conference. Together, Parts 1 and 2 will give a full overview of composite design and analysis from basic concepts to advanced techniques.
Part 1 - Coupling Closed-Form Analysis Techniques with Finite Element Analysis
Part 1 - taught by James Ainsworth of Collier Research Corporation (makers of HyperSizer) - will begin with an introduction to Lamintated Plate Theory and industry standard failure analyses. James will then discuss different FEA methods and approaches to structural optimization. View the Class Syllabus here.
Part 2 - Introduction to Advanced Analysis Techniques for Composite Materials and Structures
Part 2 - presented by yours truly - will then dive into more advanced composite analysis concepts. We will start with a discussion of what makes composites analysis challenging and the failure modes associated with composite materials. Some time will be spent examining progressive failure of composites - looking at damage initiation versus damage evolution and the process of determining ultimate failure of a composite structure. We will then put that knowledge to work in discussing analysis techniques. I will review some Finite Element Analysis basics as well as some advanced topics to consider. Finally, we'll put it all together with an example problem. View the Part 2 class syllabus here.
Whether you are new to composites design and analysis, or need to reinforce your skillset, consider joining us at the SAMPE Composite Design and Analysis tutorials.
Interested in attending these tutorials?
Monday, May 6, 2013
Room 103C
Part 1: 9:00 - 12:00 PM
Part 2: 2:00 - 5:00 PM
Rates (per tutorial):
- With conference registration: $175
- Tutorial only / one-day registration / exhibits only $220
- Full time student: $75
- Included with Premium Registration
See you there!
Rick
With the SAMPE 2013 Conference in Long Beach, CA approaching, I would like to invite all of our readers who are attending the conference to visit our booth (L32) and say hi to our team. We would love to discuss composites analysis, our transition to Autodesk, or anything else that comes to mind at the time!
Our lead blogger, Dan Milligan, will be giving a presentation titled, "Using Finite Element Analysis to Determine Composite Laminate Design Allowables," at 2:50 PM on May 7 in room 102C.
From 2 to 5 PM on Monday, May 6, I will be instructing a 3-hour tutorial titled, "Composites Design & Analysis - Part 2." This is the second half of a 2-part tutorial and focuses on advanced analysis of composite structures using finite element analysis. The tutorial covers a wide range of methods and considerations for simulating progressive failure of composites.
If you are active in the composites industry and have not attended a SAMPE conference, I strongly encourage you to attend as this is the premier advanced materials conference with hundreds of companies in the exhibit hall and over 300 presentations that include the latest in composite research and technology.
We hope to see you there!
The Composite Engineer's Pinboard is sort of like the ever popular Pintrest website...for composite engineers. My intention with this collection of blog posts is to provide a compact but handy reference for various composite engineering topics that I have hanging (pinned) around my own office and reference quite often. This week's topic will discuss how composite ply stresses in the laminate coordinate system are determined from ply strains in the laminate coordinate system.
Recall from the previous post on the reduced stiffness matrix ([Q]) that:
That is, composite ply stresses in the lamina coordinate system can be determined from ply strains in the lamina coordinate system using the [Q] matrix.
The Transformed Reduced Stiffness Matrix (Q-bar Matrix) allows the composite ply stresses in the laminate coordinate system to be determined from ply strains in the laminate coordinate system using the following equations:
The Q-bar matrix calculations can get quite cumbersome and so laminate analysis software such as Helius:CompositePro is available to perform these calculations for each ply in a composite laminate:
Previous Composite Engineer's Pinboard entries:
Our team at Firehole is pleased to announce that we have been acquired by Autodesk, Inc.
As part of this acquisition, Autodesk will incorporate the Helius line of products into the Autodesk portfolio; this includes CompositePro, Helius:MCT and Helius:Fatigue. Firehole's solutions will further enhance the Autodesk Simulation portfolio, which includes Autodesk Simulation Mechanical, Simulation CFD, Simulation Moldflow, and Simulation 360. This is in addition to Autodesk’s more well-known 2D and 3D design tools like AutoCAD® and Autodesk Inventor®.
Over the coming months, we will be integrating Firehole into Autodesk. In the meantime, you can still find us at www.firehole.com.
We are extremely excited about the opportunities in joining Autodesk. Their leadership in the simulation software industry can only enhance our current products and provide important synergies moving forward in the area of composite material design and analysis. We look forward to bringing a composites flavor to their vision: Helping People Imagine, Design and Create a Better World.
Cheers,
The Firehole Team
Engineers have many factors to take into consideration when designing a composite structure, and this includes determining the natural frequency of the structure. The natural frequency of a structure is important because each structure has a unique frequency of vibration (hertz, or cycles/second) that will cause the structure to vibrate with increasingly severe amplitudes, sometimes leading to catastrophic failure. The Tacoma Narrows Bridge is one of the most widely cited examples of this:
By determing the natural frequency of a composite structure, engineers can discover the "danger zone" of frequencies that structure is not suited for, or can make design changes if the vibrations are part of the loading or environment of the structure. To perform a natural frequency extraction for a composite structure is Abaqus very simple.
A basic natual frequency extraction step is specified by:
*FREQUENCY
(Number of eigenvalues to be calculated), (minimum frequency of interest), (maximum frequency of interest)
The number of eigenvalues is really the number of natural frequencies desired to be determined for a structure. There are multiple natural frequencies of a structure, and typically only the first 4 or 5 are physically relevant. Higher values for mode shapes are mathematically possible but are nearly impossible for the structure to deform into.
The minimum and maximum frequencies of interest can be omitted and Abaqus will search for the number of eigenvalues specifed across all frequencies. However, if a minimum and maximum frequency is specified, Abaqus will find the number of exact number of eigenvalues specified within that range, even if they are more unlikely to occur than mode shapes at other frequencies.
It is neccessary to define a material density as part of the composite material defintion using the *Density keyword. Typical unidirectional composite densities are 1.6E6 g/m^3 for carbon/epoxy and 2.0E6 g/m^3 for glass/epoxy.
When viewing the results of a natural frequency extraction, there are 3 results provided by Abaqus:
-
Natural Frequency (labled "Freq" in Abaqus/Viewer): The calculated natural frequency for a particular mode. The modes can be viewed by clicking forward and backward through the multiple simulation frames.
-
Mode Shape: This is the deformed shape the structure will assume during vibration at the calculated natural frequency. The mode shape displayed by Abaqus as a deformed shape is not the true dimensions for the mode shape, as there are not true deformation scales in a natural frequency extraction, but instead is a relative deformation that shows how much one node deflects compared to the other nodes (scaled to fit nicely on the screen).
Value = (Freq * 2 * pi)^2
This post was intended to be a brief tutorial on natural frequency extraction of composite structures using Abaqus and there is more information available through the Abaqus documentation and vibration analysis textbooks.
Firehole recently featured our first guest blogger, DeWayne Howell, the original author of Helius:CompositePro and now president and CEO of Peak Composites. DeWayne wrote a 2-part series on composite part pricing.
Converging on Composites features topics like tutorials on how to best create composite material FEA models, tips on using CompositePro to advance your composite designs, quick videos on topics like converting your metal design to composites, and all sorts of other information, tips, and resources to help our readers better understand and succeed with composites engineering. We've covered a range of topics from modeling complex geometries to reviewing Classical Laminate Theory assumptions - or highlighting composite analysis tools like Simulia's Composite Modeler for Abaqus. We also enjoy featuring great composite design accomplishments.
We look forward to hosting more great guests on our blog.
If you would like to write on a topic - or know someone that would make a great guest blogger - we would like to hear from you.
Not a subscriber?
We invite you to subscribe and receive the posts in your mailbox, comment on the posts, share them on LinkedIn, Twitter, or Google+, contribute with your own experience and suggest topics .
We look forward to hearing from you.
This is Part 2 of a 3-part guest post from DeWayne Howell. DeWayne has 24 years of experience in the design, analysis, development and fabrication of composite structures. He is the original author of the CompositePro tool suite and is now president and owner of Peak Composites, Inc. (PCI). Read DeWayne's full bio.
Part 2: What Goes Into Part Pricing?
In the previous post we concluded with a "rule-of-thumb" for approximate composite part pricing. In this post we will dig into more detail on what really goes into the cost and price of a part.
The following table shows a step-by-step breakdown of part cost activities for a typical autoclave cured component and includes an explanation of each task.(Other fabrication methods would modify this table.)
Some notes on the data columns:
- Direct costs are usually materials, consumables or any other hard cost that feeds into the fabrication process.
- Usually a shop will budget the labor time required for each process step based upon experience or standard times for certain tasks. Depending upon the quantity of parts being produced, the labor time will change based upon typical learning curves; that is, lower volume parts have higher labor hours and higher volume parts require lower labor hours, due to natural laws of efficiencies.
- The unburdened labor rate will depend upon who is doing the work; for example, a level 1 technician doing the layup will have a lower labor rate than the machinist doing the post processing or an engineer.
- A labor cost per task is simply (Labor Cost)=(Labor hrs)x(Labor rate). Once direct cost and labor costs for each item is defined a total cost is tallied. These costs are the recurring costs for a part and do not include the non-recurring costs like part development and tooling. Non-recurring costs are usually billed separately or are amortized into the first year production price.
But wait! This only determines the cost and not the price to be paid for the part.
To price the part additional factors are applied to the costs. These factors are overhead, general and administrative (G&A), and profit.
- Overhead is applied to labor time and includes fixed costs like fringe benefits, facility rental, utilities, equipment depreciation and equipment use. (Some companies will break out equipment use time as a separate line item factored in $/hr in the table above and not include it in overhead.)
- G&A, also applied to labor time, rolls in the cost for support functions like office help and management.
- Profit is how the company makes money, and is applied to the total cost including overhead and G&A.
Each company can have a very different labor rate, overhead, G&A, and profit depending upon the industry, capital equipment, number of employees, employee experience, etc. and can vary greatly. However, a typical set of rates for a small company, based upon small business internet data, would be: Overhead 130%, G&A 25%, and profit ranging from 5% to 50% depending upon if the part is, for example, a high volume automotive (5%) or complex low volume aerospace (50%) part. Unburdened labor rates, per Salary.com, are typically: Technician $15/hr, machinist $25, engineer $35/hr.
CompositePro can be a critical tool for the quoting process. Whether you are quoting a part to sell or purchasing a part from a supplier, CompositePro can help you to calculate the right amount of material to meet structural requirements, thereby minimizing costs.
Now it would be helpful to see an example stepping a real life part through the pricing process and compare the rule-of-thumb price to a detailed price calculation. Download this short ebook to get details of both methods as well as an example illustrating each method.
This is Part 1 of a 3-part guest post from DeWayne Howell. DeWayne has 24 years of experience in the design, analysis, development and fabrication of composite structures. He is the original author of the CompositePro tool suite and is now president and owner of Peak Composites, Inc. (PCI). Read DeWayne's full bio.
Part 1: A "Rule of Thumb"
Composite structures, particularly carbon fiber structures, are notoriously known for being very high priced. It would be nice if there were a simple calculation and some "rules-of-thumb" that could help one to determine an approximate price. Given the fact that price is determined by the type of fiber, fiber form, type of resin, part size and complexity, tolerances, the fabrication method, finishing requirements and company overhead and profit margins this is a tough calculation to make. However, there are few simple steps that one can follow to make an approximation:
- Knowing the fiber/resin combination get the price per pound (material $/lb) of the composite raw material from a prepreg supplier.
- Determine the finished volume of the part to be produced.
- Calculate the weight of the part as (weight) = (volume)(density)
- Calculate the cost of the raw material in the part as (part material cost) = (weight)(material $/lb)
- Finally, apply an appropriate complexity multiplier to the part material cost to approximate a price, where (part price) = (part material cost)(complexity multiplier)
Now, based on experience and an industry survey, a general rule-of-thumb for aerospace structures is to apply these complexity multipliers to the part material cost to approximate a purchase price: 3x for relatively simple structures, 7x for moderately complex, and 15x for very complex structures. A structure is more complex when it has a multifaceted geometry, is made up of several composite components joined together, requires a lot of machining and post processing, and has very tight dimensional tolerances.
As an aid for plates, tubes and beams the "Geometry" definition forms in CompositePro will calculate the weight/length of the part as a section property when the "view section properties" button is selected. The part material cost is therefore calculated as (part material cost) = (weight/length)(part length)(material $/lb).
--> In the next post we will dive into some of the details of product pricing and offer an example. This will help to foster an appreciation for why composite components are priced as they are.
The Composite Engineer's Pinboard is sort of like the ever popular Pintrest website...for composite engineers. My intention with this collection of blog posts is to provide a compact but handy reference for various composite engineering topics that I have hanging (pinned) around my own office and reference quite often. This week's topic will discuss how composite ply stresses are determined from ply strains using the reduced stiffness matrix (Q matrix) as part of Classical Laminate Theory.
Composite ply stresses are determined from composite ply strains by multiplying the strains by a stiffness matrix ([Q]):
A reduced stiffness matrix ([Q]) is a 3x3 matrix that is used to determine the in-plane stresses from the in-plane strains. The reduced stiffness matrix is typically used instead of the full 6x6 stiffness matrix in classical laminate theory because the ply is assumed to be in a state of plane stress. Recall that plane stress means that out of plane normal (33) and shear (13 and 23) stresses and strains are negligible and are set to 0. The expanded form of the reduced stiffness matrix looks like:
where:
The v21 value can be calculated using the Composite Poisson Ratio Relationships.
Previous Composite Engineer's Pinboard entries:
If you typically work with simple geometries (read “flat parts”), then this post may not be for you. But if you have the challenge of working with complex geometries, you can benefit from the helpful features of Composite Modeler for Abaqus/CAE (CMA). If you’ve not used the CMA tool, allow me to provide a brief overview.
Originally developed by Simulayt, this powerful tool is now available through Simulia and is accessed within Abaqus/CAE. It contains several composites-specific tools that help address some of the difficulties that are commonly encountered when simulating composite parts and structures. For example, CMA generates the section properties for parts with composite layups, provides linking capabilities to design packages such as Catia and FiberSIM, and can be used to generate solid elements through shell extrusion.
BUT, the most useful functionality, in my opinion, is the draping simulation toolset – for three reasons:
- This toolset allows the user to determine precise fiber orientations for plies that are draped over complex geometry, such as doubly-curved surfaces (see the example part shown here).
- It produces visualization data that helps determine if it is even possible to drape a ply with a particular orientation over the geometry of interest.
- CMA generates flat patterns for each ply to aid in manufacturing.
Abaqus/CAE screenshot showing the CMA tab, ply pattern, and fiber orientations
Two Notes If You Are Considering CMA:
The model building workflow is a bit more complex than a standard model build due to the additional information that is required in order for CMA to generate all of its outputs. Luckily, these additional steps are relatively painless and should feel familiar to experienced Abaqus composites users.
Like Abaqus' discrete orientation feature, which assigns orientations on an element-by-element basis, CMA assigns sections on an element-by-element basis. This means pre-processing can take longer than usual if the model contains a large number of elements. The improved accuracy of your model element definitions will be worth the additional time.
If you typically work with simple geometries, the advanced features provided by CMA may not provide a benefit for you. In that case, the composite specific tools included with Abaqus are likely sufficient. However, if you work with complex geometries and wish to determine precise ply orientations and manufacturability, CMA can be a powerful tool for you.
What tool do you use to solve the problem of assigning realistic ply orientations? Is the discrete orientation tool sufficient, do you resort to user-defined orientations, or do you use a different method?
We would love to hear from you.