Important Warning

Occasionally a dynamic sensor may respond incorrectly or erratically during the course of one or more traffic load repetitions. This may be due to damage sustained in the pavement or to a variety of other causes. For example, a crack may have propagated through the concrete at the gage position and the gage response may change significantly from one load repetition to the next. Or, a gage failing due to chafing of the cable in one part of the structure may pull the output of the power supply down intermittently and affect the output from gages in other parts of the structure. When the output of a particular gage is noted to be incorrect the normal procedure is to completely disconnect the gage from the data acquisition system and ground the input to the A/D converter. However, there is usually a delay between a gage going bad and identifying that it has gone bad. Also, the second type of incorrect response (pulling down the power supply) typically occurs at isolated load offsets during a wander pattern and completely disconnecting a gage due to this problem can result in the loss of valuable data. Consequently, the problem may deliberately not be corrected immediately after having been identified. The database therefore contains records which are not correct for all or part of their lengths. In the interests of making the data available as soon as possible, the user is expected to examine the data for inconsistencies before completing the analysis of individual gage responses and reporting the results. This should not be too difficult to do because the positioning of the load is extremely precise and, when a gage is working correctly, the response in both magnitude and shape is very repeatable from one wander pattern to the next (or, indeed, over a fairly large number of wander patterns relative to the number of wander patterns to failure). One strategy for analysing the data could therefore be to examine the data in groups of 66 repetitions (every complete wander pattern). On the other hand, an unfortunate side effect of having bad records in the database is that searching all records for the peak value of a particular gage is almost certain to return an incorrect value.

We believe that the database contains a wealth of data previously unavailable, requiring just a little care on the part of the user in order to provide quality information for analysis of the test item responses during the traffic tests.

Please help us to maintain this database!  Report any errors or anomalies to Dr. David R. Brill.

Last Update: 04/14/03