人工神经网络在风干腊肠加工过程模拟控制中的应用

Application of Artificial Neural Networks for Process Simulation and Quality Control of Dry-sausage

摘要:

为了实现对自然条件下风干腊肠品质的快速检测和评价。本文通过模拟自然条件，比较不同风速、温度及湿度条件下加工得到的腊肠在理化指标、脂肪氧化、蛋白质氧化等的变化差异，利用人工神经网络模型的方法，对腊肠加工条件构建神经网络模型，实现对自然条件下加工腊肠品质的预测和模拟。结果表明当风干风速为0.80 m/s，日间温度为23.60 ℃，日间湿度40.42%，夜间温度16.89 ℃，夜间湿度54.21%时，可得到低脂高蛋白（蛋白含量为38.65%，脂肪含量为16.06%）的腊肠。所建立的人工神经网络模型r值都在0.99以上，且rmse值在0.4以下，表明该模型能够进行较准确的预测，可以实现最优加工条件的预测。该方法可为阐明加工条件对风干腊肠品质的影响，为风干腊肠规模化加工建立人工气候和质量安全控制提供理论依据。

Abstract:

The purpose of this study was to achieve rapid detection and evaluation of the quality of air-dried sausage under natural conditions. In this paper, different processing conditions (such as wind speed, temperature, humidity) were used through simulating the natural conditions, and the resulting physicochemical indices, lipid oxidation and protein oxidation of dried sausage were compared. Then an artificial neural network modelling method was used to construct a neural network model for predicting and simulating the processing conditions for dry sausage production. The obtained results showed that sausages with low fat and high protein (protein content 38.65% and fat content 16.06%) were obtained under the following processing conditions: wind speed 0.80 m/s, daytime temperature 23.60 ℃, daytime humidity 40.42%, night-time temperature 16.89 ℃, and night-time humidity 54.21%. The established artificial model r values above 0.99 and rmse values below 0.4, indicating that the model can make highly accurate predictions and can predict the optimal sausage processing conditions. Therefore, this method can describe the effect of processing conditions on the quality of air-dried sausage and provide a theoretical basis for the establishment of an artificial climate and quality and safety control for large-scale processing of air-dried sausage.