A windows-mean multi-way partial-least squares（MPLS）method was proposed to build the strip steel quality prediction model for continuous annealing process.The proposed method can deal with the problem of large time ...A windows-mean multi-way partial-least squares（MPLS）method was proposed to build the strip steel quality prediction model for continuous annealing process.The proposed method can deal with the problem of large time lag in hardness measurement of the strip steel and the difficulties in building mechanism models.Based on the average trajectories of the process variable information,the proposed method was used to realize the strip steel quality prediction.Comparing with the traditional MPLS,the proposed method can eliminate the abundant and redundant process information and reduce the number of the model input variables.Moreover,it can avoid the uneven-length modeling data problem of the traditional MPLS method.The simulation result demonstrates the performance and advantages of the proposed method for predicting the strip steel hardness.展开更多
基金Item Sponsored by National Natural Science Foundation of China(60774068)National Basic Research Program of China(2009CB320601)
文摘A windows-mean multi-way partial-least squares（MPLS）method was proposed to build the strip steel quality prediction model for continuous annealing process.The proposed method can deal with the problem of large time lag in hardness measurement of the strip steel and the difficulties in building mechanism models.Based on the average trajectories of the process variable information,the proposed method was used to realize the strip steel quality prediction.Comparing with the traditional MPLS,the proposed method can eliminate the abundant and redundant process information and reduce the number of the model input variables.Moreover,it can avoid the uneven-length modeling data problem of the traditional MPLS method.The simulation result demonstrates the performance and advantages of the proposed method for predicting the strip steel hardness.
