基于宽度回声状态网络的菜籽油加工参数自动决策方法研究  

Study on automatic decision-making method of processing parameters of rapeseed oil based on broad echo state network

作　　者：金学波[1] 张雨雷 白玉廷[1] 王小艺 张维农 刘配莲 JIN Xue-Bo;ZHANG Yu-Lei;BAI Yu-Ting;WANG Xiao-Yi;ZHANG Wei-Nong;LIU Pei-Lian(College of Artificial Intelligence,Beijing Technology and Business University,Beijing 100048,China;Beijing Institute of Fashion Technology,Beijing 100029,China;School of Food Science and Engineering,Wuhan Polytechnic University,Wuhan 430023,China;COFCO Oil and Grains Industry(Feixian)Co.,Ltd.,Linyi 276000,China)

机构地区：[1]北京工商大学人工智能学院,北京100048 [2]北京服装学院,北京100029 [3]武汉轻工大学食品科学与工程学院,武汉430023 [4]费县中粮油脂工业有限公司,临沂276000

出　　处：《食品安全质量检测学报》2023年第5期16-22,共7页Journal of Food Safety and Quality

基　　金：国家重点研发计划项目(2020YFC1606801)。

摘　　要：目的 实现菜籽油生产过程中加工参数的自动给定,研究基于人工神经网络的自动决策方法。方法 利用菜籽油加工过程的检测数据,建立一种宽度回声状态网络模型对加工参数与危害物的内在映射关系进行建模;在危害物含量要求下,利用此模型可实现加工过程参数的自动给定。结果 以脱臭工序为例的实验表明,所提方法能够有效利用已知变量自动计算出加工参数,宽度回声状态网络的计算精度优于其他几种典型循环神经网络模型。结论 所提方法可有效提升菜籽油加工过程危害物的自动控制水平,进而提升加工过程的科学性和规范性。Objective To realize the automatic setting of processing parameters in rapeseed oil production, and study the automatic decision-making method based on artificial neural network. Methods Using the detection data of rapeseed oil processing, a broad echo state network model was established to model the internal mapping relationship between processing parameters and hazards. Under the requirements of hazardous substance content, the automatic setting of process parameters could be realized using this model. Results Taking the deodorization process as an example, it showed that the proposed method could effectively use the known variables to automatically calculate the processing parameters, and the calculation accuracy of the broad echo state network was better than that of several other typical recurrent neural network models. Conclusion The proposed method can effectively improve the automatic control level of hazards in rapeseed oil processing, thereby improving the scientificity and standardization of the processing.

关 键 词：菜籽油加工 参数辨识 自动决策 宽度回声状态网络 

分 类 号：TS225.14[轻工技术与工程—粮食、油脂及植物蛋白工程]