AZFEI

Finite Element Analysis Computer Modeling

Finite Element Analysis Computer ModelingAZFEI can perform sophisticated computer stress analysis and design using state-of-the-art finite element analysis (FEA) software. With over 24 years and thousands of designs in actual field use, you can have confidence in your results. See the qualifications page and feel free to call for a free estimate of your project. Click the links above for more details.

How to Choose an FEA Vendor for Computer-Aided Design:

  1. Experience. How long have you been using FEA?

    Anyone can be trained to use the software in two weeks, but it takes years to know how to properly apply complex loads and pressures, simulate manufacturing flaws, properly remesh critical areas, assess uncertainty, and redesign the model as necessary.

  2. How many models have you simulated?

    Software proficiency is a very small part of skillful FEA use. The heart of simulation is translation of real-world components and conditions into accurate computer model representation. A skill level sufficient for real-world use would require at least several hundred models experience.

  3. What is your area of proficiency in computer design?

    Some designers are skilled in static stress design, others in impact testing, yet others in hypersonic heat transfer. No one designer can be an expert in all areas. You should be weary of those who pawn themselves off as a "one-stop-shop" for all kinds of FEA

  4. What is your real-world experience?

    If the designer has not been involved the daily design, testing, and maintenance of similar products, move on. Real-world design experience is critical. For example, in the design of a tank, it is common to include a corrosion allowance in the design. Hence even though the product may be 1/4" thick, it should be modeled as 3/16". Additionally, the product manual should require inspection at weekly intervals and this instruction is just as important as the design itself. The design of pressure vessels often requires buckling analysis for the possibility of implosion. FEA models will often yield a buckling pressure which is three times the actual real-life load. Hence, if you trusted the FEA output alone, you would have a serious disaster on your hands. FEA software should be treated as a lethal weapon and left only in the hands of those who know how to properly use it and when.

  5. Has a product you designed ever failed?

    Ask for the details, and what they have learned from it.

  6. What can you tell me about overload, fatigue, corrosion, and wear failures?

    These are the primary failure mechanisms of mechanical parts and knowledge of these failures should be part of the repertoire of any designer. If this question is met with a pregnant pause, move on.

  7. Do you write programs to assist in the design of your computer models?

    Writing macros for repetitive designs allows for quick and efficient analysis of alternate designs. If you are engaged in a long-term project, you don't want to pay for the same thing 20 times.

  8. Do you provide calculations and explanation of the design?

    Every FEA output must be checked with calculations to ensure there are no gross modeling errors. The results must satisfy intuition as well. This is why it is critical to employ a designer who has the knowledge and experience to apply proper quality control to the final design.

  9. Do you warrant your work? How long can I expect my product to last?

    Often, the designer will give you some disclaimer as to the results. Would you accept this with an automotive purchase? You are paying for something to work, and there should be a guarantee that it will. Get detailed warranty terms and understand the potential liability. All products have a lifetime. This life expectancy is estimated both from calculations and field experience and is an inherent part of the design.

  10. Do you have experience in testing the products you design?

    It would be a scary world if cars, planes, and tools were not tested and compared to the design predictions. A designer should have experience in controlled laboratory testing and validation.

  11. Have you taken field measurements of your simulations to validate your software?

    Again, buckling predictions may be way off, and large-deflection behavior, weld strength, bolted joint efficiency, and material strengths may vary. A designer must have the experience to know how the software matches up to real life considerations.

  12. What is your training and experience in failure analysis?

    A good designer is also a good failure analyst. It is critical to know how a product can fail, what the failure looks like, and what laboratory tests are required to evaluate the failure. It is always possible that an unanticipated condition causes your product to fail. How will the designer revise the original design to meet the challenge?

  13. Are you a licensed engineer?

    An engineer who passes the board exam is given the title "Professional Engineer" or "P.E.". Most products which go into public use must be designed by a P.E. or by someone working under a P.E.'s direction and control. The Professional Engineer has a higher standard of ethics and is required to meet each state's standard of adequate training and experience when stamping and certifying designs. Your designer should have the letter "P.E." after their name.

[Back to top]

FEA Design Considerations

Imagination

ImaginationOften, product failure leads to consequences not anticipated by the designer. One of the most important roles of the design engineer is to creatively imagine what could go wrong. The most exhaustive researching, interviewing, and testing may still not reveal flaws in the design. Hence, the design engineer is also responsible for imagining how the product may fail and will interact with its environment. In this case, a poorly designed pulley led to hub warpage which in turn led to bearing failure. The bearings warped, leaked, and overheated, ultimately causing a fire. As one might say, this was "hard to imagine." The deflections must be scrutinized as closely as the stresses as they may impact the function of adjoining parts.

Small and Large Scale Effects

Small and Large Scale EffectsThe steel beam connection shown was overstressed due to hydraulic loads and hence resulted in a weld failure where it was welded to a plate. These kind of connections require larger safety factors and an understanding of dynamic loading. Typically, the civil engineer is not cognizant of small-scale stresses while the mechanical engineer is not aware of large-scale loads. Experience in both areas is necessary for successful design of large machines subjected to high wind, seismic, and hydrodynamic loads.

Safe & Efficient Design

Safe & Efficient DesignMany designs are reinforced at strong points while neglected at weak areas. A designer must be aware of areas which require "no fail" design and others which require economic optimization. The yoke shown can result in far greater damage to the machine as a whole if it fails. It is not necessary to analyze every part with an exhaustive FEA analysis. A relatively inexpensive part as this should be designed like a black box, while effort at streamlining should be devoted elsewhere.

Thermal Effects

Thermal EffectsThermal loads are often neglected in the design of critical parts. The simple square plate shown shows the effect of having a hotter surface temperature on the bottom. The result is that the bottom expands more than the top. The top surface tries to constrain the expansion, hence the plate curls upward. Thermal FEA analysis is very useful in determining thermal stresses accurately. The FEA graph shows the deflections of the plate, indicating that the corners curl upward the most. A constraint at these areas could result in an unexpected failure. The old saying, "garbage-in, garbage-out" holds strong in the realm of computer design. The output is useless if the actual or potential conditions are not modeled.

Large Deflection Effects

Large Deflection EffectsAn FEA software user may model a large tank using standard "linear static" analysis. This may be good as a start, but the large deflections caused in thin-shell structures will greatly change the stress magnitude and location. The "bulge" shown in red above may also cause high tensile loads on the tank which can tear apart the end restraints. As a result, very large loads on the connections will not be evident and failure at these locations will be likely. The proper type of modeling is termed "non-linear static" analysis and considers these real-life effects. The software and the designer must be capable of modeling these real-world conditions.

[Back to top]

FEA Experience Snapshot

By now, you should have some idea that FEA design is a very intricate undertaking and requires thousands and models and decades of experience to ensure proper design quality. Following are some real-world examples of product design and failure analysis using FEA.

Wastewater, Oil, and Chemical Tanks

Wastewater, Oil, and Chemical TanksShown here is a settling tank modeled using non-linear static analysis. The tanks, typically 20 ft high and wide, were subject to wind, earthquake, seismic hydrodynamic and thermal shock loads. These tanks were custom designed with various thicknesses, coatings, chemical environments, liquid/solid contents, abrasive media, and heights and installed in every state in the U.S. Various shapes have been employed, including cylindrical, square, and pyramid shapes. The tanks were often subjected to internal pressure and outdoor temperatures of -40 F as well. The software also allowed proper modeling of these stresses and shell buckling due to wind pressures or vacuum. During the design and modeling phase, the tanks were filled with water and the measured deflections compared to the computer results. Precise agreement was found and confirmed the accuracy of the modeling.

Aeration Panels

Aeration PanelsPlastic aeration panels used in water treatment plants must sustain large buoyant loads due to a flexible inflated membrane attached to the panel. The air is typically hot and reduces the plastic strength over time as well as inviting biological growth and associated corrosion. These panels were modeled and tested to validate the accuracy of the computer simulations. Inputting the long-term property data for the plastics into the computer model, the panel performance was accurately predicted to beyond 10 years. Often, as in the case of these panels, deflections are more important than stresses. Only computer modeling or actual testing can determine these deflections, and it is far easier to run a computer simulation than full-scale expensive testing.

Aquarium Failure, AZ

Aquarium Failure, AZA large acrylic aquarium failed resulting in extensive flood damage at an upscale residence. A variety of theories were proposed to explain the sudden failure after five years of problem-free performance. The solution to this highly complex dilemma could only be determined by detailed computer analysis of the steel frame and aquarium. A variety of simulations were provided with varying support conditions to demonstrate that the tank must have contained an inherent manufacturing defect. The quantitative and qualitative nature of the results successfully argued the case against rather imaginative theories posed by the defense.

Hydrofracturing Tank Failure, NM

Hydrofracturing Tank Failure, NMA large water tank failed at the seam, causing massive erosion, damage to trucks and trailers, and minor injuries of the workers. Inspection of the failure required close scrutiny of the welds, connections, drawings, and construction methods. Stress calculations combined with computer simulation of the tank demonstrated that the stresses in the weakest joint exceeded allowable levels by a factor of six. Litigation of the claim was quickly abandoned after issue of the report.

Pipe Flange Failure

Pipe Flange FailureThe flange shown was subjected to large lateral loads which ultimately stressed the weld until failure. The failure was due to a combination of corrosion, unanticipated loads, and a poor weld. Multiple contributing factors are often to blame for failures of otherwise properly designed equipment. It is important for the designer and failure analyst to consider all failure modes carefully and avoid "rushing to judgment."

Solar Panel Failures, AZ

Solar Panel Failures, AZThe goal here was to identify the source of widespread glass panel failures in a large array of solar panels (photovoltaic cells). The manufacturer of the panels and the supplier of the mounting racks were at odds with each other, each claiming the other was at fault for the mysterious cracking. Additionally, a third-party engineering consultant firm was brought in, yet after 1.5 months of collaboration and experiments, there was still no consensus amongst the experts as to the root cause of the failures. With accusations flying and litigation looming on the horizon, quick and accurate resolution was required for this conflict between several solar industry giants. After field measurements and documentation of the various failures, the solar panel structure was modeled as shown above. The modeling, employing gravity and thermal loads, revealed that the stresses in the glass were far below the expected failure levels. Along with calculations and fractographic interpretation of the glass failures, it was proven beyond doubt that the glass was defective. The analysis of this failure required knowledge of structural engineering, heat mechanics, glass fractography, and FEA computer simulation to properly isolate the cause.

[Back to top]