Coefficient of Friction 2

[ChiEngr]

Hello,

I was given a soils report for a project I am working on which stated that for sliding checks of footings, the friction factor given should be multiplied by the dead load to obtain my sliding resistance. Does anybody else use only the dead load when checking sliding resistance? It seems to me that if you have other vertical downward loads acting on the foundation, that you should be able to include those forces in the sliding resistance check unless I am grossly mistaken. Does anything have any thoughts on this?

 

[canwesteng]

if they are part of the load that causes the sliding, then sure. vertical platform live load - can't be considered. vertical load occurring with lateral wind load - could be considered (although would be unusual for this to be helping you).

 

[wth]

In my area (EC with national annex) favorable variable loads are not included in stabilty calculations, we even reduce DL by a factor of 0.9, while increased the dominant variable load by 1.5.

So unless you have a special case where the other load produce both horizontal and vertical reactions I would not included it. For instance, you can't assume snow loads when checking sliding from wind.

 

[JStephen]

The biggest factor I'm aware of that fits this is the contents weight in storage tanks. 90% or more of the seismic load is due to the contents, and the contents are also used to resist seismic forces when appropriate. (See ASCE 7-16, 15.7.6.1.5, definition of "W", for example).

 

[HTURKAK]

I wish not to disappoint you but your approach is not correct . If you are checking the foundation against sliding , we need to speak the lateral force effect . Horizontal forces could develop due to wind, seismic , blast, lateral earth pressure , or structure itself develops horizontal reaction.

You SHALL only consider dead load when checking sliding resistance for wind, earth pressure, and blast loading.
Consider the case Mr STEPHEN discussed. The storage tank will subject to the same wind loading regardless the tanks is empty or full . When you include LL ( in this case liquid content ) , if the tank is empty , it may slide.


According to the grace of God which is given
unto me, as a wise masterbuilder, I have laid the foundation, and another buildeth thereon. . . .
I Corinthians 3:10

 

[ChiEngr]

I should have been more specific in my original post. Suppose you are designing foundations for a pre-engineered metal building. If snow load is causing outward thrust and lateral force on the supporting footings, I believe that you should be able to include the vertical snow load causing the frame action thrust in your resisting gravity load. I don't see why the geotechnical report that I was given would limit me to using only the dead load for a condition like I just explained.

 

[SE2607]

The storage tank will subject to the same wind loading regardless the tanks is empty or full . When you include LL ( in this case liquid content ) , if the tank is empty , it may slide.

I agree with you that an empty tank subjected to wind loads is probably the critical load combination, but, to be clear, in the case of the tank being full of LL and considering the load combination D+0.75(LL+0.7E), couldn't one use the friction factor multiplied by the vertical component of D+0.75(LL+0.7E) to provide sliding resistance?

For the minimum seismic condition (0.6D-0.7E), there is no LL acting vertically or horizontally, so frictional resistance to sliding would be provided by the friction coefficient x (0.6D-0.7E).

Accordingly, I believe the answer to the OP's question is that the friction factor can be applied to all vertical forces included in each load combination to resist sliding.

 

[andrews2]

For the case you described with the snow load causing thrust, I don't see why you can't use the snow load's contribution to friction. Hard to have the horizontal load from the snow without the vertical component present as well.

 

[Tomfh]

Say the only load is dead load (or snow load or whatever), and this is likewise the only load causing sliding. Factor the dead load up for calculating sliding force, and factor it down for calculating resistance.

 

[andrews2]

"Factor the dead load up for calculating sliding force, and factor it down for calculating resistance."

I disagree. There's no load combination that uses different factors for the same load type in different directions (except seismic). You should just use your typical service level load combinations and apply the appropriate factor of safety for sliding.

 

[human909]

Say the only load is dead load (or snow load or whatever), and this is likewise the only load causing sliding. Factor the dead load up for calculating sliding force, and factor it down for calculating resistance.

That makes zero sense. We use factors to account for probabilistic variations. There is a probability of zero that your dead load is both higher and lower than the nominal value at the same time.
(Unless of course the dead load is on Schrodinger's cats.)

I perform seismic analysis on silos all the time. You better believe I use the full live load for calculating both the lateral force and the stabilising force.

 

[Tomfh]

That makes zero sense … We use factors to account for probabilistic variations. There is a probability of zero…. You better believe I use the full live load for calculating both the lateral force and the stabilising force.

The code factors, whilst accounting for probabilistic variation, also provide a safety factor to the overall design, over and above any certainty we may have about the loads in question.

 

[human909]

The code factors, whilst accounting for probabilistic variation, also provide a safety factor to the overall design, over and above any certainty we may have about the loads in question.

In general that is not how LRFD/LSD works. It explicitly uses probabilistic approaches in determining factors and does generally does NOT add "safety factors". I say 'generally' as some niche areas of LRFD still uses some sort of 'safety' factor. Other like US/Euro use serviceability wind loads that then get scaled up to ultimately which is a bit of an oddity.

True limit state design only has probability variance in its factors not some magical safety factor.

Say the only load is dead load (or snow load or whatever), and this is likewise the only load causing sliding. Factor the dead load up for calculating sliding force, and factor it down for calculating resistance.

I'll repeat myself here. Even with your spurious safety factor comment this makes ZERO sense.

There is no world in structural engineering where something weight 1.2G and 0.9G at the same time. Yes you should allow for probabilistic variance in the coefficient of friction, but not in the dead load within the same combination.

 

[HTURKAK]

The answer for ground supported storage tanks is NO !. The relevant load combinations copy and pasted from API 650

for wind ;
- Wind and Internal Pressure: DL + W + Fp Pi
- Wind and External Pressure: DL + W + Fpe Pe

Sliding Resistance for seismic ;

Vs = μ(Ws + Wr + Wf + Wp)(1.0 – 0.4Av)
In this case content wt is included.

For the cases the OP clarified , the horizontal force developing at supports due to gravity loads ( D+S ) proportional with vertical loads for linear systems and the friction factor can be applied to all vertical forces to resist sliding. However , the same reasoning will not be true for wind loading . I prefer tie rods for large horizontal forces rather than trusting friction.

IMO, the designers should keep Murphy's Law in mind If something can go wrong in more than one way, it will always go wrong with the worst possible outcome.)


According to the grace of God which is given
unto me, as a wise masterbuilder, I have laid the foundation, and another buildeth thereon. . . .
I Corinthians 3:10

 

[Tomfh]

There is no world in structural engineering where something weight 1.2G and 0.9G at the same time.

Example, say you have a retaining wall where the entire overturning load is due to the weight G of the soil, and the resistance is also provided entirely by the weight of the soil. Would you apply the same factor to both?

 

[human909]

Example, say you have a retaining wall where the entire overturning load is due to the weight G of the soil, and the resistance is also provided entirely by the weight of the potentially different soil. Would you apply the same factor to both?

That sounds like a different scenario and I've edited your response to highlight the difference. The unknown density/weight providing the overturning load is a different from the weight providing the resistance. So there is clearly a degree of uncertainty.

In the case of sliding ground shear caused by the seismic mass of the structure being excited. This is exactly the same mass as is providing the normal force on the ground. No matter how uncertain you are of the value, they are still the SAME value.

As I said when it comes to storage vessels the live load can vary considerably. And you would be foolish in many case not to assume a full live load condition. You'd be equally foolish to design a foundation against overturning by using the full live load and only the dead load. That is impressive cognitive dissonance.

 

[andrews2]

The code factors, whilst accounting for probabilistic variation, also provide a safety factor to the overall design, over and above any certainty we may have about the loads in question.

This is true, so why do you feel the need to create your own load combinations and apply an even greater factor of safety? The factor of safety is partially to account for uncertainty, but we can already say with absolute certainty that the dead load will not be two values at the same time as they are in your method.

In general that is not how LRFD/LSD works. It explicitly uses probabilistic approaches in determining factors and does generally does NOT add "safety factors". I say 'generally' as some niche areas of LRFD still uses some sort of 'safety' factor. Other like US/Euro use serviceability wind loads that then get scaled up to ultimately which is a bit of an oddity.

All the sliding checks I do for foundations are done using service level loads. I'm not sure where you practice, but this is the norm in the US. The geotechnical engineer typically provides the minimum factor of safety for sliding and overturning (Usually about 1.5). Service load combinations use either a 1.0 or a 0.6 factor on DL, and it has become office standard to not apply a FS if using a combination with the 0.6 factor applied to the DL. The reasoning for that is that it seems reasonable that reducing your resisting load by 40% is already an adequate factor of safety.

Also the US has been using ultimate level wind loads since the release of ASCE 7-10 14 years ago.

 

[human909]

Sorry andrews2, it seems I misspoke and was incorrect about a couple of factors that you have corrected me on.

All the sliding checks I do for foundations are done using service level loads. I'm not sure where you practice, but this is the norm in the US. The geotechnical engineer typically provides the minimum factor of safety for sliding and overturning (Usually about 1.5). Service load combinations use either a 1.0 or a 0.6 factor on DL, and it has become office standard to not apply a FS if using a combination with the 0.6 factor applied to the DL. The reasoning for that is that it seems reasonable that reducing your resisting load by 40% is already an adequate factor of safety.

Fair enough. I don't know the specifics about US as I'm Australia.

I do know that we structural engineers here do get frustrated with our geotechnical peers as many of many are still using ASD, and some seem to have no knowledge of LRFD when you question them on it. (The better ones make it clear what the ultimate resistance factors are and the recommended reduction factor.)

Also the US has been using ultimate level wind loads since the release of ASCE 7-10 14 years ago.

True. I apologise for not checking my facts. I should have just said EU codes.

 

[Tomfh]

The factor of safety is partially to account for uncertainty, but we can already say with absolute certainty that the dead load will not be two values at the same time as they are in your method.

Yes the loads are not literally two different loads at once. I'm referring to within the context of limit state design where we increase destabilizing effects, and discount stabilizing effects.

If you use a factor of say 1.0G for stabilising effects and 1.0G for destabilising effects, your Stabilizing effects equals destabilising effects from the outset, i.e. your structure is at the point of instability. The code requires a certain minimum level of separation between Load and Resistance, regardless of how well defined the loads and resistances may be.

 

[human909]

Yes the loads are not literally two different loads at once. I'm referring to within the context of limit state design where we increase destabilizing effects, and discount stabilizing effects.

EFFECTS being the key word here. Don't start discounting G just for the sake of it when it is the same G for the load and the resistance.

We instead discount factors on things like the coefficient of friction. And take extreme values on the loads such as seismic loads or otherwise.

The code requires a certain minimum level of separation between Load and Resistance, regardless of how well defined the loads and resistances may be.

Well the code I use usually doesn't. It uses the Factored-Load <= Factored-Resistance, inequality. The separation is given by either:
[ul]
[li]A. Factors due to uncertainty
[/li]
[/ul]
[ul]
[li]B. Ultimate extreme values eg extreme winds or extreme seismic (1/500 year events)
[/li]
[/ul]

Oh and in most of my designs I have an additional factor called the "I like to sleep peacefully at night factor".