Concrete mix design submittal procedure - comments please? 3

[Dennis59]

Hello.
I know this subject has been talked about somewhat from time to time. Still, I have had such pathetic experiences with getting contractors/concrete suppliers to submit what I feel is a "proper and complete" mix design submittal, that I am compelled to bring this up again.

I am a structural design engineer in the Upper Midwest, USA. I have dealt with ready mix plants and contractors large and small, and it seems that few of them know what is expected of them when it comes to submitting a concrete mix design - in spite of the fact that ACI 301-99 and 318-02 both contain flowcharts that make it pretty easy.

I would appreciate it if any of you good folks who have time and interest would read through my verbal explanations and see if you think I am on track. For what it's worth, I can not remember the last time a concrete submittal came to my desk that followed the flowchart and included all the information I needed in order to approve or reject the mix(es). And when I call the contractors and/or ready mix plants (who should KNOW ACI!), they seem essentially clueless as to what I'm asking for and why I need it. To most of them, it seems that I should be satisfied with a one-page list of ingredients and Joe's signature at the bottom stating that the mix 'should meet the requirements of the job'. (and sometimes "Joe" is a P.E.)


Here's the way I read the ACI:

Per ACI 318-02 and 301-99, there are 2 steps to getting the mix we need/want.

Step 1: Step 1 includes calculating a standard deviation for one particular batch plant - the one that will be producing concrete for the job. Probably only one standard deviation (but maybe more if there is more than one specified f'c for the job) will be needed. Then knowing the standard deviation for the batch plant, calculate an f'cr for each and every proposed mix. In essence, based entirely on the past performance of the actual concrete plant that will be producing concrete for the job in question, determine just how high we have to shoot in order to get concrete that will probably pass most of the compressive strength tests. This is f'cr. In other words, for the same exact ingredients, a poorly operating and inconsistent plant (say one with a poorly functioning scale) will have a wider scatter of break strengths and thus a higher standard deviation. This will force that poorly-operating plant to shoot for a higher average strength than a plant that produces more consistent concrete. A standard deviation calculated for some plant OTHER than the one that will be producing concrete for the job is MEANINGLESS to Step 1.


Step 2: Design each mix to meet or exceed the f'cr value that is pre-calculated for the mix.



Step 1 involves "standard deviation test records". Step 2 may also involve test records, but the test records for Step 1 and 2 need not be for the same concrete. That's the part that everyone gets wrong, I think. Something else that we sometimes get wrong, I think, is believing that 'required average strength' is somehow determined by trial batches. This is not correct. 'Required average strength' is determined before trial batches are ever initiated.

For Step 1, all we need is to find out how consistently the batch plant has produced concrete in the past - ANY concrete that was 'sort of similar' to the concrete we're trying to design. That is done by calculating a standard deviation for the compressive strength test records of that 'sort of similar' mix and then saying that this represents the consistency of the plant itself, sort of. ACI only requires a few things of the "standard deviation test records". They need to be for concrete that is within 1000 psi of the concrete we want, and they need to be from concrete that was produced with similar materials and quality controls as what we're expecting for the concrete to be produced. Per ACI 301, these tests should be from within the last year, and encompass a time period of at least 60 days. ACI 318 doesn't seem to care about the time period. Practical considerations would dictate that they be relatively recent. (We're after a fair picture of how consistently the plant is operating today, not 10 months ago.)

30 tests are required in order to accurately calculate the standard deviation of the plant. If there are less than 30 tests available, but more than 15, then the calculated "s" needs to be amplified. If there are less than 15 tests (which I find unimaginable), then forget about "s" and calculate f'cr with Table 4.2.3.3.b (ACI 301-99) or Table 5.3.2.2 (ACI 318-02). (That's the Readers Digest Condensed Version - really have to read through several paragraphs of ACI to get all the specifics. ACI 301 is worded differently than 318 as far as the number of tests is concerned, but in the end, they both really are the same - I find it confusing.)

(And here is another point - a "test" is always the average of two 28-day breaks. I have received many submittals with only one 28-day break for each 'test', and the contractor is angry with me when I explain that this is not a valid "test" per ACI's definition.)

Once we've calculated "s" (or decided that there is not enough info to calculate "s"), then we calculate f'cr.
If we have "s", we can calculate f'cr with either 4.2.3.3.a (ACI 301) or with Table 5.3.2.1 (ACI 318). ACI 318 differs slightly from ACI 301 in that 318 has a special equation for f'cr if f'c is over 5,000 psi.

If we don't have "s", then use Table 4.2.3.3.b (ACI 301) or Table 5.3.2.2 (ACI 318) to calculate f'cr for whatever mix we're trying to design. The tables are identical for f'c<=5,000 psi, but differ for f'c >5,000 psi.

STEP 1 Example 1: Two mixes are specified for the job - a 4,000 psi mix and a 3,000 psi mix. The actual ready mix plant that will be supplying concrete has 30 test records for a mix that was designed for 3,000 psi, and was produced with similar quality controls as we're expecting in the job in question. The standard deviation of those 30 test records is "s"=550 psi.

Since the design f'c of the "standard deviation test records" is within 1000 psi of both of our new mixes, we can use the same "s" to calculate both f'cr values.

For our 3,000 psi mix, f'cr = f'c + 1.34s = 3000 + 1.34*550 = 3737 psi.
OR
For our 3,000 psi mix, f'cr = f'c + 2.33s - 500 = 3000 + 2.33*550 - 500 = 3781 psi. <<Controls, since we have to use the larger

For our 4,000 psi mix, f'cr = 4000 + 1.34*550 = 4737 psi.
OR
For our 4,000 psi mix, f'cr = 4000 + 2.33*550 - 500 = 4781 psi. <<Controls, since we have to use the larger



STEP 1 Example 2: The same 2 mixes are spec'd, but all they can find to submit is 20 tests from a 3,000 psi mix, and the "s" is 750 psi.

Since the design f'c of the "standard deviation test records" is within 1000 psi of both of our new mixes, we can use the same "s" to calculate both f'cr values. "S" needs to be amplified since there are less than 30 tests. Amplifier is 1.08.

For our 3,000 psi mix, f'cr = f'c + 1.34*1.08s = 3000 + 1.34*1.08*750 = 4085 psi.
OR
For our 3,000 psi mix, f'cr = f'c + 2.33*1.08s - 500 = 3000 + 2.33*1.08*750 - 500 = 4387 psi. <<Controls, since we have to use the larger

For our 4,000 psi mix, f'cr = 4000 + 1.34*1.08*750 = 5085 psi.
OR
For our 4,000 psi mix, f'cr = 4000 + 2.33*1.08*750 - 500 = 5387 psi. <<Controls, since we have to use the larger

>>> Here is a question for an ACI expert: For STEP 1 Example 2, the f'cr calculated with "s" is larger than if the ready mix plant had NO data at all! I have exaggerated and used a very bad (high) value of "s". In a case like this, would it be permissible to use the 'data not available' tables and just add 1,200 to f'c? <<<


NOW (and only now) we can go to Step 2 and see if the contractor is proposing mixes that will satisfy the spec. No matter whether he uses field test records or whether he actually produces trial batches, the average strength he needs to prove is f'cr (as just calculated above) - NOT f'c.
For what it's worth, I do not remember EVER dealing with a contractor who understood that f'cr is the number to meet.


If the contractor wants to use past history test records, all he needs is 10 tests, provided that they're all from a certain short period of time (318 says 45 days, and 301 says 60 days). There is no reason for him to submit more than 30 tests. Obviously, the importance of a large number of tests is much less for this part of the process than it was in the calculation of standard deviation. Other than 'quantity of tests', though, the other requirements about 'similar materials and quality controls as what we're expecting for the concrete to be produced' are still applicable.

If the ready mix plant can't come up with 10 tests, trial batches are required.
Sometimes, the test records that were used to establish "s" could also be used to qualify the mix in question. But where the standard deviation is large, I think it is unlikely that the average strength will also be large enough to meet f'cr. And if the average is less than f'cr, we have to reject it.

STEP 2 Example 1: For 3000 psi mix, Contractor hopes to use same test records that he submitted for standard deviation, and average f'c is 3,400 psi. He's thinking he has it made, but this does NOT meet f'cr of 3781.

For 4,000 psi mix, he submits 15 records from a similar mix, and average f'c is 4,800. It's better than 4781, so OK.


STEP 2 Example 2: Contractor submits same as above - and both mixes need to be rejected. Even though the average strengths seem fine, the variation in the plant is so bad that it is very likely that some tests will not meet design f'c.



Is it just me or do you folks also have trouble getting contractors and concrete suppliers to understand all this?

Am I understanding it right for that matter?

Any comments will be appreciated.

Thanks for reading my ranting novel.

 

[henri2]

In my neck of the woods, major ready mix companies have technical experts who thoroughly comprehend ACI 211, ACI 301, ACI 318, job specifications, and applicable local code provisions for mix designs; in addition, these individuals have a good feel for the way a concrete mixture will perform in terms of strength and workability as they are very familiar with source of materials and fresh/hardened properties of concrete mixtures based on these materials.

I had posed a question regarding mix submittal practice. Check this thread

http://www.eng-tips.com/viewthread.cfm?qid=141432

ACI 318 recognizes 3 mix design methods. Each approach has several steps. The three methods are:

1. Field experience--- relevant sections are 5.3.1 through 5.3.3.1. This approach is based on statistics/standard deviation and is the most desirable. The flow chart you refer to is shown in Figure R5.3

2. Trial Mixtures---relevant section is 5.3.3.2. Section 5.3.3.2(f) indicates that the f'cr by this method shall be based on Table 5.3.2.2. Field or laboratory trial batches will satisfy this criteria unless the enforcing jurisdiction has other provisions. At least three trial batches must be evaluated with this method. If memory serves me correctly, trial laboratory trial batches must be performed over three consecutive days and average compressive strength test results are based on the average of three individual test results from a sample. I think ASTM C 192 deals with this. I do not have a copy so please verify.

3. Proportioning without field experience or trial mixtures---Section 5.4. RDP approval is required and the f'c shall not exceed 5,000 psi. In the 95 and previous editions of ACI 318, an empiral approach was utilized that was based on a table which provided a relationship between f'c and w/c ratio

Many mix design submittals from ready mix companies and CMT labs are just 1 page and will include information such as:

-Name of ready mix company and plant address
-Project address
-Date
- Name of purchaser
-Source of materials and applicable ASTM designations
-Mix design designation
-Aggregate gradation (individual and combined)
-Mix proportions based on 1 cu.yd SSD (i.e. weights of materials including admixture dosage rate and absolute volume of ingredients including air)
- MSA, f'c, f'cr, slump, unit weight, strength test results (actual or based on interpolation), s (if applicable), w/c ratio
- method of placement (pump or regular tailgate placement etc)
- location in structure
-Mix design methodology (field experience, trial batches etc..with appropriate code and or spec section)

Many companies will attach back up data to validate the 1 page mix design submital..usually when stipulated by job specifications.

Some agencies have provisions for mix design submittals...and indeed, a one page submmital may satisfy it. An example is the Los Angeles City Dept of Bldg and Safety

http://www.ladbs.org/faq/info bulletins/building code/IB-P-BC 2002-062 Concrete Mix Design.pdf

 

[GeoPaveTraffic]

Dennis59,

While I'm not an ACI expert by any stretch, your description of the process is essentially complete. I also cannot get "good" submittals from contractors, however, the suppliers have figured out how high they must aim to meet our requirements based on past experience. That may also be the case in your area, don't know. In the end, get the right concrete is more important than how the suppliers figure out how high f'cr should be. I find it disappointing that they have so little control of the mix supplied to the job that the apparent f'cr is 1000 psi above the required f'c, but that appears to be the state of the practice here in the mid mid west.

As a side note, depending on how the specification is written, the number of samples taken during the job can have a large impact on f'cr. I'm in the process of evaluating several years worth of data for paving projects where we have 40 to 80 tests per contract. Because our specifications are written as below, f'cr increases as the number of test taken increases. I'm currently trying to rewrite our spec to eliminate, or at least reduce, this impact.

If any compressive strength is deficient from the ultimate compressive strength required (4000 psi), the Contractor will be required to reimburse the City for the deficiencies in strength as noted in the following schedule. A deduction will be made to the contract unit price for “Removal and Replacement of P. C. C. Pavement” for that area of pavement represented by the failing test. The deduction shall be a linear deduction varying from no (0) deduction at 4,000 psi to seventy-five percent (75%) at 3,000 psi and one-hundred percent (100%) deduction for any test result less than 3,000 psi. Alternately, as determined by the Engineer, concrete represented by strengths below 3,000 psi shall be removed and replaced at the sole cost of the Contractor. The amounts due as penalties under this section shall be immediately withheld from any sums due the Contractor as soon as the extent of the deficiencies are determined.

Deductions for deficient concrete strength will only be imposed if any one of the following conditions is met:

1) Any two (2) sets fail to exceed a twenty-eight (28) day strength of 3,250 psi.

2) Any three (3) sets fail to exceed a twenty-eight (28) day strength of 3,500 psi.

3) Any five (5) sets fail to exceed a twenty-eight (28) day strength of 3,750 psi.

4) Any ten (10) sets fail to exceed a twenty-eight (28) day strength of 3,900 psi.

 

[DRC1]

I have never heard of anybody doing this. Most plants do not keep data on mix designs, esp. custom mixes. Specifications require compliance to f'c which is what most (if not all) inspecting engineers require. No plant is going to go through this exercise to quote this material and unless it is a large order, they would not be interested in it to provide it for approval. Most concrete mixes break at 500 psi or more over design strength. If I am concerned about quality and consistancy, I use a standard DOT mix from a DOT approved plant. These mixes and plants have a long history of testing and acceptance. It is a readily available mix and does not delay the approval process. Remember your contractor wants to build a job, not develop and test new mix designs.

 

[GeoPaveTraffic]

DRC1,

You are right that it is a lot of work, however, it is also what ACI requires. So if you are going to design by ACI... That said, the plant can analize the data they have on standard mix designs and submitt them for approval. A plant that goes through this process may very well not need to have 500 psi more strength in the mix than is required and therefore will be able to produce an acceptable product for less money. That is a situation in which everyone wins.

 

[BigH]

In the US and Canada - Europe too, I suppose - you have this luxury of ready mix plants having standard mixes. We go through a process every time on our jobs overseas - whether in China, Laos, India, Indonesia. Every job I've worked on for the last decade or so have had no ready mix plant within reach nor mixes with history. Consider yourselves all very lucky.