Process Monitoring and Visualisation Using Self-Organizing Maps

Analysis and control of complex nonlinear processes constitutes a diicult problem area. In complicated systems, as in chemical processes, it is not possible to predict all possible error types in advance. The self-organizing map algorithm can be used to investigate complex dependencies between various process parameters as well as input and output variables. It can also be utilised to create systems to monitor complicated, dynamical processes and to visualise the process development. 2. Introduction The self-organizing map (SOM) algorithm 8]]9] creates a mapping from input patterns to map units. The key point in the applicability of the SOM algorithm is the topological nature of the mapping; similar signal patterns are mapped to nearby locations on the map. Such a mapping can be applied to pattern sequence analysis by nding the mapping locations of subsequent patterns and observing the trajectory, that is the curve of the locations in time. The pattern sequence in multi-dimensional input space can in that way be transformed into a two dimensional trajectory which can be used, for example, in process monitoring applications. In the following, the use of the self-organizing map in process state monitoring is described. Some applications in visualisation of the process operation

[1]  L Leinonen,et al.  Acoustic pattern recognition of fricative-vowel coarticulation by the self-organizing map. , 1993, Folia phoniatrica.

[2]  Ossi Taipale,et al.  Optimization and simulation of quality properties in paper machine with neural networks , 1994, Proceedings of 1994 IEEE International Conference on Neural Networks (ICNN'94).

[3]  Olli Simula,et al.  Process State Monitoring Using Self-Organizing Maps , 1992 .

[4]  Olli Simula,et al.  Representation and Identification of Fault Conditions of an Anaesthesia System by Means of the Self-Organizing Map , 1994 .

[5]  Juha Tuominen,et al.  PROCESS ERROR DETECTION USING SELF-ORGANIZING FEATURE MAPS , 1991 .

[6]  Teuvo Kohonen,et al.  The self-organizing map , 1990 .

[7]  T. Kohonen.  Self-organized formation of topology correct feature maps , 1982 .

[8]  Jie Lin,et al.  Nuclear power plants transient diagnostics using LVQ or some networks don't know that they don't know , 1994, Proceedings of 1994 IEEE International Conference on Neural Networks (ICNN'94).

[9]  Dagmar Niebur,et al.  Unsupervised Neural Net Classification of Power System Static Security States , 1991 .

[10]  L Leinonen,et al.  Self-organized acoustic feature map in detection of fricative-vowel coarticulation. , 1993, The Journal of the Acoustical Society of America.

[11]  Alfred Ultsch,et al.  Self Organized Feature Maps for Monitoring and Knowledge Aquisition of a Chemical Process , 1993 .

[12]  Olli Simula,et al.  Monitoring of an Anaesthesia System Using Self-Organizing Maps , 1994 .

[13]  L Leinonen,et al.  Dysphonia detected by pattern recognition of spectral composition. , 1992, Journal of speech and hearing research.

[14]  L Leinonen,et al.  Acoustic pattern recognition of /s/ misarticulation by the self-organizing map. , 1993, Folia phoniatrica.

[15]  R. Fischl,et al.  Application of neural networks to power system security: technology and trends , 1994, Proceedings of 1994 IEEE International Conference on Neural Networks (ICNN'94).

[16]  Heikki N. Koivo,et al.  Application of artificial neural networks in process fault diagnosis , 1991, Autom..

[17]  Peter Weierich,et al.  Unsupervised Detection of Driving States with Hierarchical Self Organizing Maps , 1994 .