基于双取样点的立锅间歇蒸煮过程卡伯值软测量模型

建立了一个基于双取样点的间歇蒸煮过程卡伯值软测量模型.该模型同时利用了大量脱木素起始阶段和残余脱木素起始阶段两个取样点的有效碱浓度,降低了有效碱浓度测量误差对卡伯值预测精度的影响,提高了模型的鲁棒性,比其他单独利用初始有效碱浓度或某个过程取样点有效碱浓度的模型更能反映蒸煮过程卡伯值的变化.通过用工厂实际数据对模型的检验,新模型的优点得到了证实.

A New Kappa Number Soft Sensing Model in Batch Kraft Pulping Process

Present models used in cooking process include only effective alkaline (EA) of a sample point, so the EA error contributes much to the model errors. This paper derives a new model in which the EAs at the beginning of bulk delignification phase and residual delignifi- cation phase are used as the information. The application of double sample points reduces the impact of EA error and improves the robust of the model. The model can better predict the change of kappa number. The test with data from a pulp mill testifies the merits of the model.