In salmon recovery programs it is commonly necessary to chill incubation and early rearing temperatures to match wild development times. Chiller failure can result in rapid temperature changes and increased deformities and mortality during the early development stages of sockeye salmon. Two common types of chiller systems are shown in Figure 1. The most common failure modes are (a) chiller failure due low voltage, failure of compressor or internal pumps, and low refrigerant levels) or (b) failure of external recirculation pumps. The magnitude of the temperature change depends strongly on both the type of chiller system and the failure mode. The objective of this research is to document the temperature variation of the two types of chiller systems and develop design criteria for chiller systems used in salmon incubation/early rearing. Three types of chiller failure were studied: (a) Type 1, chiller failure (NR), Type 2, chiller failure (CF), and Type 2, pump failure (PF).

Representative temperature changes are presented in Figure 2 for the three failure modes. The most rapid temperature in the first 120 minutes occurred for NR followed by PF and CF. The maximum temperature for CF was significantly higher than well water supply. The maximum temperature changes for 30-, 60-, and 90-minutes were estimated. The temperature response of incubation systems was modeled using tracer analysis. It was assumed that the reservoirs and incubators could be considered ideal continuous-flow stirred-tanks reactors (CFSTR), the volume of the connecting pipes could be ignored, and heat transfer could be ignored. The Type 1 system was modeled as a single CFSTR while the Type 2 system was modeled as two unequal sized CFSTR in series. The models were evaluated in terms of theoretical and measured mean hydraulic residence times. The models could be used to accurate predict temperature results as a function of system characteristics.