Impacts of chiller failure on temperature change in isolation incubators for salmonids

Mechanical chillers can be used to slow the development of salmon eggs and fry. Chiller failure can result in a rapid temperature increases that may adversely impact salmon development. In this study, three types of chiller failure were simulated: (1) CF − failure of chiller, (2) PF − failure of recirculation pump, and (3) NR − chiller failure for a chiller system without a coldwater reservoir. Temperatures were monitored at 38 locations at the Burley Creek Hatchery using 4-channel loggers (Onset, Model U12-008) and Hobo pendant loggers (Onset, UA-001-64). The maximum temperature responses for 30-, 60-, and 90-min intervals were determined for both failure and restart. For the 30-min period, the maximum ΔTs were equal to 3.37 °C for NR, 2.62 °C for PF, and 1.79 °C for CF. The magnitude of the ΔTs were larger for restart compared to failure. The response of the Hobo loggers were very close to the 4-channel loggers even though their time response was significantly slower. The PF and CF failure modes were modeled as two unequal sized CFSTR (coldwater reservoir and incubator) in series and NR mode was modeled as a single CFSTR (incubator). Theoretical and measured mean hydraulic resident times were used to estimate the both deviation between the actual temperature and the modeled temperatures as well as the maximum temperature increases at 30-, 60-, and 90-min intervals. The PF-failure and NR-restart were quite good CFSTRs (stagnant regions of about 9%), while the remaining failure modes had poorer performance (stagnant regions ranging from 25 to 35%). If the theoretical mean hydraulic residence times are used for design, these values must be multiplied by the appropriate reactor correction factors to estimate the size of physical coldwater and glycol reservoirs needed.