Lognormal control charts for moisture content of kiln-dried lumber.

摘要

This paper focuses on the application of statistical process control principles to monitor the lumber kiln-drying process through the use of innovative quality control charts. Three Lognormal control charts are proposed to monitor quality characteristics that follow a three-parameter Lognormal statistical distribution. The first two control charts, called the "scale chart" and the "chart for geometric means," monitor the central tendency of the process. The third chart, called the "shape chart," monitors the process variability. Practical procedures are presented for calculating centre lines and control limits, and for plotting the data on the charts. A rationale is given for using geometric means rather than arithmetic means for assessing process' central tendency. The choice of parameters to be monitored on control charts, along with parameter estimation issues, are discussed. A succinct comparison with the customary "Normal" charts is also included. The methods presented were tested on a data set of Douglas-fir (Pseudotsuga menziesii) lumber collected from a production facility in British Columbia, Canada, for which the statistical distribution of moisture content measurements was determined to be well modelled by a three-parameter Lognormal distribution.