Engineering Brief # 34A

May 13, 2002

Subject: INFORMATION : Engineering Brief No. 34A,

"Referee Testing of Hardened Portland Cement
Concrete Pavement--Percent Within Limits Revision"

From: Manager, Airport Engineering Division, AAS-100

To: All Regions

ATTN: Airports Division Managers

Engineering Brief No. 34A, "Referee Testing of Hardened Portland Cement

Concrete Pavement--Percent Within Limits Revision", copy attached,
suggests a process to use to substantiate the strength of portland cement
concrete pavement when the results of normal flexural strength acceptance
tests are suspect and when the acceptance criteria is based on percentage
of materials within specification limits (PWL).

The conduct of referee tests should not be done on a carte-blanche basis.

Referee testing should only be considered when there is evidence to
confirm that the standard flexural strength test results are not
representative of the in-place material. The suggested method is a
revision to Engineering Brief No. 34, which was originally used by the U.
S. Navy, and subsequently by the Federal Aviation Administration (FAA).
Engineering Brief 34 as previously published is now cancelled.

Devising a plan for referee testing is not the responsibility of FAA.

However, any plan must be approved by the FAA to establish whether the
results of the referee testing may be used to determine pay factors for
the material in question when Airport Improvement Program rules apply.
The program to establish all of the criteria for random sampling,
testing, and acceptance should be developed and accepted by all parties
prior to physical sampling and testing.

Rick Marinelli

Attachment
 ENGINEERING BRIEF NO. 34A

REFEREE TESTING OF HARDENED

PORTLAND CEMENT CONCRETE PAVEMENT
PERCENT WITHIN SPECIFICATION LIMITS (PWL) REVISION

Purpose

The purpose of this Engineering Brief is to present a method of testing

hardened portland cement concrete pavement when doubt exists as to the
validity of the standard flexural strength test results and when the
acceptance criteria is based on percentage of materials within
specification limits (PWL). The method does not purport to give an exact
strength, but rather, a strength that is consistent with FAA design and
acceptance assumptions. This method is not to be routinely used in lieu
of standard flexural strength acceptance testing.

Background

The standard test for acceptance of portland cement concrete strength is

ASTM C 78, Flexural Strength of Concrete (using simple beam with
third-point loading). Item P-5Ol, Portland Cement Concrete Pavement,
provides guidance on the interpretation of test results and pay factors
for flexural strength acceptance testing. Sometimes disputes arise as to
the validity of the flexural strength test results. Improper specimen
preparation or curing, mishandling of test beams, improper testing
techniques, uncalibrated testing machines, etc. are often cited as
reasons for questioning the flexural strength test results.

Process

The process described below is suggested and can be used to establish the
flexural strength of hardened concrete with reasonable confidence and
applied to the PWL acceptance criteria for portland cement concrete
pavement. The process contains information regarding test method,
testing and acceptance program, and application. The process assumes
that acceptance testing, presumed to reflect the in-situ product, has
been completed prior to the possibility of referee testing.

1. Test Method. When it can be established that the validity of

flexural acceptance test results is not sufficient, the hardened concrete
may be cored, moisture conditioned, and tested to determine the splitting
tensile strength in accordance with ASTM C 42, Obtaining and Testing
Drilled Cores and Sawed Beams of Concrete.

The tensile split test is recommended as a referee test since the failure
mode is somewhat similar to flexural failure. A correlation between
compressive strength and flexural strength could be used, however it is
felt that the difference in the failure modes between compressive and
flexural failure make compression testing less attractive as a referee
test. ASTM C 42 is recommended because it provides for proper moisture
conditioning prior to testing in accordance with the applicable
provisions of ASTM C 496, Splitting Tensile Strength of Cylindrical
Concrete Specimens.

2. Testing and Acceptance Program. Before initiating physical sampling

and testing, the scope of the testing program must be developed and
accepted by all involved parties and approved by the FAA. This program
becomes a binding agreement between the involved parties and should
include the following items:

a. Basis for Retesting. The project contract documents provide an

industry recognized procedure of acceptance testing. The acceptance
testing procedure contains provisions and safeguards that customarily
produce results representative of the product placed in the field. If a
party to the contract can provide concise, persuasive evidence that the
acceptance testing procedures did not produce results indicative of the
field product, such evidence would be the basis for referee testing.

Prior to retesting, it is important and required to determine why the

original acceptance testing is incorrect. The reason the original tests
are incorrect can greatly influence the method and scope of the referee
testing program. There are hypothesis tests to determine if two sets of
data represent the same population (F-test) and if the two sets of data
represent the same average level of production (t-test). If the test
data from acceptable lots and from questioned lots are compared using
these hypothesis tests, then a 5 percent significance level is
recommended.

b. Cost for Referee Testing. Only the original acceptance testing

is eligible as a project cost. In the event referee tests are necessary,
the additional cost for establishment of the referee testing program, the
physical sampling and testing, and concrete repair at sampling locations,
are not eligible for Federal reimbursement. The cost associated with
referee testing will usually become the responsibility of the party
causing the need for additional testing. It is recommended that the
party(ies) responsible for the costs be identified in the referee testing
program. Regardless, no cost associated with referee testing is eligible
for Federal reimbursement.

c. Test Methods. Since they are not addressed in the official

project documentation, all necessary sampling and testing procedures need
to be identified. Test methods for items such as specimen extraction,
handling, transporting, testing, etc. must be identified prior to
physical testing. ASTM D 3665, Random Sampling of Construction
Materials, and ASTM C 42, Obtaining and Testing Drilled Cores and Sawed
Beams of Concrete, should be identified at a minimum.

d. Testing Frequencies, Sampling, and Coring. In most cases,

rejection of the standard flexural test results leaves little or no
information available for the lot or sublot in question. As a result,
referee testing must reproduce the statistical sampling program
originally specified in the project documents. The required number of
referee test specimens is often heavily debated due to cost of testing,
pavement closure, and damage to the pavement. As a general rule, the
sampling and test frequencies established in the original document should
be maintained. Sampling locations should be determined in accordance
with ASTM D 3665.

Only one core sample should be drilled from each sublot in accordance
with ASTM C 42.

It may be possible to obtain more than one test specimen from a core. No
more than two specimens per core, meeting the minimum height requirements
 of ASTM C 42, should be tested, and the average of the two tests should
be used as the strength result.

Core diameter should conform to the following:

For plain concrete, core diameter should be 5.75 inches +/-
0.25 inches.
For reinforced concrete, core diameter may be reduced to 4
inches +0.25 / -0.0 inches, to avoid inclusion of the
reinforcement steel in the core samples.

e. Estimating Factors. . The recommended correlation between

splitting tensile strength and flexural strength is given in Equation 1
as follows

Flexural strength (psi.) = 1.02 x Splitting tensile strength (psi.)

+ 117 psi. Eqn. 1

Equation 1 can be used with 85 percent confidence that the estimated

flexural strength is at least as strong as the strength derived from the
original regression analysis detailed in CONCRETE STRENGTH RELATIONSHIPS,
Miscellaneous Report Number S-74-30, published by the Army Engineer
Waterways Experiment Station, December 1974.

f. Factors Affecting the Strength of Samples. The expression in

paragraph 2e above, does not address the complication that various
factors can affect the strength of the samples. Items such as age of the
specimen, size of the specimen, fly ash or slag additives that result in
later age strength gains, may require numerical adjustments to the
physical results. There are several possible schemes to estimate the
strength of older specimens with the age required by project
specifications. An accepted industry standard may be applied to adjust
the samples or the program may include additional testing of previously
unquestioned concrete for the project to provide a relative comparison.

As an alternate to a detailed investigation into the factors affecting

the later age strength gain, a method is presented below that applies a
variable confidence interval to the original regression over a time
period. Table 1 below varies the confidence interval from 85 percent to
95 percent over a seven month period. The table shows three age-strength
intervals.

TABLE 1 FLEXURAL STRENGTH ESTIMATE - AGE AND CONFIDENCE INTERVAL

In-Place Flexural Strength (MR) Estimate Confidence
Concrete Age where T = Splitting Tensile Strength from Interval
(in days) ASTM C 496 (percent)
up to 60 days MR (psi.) = 1.02 x T + 117 85 percent
60 to 276 days MR (psi.) = 1.02 x T + 117 - 0.25 x (Age 85 to 95 percent
- 60)
over 276 days MR (psi.) = 1.02 x T + 63 95 percent
Note: For a given strength, the estimates assumes a gradual 54 psi. strength
gain after the first 60 days.
The three intervals assume a gradual strength gain with time after
initial curing. For the time period between the 28-day strength and 60
days in place, no increase in the confidence interval is applied and no
strength increase is applied. After 60 days in place and up to nine
months in place, the confidence interval is gradually increased from
85 percent to 95 percent. After nine months in place, no additional
strength increase is applied. For example, if a pavement has a splitting
tensile strength of 523 psi. at 60 days, it correlates to 650 psi
flexural strength. At 150 days, the above estimates assumes a splitting
tensile strength gain of 22.5 psi higher and the 545.6 psi. strength at
150 days would continue to correlate to a flexural strength of 650 psi.
at 60 days At nine months and beyond, the total splitting tensile
strength gain assumed by the expression is 54 psi.

The flexural strengths derived using the agreed to correlation expression

may be used in accordance with pay adjustment schedule guidance contained
in the project documents, with no additional adjustments, in lieu of the
original questionable flexural strength tests for projects evaluated
using PWL concepts

g. Combining Referee Tests with Original Flexural Tests. It is

not advisable to combine referee tests with the original flexural tests
since different test methods are used.

3. Application. The conduct of referee tests should not be done on a

carte-blanche basis. Reasonable doubt as to the validity of the normal
flexural strength acceptance tests must be established. Normally, it is
the responsibility of the engineer, testing lab or contractor to propose
a method of referee testing to either prove or disprove the flexural
strength results.

Originally Signed By
Rodney N. Joel, P.E.
Civil Engineer/Airfield Pavements