Study of Long-Term Pavement Performance (LTPP): Pavement DeflectionsChapter 6. Noted Anomalies and Other Potential Data
Problems
Data Screening Anomalies or Spurious Data
During the process of identifying the data errors described in
chapters 3, 4, and 5, other potential errors or anomalies were also
identified. In most cases, however, these anomalies could not be
verified or rectified; thus, they were not recommended for specific
changes in the database. These unverified errors or anomalies are
discussed in this chapter, in one of the two following
categories:
- General or global notes about the types of problems or spurious
data encountered (some specific instances are also covered
elsewhere in this report).
- Specific notes referring to each spurious day file (day
file-specific notes only).
Precautionary Notes Referring to the FWD
Load-Deflection Tables
Table 7 consists of a list of the most important, general data
anomalies noted during the course of analyzing the FWD
load-deflection data. Some of these anomalies have been addressed
in the foregoing chapters as well.
Table 7. General data anomaly notes of unchanged records
or files.
| Note 1 |
In all cases where unbound material tests (designated as lane
S* or G*) were conducted, these data have not been recommended for
changes or flags in the level E database. It should be noted that
the quality and repeatability of this data is an order of magnitude
or so poorer than the rest of the level E load-deflection data.
Therefore, unbound material test data have not been as thoroughly
screened as the remaining load-deflection data in the data
tables. |
| Note 2 |
It is possible to interchange one dedicated FWD field computer
for another, such as when there are two FWDs and corresponding
system processors in the same region. This may or may not have
occurred in practice; if so, the load and deflection readings will
be affected to an unknown degree (possibly several percentage
points), but such effects probably are not detectable using the
available screening tools. |
| Note 3 |
In many cases, the configuration number indicates PCC joint
testing positions (e.g., J4, J5, C4, or C5) as 0, 305, 305, 457,
etc.; in each of these cases the value +305 denoted for d2 should
be, or should already have been, changed to –305, corresponding to
the position of the sensor behind the FWD loading plate. |
| Note 4 |
There were many stationing errors where the plus (+) or minus
(–) sign was recorded and stored in the level E data tables
opposite of what it should have been. Most, but probably not all,
of these errors have been identified and recommended for change in
the data tables. |
| Note 5 |
In Region 3, States 35 and 40 (New Mexico and Oklahoma), in
many cases when PCC testing was conducted on different dates, there
appear to be large deflection differences in the lane 1 tests from
one test date to the next (up to a factor of 2 or 3, or more).
Quite possibly, this could be due to liftoff or slab warping
resulting from extreme thermal gradients. If so, this is a
phenomenon worthy of further investigation and, possibly, altering
testing protocols for jointed PCC pavement in some areas. |
| Note 6 |
In Region 2, on some of the SPS–5 sections denoted 27–050*, it
was noted that the variation from date-to-date in deflections was
very large. This variation may in fact have been due to an error in
section identification. However, it was not possible to determine
the actual cause of this potential set of anomalies. Therefore, the
FWD data associated with these SPS–5 sections have not been
recommended for changes or flags in the level E database. However,
such anomalies or potential site/section errors could be identified
with certainty by installing a low-cost global positioning system
on the affected FWDs for future testing. |
| Note 7 |
In Region 4, there are a large number of PCC joint efficiencies
measuring well over 100 percent; however, only those over 113
percent were noted in a separate data table. None of these has been
corrected, and no data have been removed or flagged in the level E
data tables. |
| Note 8 |
In Region 3, there are a small number of PCC joint efficiencies
measuring over 100 percent; however only those over 105 percent
were noted in a separate data table. None of these has been
corrected, and no data have been removed from the level E data
tables. |
| Note 9 |
In Region 2, it appears that there are many cases where the PCC
joint was placed between d3 and d4 instead of d1 and d3 as per
protocol, for lane J4 tests. However, only the most extreme cases
of these were noted in a separate data table. None of these has
been corrected, and no data have been removed from the level E data
tables. |
| Note 10 |
In Region 1, it appears that there are some cases where the PCC
joint was placed between d3 and d4 instead of d1 and d3, for lane
J4 tests. However, only the most extreme cases of these were noted
in a separate data table. None of these has been corrected, and no
data have been removed from the level E data tables. |
| Note 11 |
In Region 4, section 04–0502, on all six dates of testing there
was a bizarre pattern of deflection development between dates. The
FWD load-deflection table results are nevertheless possible; thus
no errors have been reported, and no data have been corrected,
removed, or flagged in the level E data tables. |
| Note 12 |
In Region 4, section 04–0509, on all seven dates of testing for
this section (intensive rehabilitation: ~50.8-mm (2-inch) overlay),
the data is somewhat peculiar. Nevertheless, since FWD testing was
possibly (but not definitely) conducted on the correct test section
before, during, and after rehabilitation given these strange
results, no errors have been reported, and no data have been
corrected, removed, or flagged in the level E data tables. |
| Note 13 |
In Region 4, section 06–b420, on all four dates of testing for
this section there were many test points (whether J1, J4, J5 or J6,
etc.) that show strange or bizarre deflections, such as
nondecreasing, too large, or too small, etc. However, not all test
points were unusual; thus no errors have been reported, and no data
have been corrected or removed from the level E data tables except
through other procedures mentioned in other sections of this
report. |
| Note 14 |
In Region 4, section 08–0214, on all three dates of testing for
this section, a plausible pattern of deflection development was
noticed between test dates, but the magnitudes of the deflection
readings were vastly different (by as much as a factor of 3). These
seem unlikely, but since the results are possible, no errors have
been reported, and no data have been corrected, removed, or flagged
in the level E data tables. |
| Note 15 |
In Region 4, section 08–0220, on all three dates of testing for
this section, a plausible pattern of deflection development was
noticed between test dates, but there were appreciable differences
between dates, especially in the deflections between stations 100
and 350. Nevertheless, no errors have been reported, and no data
have been corrected, removed, or flagged in the level E data
tables. |
| Note 16 |
In Region 4, section 08–2008, on all dates of testing for this
section, it appears that there may have been an overlay placed
between May and October 1991, but the construction number remains
the same throughout (CN = 1). This is probably (but not definitely)
an error in the level E database value of CN, which should be
considered in any subsequent data analyses conducted on the test
section. |
| Note 17 |
In a handful of cases, it is possible that lane designation
errors occurred during an entire day of testing, although this
particular category of data errors was not specifically screened.
One such instance (on two consecutive dates of test) was identified
and corrected, as discussed in chapter 5 of this report. |
| Note 18 |
In the course of screening for FWD load-deflection errors and
anomalies, it was noticed that some of the operator comments had
not been uploaded to the database, while others were recorded for
general DataPave or level E use. It is suggested that the remaining
comments be uploaded together with the deflection data as both
become available. Such comments often yield important information
about data collection problems or anomalies that can be quite
useful to the analyst. |
|
