Professor Ehsani's development of PileMedic for repair of columns started in the early 1990s where he and his associates at the University of Arizona received a research grant from the U.S. National Science Foundation to study retrofit of bridge columns with FRP. A total of 15 specimens were constructed and tested. The findings of that study were fully detailed in the following three journal articles; it is recommended that those who desire an in-depth understanding of the subject, read these articles that are available on this website:
Testing of Deficient Columns
To simulate and understand the behavior of "older" and poorly-designed bridge columns, concrete columns were constructed and tested. As shown below, each column was subjected to an axial load of 100,000 pounds. hen the top of the columns was subjected to reverse cyclic loading to simulate the earthquake motion. The graph below shows how these specimens fail. After a couple of loading cycles, the column can no longer resist the applied load; this is shown from the graph where the maximum load at the end of each cycle is less than that at the end of the previous cycle. .
During these tests, columns suffer significant damage at their base (where the column is connected to the footing). The photo shows the column after it has failed; it is evident that the longitudinal steel bars are buckled and are no longer straight; this would indicate that the column had become slightly shorter by the end of the test!
Not withstanding the shortening of the column, we were interested in showing if such a failed column could be saved by wrapping it with FRP jackets. Such a repair, if successful, could be useful in determining the faith of columns in buildings and bridges that are damaged during an earthquake. Engineers are often faced with deciding whether a structure can be repaired and safely put back in service or if it has to be totally abandoned and demolished.
Repair of Damaged Columns & Retesting
All columns in the study were repaired and retested. Repair consisted of removing any loose concrete and packing the void with a concrete mix. Since full compaction of concrete would be difficult to achieve under such circumstances, there would be voids within the concrete at this stage. As described later, these would be filled with an epoxy resin.
In the early 1990s since we had not developed the PipeMedic laminates, we manufactured a somewhat similar but inferior product as shown in the figure on the left below. Unidirectional glass fabric in 6-inch wide bands were saturated with epoxy resin and wrapped several turns around a mandrel representing the shape and size of the column. Mylar sheets were used between the layers to prevent bonding of the saturated fabric to itself. The mandrel and fabric were placed in an oven and cured. Once removed, the result was the laminates that are shown in the photo below.The laminates were wrapped around the column by brushing a layer of epoxy between the layers to create a multi-layered system. Typically four 6-inch wide bands were epoxy-bonded to the lower 2 feet of the column. Then as shown epoxy resin was injected into the deteriorated column to fill the voids in the concrete. These columns were subsequently tested. For the column described earlier, the results of the testing after repairs are shown below. The hysteretic (load vs. displacement) response of the repaired specimen is shown in green line. For ease of comparison, the response of the specimen before the repairs that was shown above is repeated in dashed line here. As can be seen, the repaired specimen continues to carry more load with each new cycle of loading. In fact, testing had to be stopped because we reached the maximum pull and push displacement that the testing equipment could provide (+/- 5 inches) without being able to fail this specimen! This test clearly demonstrated the effectiveness of this system for repair of damaged column -- making them much stronger than their original undamaged strength.
Testing of Retrofitted Columns
As a part of this study, we also tested a specimen identical to the control one. But in this case, the specimen was retrofitted with FRP bands in a similar manner and tested without any prior testing and damage. This hysteretic response is shown in red. Again, for comparison the response of the control specimen is shown in dashed line. It is clear that the retrofitted specimen also performs very well -- as expected.