Statistical process control in food processing

Abstract

When used effectively, statistical process control can be a powerful tool for process improvement. Procter and Gamble's plant at Portsmouth, VA, produces Jif™ peanut butter. Level 2 monitoring is done using a MODCOMP II/45 Classic™ computer running SETCON™ and SETCON-SPC™. Real-time statistical process control was installed at the Portsmouth facility in 1987. Pareto analysis led to decreased unscheduled downtime. A capability study led to peanuts being roasted within specifications and with less variability. Studying XBAR/R control charts after normal process start-ups led to process modifications that resulted in better control of salt content in the final product. This article focuses on improvements made to the roasting and blending/grinding parts of the process. Peanuts are roasted and fed into a surge tank. From there, the peanuts are fed into grinders. Salt and other minor ingredients also are fed into the grinders. Each ingredient is fed through its own chute. All ingredients are ground into a paste (peanut butter), which is held in a second surge tank until the paste can be fed to the packing line.