Research On Drying Technology Of Basa Fish And Establishment Of Artificial Neural Network Model

Basa fish?Pangasius bocouti?is fresh and nutritious,and belongs to freshwater cultured carp,which has a large annual output and is well received by consumers.With the improvement of people's living standards,new demands have arisen on aquatic products,requiring them to be convenient to eat,nutritious,safe and healthy.This paper firstly explores the nutritional composition evaluation and safety analysis of Basa fish,and uses Basa fish as the experimental raw material to seek the key control points in the process of drying Basa fish fillets through the hot air drying process,and enhance the value of the deep processing value of Basa fish.Secondly,this paper evaluates the quality of instant Basa fillets and determines the drying process of the best instant Basa fillets by comprehensive evaluation.On the other hand,based on the drying process control point of instant Basa fillets,it is expected to establish an artificial neural network model suitable for the drying process of instant Basa fillets,and provide theoretical guidance for the application and promotion of dried instant fish fillets.In this paper,the nutritional composition analysis of Basa fish was firstly carried out,and the basic nutrients,amino acid content,fatty acid content and mineral element composition of Basa fish were determined.The study found that the crude protein content in the muscle of Basa fish is 9.8%,the crude fat mass fraction is 1%;the total amino acid content in the basa fish is 85.29g/kg,and the essential amino acid ratio is 42.34%;the mineral elements are detected.There are 8 kinds,among which Na,K,P,Mg,Ca and other elements are especially rich in content.According to the sensory evaluation method,the sensory evaluation of instant Basa fish fillets with different moisture levels was carried out to determine that the final moisture content of the instant Basa fillets was controlled to 20%.The effects of drying temperature,relative humidity of air and drying wind speed on drying efficiency were analyzed by plotting the drying kinetics curve of instant Basa fish fillets.The study found that the drying temperature and drying wind speed have significant effects on the drying efficiency,and the time taken to reach the drying end point can reduce the drying time by 71.42%and 42.85%,respectively.Secondly,the instant Basa fillet products prepared under different drying conditions were measured for color,moisture distribution,microstructure and sensory evaluation.The research results show that the best drying conditions is air relative humidity of 20%,dry wind speed of 16m/s,drying temperature of 80?.This paper studies how to design and establish an artificial neural network model for the drying process of instant Basa fillets,and to evaluate its predictive ability.The BP neural network model was selected and the grid topology structure was 4a-12b-1c.The 424 sets of experimental data were collected.The weights and deviations of network learning are determined and the network is operated on this basis.The predictive capability coefficients of regression model in network performance are R _training=0.99487,R _test=0.99665 and R _verification=0.99551,respectively.Comparing the predicted value of the network model with the real value and fitting linearly,it is found that the fitting coefficient reaches 0.9965,and the predicted result is acceptable.