Evaluation of process control effectiveness and diagnosis of variation in paper basis weight via multivariate time-series analysis

Multivariate time-series techniques are used to analyze the effectiveness of basis-weight control on a paper machine. Basis weight and four other process variables were collected from a production paper machine under three control conditions, ranging from no computer control to the normal operating basis-weight control strategy. Process control effectiveness is measured by comparing the observed output variation with an estimate of the theoretical minimum variation obtained from autoregressive moving-average vector (ARMAV) time-series models. To diagnose sources of variation in the process, the dynamic effects and interactions of the process variables are evaluated using the analysis of dispersion (AD) and spectral estimates obtained from the ARMAV models are used to diagnose sources of periodic variation in the process.