Factor of safety in machine design with dynamic & shock loading 1

[RF-ME]

A question I have been wanting to put out there to other engineers for a while...

For structures with live/dead loads that are known values, the design process is 'fairly' straight forward depending on the code/standard being used.

In this instance I am going to use earth moving machinery/implements as the focal point.

In the situation where not all loads and load factors are known (shock loads, etc.) what have people used as load factors, or factors of safety in their design. the machines self weight will be knows, as well as the material it is moving. so the combined loading on the machine can be approximated. as mentioned above, it is the dynamic loads / allowance for operator abuse, etc. that is the unkown.

In Aus, some transport design codes call for a factor of safety of 3 for HWY use, and up to 5 for off road use, to account for these factors.

In the absence of test data to base decisions off, I was leaning towards using a FOS of 3 (Stress allowable = Fy/3) for non critical areas, and FOS 5 (Stress allowable = Fy/5) for critical areas.

I would then go about analysing stress conditions using FEA.

would welcome any thoughts or suggestions based on similar experience/situations.

 

[GregLocock]

Put it this way. I have measured more than 6*dead load in the load path on production cars in events that they are expected to survive ie not crashes.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[RF-ME]

Hi Greg,

interesting. so are you able to shed any insight in what you choose as the design basis in those situations? i.e. if testing shows 6 x dead load, do you then use that, or is it far to impractical?

I would love to be able to conduct physical testing, and down the track we will, but choosing the starting point for the prototype is proving tricky... and not something to be taken lightly due to the investment required to build it.

cheers for the input!

RF

 

[MintJulep]

For earth moving equipment there is little incentive to make it "as light as possible", so the old adage "when in doubt, make it stout" applies.

There is a broad population of existing equipment for you to eyeball and get a sense of what works.

Also, there are many YouTube videos featuring the abuse cases for you to study.

 

[GregLocock]

Rent something similar and strain gage it. Those quasi static loads are used for initial sizing of components, but the production parts are optimised using measured loads, but they still use a FoS.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[RF-ME]

Thanks Greg, fair suggestion.

 

[CWB1]

Not to nitpick but FoS isnt normally a design consideration. Typically corporate design standards are established for specific product use-cases based on warranty/service data, customer surveys, and testing conducted to map the extremes for a hypothetical product line. Those standards might include a small FoS but ultimately, its all tested science not guesstimation or anything else that will cause the legal dept and management nightmares. In application they're all specific pass/fail criteria.- shock, vibe, temp, and other limits. A similar, related, and common mistake is treating society "standards" as how-to guides without the context of corporate standards/guides or experience bc they're rather misleading, hence why its unethical and often illegal to work unsupervised outside your experience.

Not sure what your project is, but a lot can be learned simply by reviewing design limits published in the major OEMs' aftermarket-upfitter guides/documentation and contacting customer service with questions. The construction and ag industry are prob the most willing to share but automotive, rail, and other industries do so as well bc it drives sales.

 

[human909]

In Aus, some transport design codes call for a factor of safety of 3 for HWY use, and up to 5 for off road use, to account for these factors.

In the absence of test data to base decisions off, I was leaning towards using a FOS of 3 (Stress allowable = Fy/3) for non critical areas, and FOS 5 (Stress allowable = Fy/5) for critical areas.

At risk of repeating what others have said. A FOS is not an appropriate approach to a 'general' impact load. Those FOS seem more suitable to static loads movement. Eg lifting or tie down restraints.

Dynamic loads can readily be 5x, 10x, 20x, 100x the dead load. It is all about energy absorption. Without further information we really can't provide further guidance.

 

[RF-ME]

thanks for your thoughts all, human909, no dramas with you exploring/expanding on what others have said, appreciate you taking the time to contribute.

I expand on the explanation of where I was coming from with my FOS comment. my interpretation is that over time, the transport industry has found that applying an Fy/3 for strict Hwy use, and Fy/5 for off-road use, accounts for the different loading conditions experienced by the typical operator operating a truck/trailer combo, so in this example, it has become an industry practice to design based on known static loads, and account for the different loading conditions using the Fy/3 or Fy/5 principle.

what I have perhaps poorly articulated is that I see a direct correlation in the off-road use cases in that industry, to the project I am involved in, and was potentially looking to apply that thinking for our prototype design. the other thing I was thinking of doing was applying loads to a design that is known to 'work' in practice, and use that to establish a limit case to apply to design of similar elements.

I agree that if resources were no barrier, instrumenting and collecting our own raw data, which down the track we will do, would be the best approach. For the first prototype build, we will be sizing based on experience with similar equipment, and running some studies based on internal reasoning of loads.

I would be interested to know how others in the industry are conducting analysis of dynamic loads based on the collection of test data.

cheers,

 

[GregLocock]

Bit easier in the car business, we have wheel force transducers that measure all 6 force/moments into each wheelhub, and also strain gage or use load cells on some components. Also, for square edge pothole, we have an instrumented pothole.

I then take the time histories of the forces recorded on the durability tracks, apply them to a half vehicle multi body dynamics model, and use that to generate time histories for each mounting point or joint in the car. These then get bundled off to the roadloads guys who do things to them to create load profiles for their rigs, and the FEA guys dos something similar to use them for strength/fatigue testing their models.

It's an expensive and time consuming process, and assumes the measured car is representative of the MBD and FEA model. The Holy Grail would be to scan each road surface and use that to drive the MBD model directly. That has been an on-going research project for 15 years, and has yet to pay dividends.

An alternative approach is a more sophisticated version of 321 which has been used to design race cars for decades. That is apply 3g vertical, 2g longitudinal and 1g lateral. So we have a bunch of 'highest g ever recorded' in various events and use them. Less than ideal but gets you going.




Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[RF-ME]

fascinating insight Greg, thanks!