Development of Neuro-Intelligent Sensor and its application to quality control of whiskey aroma

神经智能传感器的研制及其在威士忌香气质量控制中的应用

基本信息

批准号：
02555087
负责人：
MORIZUMI Toyosaka
金额：
$ 6.91万
依托单位：
Tokyo Institute of Technology
依托单位国家：
日本
项目类别：
Grant-in-Aid for Developmental Scientific Research (B)
财政年份：
1990
资助国家：
日本
起止时间：
1990 至 1991
项目状态：
已结题

来源：
https://kaken.nii.ac.jp/grant/KAKENHI-PROJECT-02555087/
关键词：
Odor sensor Neural network Pattern recognition Sensory test Whiskey Olfaction lipid Quality control におい化学センサ

项目摘要

An odor sensor was developed for the application of quality control of whiskey aroma in the present study. The output patterns from plural quartz-resonator sensors with partially overlapped specificity are rcoonized by an artificial neural network in this odor sensing system.It was indispensable to discriminate closely similar aromas of whisakeys in whiskey quality control, and an odor sensor would be quite promising for that discrimination. Therefore, the research was performed for the purpose of realizing the sensor with a high identification capability. In order to raise the capability, sorption membranes suitable for identifying whiskey aromas were selected, and the data variation was reduced by improving the measurement system.In the preliminary stage, the membranes were selected among stationary phase materials for gas chromatography and cellulosic ones using a statistical method. Later, several lipid membranes, which are the components of a biological cell, were added to the mem … More brane set.The measurement system was modified as follows. Stainless steel was adopted as the material of the sample flow system to reduce the adsorption on the internal wall. A mass-flow controller was introduced to control the air flow rate and keep the aromatic vapor supply to the sensor cell constant. A thermoregulator was used to stabilize the temperature of the sample bottles and sensors so that the vapor pressure and the sensor sensitivity could be kept constant. As a result of these system modifications, the data variations of the sensors were reduced to 0.3%. After the membrane selection and the system modification, the recognition probability of whiskey aroma, which were difficult to identify before this study, was raised to more than 90%.The prototype system for an industrial use was set at a whiskey manufacturer, and the recognition probability was found almost 90% when the whiskey identification was performed. The data necessary for developint the system furthermore is now being gathered. Less

本研究中开发了一种用于威士忌香气质量控制的气味传感器，该气味传感系统中的人工神经网络对具有部分重叠特异性的多个石英谐振器传感器的输出模式进行了识别。威士忌的气味非常相似，在威士忌质量控制中，气味传感器非常有前景。因此，为了实现具有高识别能力的传感器，进行了研究。前期采用统计方法在气相色谱固定相材料和纤维素固定相材料中选择膜，并通过改进测量系统来减少数据变化。膜组中添加了作为生物细胞组成部分的几种脂质膜。测量系统采用不锈钢作为样品流系统的材料，以减少内部的吸附。引入质量流量控制器来控制空气流量并保持传感器单元的芳香蒸气供应恒定，并使用温度调节器来稳定样品瓶和传感器的温度，从而保证蒸气压和传感器灵敏度。通过这些系统修改，传感器的数据变化减少到0.3%。在进行膜选择和系统修改之后，威士忌香气的识别概率在本研究之前是难以识别的。提升到90%以上。原型机威士忌制造商设置了工业用途的系统，进行威士忌识别时发现识别概率几乎为 90%，而且目前正在收集开发该系统所需的数据。