The acoustic resonant analysis method is an interesting technique for material and structural integrity testing in on-line quality control. At this time only a few applications have been developed for objects with uniform geometry and homogeneous material properties. Often the resonant frequency of a fundamental mode is compared with a certain threshold value and an accept/reject classification is made. No more information is available about the seriousness of the mechanical damage.
This project extended the possibilities of the technique towards the classification of objects with more variable shape and material properties. This progress was achieved by the development of new signal algorithms. These algorithms were based on the information extraction of the whole frequency spectrum of the response signal and on the intelligent use of multiple measurements on one object. More detailed information was generated about the type of damage that is observed.
The technical feasibility study was performed on two test objects, namely chicken eggs and vitrified clay pipes.
- About 5 to 15% of all eggs produced have eggshell defects. These eggs must be sorted out at the moment of grading and packaging. Traditionally this is done by human visual inspection. Higher sorting speeds (>3 eggs per second on each lane of a 12 rows conveyor) require automated detection systems. Currently developed systems are too expensive when applied to low capacity machines (<8000 eggs/hour) and to next generation very high capacity machines (>120 000 eggs/hour). This high cost is mainly due to the high number of measurements per egg (24 hits/egg). The current crack detection rate of about 80-85% must also improve to 90-95%. Based on the new developed technology it proved that two to three measurements per egg can guarantee the required accuracy. No mechanical and electronic limitations were expected for real time measurement and analysis.
- In the case of the vitrified clay pipes, invisible hairline cracks create a lot of problems with leackage of sewage water. The end-user in this project has an annual cost of damage claims of 375 KEuro. Traditional quality control is based on visual and acoustic (tapping the pipe) inspection by operators. A technical feasibility study performed on a limited number of intact and defective pipes indicated that both information of the resonant frequencies and damping properties have to be used in order to detect these defects. The measuring and analysis time was much larger than for eggs (20 seconds/pipe) and hence more sophisticated algorithms can be used.
This newly developed technology was expected to have a number of applications in terms of the automated quality inspection of rigid structures in a production environment.