Non parametric on-line control of batch processes based on STATIS and clustering
نویسندگان
چکیده
Batch processes are widely used in several industrial sectors, e.g. food and pharmaceutical manufacturing. Process performance is described by variables which are monitored as the batch progresses. Data arising from such processes are usually monitored using control charts based on multiway principal components analysis. In this paper we propose a non parametric quality control strategy for monitoring batch processes with fixed as well as variable duration. In our proposition, data sets associated to batches are reduced using the STATIS method. Monitoring of batch performance is accomplished directly on principal plane graphs, from which non-parametric control regions are derived through convex hull peeling. This general approach allows off-line monitoring of batch processes as well as on-line monitoring after a constrained clustering step based on multivariate extension of W.D. Fisher’s algorithm is carried out. A real example of batch process with fixed duration illustrates the proposed method. Résumé : Les procédés par lots sont largement utilisés dans le secteur industriel notamment dans l’industrie agroalimentaire, chimique ou pharmaceutique. Le suivi de tels procédés est effectué à travers un ensemble de variables caractéristiques du procédé prélevées par un échantillonnage en ligne au fur et à mesure de son déroulement. Le procédé est contrôlé à travers des cartes multivariées basées sur une analyse en composantes principales particulière (multiway principal component analysis). Nous proposons une approche du contrôle de qualité des procédés par lots basée sur la méthode STATIS et des régions de contrôles non paramétriques obtenues à partir d’enveloppes convexes. Cette approche générale peut être utilisée pour le contrôle en fin de fabrication des procédés par lots ainsi que pour le contrôle en cours de fabrication après une étape de classification sous contrainte basée sur une extension multivariée de l’algorithme de W.D. Fisher. La méthode proposée est illustrée sur des données réelles issues d’un procédé par lots à temps fixe.
منابع مشابه
Batch Process Monitoring by Three-way Data Analysis Approach
We propose a non parametric quality control strategy based on the three way method STATIS and convex hull peeling for monitoring batch processes with constant or variable duration. The method is illustrated on a simulated data set and on a real one.
متن کاملA robust wavelet based profile monitoring and change point detection using S-estimator and clustering
Some quality characteristics are well defined when treated as response variables and are related to some independent variables. This relationship is called a profile. Parametric models, such as linear models, may be used to model profiles. However, in practical applications due to the complexity of many processes it is not usually possible to model a process using parametric models.In these cas...
متن کاملOn the non-parametric multivariate control charts in fuzzy environment
Multivariate control chats are generally used in situations where the simultaneous monitoring or control of two or more related quality characteristics is necessary. In most processes in the real world, distribution of the process characteristics are unknown or at least non-normal, so the non-parametric or distribution-free charts are desirable. Most non-parametric statistical process-control t...
متن کاملA Non-parametric Control Chart for Controlling Variability Based on Squared Rank Test
Control charts are used to identify the presence of assignable cause of variation in the process. Non-parametric control chart is an emerging area of recent development in the theory of SPC. Its main advantage is that it does not require any knowledge about the underlying distribution of the variable. In this paper a non-parametric control chart for controlling variability has been developed. I...
متن کاملProposing a Novel Cost Sensitive Imbalanced Classification Method based on Hybrid of New Fuzzy Cost Assigning Approaches, Fuzzy Clustering and Evolutionary Algorithms
In this paper, a new hybrid methodology is introduced to design a cost-sensitive fuzzy rule-based classification system. A novel cost metric is proposed based on the combination of three different concepts: Entropy, Gini index and DKM criterion. In order to calculate the effective cost of patterns, a hybrid of fuzzy c-means clustering and particle swarm optimization algorithm is utilized. This ...
متن کامل