A qualitative simulation-based learning environment: how to enhance causal understanding of complex phenomena in large-scale plants

摘要

The authors describe a framework of a qualitative simulation-based learning environment that focuses on causal understanding of complex phenomena in large-scale plants. Training styles such as study with manuals and exercise with a numerical simulator have been developed. However, besides such training styles, a learning environment should provide the functions that enable operators to achieve deep understanding of target plants. The qualitative simulation-based learning environment aims at providing operators a thinking tool for understanding, where human-computer interaction is also designed based on this notion. To realize this environment in a computer, two problems on qualitative reasoning have to be resolved. One is how to construct qualitative models of adequate grain size as target plants become complex, and the other is how to prune spurious qualitative behaviors. New techniques on qualitative reasoning are proposed, which are called qualitative reasoning with association mechanisms to qualitative information. Its effectiveness is discussed through applying this framework to large-scale power plants.