An experimental and analytical study of the process of fatigue crack growth and coalescence in the region of a notch has been performed. The experimental program consisted of constant amplitude fatigue tests of 24 steel bar specimens. Each bar had a 135° vee notch machined across the middle.

An important part of the experimental program was the monitoring of the fatigue cracks by the Alternating Current Potential Drop (ACPD) method. To better understand this technique, a preliminary part of the test program was a study of the electric field induced in the specimen by the ACPD equipment. This study showed that the field is not uniform. This has led to the development of a data collection probe array that compensates for this lack of uniformity and yields crack growth measurements that are accurate even before being calibrated with visible beachmarks after the completion of the test.

The results of the experimental program are examined and compared to existing design codes that prescribe a method for dealing with adjacent cracks. It has been found in the present study that pre-coalescence interaction between adjacent cracks is not evident and, in fact, that coalescence itself is not an instantaneous process as the design codes suggest. Accordingly, a new crack shape development model has been developed that describes the coalescence process and accounts for the fact that the specimen has a finite width and the crack will eventually break through the edges and grow as a straight fronted crack. It has been concluded that the codes are over-conservative and are not realistic in the manner in which they treat the problem of adjacent fatigue cracks.

The experimental results were compared to computer predictions of fatigue crack growth using software that performs an integration of the Paris crack growth law and incorporates the new crack shape development and crack coalescence mode. The predictions were found to agree reasonably well with the lives actually achieved by the test specimens.