Predicting strength characteristics of paper in real time using process parameters

Abstract

Online paper strength testing methods are currently unavailable, and papermakers have to wait for manufacture of a complete reel to assess quality. The current methodology is to test a very small sample of data (less than 0.005%) of the reel to confirm that the paper meets the specifications. This paper attempts to predict paper properties on a running paper machine so that papermakers can see the test values predicted in real time while changing various process parameters. This study was conducted at a recycled containerboard mill in Chicago using the multivariate analysis method. The program provided by Braincube was used to identify all parameters that affect strength characteristics. Nearly 1600 parameters were analyzed using a regression model to identify the major parameters that can help to predict sheet strength characteristics. The coefficients from the regression model were used with real-time data to predict sheet strength characteristics. Comparing the prediction with test results showed good correlation (95% in some cases). The process parameters identified related well to the papermaking process, thereby validating the model. If this method is used, it may be possible to predict various elastic moduli (E11, E12, E22, etc.) in the future as the next step, rather than the traditional single number “strength” tests used in the containerboard industry, such as ring crush test (RCT), corrugating medium test (CMT), and short-span compression strength test.