| NDT.net - March 2001, Vol. 6 No. 03 |
 12th International Symposium on Nondestructive Testing of Wood | Acoustic emission study of within-ring internal checking in radiata pine-Wood NDT 2000Booker, Rudolf, Haslett, Tony NScientist, Forest Research Institute Ltd. Rotorua, New Zealand Sole, John A Corresponding Author Contact: Email: Rolf.Booker@forestresearch.co.nz |
During the last eight years within-ring internal checking has become a serious problem during the drying of radiata pine sapwood boards used for appearance grade timber. A within-ring internal check extends radially through the earlywood of a single annual ring and does not penetrate through the latewood on either side. Hence its radial length is limited to the thickness of the earlywood of a single annual ring, which normally lies in the range 3 to 12 mm. The length of a check can be anything from 10 mm to the complete length of the board. These checks normally do not penetrate to the surface of the boards and hence remain invisible until re-machining occurs. Consequently boards containing within-ring checks may already incur considerable transport and manufacturing costs before the problem is detected. Before it is possible to develop drying schedules that minimise the development of within-ring internal checks, it is essential to know during what stage of drying the checking in radiata pine sapwood occurs.
A series of 14 radiata pine sapwood boards believed to be very susceptible to within-ring internal checking were sawn into 600 mm lengths (2 ft) with an average moisture content of 146%. These boards were randomly assigned to 5 different kiln charges. These kiln charges were dried for periods of 6, 12 and 20 hours (3 charges) according to a schedule of 120oC dry bulb and 70oC wet bulb with a heat up time of two hours. The boards were then sectioned to assess the severity of internal checking immediately after drying was terminated.
Four boards from the six hour charge were fitted with acoustic emission equipment prior to drying and the development of internal checking was monitored. Three pins that were an integral part of the transducer casing were pressed into the wood to achieve good contact without the need for constant force clamps.
The total number of checks that were counted after drying for 6, 12 and 20 hours was 93, 73 and 39 checks respectively. This shows that internal checking occurs mainly during the first six hours of drying. This was confirmed by the acoustic emission data that showed that almost all the internal checking occurred within the first six hours of high temperature drying, at the end of which the average moisture content was 72%. The maximum rate of checking occurred 1.6 to 2 hours after the start of drying, ie. when the kiln first reached its operating temperature and lasted for about half an hour.
It is concluded that within-ring internal checking in radiata pine sapwood begins as soon as the wood starts to dry. The checking process was practically complete when the sapwood moisture content reached 79%, which is well above the fibre saturation point of 30%. This indicates that within-ring internal checking in radiata pine is caused by water tension and not by differential shrinkage.
KeywordsIn the last eight years within-ring internal checking has become a serious problem when high temperature dried radiata pine boards are used for appearance grade timber (McConchie 1999). Figure 1 shows the photograph of a moulding machined from a board with internal checks. During drying radial checks form that are confined to a single earlywood layer between two latewood bands. Frequently only a few annual rings in a tree are severely affected by within-ring internal checks, while adjacent rings are completely unaffected. While the checks extend only through the thickness of the earlywood of a single annual ring, the checks are very clearly visible after they have been exposed by machining (Figure 1), as the earlywood is frequently 12 mm wide for radiata pine. Along the grain direction the length of the checks can vary from 10 mm to several metres. On drying within-ring internal checks only form in sapwood and not in heartwood. The checks seldom reach the board surfaces, and hence remain invisible until the boards are further processed by which time the boards may already have incurred considerable transport and re-manufacturing costs.
If a kiln schedule is too severe, several types of checking may occur in radiata pine boards after their moisture content (mc) drops below the fibre saturation point, such as surface checking, multi-ring internal checking and end-checking. These checks are caused by differential shrinkage between the surface and the core. In contrast, in eucalypts and many other hardwoods within-ring internal checking and collapse occur at the start of drying and hence well above the fibre saturation point (Tiemann, 1913). The driving mechanism for collapse and internal checking in hardwoods is water tension in the water columns in the cells (Tiemann, 1915). Booker (1994) has proposed a mechanism for the formation of internal checking in radiata pine that is similar to that for eucalypts in that it predicts that within-ring internal checking in radiata pine occurs well above the fibre saturation point.
To develop kiln schedules to reduce the incidence of internal checking in radiata pine, it was essential to determine whether drying conditions in a kiln should be made milder above or below the fibre saturation point. Consequently the following investigation was carried out to determine at what stage of drying internal checking occurs.
Fig 1:Profiled product machined from a radiata pine board containing within-ring internal checking. Removing the surfaces has exposed the checks that were originally hidden. |
A series of random width 50 mm thick boards was obtained from a sawmill that was known to have problems with within-ring internal checking. The sawmill selected boards they believed to be particularly susceptible to this problem. Fourteen of these boards were sawn into six 600 mm lengths labelled A to F. Six charges were made up from these lengths, randomly distributing the lengths A to F among the charges. (Subsequent work has confirmed this was a good precaution, as internal checking severity decreases sharply in the tree with height from the butt end). A 30 mm section was cut between all 600 mm boards to yield density and moisture content values. The 600 mm boards were end-coated with two layers of epoxy paint to reduce end-checking and weighed just prior to being placed in the kiln. Each was re-weighed immediately after coming out of the kiln.
A small experimental high temperature kiln was used for drying the charges at a schedule of 120/70oC. A heating-up time of 2 hours was selected, which is normal commercial practice. Each kiln charge was dried for one of the following periods: series A for 6 hours, series B for 12 hours, and series C, D and E for 20 hours. Series F was accidentally destroyed during a kiln malfunction. Series D and E were used to determine the effect of reconditioning on the number of checks as well as the effect of a delay between sawing and drying on the number of checks.
After removal from the kiln the boards from series A, B and C were immediately weighed and cross cut into sixteen 33 mm wide strips as shown in Figure 2. The number of checks was counted on one cross-cut surface of each strip (arrowed in Fig. 2) immediately after being sawn, ie. while the board was still hot. Strips number 5 and 10 were weighed and oven-dried for 24 hours and then reweighed to determine the moisture content immediately after drying. The moisture content of the two strips was averaged to give the board moisture content. In addition the boards were assessed for collapse.
Locan-AT acoustic emission equipment manufactured by Physical Acoustics Corporation was hired for the acoustic emission work. Its main drawback was that it could only record continuously for 6 hours.
Boards 1, 2, 3 and 4 of the six hour kiln charge were instrumented with locally made acoustic emission sensors each of which consisted of a piezoelectric element in an aluminium container with three pins for attaching the sensors to the wood (Figure 3). Each piezo-electric element is a Philips PXE5 disc, 25 mm in diameter and 10 mm high, surmounted by a 7 mm high lead inertia member that reduced the resonance frequency from 212 kHz to about 89 kHz. This design ensured excellent coupling between the wood and the transducer at all temperatures without the use of coupling jellies or constant pressure clips. The transducer signals were carried outside the kiln by co-axial cable and amplified by B&W preamplifiers with AC75L filters with a bandpass range of 45 to 90 kHz.
Fig 2: Diagram showing the way the boards were sectioned. The arrows indicate the surfaces on which the checks were counted.
| 
Fig 3: The acoustic emission transducers with pins to hold them to the wood surface |
Results of sectioning experiments
Table 1 indicates the annual ring orientation (flat sawn F or quarter sawn Q), basic density, green % saturation, as well as % moisture content of the boards before and after the three periods of drying. Table 2 shows the checking severity. The
number of checks in Table 2 is the total number of checks counted on the faces
of strips 4 to 14 in Figure 2. This is a measure of the checking in the interior of the boards, and excludes the strips in the outer 100 mm that may contain end-checks. After drying for 20 hours, charge C was immediately sawn into strips for assessment, charge E was steam reconditioned before being stored for 5 weeks, while charge D was stored for 5 weeks before being sectioned.
Column 6 in Table 1 shows the % saturation. When the % saturation is greater than 87% this indicates that the board consists of sapwood only. The high % saturation and green moisture content values indicate that all the boards consisted primarily of sapwood.
Clearly the boards that were most susceptible to internal checking are boards number 1, 2, 6, 8 and 14, while boards 4, 5, 10, 12 and 13 were least susceptible. By comparing the densities of these boards it is clear that basic density is no indicator of the propensity to check. This agrees with earlier results reported by Miller et al (1992).
The collapse severity was assessed and ranges from 0 (none) to 4 (severe)[hese are relative terms. What would be called severe collapse in radiata pine may well be called low to medium if it occurred in some of the eucalypt species.] (Table 2). It is clear that collapse can also be a problem during the high temperature drying of radiata pine. For 8 out of 42 board sections in the first 3 charges the cells were strong enough to resist both internal checking and collapse. For the others there seem to be two competing tendencies. No collapse occurred in board sections that had more than 12 checks. For the six board sections that had both severe (grade 4) collapse and some internal checking the average number of checks was only 3.3. There is a tendency for collapse and internal checking to compete with each other, when internal checking is high collapse tends to be low and vice versa. This agrees with the mechanisms proposed by Booker (1994).
| Board number Charge: | Saw pattern | Density, kg/m3 | Width | Green % sat | Moisture content | |||||
| value | Stdev | mm | value | Stdev | green | 6 hrs | 12 hrs | 20 hrs | ||
| A | B | C | ||||||||
| 1 | F | 418 | 5 | 417 | 93 | 1 | 161 | 81.3 | 18.3 | 3.8 |
| 2 | Q | 440 | 9 | 192 | 78 | 4 | 125 | 60.4 | 17.4 | 7.0 |
| 3 | Q | 395 | 8 | 182 | 91 | 2 | 170 | 89.8 | 20.8 | 5.1 |
| 4 | Q | 388 | 9 | 184 | 86 | 4 | 164 | 65.1 | 20.5 | 10.4 |
| 5 | Q | 428 | 45 | 168 | 90 | 3 | 153 | 66.4 | 18.0 | 5.8 |
| 6 | F | 414 | 5 | 474 | 93 | 2 | 162 | 85.2 | 22.7 | 4.3 |
| 7 | Q | 413 | 16 | 194 | 67 | 11 | 117 | 61.7 | 28.0 | 9.5 |
| 8 | Q | 456 | 7 | 164 | 77 | 3 | 118 | 71.3 | 21.7 | 7.3 |
| 9 | Q | 404 | 5 | 168 | 90 | 4 | 163 | 86.4 | 36.9 | 10.8 |
| 10 | F | 406 | 12 | 207 | 60 | 13 | 108 | 45.1 | 9.9 | 1.6 |
| 11 | Q | 387 | 13 | 186 | 83 | 4 | 159 | 86.1 | 26.4 | 8.0 |
| 12 | Q | 402 | 4 | 171 | 91 | 2 | 165 | 76.5 | 33.5 | 6.3 |
| 13 | F | 410 | 7 | 167 | 89 | 4 | 157 | 70.3 | 16.2 | 1.6 |
| 14 | Q | 443 | 25 | 174 | 84 | 4 | 128 | 59.8 | 28.1 | 5.7 |
| Av | 415 | 84 | 146 | 72 | 23 | 6 | ||||
| St.dev | 21 | 11 | 24 | 13 | 7 | 3 | ||||
| Table 1: Moisture content after different drying times | ||||||||||
The average board moisture content after six hours of drying was 72±13%. Within-ring internal checking occurs only in the sapwood, while the board average mc is based on boards that also contained some heartwood. A better sapwood average mc after 6 hours of drying is obtained by averaging only the sapwood boards, in other words boards with a % saturation greater than 87%. This gives a sapwood moisture content after 6 hours of drying of 79±9%. The total number of checks after drying for 6 hours was 93. So after 6 hours of drying 93 checks had occurred in the sapwood while the sapwood moisture content was still very high at 79%. This shows that for radiata pine internal checking occurs well above the fibre saturation point of 30% mc.
For the three charges that were dried for 6, 12 and 20 hours and immediately sawn into sections the total number of checks was 93, 67 and 39 respectively. There are two possibilities for this decrease; either some of the checks closed up after drying below the fibre saturation point and became impossible to see, or it is a statistical aberration introduced by the fact that internal checking is extremely variable. While all boards in the five charges were end-matched, it is now known that internal checking decreases rapidly with height in the tree (McConchie and Booker, unpublished). The boards were randomly assigned to the different charges to counteract this effect. Charge E was dried for 20 hours, steam reconditioned to relieve the stresses in the wood and then sectioned five weeks later, while charge D was stored for 5 weeks without being reconditioned. Reconditioning is the recommended treatment for preventing the development of checks after drying by eliminating built-in drying stresses. Yet the total number of checks in the reconditioned boards was 147, while in the non-reconditioned boards of charge D the total number of checks was only about half that at 76. Out of the 147 checks 145 occurred in only two boards. It is concluded that because of the small number of specimens and the variability of internal checking no conclusions can be drawn regarding variations between the number of checks for the five treatments. To do so a much larger number of specimens would be required. However, as the number of checks after 6 hours is similar to that for the longer drying times, it is clear that most if not all the internal checking occurs during the first six hours of drying. The acoustic emission results below confirm this, and will pin-point a half hour period starting from shortly before the time the kiln reaches its operating temperature as the most active checking period.
| Board number Charge: | Sawing pattern | Number of checks after: | Collapse severity | ||||||
| 6 hrs | 12 hrs | 20 hrs | 6 hrs | 12 hrs | 20 hrs | ||||
| A | B | C | D | E | A | B | C | ||
| 1 | F | 9 | 20 | 6 | 30 | 68 | 0 | 0 | 0 |
| 2 | Q | 22 | 12 | 11 | 4 | 0 | 0 | 2 | 4 |
| 3 | Q | 0 | 7 | 0 | 0 | 0 | 0 | 0 | 4 |
| 4 | Q | 0 | 0 | 0 | 0 | 0 | 2 | 4 | 4 |
| 5 | Q | 0 | 0 | 0 | 0 | 0 | 2 | 0 | 1 |
| 6 | F | 32 | 6 | 1 | 34 | 77 | 0 | 0 | 0 |
| 7 | Q | 0 | 4 | 2 | 0 | 0 | 2 | 0 | 0 |
| 8 | Q | 10 | 8 | 3 | 3 | 0 | 0 | 2 | 2 |
| 9 | Q | 1 | 0 | 0 | 0 | 0 | 2 | 0 | 0 |
| 10 | F | 6 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| 11 | Q | 0 | 1 | 4 | 2 | 1 | 2 | 4 | 4 |
| 12 | Q | 0 | 0 | 0 | 0 | 0 | 2 | 2 | 4 |
| 13 | F | 1 | 0 | 0 | 0 | 0 | 0 | 0 | 0 |
| 14 | Q | 12 | 9 | 12 | 3 | 1 | 0 | 0 | 2 |
| Total: | 93 | 67 | 39 | 76 | 147 | 10 | 14 | 25 | |
| Average: | 8 | 6 | 4 | 7 | 13 | ||||
| St dev. | 10 | 6 | 4 | 13 | 29 | ||||
| Table 2: Checking severity after drying | |||||||||
Results of Acoustic emission studies
Four boards from the six hour charge were monitored with acoustic emission equipment, boards 1, 2, 3 and 4. The graphs of accumulated acoustic emission hits versus time for boards 1 to 3 are shown in Figures 4 to 6. Note that the scales for the three graphs are very different, as board 1 was much more active than boards 2 and 3. The rate of emissions for boards 1 and 2 is shown in Figures 7 and 8. The AE rate for board 3 was so low as to be meaningless.
Fig 4: Graph of total acoustic emission events versus time (seconds) for board 1.
| 
Fig 5: Graph of total acoustic emission events versus time (seconds) for board 2. |
Fig 6: Graph of total acoustic emission events versus time (seconds) for board 3.
|
The number of checks in the boards and the number of AE hits are compared in Table 3. No acoustic emissions were detected in board 4, and on sectioning it was found this board had no internal checks and no end checks. In board 3 no checks were detected either, which does not mean that none occurred in the wood between the sawcuts that are 33 mm apart. It is very probable that the 22 AE hits were caused by a small check that remained undetected.
Considerable internal checking activity occurred in boards 1 and 2. This was accompanied by end-checking as well, in spite of the double layer of epoxy paint to prevent it. The same is true for the other boards in Table 2; boards with many internal checks had a tendency to have many end-checks and vice versa.
The 1607 hits in channel 1 were generated by 9 internal checks in the interior 400 mm of board 1, or 24 checks in the total 600 mm length if we include the 15 checks occurring in the outer 100 mm on each side. However, as the sensors were placed halfway between the board ends, they should be much more sensitive to the true internal checks than to the end-checks.
| Board number | No. of interior checks | Total no. of checks | No. of AE hits |
| 1 | 9 | 24 | 1607 |
| 2 | 22 | 59 | 220 |
| 3 | 0 | 0 | 22 |
| 4 | 0 | 0 | 0 |
| Table 3: Comparison of acoustic emission and sectioning results | |||
In contrast with the 1607 hits on channel 1, channel 2 that was connected to board 2 registered only 220 hits. These were generated by 22 checks in the interior or 59 checks in the total length of board 2. No attempt was made to equalise the sensitivity of the AE channels, and it is very probable that the difference in the number of hits per check for boards 1 and 2 was caused by differences in channel sensitivity.
None of these considerations affect the rate of generation of AE hits as a function of time, as the sensitivity of a given channel does not change with time. Figures 4 to 6 show that the accumulated number of AE events started to rise at a slow rate from the moment the kiln was switched on. Then 1.5 to 1.6 hours later the accumulated number of events suddenly started to rise linearly for 0.5 hours, after which the rate of accumulation decreased. After 6 hours the rate of event accumulation in board 1 was very low, and was about 95% complete, while in board 2 AE activity had ceased. This is shown more clearly in Figures 7 and 8 which show graphs of the rate of generation of AEs versus time for boards 1 and 2. While for board 1 the rate of AE hits was largest 1.9 hours after the start, board 2 had two maxima at 1.6 and 1.8 hours, and board 3 had a maximum at 2.0 hours. This means that internal checking started at a low rate from the moment the kiln was switched on. When the kiln was close to reaching its set point the number of checks increased linearly for half an hour, after which it decreased until it was practically complete after six hours of drying.
Fig 7: Acoustic emission rate versus time (seconds) for board 1.
| 
Fig 8: Acoustic emission rate versus time (seconds) for board 2. |
Two concurrent studies were carried out to determine at what stage of drying within-ring internal checking occurs. Kiln charges were dried for periods of 6, 12 and 20 hours in a kiln with a heating-up time of two hours. The boards were immediately cut into strips without waiting for them to cool down, and the number of checks were counted. This showed that internal checking occurs primarily within the first six hours of drying, when the wood is still well above the fibre saturation point.
Four of the boards were fitted with acoustic emission sensors. These showed that within-ring internal checking began as soon as the kiln began to heat up. Checking activity was at its maximum for a period of half an hour starting shortly before the time the kiln reached operating temperature. Checking activity then decreased and after six hours in the kiln at least 95% of the total number of checks had formed.
These results support the theory by Booker (1994) that within-ring internal checking is caused by tension in the liquid water columns in the cells during drying. The first few hours of drying in a kiln are the most critical for the development of within-ring internal checking. Hence the initial drying conditions are the most critical for the development of milder kiln schedules to reduce within-ring-internal checking.
| © NDT.net - info@ndt.net | |Top| |