FEA to physical testing 4

[MechLAD]

I work for a mid-sized company that recently brought the engineering (me and 2 others) in house. It has been my duty to incoporate a structural analysis and physical testing process on all products. (Front-end loaders, Backhoes, R.O.P.S., etc.)
Most products are relativly straight forward, except the backhoe. This gets used in so many different ways and it is used in so many different environments that it makes a comprehensive analysis very difficult to test.
Does anybody know of any source of info (websites, papers, books) that address physical testing of mechanisms in corralation to FEA results?
Does any body have any "best practices" or "rules of thumb" they could share regarding FEA -> physical testing process.
Any information would be appreciated
Thank you

 

[gwolf]

Stress analysis is 10% stress and 90% figuring out what the loads really are. The case of the backhoe is a great case in point - no amount of FEA will help if you don't know what the loads are in this extremely variable application.

In aerospace the subject of wind gust loads during flight is another good example - nobody really knew what the loads were either in terms of frequency or magnitude. The aerodynamic/environmental mechanisms were so complex that no predictive tool could help. The solution was that many planes were flown with flight recorders in the early days to simply measure the gusts.

I suggest that in the case of the backhoe your first point of attack should be to strain gauge a few machines and make some FE models to relate the strain gauges to what happens in the machinery. This will allow you to deduce what the loads are and develop some design loads for backhoes.

Hope this helps.

Gwolf.

 

[JStephen]

Can you could base your FEM design on the maximum hydraulic cylinder capacities, rather than applied external loads?

 

[mloew]

Hire a former engineer from Cat or Deere. They know what they are doing.

Best regards,

Matthew Ian Loew


Please see FAQ731-376 for tips on how to make the best use of Eng-Tips Fora.

 

[BillPSU]

Backhoe loaders see all sorts of external loads such as when you push snow or dirt and the bucket runs into a non-movable object like a rock or a curb. Boom and dipper assemblies see the normal loading for digging but also see lifing loads which sometimes are dragging loads followed by a falling load. Another problem is they are weldments which are not the same as a solid part model. Getting model correlation with the real world part must be done.

Build your first model apply a load on the model and compare the model to strain gage results from the same part under the same load.

 

[GregLocock]

Mathew, hmmm. I don't see enough broken backhoes to make me think that there is much finesse to the structural design.

I'd guess they design them to the hydraulic forces *5, the extra weight of metal would account for the unauthorised mods and abuse that occur in use, and downtime would far outweigh the extra running cost of the additional weight.

Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[corus]

In general you'll find that FEA, for calculating stresses, is accurate to within 10% in my experience. For thermal analysis the accuracy is much greater. Remember that the results you get are being corrleated with material properties that themselves are only approximately correct. In general all my work is correlated/validated/verified one way or another, or else nobody would believe the answers.
For backhoes and the like, use the maximum hydraulic pressures available in the system and add in a factor for impact etc. from design codes. Using this method fatigue assessment of welds from FEA results seemed fairly reliable if you take out the factor of bad welding practice.

corus

 

[MintJulep]

What is the goal of this "structural analysis and physical testing process", and how many product development cycles do you have to implement it? Do the bosses understand or even know what they want?

It sounds like you want to test to validate the FEA model and results, correct? The block diagram of such a process is pretty simple. Define load cases. Build FEA model. Apply BCs and load cases, crank analysis. Build physical prototype. Apply BCs and loads, measure strain and displacement. Compare test resutls to FEA results. Figure out where you screwed up. Refine model. Rerun analysis. Compare. Repeat until you get satisfactory correlation for all load cases simultaniously (that, no tweaking for each case).

Now you "know" how to model "that sort of part", and can be reasonably certain that any FEA results that come out of that model will be fairly accurate. Now what? If the next product has a similar "that sort of part" you have a head start on the model, and your validation testing goes faster. Eventually you may even get to the point where you are suffiently confident in your modeling techniques to do away with the validation testing.

 

[ccw]

gwolf's post is a chilling reminder about the Airbus crash in Rockaway NY few years back. Was there another load created in the rudder command / yaw damper control system that was not in the set of loads the Airbus structural engineers used? Doing a scan on the internet about that accident seems to lead to uncertainty as to final determination of root cause.

The backhoe industry is mature and competitive. In those kinds of industries one normally finds designs that barely, but adequately, survive the worst case set of loads (both magnitude and frequency for a given set of materials) that could be achieved. The retired Cat or Deere engineers would probably carry that set of loads in their design kit bag.

All of the failed backhoe designs from misguided finesse probably occured when Greg was a teenager ;-).

 

[GregLocock]

Harrumph.

The Airbus rudder/fin failure was because for their own inscrutable reasons the FAA load for that part consists of a full application of rudder and return to straight ahead.

So if you waggle the stick the fin falls off. Oo er.



Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[BillPSU]

I used to work for Allis Chalmers in the division where backhoe loaders were built. I was a manufacturing engineer not a product engineer and I was young so I didn't see all the hand calculations which were made to the product but I did see a great deal of life cycle testing. Hand calculations, FEA, and computer simulations get you in the ball park assuming you know the loading of the structure however knowing the loading is very difficult due to dynamic loading and outside forces which a user can apply to a machine.

The solution of hiring an ex Deere or Cat engineer may seem feasible to mloew but the original engineers which designed the structures have probably retired and the current designs are still based on their early work. Those people are not available.

 

[amorrison]

My ancient 1973 SAE Handbook says their exists a "Hoe Dipper (Backhoe Bucket) Rating - SAE J296"
prepared by "Construction and Industrial Machinery Technical Committee".
The standard itself isn't much - one page - describes how bucket volume (material heaped over bucket "top" in center)is to be measured.
Might be worth searching SAE to see what is available currently on backhoes.

 

[gwolf]

Stress analysis is 10% stress and 90% figuring out what the loads really are. The case of the backhoe is a great case in point - no amount of FEA will help if you don't know what the loads are.