Identification of critical situations in cement production cycle

2015. №4, pp. 281-291

Kuzenkov A.N. - post-graduate student, Mordovia State University, 39 B. Khmelnitskiy St., Saransk, 430005, Russian Federation,

Author made the analysis of the classifier data for the developed system of quality control (QC) based on Shewhart control charts. For implement an effective classification of critical situations, we justified target-conformity application of support vector. We accumulated the base of possible solutions in special cases when process of cement production goes out from control limits. For pro-programmatic implementation of support vector, we used library application LIBSVM. We described procedures for the preparation of the accumulated data for training models of SVM. We determined optimal parameters of created models nuclei to ensure accurate classification. We created SVM models for all monitored parameters. We created HMI display effective solutions to prevent critical situations for the operator.

Key words: контрольные карты Шухарта, помол цемента, статистическая управляемость, база знаний, контроль качества, машина опорных векторов, классификация, кросс-валидация, Shewhart control charts, grinding of cement, statistical controllability, knowledge base, qual