Research On The Prediction Method Of Beef Drying Moisture Content And Energy Consumption Based On Improved Neural Network

The monitoring and control of moisture content and energy consumption in drying process is of great significance to the processing and production of meat products.Based on the research of drying characteristics of meat samples,the intelligent information processing technology and advanced drying technology are combined to accurately calculate moisture content and energy consumption under the condition that the drying process is not finished,which will provide a strong guarantee for the drying quality of meat and its products.In this thesis,the prediction method of moisture content and energy consumption in microwave drying of beef is studied using the nonlinear and adaptive advantages of artificial neural network.Beef sirloin is selected as the research object,and six different microwave power levels are set to study the drying characteristics of beef.Pearson and random forest correlation analysis were established to obtain the correlation and characteristic contribution rate among various factors.Microwave power,sample mass and drying time were determined as the input variables of neural network prediction model,and moisture content and energy consumption were the output variables.Considering the local optimization defects of traditional BP neural network,Sparrow Search Algorithm(SSA),Particle Swarm Optimization(PSO)algorithm,Simulated Annealing(SA)algorithm and Genetic Algorithm(GA)were selected to test the performance of the algorithm.Taking the global search ability,convergence speed and stability as evaluation index,the sparrow search algorithm was finally selected to optimize the inertia weight w and threshold b of BP neural network,and SSA-BP neural network prediction model was established.The results show that the goodness of fit of moisture content and energy consumption is 0.9995,0.9837,and root mean square error is 0.3331,0.2267,and mean absolute percentage error is 0.6151,4.5251.Compared with BP and GA-BP prediction models,SSA-BP neural network has a strong prediction advantage in the prediction of beef drying index.In order to improve the theoretical guidance and application value of SSA-BP drying prediction model for the actual drying process,the visual prediction platform of moisture content and energy consumption of meat microwave drying was designed by using MATLAB platform,and the application from algorithm model to system visualization was realized,so as to facilitate the real-time monitoring and evaluation of moisture content and energy consumption in meat industrial production process.