ACI 318-19 Requirement that beams and slabs must be tension-controlled

I might be mistaken... But, this has always been a requirement, yes? That is the purpose of imposing reinforcement and strain limits based on the area of concrete. The purpose is so that the reinforcement begins to yield and give warning signs of collapse in lieu of a sudden concrete blowout failure on the compression face of a slab / beam.

It was, then it wasn't. In ACI 318-14 phi was determined based on strain.

Best I can find in the code to this effect so far is 8.3.3.1 and 9.3.3.1

Phi has been strain determined for awhile, 318 just used to allow you to dip into the transition zone with et = 0.004 being the lower bound. Commentary for 9.3.3.1 mentions this change without much explanation. Commentary for 8.4.2.2.4 mentions capturing new higher grade reinf.

It has always required flexural members to be tension controlled.

The only difference is a strain limit rather than a neutral axis depth limit. A strain limit does not stop it from being tension controlled.

I think I traced the reference I heard to the sections listed above. In the past, you could have a compression controlled beam section but with smaller phi. Now the practical limit for strain is .003 + yield strain for the steel.

You could never have a compression controlled non-prestressed beam, there was a limit on the extreme tensile strain of 0.004 (think this was introduced in 318-02)

DTS491, I implore you to abandon this compression controlled method you have apparently been using. This is against the industry standard, and for good reasons that I stated above. I am not saying your method is incorrect by Code but there's a reason that the majority of engineers design for tension controlled beams/slabs.

WesternJeb, see Table 21.2.2 in 318

Please read the second half of my comment above. I am not saying you're doing something wrong per Code, but this isn't good practice. Compression style failures in flexural members give little to no warning before sudden failure. It is bad practice, regardless of it is acceptable for design or not. This is about the structure in 50 or 60 years as it begins to degrade and people start to inspect it for signs of distress.

ACI 318-19...

The beam requirements always superseded the phi factor table and required a minimum extreme tensile strain of 0.004, revised in 19 to be 0.003+ey. Compression controlled non-prestressed beams have never been allowed per code.

edit: non-prestressed beams with Pu >= 0.10 f'c Ag were excepted from the strain limit.

Celt83, yes that's the new requirement I was referring to. That is new in the 19 code.

DTS419, that requirement is not new the strain limits have been in place since at least ACI 318-02. The change in ACI 318-19 is that you are no longer allowed to dip into the transition zone (0.004 up to 0.005 tensile strain with 60 ksi steel).

relevant section of ACI 318-02 through 318-11:


The section moved to 9.3.3.1 in ACI 318-14:


And then in ACI 318-19 9.3.3.1 was revised to enforce "tension controlled" sections:

That's right, the "tension-controlled" part is "new" in 19.

DTS419 just clarifying as you said this:

DTS419 said:

...you could have a compression controlled beam section but with smaller phi....

Click to expand...

my post is illustrating that compression-controlled beam sections were not acceptable per code from 318-02 through 318-11. 318-14 revised the language to allow any member with net compressive Pu above the lower bound of 0.1 f'c Ag to use a lower phi, in my experience it has been very rare for beam/slab sections to satisfy this requirement.

Good design practice (what we are supposed to have been taught and do every day) has always recommended against compression controlled design of flexural members.

Codes have always (to my knowledge) reduced the capacity of compression controlled members to make them safer by reducing the phi factor in capacity reduction type codes. Material factor codes do this automatically as the concrete factor (normally 1.5) dominates compared to the steel factor (normally 1.15).

If compression controlled, simplified calculation methods should not be used as the steel stress cannot be assumed to be fsy.

Newer higher strength steels and high strength more brittle concretes have made the old phi logic and the logic of allowing a maximum reinforcement ratio of .66 - .75 pbal unusable, so strain limits were introduced.

If you have to have a compression controlled member, it is important to have properly detailed and restrained compression reinforcement and to use strain compatibility calculations to determine actual capacity if the steel ratio is above pbal.

Codes are trying to adjust to the newer materials and normally are a little behind, but we have to accept that they need to change and account for the fact that technology might be in front of them.

And please avoid compression controlled flexural members especially in critical members. Extra concrete is a better solution.