A Three-dimensional Optimal Sawing System for Small Sawmills in Central Appalachia

—A three-dimensional (3D) log sawing optimization system was developed to perform 3D log generation, opening face determination, sawing simulation, and lumber grading. Superfi cial characteristics of logs such as length, large-end and small-end diameters, and external defects were collected from local sawmills. Internal log defect positions and shapes were predicted using a USDA Forest Service model. A 3D virtual reconstruction of a log and its internal defects was generated using 3D modeling techniques. Heuristic and dynamic programming algorithms were developed for opening face determination and sawing optimization while grading procedures were programmed based on the National Hardwood Lumber Association rules. Th e system was validated through comparisons of lumber results generated by the system and by sawmills. Our preliminary results indicated that a signifi cant gain in lumber value can be achieved using this optimization system. Th is study will help small sawmill operators improve their processing performance and understand the impacts of defects on lumber grade, resulting in improved industry competitiveness. 1Graduate Research Assistant (WL), Associate Professor (JW), Division of Forestry and Natural Resources, West Virginia University, 322 Percival Hall, Morgantown, WV 26506; and Research Scientist (RET), U.S. Forest Service, Northern Research Station, 241 Mercer Springs Road, Princeton, WV 24740. JW is corresponding author: to contact, call (304) 293-7601 or email at [email protected]. INTRODUCTION Currently, hardwood sawmills in the central Appalachian region face many challenges, including decreasing log size, low-quality logs, limited availability of logs, and pressure from foreign competition (Milauskas and others 2005). In addition, a weak global economy and the housing slow down have had impacts on the hardwood products industry in the region. All of these factors are pressuring hardwood sawmills to use more effi cient processing methods that increase the value or volume of lumber produced. Studies have shown that the value of hardwood lumber can be increased by 11 to 21 percent through optimal sawing strategies gained through the ability to detect internal hardwood log defects (Sarigul and others 2001). Existing nondestructive methods (x-ray or computed tomography) do not lend themselves to fast, effi cient, and cost-eff ective analysis of logs and tree stems in the mill, especially for small sawmills. A three-dimensional (3D) laser-line scanner has been developed by the U.S. Forest Service in cooperation with Virginia Tech, and it uses relatively low-cost equipment ($30,000 plus integration labor cost) (Th omas 2006). Th e Forest Service also has developed models to predict internal defect characteristics based on external defect measurements. A recent study has shown that the models can predict approximately 81 percent of internal red oak knot defects (Th omas 2010). Log breakdown procedures must be combined with the scanned information to be successful at improving lumber value, otherwise it is not guaranteed (Occeña and others 1997). Software can generate hypothetical logs and defects, and simulate their breakdown. Over the past few decades, many log breakdown models