Autoregressive total projection to latent structures for process monitoring

A new autoregressive total projection to latent structures(AR-TPLS) is proposed in this paper, the input and output data spaces are projected to four subspaces, a principal subspace and a residual subspace generated by the predicted value of quality variables, a principal subspace and a residual subspace generated by the residual of process variables with corresponding statistics established to monitor quality variables and process variables unrelated to quality variables. The new method not only avoids the complex solving process of nonlinear iterative partial least squares algorithm (NIPALS) in projection to latent (PLS) and total projection to latent structures(T-PLS) proposed by ZHOU, but also overcomes the problem that process residual of modified PLS proposed by YIN still has large variations which are not proper to be monitored using Q-statistic. TE shows the effectiveness of the proposed method.