0968-A2

Effect of high temperature drying on bending and shear strengths of western hemlock lumber

Massamba Thiam ¹ , Dr. Michael R Milota ² , Dr. Bob Leichti ²

Abstract

Three hundred pieces of N°2 and Better nominal 42 by 145 mm (nominal 2 by 6 inches) by 2400 mm western hemlock, dimension lumber was tested in bending and shear after drying at conventional (82°C) and high (116°C) temperatures. The shear strength was reduced by 6.4% and there was no significant effect on the moduli of elasticity or rupture. The higher drying temperature resulted in a drying time of approximately 24 hours, 50 percent less that the conventional temperature.

Introduction

High temperature kiln drying of lumber is done at a dry-bulb temperature that exceeds 100°C. In this seasoning method, the humidity in the kiln may or may not be controlled. Drying at a high temperature is one way to reduce manufacturing costs. Quality, however must be maintained because the wood costs are a significant proportion of the total cost of manufacturing. Quality is often measured by the grade recovery, particularly for visually-graded lumber. However, the intrinsic strength is given in the National Design Specification for a certain grade and species. Any strength loss due to the high temperature drying process will go undetected and the wood will be used on the predetermined strength values.

It is known that exposure to high temperature , particularly at high moisture contents, affect the strength of wood (7). The impact of exposure to high temperature on the end use is not as straight forward for dimension lumber.

The objectives of this study were to compare the effect of high-temperature drying (116°C) to conventional temperature drying (82°C) on the modulus of elasticity and the mean and fifth percentile bending strength of 42 by 145 mm, N°2 and Better, western hemlock, dimension lumber. The mean and fifth percentile shear strength parallel to the grain of small, clear, specimens were also determined.

Materials and Methods

Three hundred pieces of N°2 and Better, nominal 42 by 145 mm, 2,4 m long, western hemlock (Tsuga heterophylla) lumber were randomly selected at local sawmill. The actual green size of each piece was approximately 42 mm by 147 mm. Each set of 100 pieces was a replication.

The modulus of elasticity was non destructively determined for each piece using a metriguard model 340 E-computer. Within each replication, the 100 pieces were sorted in ascending order by modulus of elasticity and every other piece was assigned to one of the two treatments. Thus the specimens were distributed by stiffness between the high and conventional temperature drying treatments.

Stacking and loading

The lumber was stacked on 19mm stickers spaced at 0,61m. Six hundred and fifty kg of concrete (2kPa) were placed on the top of the load to reduce bow, twist, and crook. After a warm up of 2 hours, constant drying conditions were held for the remainder of the schedule, 82°C dry-bulb and 66°C wet-bulb for the conventional treatment and 116°C dry-bulb and 90°C wet-bulb for the high temperature treatment . The air velocity was 3.6m/s. Moisture content was checked with a Delmhorst RDX-1 conductance-type wood moisture meter on 20 specimens near the end of drying. The kiln was stopped when the average moisture content was 15%.

After kiln drying, the dry lumber was cooled 2 to 4 hours with the weight on. Then small specimens were placed in a chamber maintained at 21°C and 65% relative humidity with a fan for air circulation for six weeks until practical equilibrium was reached. Six samples were randomly selected from each replication and weighed every other day to verify achievement of equilibrium.

Wood properties

Before testing, the following measurements were made on the dry lumber. Pieces that did not meet the grade characteristics permitted by Western Lumber Grading rules (13) for slope of grain, annual rings, knots sizes for N°2 and Better lumber were eliminated from the study.

The length (± 1.3 mm), width, and thickness (± 0.5 mm) of each piece were measured according to ASTM D 4761 (6). Each piece was weighed (± 2.25 g). Slope of the grain was measured using a scribe over a 0.3 mm length (horizontal distance) in an area of the board representative of the general slope of the fibers. Lumber having a slope greater than 1 in 8 was eliminated from the study.

Growth rate was assessed by placing a tape measure on the end of the piece and counting the number of rings over a 25 mm distance. Pieces having less than 1.6 rings per centimetre (4 per inch) were eliminated from the study. The angle made by the annual rings relative to the face of the board (± 5°) and the distance to the pith (± 3 mm) were measured by matching the rings on a clear template to the rings of the piece. A grain angle of 0° indicates that flat sawn lumber and 90° indicates quarter sawn. The radius of curvature of the arc on the template that match the ring at the center of the piece was the distance to the pith.

Knots were measured according to the Western Lumber Grading Rules (13) to make sure sizes did not exceed 48 mm at the wide face edge and 73 mm at the wide face centreline. Pieces having major defects or having a knot greater in size than allowed by the grading rules for N°2 and Better lumber were eliminated from the study

Testing procedures

A Tinius Olsen universal testing machine was used to conduct a four-point bending test in accordance with ASTM D 4761 (6).

A small, clear, specimen near the failure was cut from each piece of lumber tested in bending for the shear parallel to the grain test. The size of the shear parallel to grain specimen was 50 by 38 by 63 mm to produce failure on 50 by 50 mm surface in accordance to ASTM D 143 (1). An Instrom testing machine was used to apply the load at a rate of 0.6 mm/min.

A second clear specimen was removed from near the failure of each piece of lumber after full size bending tests were completed. Moisture content on an oven dry basis and specific gravity at 12% moisture content were determined in accordance with ASTM Standard D4442-92 method B (5) and ASTM D 2395 (4), respectively. Similarly, the moisture content and specific gravity of each shear specimen were determined.

Analysis

The moduli of elasticity, E (Pa), and rupture, MOR (Pa) were determined using equations 1 and 2, respectively (2),

E = S *(L ³ / 4.7 bh ³ )

MOR = PL/bh ²

where,

S = slope of the load-deflection data as determined by a linear regression on the data between 15 and 30% of the maximum load (N/mm),
P = maximum load (N)
L= span of beam (m),
b = width of beam (m), and h= depth of beam (m).

where A = area of the shear plane (m ² ) and F_u = ultimate shearing load (N).

The modulus of elasticity, modulus of rupture, and shear strength of each sample were adjusted to 12% moisture content in accordance with ASTM D 1990 (3).

Results and Discussion

Table 1 shows the average moisture content taken with a conductance-type wood moisture meter at the end of drying and the drying time. The charges dried at a conventional temperature took between 46 and 48 hours whereas the charges dried at a high temperature took between 22 and 24 hours. The time savings between the two schedules is approximately 24 hours.

TABLE 1. Moisture content s(%) and drying time (hr) for each kiln charge

Wood properties

From the 300 specimens at the beginning of the study, 42 were eliminated because they did not meet the minimum requirements for N°2 and lumber given by the lumber grading rules(13). Table 2 shows number of specimens tested in static bending shear per treatment and per replication.

TABLE 2. Number of specimens included in the static bending and shear results

The mean values and coefficients of variations of annual rings per inch, slope of grain, grain angle, distance from pith, specific gravity and moisture content of each treatment at time of bending test are summarized in Table 3. After coming to practical equilibrium, the moisture content of the wood dried at high temperature was 0.7 % lower than that dried at conventional temperature. High temperature drying results in a permanent reduction of hygroscopicity.

TABLE 3. Mean values of physical properties and p-values showing no differences between the treatments.

Bending

Bending failures between the load points represented 87 % of the total failure modes. Combined bending and shear failures, which occurred between the load points and on one of the shear spans, represented about 9% of the total number of failures. Shear failure, which occurred in the clear portion of the wood represented 3 % of the total number of failures. Among the 7 specimens that failed in shear, five received the high and two the low temperature treatment. This failure mode normally does not occur in this type of test. A probable explanation might be the presence of shake in the tree ring failure during drying, the presence of which would reduce the resistance to shear.

Adjusting the bending properties for moisture content to 12% from 13.2% or 12.8% resulted in increase in the strength values of 1 to 3 %.

The mean value for modulus of elasticity was reduced by 1.2% from conventional to high temperature drying (Table 4); however, the analysis of variance indicated that this was not significant (p-value of 0.287). Similarly, it cannot be concluded that drying temperature had an effect on MOR (p=0.64) despite a reduction in the mean value of MOR by 4.7% from conventional to high temperature drying.

To specify the allowable design load it is important to know the characteristic value, an intermediate value in the development of allowable unit stress and stiffness values. For modulus of elasticity, the characteristic value is the same as the mean value. For modulus of rupture, it is calculated by using the 5% lower exclusion limit using the coefficient in Table 3 of reference 5. for wood dried at the conventional temperature, the characteristic value for MOR is 13.04 MPA whereas it is 12.01 MPA for wood dried at high temperature. These 5% lower exclusion limits were from means that were not statistically different between treatments allowing us to state that there is no significant difference between characteristic values for western hemlock dried at conventional and high temperatures.

Compared to other studies, the coefficient of variation of MOE is similar to small clear values published in the Wood Handbook (12) whereas the high variability of MOR is close to the one published from N°2 Douglas-fir lumber (10). The strength loss results are in general agreement with Kozlik (9), except that the strength loss is less severe. These bending data also confirm the work done by Salamon (11) on the effect of drying temperature up to 110°C on the strength of 0.9 m-long pieces of western hemlock.

TABLE 4. Summary of MOE and MOR for conventional and high temperatures.

Shear parallel to the grain

The data in Table 5 indicate that neither average specific gravity or moisture content at the time of the shear test were significantly different between the levels of kiln temperature. The 0.7% difference in moisture content between the equilibrated samples again demonstrates the effect of drying temperature on hygroscopicity, similar to that observed for the bending specimens.

The moisture content adjustment to 12% led to an increase of 1.2% and 0.2% in the shear strength for the wood dried at the conventional and high temperatures, respectively. Tests of skewness and kurtosis indicated that the data were normally distributed.

TABLE 5. Means values of physical properties and p-values showing no differences between the treatments for the shear specimens

The analysis of variance on the shear strength values adjusted to 12% moisture content (Table 6) indicates evidence of a difference between mean shear strengths of wood dried at conventional temperature and high temperature (p= 0.0001). the mean shear strength parallel to the grain is estimated to be between 0.50 MPa and 0.72 MPa greater in conventional than in high temperatures (95% confidence interval).

Similar to bending, the characteristic values, determined by using the 5% lower exclusion limit, are 7.08 Mpa and 6.18 Mpa (Table 6) for the conventional and high temperature treatments respectively. The characteristic values are significantly different since there were derived from the means values that were significantly different. From a practical standpoint, these results are important when wood is loaded in shear such as in strut or bolt connections or short deep beams. The coefficient of variation for shear strength found in this study (Table 6) is close to the published value (8, 12). The demonstration of reduced shear strength in this study clarifies the results of Kozlik (8) in which only one of two analyses showed a loss in shear strength and he suggested refining the testing procedures.

TABLE 6. Summary for shear strength for conventional and high temperatures.

Conclusions

The kiln drying time for western hemlock at 116°C is 50% less than at 82°C at an air circulation rate of 3.6m/s. The high temperature drying may lower the moisture content of the wood at equilibrium compared to conventional drying; however, this was not statistically significant.

The mean values for moduli of elasticity and rupture as determined using full size pieces of N°2 & Better western hemlock lumber were not significantly different between wood dried at 116°C and 82°C.

There is a significant reduction of 6.4% in the average maximum shear strength between pieces dried at 82°C and 116°C.

Further research on the effect of high temperature drying on mechanical properties will be more generally applicable if it is possible to use a larger sample size and to integrate more variables such as additional geographic sources of wood, sizes of lumber, and grade of lumber.

Further research on shear perpendicular to the grain, oblique shear, and horizontal shear will be helpful since forces that produces shearing stresses are classified according to the direction in which they act.

Bibliography

1. American Society for Testing and Materials.1995. Standard methods of testing small clear specimens of timber. ASTM D 143-94, Philadelphia, Pa. USA.

2. American Society for Testing and Materials.1995. Standard methods of static tests of lumber in structural sizes. ASTM D 198-94, Philadelphia, Pa. USA.

3. American Society for Testing and Materials.1995. Standard practice for establishing allowable properties for visually-graded dimension lumber from In-grade tests of full-size specimens. ASTM D 1990-91, Philadelphia, Pa. USA.

4. American Society for Testing and Materials.1995. Standard test methods for specific gravity of wood and wood-base materials. ASTM D 2395-93, Philadelphia, Pa. USA.

5. American Society for Testing and Materials.1995. Standard test methods for direct moisture content measurement of wood and wood-base materials. ASTM D 4442-92, Philadelphia, Pa. USA.

6. American Society for Testing and Materials.1995. Standard test methods for mechanical properties of lumber and wood-base structural. ASTM D 4761-93, Philadelphia, Pa. USA.

7. Tsoumis, G. 1991. Science and technology of wood: Structure, properties, utilization. Chapman and Hall, New York, USA.

8. Kozlik, C.J, .1967. Effect of Kiln condition on the strength of Douglas fir and Western hemlock. Report D-9, Forest Research Laboratory, Oregon State University 32 pp. USA

9. Kozlik, C.J. 1968. Effect of kiln temperature on the strength of douglas-fir and western hemlock dimension lumber. Report D-11. Forest Research Laboratory. Oregon State University, Corvallis, Oregon, USA. 20pp.

10. Leichti, R.J. and V. Eskelsen. 1992. A comparison of North American and European community test procedures for the assignment of characteristic values to lumber in structural sizes. Final Report. Limited distribution Report, Department of Forest Products, Oregon State University, Corvallis, Oregon. USA

11. Salamon, M. 1965. Effect of high temperature drying on quality and strength of western hemlock. Forest Products Journal. 15 (3): 122-126. Wisconsin. USA.

12. Wood handbook: Wood as an engineering material. 1987. USDA Forest Service, Forest Products Laboratory, Madison, Wisconsin.USA.

13. Western Wood Products Association. 1998. Western Lumber Grading Rules. Portland, Oregon. USA(13)

¹ Forest Engineer, Research Assistant. Institut Sénégalais de Recherches Agricoles / Centre National de Recherches Forestières. Route des Pères Maristes Hann BP 2312 Dakar- Sénégal
Email : [email protected]
[email protected]

² Associate Professors. Department of Wood Science and Engineering. Oregon State University