Geo-analysis software 1

Anyone? Come on! :-(

I think a lot depends on the sort of work you are doing and the experience of the people undertaking the work.

Broadly speaking I think you can split geotechnical software down into two types. Single problem software, for example slope stability, laterally loaded piles, settlement, etc, and general numerical analysis software (finite element, finite difference, limit analysis) either in 2D or 3D.

In general its easier to use and validate the single problem software, and in most cases its a good idea to have access to at least one of these software packages if you are undertaking various aspects of geotechnical design. A engineer familiar with the hand calc approach can often pick up how to use these software packages in an afternoon, or perhaps a couple of days for more complex problems.

In theory you can look at any problem with numerical analysis e.g. finite element analysis. However, the skill and experience required is significant. I would go as far as to suggest that geotechnical numerical analyst is discipline in itself, requiring ongoing work over a number of years. Firstly you need to understand how to use the software, and while this is becoming easier with increasing user friendly software, this is still a significant investment of time. You also need to have a good theoretical understanding of how the software works as well as a relatively high level of understanding of geotechnics, probably at Masters degree level or above.

Further considerations in the commercial environment is the checking and QA process, which probably requires at least two experienced engineers to cross check each other work, and how the results are reported. There are further financial questions as well, with many of these software packages representing a significant financial investment.

Hope my thoughts help. However, I think your question is very general and you might get better a better more specific response if you refer to a specific software package (or type of software) and a specific problem e.g. what do people think of software package xxxx to look a slope stability.

An additional thought; as you are probably aware most software producers offer free trial versions so you can have a play with the software before committing to buying.

@DRMABZ
Thank you very much for the input.
Ive been looking a lot at "Seattle3D" and Plaxsis 3D
Question wich raises immediately by looking at these geo analysis softwares is "software will calculate the ground settlement based on inserted soil parametars and loads, but how to use these results when analysing the superstucture"?
How to reflect this soil reaction and its settlemet to the structure and design the structure for this situation?

Settlement analysis involves much more than a computer software.

Before using any settlement software understand the following:
1) geologic formation
2) normally consolidated vs overconsolidated soils
3) soil compressibility
4) fill & surcharge placement
5) effects of drainage conditions

If you do, then try to calculate the soil consolidation of multi-layered strata by hand using soil boring information, laboratory tests, foundation geometry and structural loading.

This is a very advanced topic- If you look at the publication, "Theories of Consolidation" by the late Dr. Robert L. Schiffman and his nearly 500 page book you will begin to grasp the level of difficult involved.

Let me give you an example. Several years ago, I looked at a soils report by a reputable firm and their calculations for settlement for a multi-story condo development. The soils were clearly heavily "overconsolidated" but the calculations were based on "normally consolidated" soils. The bid package at the time was for improving the subsoils to limit the calculated total settlement. I read few approaches of the bidding design-build bidders and just grinned. There was no need to improve the soils. The building foundations were supported by driven concrete piles and the mat floor was fine as is on the overconsolidated soils.

So take a year or so and start flipping few pages on soil consolidation before you look at a settlement software. You can put quality data and be very precise, but may not be accurate. Have an experienced geotechnical engineer provide you some input until you can absorb the related soil mechanics principles.

As suggested in my reply and that of FixedEarth; this is a big and complex subject. In particular 3D geotechnical FE, as used by Plaxis , is a very complex analysis and a sub-discipline of geotechnics in its own right. Settle is a simpler program then Plaxis, as it only looks at one aspect (settlement) with a number of assumptions fixed within the program. However, I would still suggest it need a fair degree of geotechnical experience to use it, including a familiarity with alternative calculation methods.

In brief and at a high level;

> you decide on your soil parameters (which typically requires a experienced geotech to balance the quality of the data and assess parameters not directly measured)
> you decide on your soil constitutive model and analysis method e.g. small strain, large strain, coupled mechanical and fluid interaction, load controlled, displacement controlled, etc, etc.
> you look at issues such as mesh refinement, boundary location, etc
> you undertake a few analysis to characterize the problem.
> you undertake the final analysis case or cases

For a given load case you will have a displacement, which may include tilting, etc, if you have a horizontal load component. Issues such as differential settlement are typically not addressed well unless you get into some very complex analysis not well addressed in most commercial software.In geotechnical FE settlement and bearing capacity failure merge together, with the latter associated with very large displacements.

I'm probably not best placed to say how settlement/displacement is incorporated into the structural design. However, my impression is that there is general generic guidance available dependent on structure type which is used to provide an allowable settlement to compare to the geotechnical analysis. This might be, for example, as settlement that leads to plaster cracks in walls, etc. If you were doing structural FE or similar analysis, say for a steel structure, you could also take the displacement into your analysis directly. Of course any displacement you used would need to reflect a safety factor, conservative soil parameters and/or target reliability, etc.

As a side point - as geotechnical software get more and more user friendly it becomes easier for all engineers to use, including those without a specialism in geotechnics and associated theoretical background and experience. Its a bit of a pet hate of many geotech engineers so you might get a slightly hostile response to your question from some.

Done all that long time ago on my collegue classes and few times in last 10 years, when talking by hand calculation. So plase lets not get
Im asking some guidence towards understanding software on market today and its abilites. One can get lost very fast.

My question for programas like PLAXIS and SETTLE3D "software will calculate the ground settlement based on inserted soil parametars and loads, but how to use these results when analysing the superstucture"?
How to reflect this soil reaction and its settlemet to the structure and design the structure for this situation?
I read a lot on subgrade reaction modulus "k" but never calculated an egsample by hand.
Seems that all software programs use this value. But It seems that PLAXIS and SETTLE are not using these parametars?!?!
So I get lost here...

Modulus of subgrade reaction is not a "real parameter" or a intrinsic soil property. It is a load displacement behavior represented as a linear spring. You see typical values in textbooks or you could derive it from in-situ testing. However, as soil behavior is generally non-linear except at small loads, it requires some care in use.

Software such as plaxis and settle undertake an analysis based on fundamental geotechnical parameters. For example in a simple analysis you may use elastic soil properties such as young's modulus, poisson's ratio, shear modulus, bulk modulus, etc, and plastic properties such as shear strength. Soil unit weight can also be considered. More complex analysis may move away from a Mohr Columb constitutive model or include for settlement based on mechanical fluid interaction and properties such as soil permeability.

In theory you could simplify the results of this analysis to a modulus of subgrade reaction by fitting a linear modulus through the load displacement curve you produce. However, I think it would be more appropriate to just use the displacements from this analysis for the load conditions that need to be considered.

As others will no doubt suggest - its easier then ever to get an answer out of these modern software packages. However, it just as complex, or more so, to determine an appropriate solution.

@DRMABZ
Finaly I see some light!
What you described here, sound to me that granular, sandy, mainly non cohesive soils will get more accuratly described with "k" modulus, since they usuly undergo only elastic "fast" settlement, time "independent" behaviour.
This "load-displacement" behaviour of granular soil - plate load test gives this relation. Could this value be used for defining "k" modulus?
But this is never a straight line?!
We only one time had this almost ideal behavior when soil test results gave a straight line up to a certain load and then boooom , suden break. I think we basicly witnessed general shear failure and "ideal" soil behaviour.
Ofcourse bore holes were made so we had soil profile mad up of rocks and sands up to a large depth.

Re: plate load tests - Real soil behavior is not often linear over any large range of displacement. Hence why I don't like modulus of subgrade reaction. I think its main popularity is that its simple to include in structural analysis. Analysis in software such as Plaxis will generally be more rigorous.

There is a lot of things going on - but at a high level and a bit simplified, elastic stiffness is greater then plastic stiffness. The load displacement response starts to curve with the onset of plastic flow and the load displacement response "softens" up. Once you get unconfined plastic flow (bearing capacity failure) the increase in load with depth is very small and related to soil unit weight (n gamma bearing capacity term)and a small increase in bearing capacity factor (n c bearing capacity term) with depth.

I have in the past reluctantly derived a modulus of subgrade reaction from a load displacement curve. You simply draw a line that undercuts the curved line and get its gradient. Due to the curve in the real load displacement response you need to be careful, if your straight line crosses the curve then any analysis becomes non-conservative i.e. you overestimate the stiffness. I normally draw the line a little conservatively, then give a range of validity. e.g xx kN/m^2/m for 0mm to 20mm.

@DRMABZ
I found this "plot" we made long time ago from plate-load test.
We use this plot to determine the bearing capacity by the "double tangent" method. From there we work back and get the bearing capacity factors Ngama, then fi angle, and finaly determine ulitmate bearing capacity of a foundation.
We check this against allowable settlement wich was 20mm for this kind of structure. Form allowable foundation settlement we determine plate settlement read it from the graph and find the load itensity wich coresponds to this.
This is usualy the critera that governs the design couse it gives smaller net safe bearing capacity.
So for this egsample the "k" modulus of subgrade reaction would be this net-allowable bearing pressure divided by plate settlement? Something around 19 t/m2 divided by 0,012m?

...forgot picture

Sorry I don't have time to reply in detail, however a quick general reply.

The units of modulus of subgrade reaction are a little funny. The m^2 relates to area of foundation and the 3rd meter relates to gradient, per meter of penetration. So yes divide by your distance in meters then multiply by 1m. If you sketch on you line of modulus of subgrade reaction by extending back from 1m you want it to undercut your curve (i.e. be conservative) for the depth range of interest e.g. 20mm.

Incidentally the numbers look sensible for a medium dense sand, or a reasonable strength clay.

If your looking for some further background reading on the subject Foundation analysis and Design by Bowles has a reasonable amount on the subject.

Will look at the book you sugested and read more on this topic.
But what if you have to model clay soil in this way?!?
Obviously plate-load test results arent desirable thing to use in the design process!