Through-lumber-thickness density estimation by radio frequency scanning has successfully detected knots in southern pine lumber. Density differences between knots and clear wood is believed to be the major factor allowing knot detection by this radio frequency system.

Radio frequency waves are known to be sensitive to both wood density and moisture content. Estimating lumber strength with radio frequency will require separation of the density and moisture content components of the through-wood-thickness signal. Technology to achieve this has been developed for microwave signals but not for radio frequency signals. There is little motivation to develop technology to separate these density and moisture content signal components with the wood density detection capabilities of radio frequency waves are unknown.

The objective of this study was to show that lumber density and strength can be accurately estimated with radio frequency waves if the moisture content influence in the through-lumber-thickness signal can be eliminated.

Fifty pieces of 12-foot long nominal 2 x 4-inch No. 1 grade southern yellow pine lumber were randomly selected from the dry-end production of a local sawmill. Sample lumber was straight-grained, without knots or wane. Lumber was conditioned for 6 months in a 12 percent equilibrium moisture content (EMC) environment. Just prior to scanning, one-inch wide moisture content and specific gravity sections were cut 16 inches from both ends of each lumber specimen. Through-lumber-thickness measurements were made by a radio frequency capacitor at 0.250-inch intervals to obtain 256 observations. Capacitor electrodes were 0.125 inch from lumber surfaces. MOR of each piece of lumber was determined by 3-point loading.

Thirty of the 50 lumber specimens were randomly selected to develop a prediction equation relating MOR to measured capacitor voltage. In spite of a lengthy storage in 12 percent EMC environment there was some variation in moisture content among the specimens. For this reason the actual voltage values were statistically adjusted to account for the observed influence of moisture content. MOR is also known to be influenced by moisture content. MOR was therefore regressed on both the adjusted voltage values and moisture content values. The inclusion of the moisture content variable served to correct the actual MOR values for the moisture content influence. The result of this analysis was predictive Equation 1 which had an R-square of 0.59.

MOR = - 13087 + 181.9 Adjusted voltage + 1726.6 Moisture content           Equation 1

The MOR of 20 new specimens was then estimated based on the predictive equation. The estimated MOR values for these specimens were then regressed on the actual MOR values giving an R-square value of 0.58. This R-square value is slightly higher than the 0.53 value reported between actual MOR and x-ray estimated MOR for southern yellow pine. These results indicate that estimation of southern yellow pine lumber strength by radio frequency scanning is feasible if combined with accurate measurement, and adjustment for lumber moisture content.