Abstract: [Objective] The aim of this study was to optimize the ultrasonic extraction process parameters of crude fat in feed. [Method] Using the grass carp expanded compound feed as experimental material, a single factor experiment design was adopted to investigate and analyze the effects of 4 factors （liquid to material ratio, soaking time, ultrasonic duration and ultrasonic frequency） at different levels on the determination of crude fat content in feed. In addition to the single factor experiments, a 4-factor and 3-level L9 （3⁴） orthogonal experiment was conducted to optimize the ultrasonic extraction process parameters of crude fat in feed. The crude fat content of grass carp expanded compound feed was measured using the optimized ultrasonic extraction process, and the results were compared to those obtained using the Soxhlet extraction method. The process was repeated 6 times to verify the optimal ultrasonic extraction process parameters. Following the determination of the ultrasonic extraction process parameters, the crude fat contents of 13 different feed samples were measured using both the ultrasonic extraction method and the Soxhlet extraction method with 6 repeated tests. The differences in crude fat content of the feed raw materials and common animal feed products between the two extraction methods were then analyzed and compared. [Result] The factors affecting the determination of crude fat content in feed was in the order of ultrasonic frequency>liquid to material ratio>ultrasonic duration>soaking time. The optimal ultrasonic extraction process parameters were liquid to material ratio of 37.5 mL/g, soaking for 0 min, ultrasonic duration for 30 min, and ultrasonic extraction twice. The crude fat content of grass carp expanded compound feed determined by the optimal ultrasonic extraction process was 4.76% （n=6）, whereas that determined by the Soxhlet extraction method was 4.77% （n=6）, proving that the ultrasonic extraction process was stable and feasible. The differences in crude fat content determination results of 13 different feed samples between the ultrasonic extraction method and the Soxhlet extraction method were not significant （P>0.05）. [Conclusion] The optimal process parameters for ultrasonic extraction of crude fat in feed determined in this study were stable and feasible.

Key words: feed, crude fat, ultrasonic extraction, Soxhlet extraction method