Reducing lumber thickness variation using real-time statistical process control

A technology feasibility study for reducing lumber thickness variation was conducted from April 2001 until March 2002 at two sawmills located in the southern U.S. A real-time statistical process control (SPC) system was developed that featured Wonderwareo human machine interface technology (HMI) with distributed realtime control charts for all sawing centers and management offices. The thickness data were distributed to all PCs at the instant of measurement by way of wireless transmitters attached to calipers. There was statistical evidence in the study that suggested that the average thickness of 4/4 ash (Frainus caroliniana) declined by 0.060" and the average thickness of 4/4 cottonwood (Populus deltoider) declined by approximately 0.035" at Mill #2. Lumber produced at the resaw machine center generally had the largest standard deviation for all species at both sawmills. The large variation at the resaw was the most limiting factor for further target size reductions. The average overrun for Mill #1 for all species increased by approximately 2.6 percent after the installation of the real-time SPC thickness improvement system. Unfortunately, a change in log grading procedures confounded the overmn results at Mill #2 so that the impact of the SPC implementation on lumber recovery could not be determined. The average ccCommon and Bettery7 lumber grade recovery increased 3 percent at Mill #1 after installation of the real-time SPC system due to a reduction in wedge-shaped, undersized lumber. The average daily recovery of "Common and Better" lumber at Mill #2 increased by 4.4 percent after installation of the real-time SPC system. The estimated annual financial cost savings realized by the two mills combined was approximately