Identification of Hybrid Systems with Application to Fault Detection of a Reverse-Flow Reactor System

In this work, the suitability of several techniques for monitoring a reverse-flow reactor system are studied. Reverse reactor systems are operated by periodically reversing the direction of the flow inside the reactor [1],[2]. One benefit of a reverse-flow configuration is that, for exothermic reactions, the system exhibit a heat-trap effect [3] where the temperature profile across the reactor achieves a maximum near the centre of the reactor. This enhanced temperature profile allows the combustion of lean feed streams that could require pre-heating for combustion in a traditional, one-way reactor system. The reverse-flow reactor system under study in this work is used for combustion of low concentration methane streams and has been modeled in the literature [3]. The data used for this work was obtained from the CANMET Energy Technology Centre at Varennes, Quebec Canada. Data from the reactor system under three distinct operating conditions will be studied. The three conditions are distinguished by the frequency at which the feed direction is switched and the concentration of methane in the feed stream. Table 1 lists the switch time and concentration associated with each condition. Plots of temperature versus time for selected thermocouples are shown in Figure 1. As can be seen from Figure 1, the switching flow direction induces temperature oscillations in the reactor. Furthermore, the period of the temperature oscillations is related to the period of switching. Less clear, however is the relationship between the feed concentration and the reactor temperature. As can be seen from Figure 1, the general trend of the mean temperature in the reactor is flat when operating at a feed concentration of 0.3% methane. However, when operating at a feed concentration of 0.7% methane, the averaged temperature increases with time.