PhD Theses

Safe operation of thermally sensitive chemical reactors remains a crucial engineering issue. Thermal runaway can result in serious consequences, such as the explosion of the reactor, therefore; engineers must know about reactor runaway in detail, and they must know the possible causes and consequences. Thermal runaway occurs mainly due to loss of temperature control, but many chemical accidents initiated by thermal runaway can be foreseen by an appropriate analysis of thermal process data. So-called thermal runaway criteria can be used to predict the development of reactor runaway which can be geometric-, stability- and sensitivity-based methods. 

Runaway criteria classify the different states of a reactor operation as non-runaway or runaway based on a critical equation. However, these runaway criteria indicate the development of thermal runaway in different states, and there is no a fully general method or theory which can be applied with the highest reliability and with the shortest indication time for every reactor and reaction system. Each criterion has its truth about the runaway. Despite of it I developed two new thermal runaway criteria (Modified Slope Condition, MSC, Modified Dynamic Condition, MDC), and their performances were compared with the most frequently applied runaway criteria based on their reliability and their indication time. MDC criterion came out as the most reliable while the indication time is in the midfield. Since runaway criteria do not consider system specifics (such as Maximum Allowable Temperature, or maximum process pressure), I applied a genetic programming-based methodology to identify system-specific critical equations, which outperforms the conventional criteria.