黃麴毒素在農作物加工、貯存及物流過程若管理不慎,在台灣濕熱天氣下容易產生,其中又以黃麴毒素B1 肝毒性最高。本研究根據台灣衛生福利部針對食品中黃麴毒素B1檢驗方法,設計一可由一般民眾進行現場花生樣品檢測前處理之儀器。設計過程首先針對台灣、美國、歐盟對於黃麴毒素B1檢驗方法做綜合性評估,藉由相互比較得出最適化流程,接著將流程轉換成設計規格,設計了針對固態樣品之 自動化前處理系統。檢測對象設定為花生、玉米等固態樣品,在前處理過程之前除了要確認取樣結果是否具代表性外,還需將樣品粒徑減少至一定程度並且與萃取液充分混合才能確保萃取結果之有效性,根據上述需求,設計包含:樣品重量感測、流道、研磨以及過濾系統。樣品重量感測系統以應變規配合濾波電路達到公定檢驗方法之要求,在感測25 g 樣品時能將誤差控制於± 0.025 g ( ± 0.1 %) 以內,且藉由電磁鐵實現了自動化進樣之程序;流道系統以蠕動幫浦配合電磁閥實現萃取液流量、清洗液及潤洗液之控制,並以流速計將萃取液體積量之誤差保持在± 1 mL 內,確保萃取效率不會有明顯變異性;研磨系統則以高速剪切混合之理念設計並以壓克力製造研磨刀頭,將71.53 ± 4.641% 之樣品磨碎至粒徑1 mm 以下;過濾系統以圓筒濾紙為核心設計半自動化過濾,僅需手動丟入、取出濾紙即可完成過濾,並配合抽氣馬達將過濾時間從30 分鐘縮短至5 分鐘。閥門機構由於製造能力之限制,使得自製虹膜閥無法應用於系統中,最後改以球閥取代。整體自動化流程符合公定法之規定,然在重複性試驗中,重量感測系統會因製造公差對量測結果產生超過規定之誤差。針對單次檢測後之清洗流程進行評估,在現有架構下之清洗性能有限,因此在根據上述性能表現以及失效情形進行設計回顧,提供後續改良之依據。
Aflatoxins (AFs) are poisonous carcinogens commonly produced in a mismanaged food processing, storage and logistics, especially in the hot and humid weather of Taiwan. Among all types of AFs, aflatoxin B1 (AFB1) possesses the highest hepatotoxicity. This study designed a sample (peanut) pretreatment apparatus for in-field test which operated by lay person AFs. All processes followed the regulation announced by the Ministry of Health and Welfare of Taiwan. Therefore, this study made a comprehensive evaluation of the AFs inspection methods in Taiwan, the United States, and the European Union to obtain the most suitable pretreatment process. This process was then converted into design specification, and an automated pretreatment system for solid samples was designed accordingly. As this system targeted at solid samples such as peanuts and corn, comfirmation of the sampling representativity was an important premise before the pretreatment process. It was also necessary to reduce the sample particle size to a certain scale in order to obtain an homogeniously mixed extraction for a valid test result. Based on the requirements above, the designed processes included: sample weight sensing, flow, grinding and filtering system. The sample weight sensing system met the requirements by using strain gage type load cell with filter circuit, and the error wass controlled within ± 0.025 g (± 0.1 %) when measuring a 25 g sample. Solenoid was used for the automatic sample input procedure. By using peristaltic pump with solenoid valve, the flow system achieved the flow rate control of extraction liquid, cleaning solution and rinsing solution. The system is able to keep the error of the extraction solution volume within ± 1 mL, ensuring that there would be no significant variability in the extraction efficiency. The grinding system was designed with the concept of high speed shear mixer and made of acrylic. This system reduced 71.53 ± 4.641% of the sample size to less than 1 mm. The filter system used cellulose thimble as filter paper to achieve semi-automatic filtration (only need to manually throw in and take out the thimble), and further combined with a vacuum motor to shorten the filtration time from 30 min to 5 min. The cleaning effeciency was evaluated after a thorough cycle of the pretreatment process. The self-made iris valve couldn’t apply in system due to manufacturing limitation. Ball valve was then used and effectively control the flow path. The overall automation procedure complied with regulation in Taiwan. In repeated test, weigh sensing might produce unbearable error due to dimensional tolerance Under the current version of automated pretreatment system, the cleaning performance was limited. Therefore, the design review was based on the above performance and failure mode to provide a basis for future improvements.