Statistical Process Control Fundamentals in Packaging Substrate Converting
Statistical process control in substrate converting relies on sub-grouping across machine direction variance and isolating non-normal barrier defects.
Non-Normal Distributions
Data Asymmetry
Statistical datasets that do not follow the classic symmetrical bell curve are common when measuring physical paper and board properties. These non-normal distributions often occur in measurements of tear resistance, Cobb value, or surface roughness, where the data has a natural lower limit of zero or is skewed by process boundaries. Recognizing this asymmetry is necessary for selecting the correct statistical analysis tools.
Quality Analysis
Applying standard control limits to skewed data leads to incorrect conclusions about the stability of the board manufacturing process. For example, if a quality engineer assumes a normal distribution for a skewed variable, the system may generate frequent false alarms or fail to detect actual process drifts. To prevent this, data must be transformed using mathematical operations such as the Box-Cox method to stabilize the variance and ensure that control charts accurately reflect the performance of the mill.
Process Control
Correct analysis of skewed data ensures that process adjustments are only made when a genuine change has occurred. This prevents unnecessary adjustments to the headbox or the starch press, which could introduce more variability into the paperboard. This statistical discipline is essential for maintaining the high uniformity required by modern converter plants.