Recognition of Acoustic Emission Patterns from Mixed Mode Wood Fracture.

Automatic, reactive control of wood drying to maximize drying rate and minimize drying defects would be possible if the development of internal stress associated with micro and macro failure processes due to shrinkage could be detected in real time. Assuming that AE signals due to micro and macro failures during wood fracture testing are the same or similar to signals produced by drying stresses and check formation, it was decided to collect AE signals during wood fracture testing under the type of conditions of moisture content and temperature which might be found during the kiln drying process, and investigate if they may be useful in automatic, reactive kiln control. In particular, it was intended to determine if there were AE patterns associated with specific load levels leading to wood fracture which could give early warning of impending failures. AE signals and load were recorded during fracture testing of Pinus ponderosa and Quercus kelloggii. Single -edge notch tension specimens in the TL orientation were tested in mixed mode (Modes I and II) to determine if there are AE patterns associated with particular loading stages. Tests were made at three levels of temperature--20, 40, and 60 ^circC--and two levels of moisture content--12 and 18%. It was found that (a) maximum event rate increased with increasing load to maximum load and beyond, (b) temperature had a significant effect on number of events to maximum load, (c) moisture content had a significant effect on number of events to conclusion of test, (d) AE signal patterns could be successfully classified by cluster analysis and canonical discriminant analysis, (e) temperature, moisture content, and their interaction had a significant effect on features of AE signal patterns. The AE signal patterns showed very little relationship to stress levels in wood fracture. Pattern recognition of single AE signals therefore does not hold much promise for application to monitoring and control of the kiln drying process. Recognition of key features such as maximum event rate and their critical values therefore appears to be the more useful approach.