@article { author = {Ayari, F. and Mirzaee- Ghaleh, E. and Rabbani, H. and Heidarbeigi, K.}, title = {Implementation of a Machine Olfaction for the Detection of Adulteration in Cow Ghee}, journal = {Journal of Agricultural Machinery}, volume = {10}, number = {2}, pages = {129-139}, year = {2020}, publisher = {Ferdowsi University of Mashhad}, issn = {2228-6829}, eissn = {2423-3943}, doi = {10.22067/jam.v10i2.67524}, abstract = {IntroductionOne the most important discussions of the world community is the importance and the role of edible oils in the nutrition and physical health of individuals, especially in the prevention of cardiovascular disease. One of these oils, used in cooking, is cow ghee. Cow ghee should be free of vegetable oil, animal fat, mineral oils, flavored additives and any other external ingredients. It is hard to find a technique that can easily and reliably measure the quality of the oil. So far, no special machine or system has been designed or built to distinguish the pure cow ghee from the adulterated ones. Electronic nose is a new method that has recently been considered by researchers in agriculture especially in the field of food quality. Because of high ability of e-nose system, in this research, this system was used for the detection of pure cow gee from the adulterants ones.Materials and Methods An olfactory machine system based on eight MOS sensors was designed to detect pure cow ghee from the adulterated with various proportions of vegetable oil and animal fat. Designed system includes data acquisition system, sensors, sensors chamber, sample box, power supply, connections, electric valves, air pump and air filter. The sensor array was consisted of the 8 MOS sensors that each of them react to specific volatile compounds. These sensors are widely used in olfactory machines because of their high chemical stability, high durability, low response to moisture and affordable prices. These are the most commonly used sensors in electronic nose system. To prepare samples with different percentages of adulteration, animal body fat and refined vegetable oils were added to pure cow ghee. In order to carry out the experiments, the sample was placed in sample box and in the baseline correction step (200 seconds), clean air was passed through the sensors to transmit the response of sensor array to steady state. At the injection step (180 seconds), the sample headspace was transmitted and passed through sensors chamber.  Output voltage of each sensor depends on the type of sensor and its sensitivity. At the cleaning step (120 seconds) the clean air was passed through sensors to get the sensor array responsive to a stable state. Also, at this step the pump removed the odor remaining inside the sample container and system was prepared for the next test. The signals obtained from the sensors were recorded and then pre-processed. Results and DisscussionPCA and QDA analysis were used for detection the differences between pure cow ghee and adulterated ones. The data obtained from the signals processing with fractional method were used as input of PCA. The PCA results showed that the total variance between pure cow ghee and mixture of cow ghee with animal's fat was 97%. Also score plot of cow’s ghee and its mixture with vegetable oil showed the total variance of 96% between different samples. Sensors are the main components of an electronic nose system therefore it is necessary to select the best sensors to detect differences between samples. The loading plot was obtained to show the role of sensors in e-nose system and demonstrates that the selected sensors have a high degree of complementarity. Based on confusion matrix obtained from QDA analysis, pure samples were detected from vegetable oil and animal fat samples with correct classification rate of 95.24 and 97.15, respectively.Conclusion An eight-sensory olfactory machine system (MOS) was designed to detect pure cow ghee from the presence of vegetable oil and animal fat oil. In PCA analysis, the variance between samples was 97% and 98%, respectively. According to the results the radar graph of PCA analysis, it can be concluded that the sensors No 2 (TGS822), 3(MQ136), 4(MQ9) and 8(TGS2620) have the highest and sensor 6 (MQ135) has the lowest ability in classification. The MQ135 sensor reacts to the detection of ammonia, benzene, and sulfide. In other words these gases did not play important role in separating of cow ghee from other mixed oils.}, keywords = {Semiconductor sensors,cow ghee,Machine olfaction,Principal component analysis}, title_fa = {پیاده‌سازی سامانه ماشین بویایی به‌منظور تشخیص تقلب در روغن حیوانی گاوی}, abstract_fa = {تقلب در محصولات لبنی نه تنها تهدیدی جدی برای سلامت انسان است بلکه زیان‌های اقتصادی متعددی را نیز به دنبال دارد. ازجمله تقلب‌های رایج در روغن حیوانی گاوی، ترکیب کردن آن با روغن نباتی و روغن دنبه است. در این پژوهش، یک سامانه‌ی ماشین بویایی بر پایه هشت حسگر نیمه‌هادی اکسـید فلـزی ساخته شد و قابلیت آن در تشخیص مقادیر مختلف ترکیب روغن نباتی و روغن دنبه در روغن حیوانی گاوی خالص (10، 20، 30، 40 و 50 درصد) مورد بررسی قرار گرفت. بردار ویژگی‌ها از سیگنال پاسخ حسگرها به ترکیبات فرار و معطر انواع روغن‌ها، استخراج و به‌عنوان ورودی مدل تشخیص الگو استفاده شد. هم‌چنین جهت طبقه‌بندی ویژگی‌های استخراج‌شده از روش تحلیل تفکیک درجه دوم (QDA) استفاده شد. نتایج حاصل از آنالیز مؤلفه‌های اصلی با دو مؤلفه‌ی PC1 و PC2، به‌ترتیب واریانس 98 و 97 درصد را برای ترکیب روغن حیوانی با روغن نباتی و روغن دنبه نشان داد. همچنین نمودارهای لودینگ و رادار نشان داد که بوی روغن حیوانی گاوی بیش‌ترین و کم‌ترین تأثیر را به‌ترتیب روی حسگر TGS822 و حسگر MQ135 دارد. همچنین بوی روغن نباتی و روغن دنبه بیش‌ترین و کم‌ترین تأثیر را به‌ترتیب روی حسگرهای MQ136 و MQ135 داشت. با توجه به نتایج به‌دست آمده از نمودار رادار مشخص شد که حسگر MQ135 کمترین نقش را در طبقه‌بندی دارد. هم‌چنین براساس نتایج حاصل از طبقه‌بندی، دقت طبقه‌بندی برای روغن حیوانی مخلوط با روغن نباتی و روغن دنبه به‌ترتیب برابر 24/95 و 15/97 درصد به‌دست آمد.}, keywords_fa = {حسگرهای نیمه هادی,روغن حیوانی گاوی,ماشین بویایی,مولفه های اصلی}, url = {https://jame.um.ac.ir/article_34140.html}, eprint = {https://jame.um.ac.ir/article_34140_0837c24e010625bca3cd82f916b21cb4.pdf} }