The use of acoustic emission in improving hardwood timber seasoning productivity

The Tasmanian sawmilling industry typically does not season eucalypt material specifically for furniture applications because of its highly refractory seasoning characteristics. Boards cut with wide faces parallel to the growth rings ('backsawn or 'flatsawn' boards), in particular, shrink significantly with drying and often experience surface checking (seasoning 'degrade'). It is standard practice to season timber batches under schedules designed to produce relatively high-grade structural material and sell the small proportion of boards that do not suffer seasoning degrade as furniture grade material. The aim of this study was to investigate optimum seasoning of backsawn material specifically for furniture applications to take advantage of the premium prices paid by furniture manufacturers for degrade-free material. Acoustic emission is a well known non-destructive testing tool. Acoustic emission results from stress waves generated by irreversible stress release activity such as the growth of small cracks in a solid material under load. Acoustic emission is typically used in homogeneous materials such as plastics and metals to determine the structural integrity of components in situ. Acoustic emission was first measured in timber subject to external load over thirty years ago and has subsequently been measured in drying timber specimens where differential drying rates between the surface and centre of a sample may cause appreciable drying stresses. Acoustic emission appeared to be related to the severity of the drying conditions and the development of surface checking (Becker 1982, Noguchi et al. 1987). Under harsh drying regimes, surface checking occurred and a significantly higher AE rate was measured than in the same species subject to mild drying conditions. The 'AE rate' is a measure of the occurrence of high-energy bursts of acoustic waves during a particular time-period - the higher the AE rate, the more energetic bursts are measured. On this basis, it appeared feasible to measure acoustic emission in sample boards in a kiln during drying and adjust the drying conditions as the AE rate approached levels previously determined to correspond to surface checking. Various AE-based kiln control systems were reported in the literature (Honeycutt et al. 1985, Noguchi et al. 1987) but it appears acoustic emission was not satisfactorily related to surface check development. These control systems controlled the drying conditions based on arbitrary AE rates that were nominally selected as 'safe' without optimising the drying process. The heterogeneous structure of timber posed significant obstacles to forming a clear picture of the stress release processes occurring in timber. Acoustic emission measured in drying Tasmanian eucalypt boards was employed in this project as an online measure of the magnitude of the stresses during drying. Numerous batches of eucalypt boards were obtained from a sawmill during the project and dried under harsh conditions to induce surface checking. It was determined that the AE rate generated at the onset of surface checking in backsawn and quartersawn boards was effectively constant (within the variability of the material). This 'AE checking threshold' was independent of basic density and clearly independent of bulk stiffness (since the bulk tangential stiffness is effectively half the radial stiffness in the species studied). It appeared that the acoustic emission phenomenon was independent of the timber bulk material properties. It is proposed that acoustic emission waves propagate from irreversible slips or dislocations in the crystalline cellulose regions of cell walls under high stress. These slips are sudden, energetic stress release events which propagate stress waves from the local site. The amorphous regions of the cell structure behave in a rather plastic manner with local stress expected to be consumed by ductile flow processes that do not generate elastic waves. Dinwoodie (1968) reported the existence of such crystalline slips in spruce timber under external compression. Siau (1984) reported that the proportion by mass of cellulose in normal wood (as opposed to tension and compression wood) was remarkably constant and apparently independent of species. Further, the proportion of crystalline cellulose is also relatively constant (Siau 1984). This appears to be directly related to the remarkably constant AE rate measured at the onset of surface checking in this study. This finding led to the measurement of acoustic emission in other species. It was hoped that the same AE 'rate would be measured at the onset of checking in different woods. Acoustic emission was measured in backsawn radiata pine (Pinus radiata) and myrtle (Nothofagus cunninghamii). The acoustic emission measured in radiata pine was significantly different from the acoustic emission measured in the eucalypt material, with relatively 'massive' AE rates detected in boards free of seasoning degrade. This behaviour was attributed to brittle failure in the resin canals, structural elements not present in the eucalypt genus. The characteristic AE responses measured in drying myrtle boards resembled the behaviour measured in the eucalypt boards, apparently due to the more similar structure. Surface checking was detected at approximately the same AE rate as that measured in the eucalypt boards. Much of the improved understanding of the acoustic emission phenomenon developed in this study was facilitated by the existence of a one-dimensional nonlinear drying model developed by Oliver (1991). Oliver wrote KILNSCHED (KILN SCHEDULING PROGRAM), a computer program based on this model, which simulates the drying behaviour of a single board, with arbitrary bulk material properties, subject to arbitrary drying conditions. KILNSCHED is particularly suited to low temperature drying of eucalypt materials. Kiln drying trials quickly revealed that 'green' Tasmanian eucalypt boards invariably suffered surface checking within 24 hours when subject to drying temperatures as low as 23°C dry bulb temperature and 21°C wet bulb temperature (at 0.5 m/s airspeed). Such temperatures are significantly lower than those often employed in drying material direct off-saw in the timber industry but none-the-less are considered 'harsh' in this study. At the commencement of this study, KILNSCHED was used in a purely predictive mode to assess drying schedules prior to drying. The author of this thesis modified KILNSCHED to simulate drying using the real-time drying temperatures measured in the experimental kiln. This allowed measured acoustic emission during drying to be compared with the drying behaviour calculated with KILNSCHED. This modification revealed that Tasmanian eucalypt timber is far more sensitive to small temperature fluctuations than was previously expected. The reader must keep this material sensitivity to temperature and temperature change in mind at all times when reading this thesis. The AE rate measured during drying was successfully related to the instantaneous strain at a board surface calculated with KILNSCHED using the measured drying conditions. Instantaneous strain is the strain component employed as the failure criterion in the drying model. This enabled the author to place considerable confidence in behaviour calculated with KILNSCHED and the 'optimum drying' program SMARTKILN discussed below. This author modified KILNSCHED to incorporate an optimisation algorithm that determined the optimum drying conditions required to dry the timber in the minimum time at a preset arbritrary maximum surface instantaneous strain. The resulting program, SMARTKILN, develops drying schedules to maintain the calculated surface instantaneous strain at a preset 'Control Strain' below the ultimate surface instantaneous strain. Together, SMARTKILN and acoustic emission measurement form the basis of Clever Kiln Controller®, a kiln control system to dry Tasmanian eucalypt timber in the minimum time with minimum degrade. In Clever Kiln Controller, SMARTKILN uses datalogged real time drying conditions to simulate the drying behaviour of a sample board in the kiln. Calculated drying behaviour is continuously compared with measured drying behaviour (measured AE rate and moisture profiles measured by regularly slicing sample boards). Provided the calculated and measured drying behaviour are satisfactorily matched, the optimum drying schedule developed by SMARTKILN is applied to the kiln. When the AE rate approaches the AE checking threshold, the drying conditions are automatically ameliorated to prevent surface checking. Subsequently, SMARTKILN develops a refined optimum schedule to incorporate the new datalogged drying conditions. This study has developed the understanding of acoustic emission from the existing (often misleading) information recorded in the literature to a level that could be incorporated with current knowledge of the behaviour of drying timber. The thesis describes how this was achieved. Various aspects of the understanding have been reported in papers published, in press or under review. The final result is the development of a commercial kiln controller which is described and already implemented in some selected experimental kilns at present restricted to eucalypt materials