甘蔗糖业
甘蔗糖業
감자당업
Sugarcane and Canesugar
2015年
5期
38-43
,共6页
邓广华%蔡碧云%吴小员%农立忠%黄立新
鄧廣華%蔡碧雲%吳小員%農立忠%黃立新
산엄화%채벽운%오소원%농립충%황립신
白砂糖%色值%pH%关系模型%测定
白砂糖%色值%pH%關繫模型%測定
백사당%색치%pH%관계모형%측정
Whitegranulated sugar%Color value%pH%Relation model%Determination
论文研究测定不同等级的、一定pH范围的糖液的白砂糖样品的色值,发现除精制糖外,其色值均随糖液的pH值增大而升高,均呈现明显的指数关系,其中二级白砂糖为y =5.996e0.4955 x,一级白砂糖为y =1.154e0.6528x,相关性好;精制糖和优级糖的糖液在紫外区有较高的吸收值,对其色值与它们的40%(w/w)浓度的糖液在220、240、260、280、300 nm 这5个波长处的吸光值进行一元线性回归模拟分析,发现色值与糖液在220 nm处的吸光值线性关系较好,模型相关系数较大,预测的标准偏差较小。采用多元线性回归的方法处理5个波长处糖液的吸光度,得到相关系数R2=0.965的白砂糖色值的多元线性回归测算模型C=60.334A220+16.657A240-69.062A260-20.220A280+47.662A300-50.207,其预测精度比一元线性回归的更好,模型的鲁棒性更高。
論文研究測定不同等級的、一定pH範圍的糖液的白砂糖樣品的色值,髮現除精製糖外,其色值均隨糖液的pH值增大而升高,均呈現明顯的指數關繫,其中二級白砂糖為y =5.996e0.4955 x,一級白砂糖為y =1.154e0.6528x,相關性好;精製糖和優級糖的糖液在紫外區有較高的吸收值,對其色值與它們的40%(w/w)濃度的糖液在220、240、260、280、300 nm 這5箇波長處的吸光值進行一元線性迴歸模擬分析,髮現色值與糖液在220 nm處的吸光值線性關繫較好,模型相關繫數較大,預測的標準偏差較小。採用多元線性迴歸的方法處理5箇波長處糖液的吸光度,得到相關繫數R2=0.965的白砂糖色值的多元線性迴歸測算模型C=60.334A220+16.657A240-69.062A260-20.220A280+47.662A300-50.207,其預測精度比一元線性迴歸的更好,模型的魯棒性更高。
논문연구측정불동등급적、일정pH범위적당액적백사당양품적색치,발현제정제당외,기색치균수당액적pH치증대이승고,균정현명현적지수관계,기중이급백사당위y =5.996e0.4955 x,일급백사당위y =1.154e0.6528x,상관성호;정제당화우급당적당액재자외구유교고적흡수치,대기색치여타문적40%(w/w)농도적당액재220、240、260、280、300 nm 저5개파장처적흡광치진행일원선성회귀모의분석,발현색치여당액재220 nm처적흡광치선성관계교호,모형상관계수교대,예측적표준편차교소。채용다원선성회귀적방법처리5개파장처당액적흡광도,득도상관계수R2=0.965적백사당색치적다원선성회귀측산모형C=60.334A220+16.657A240-69.062A260-20.220A280+47.662A300-50.207,기예측정도비일원선성회귀적경호,모형적로봉성경고。
Measuring of the color value of different-class white granulated sugar was studied at a certain range of pH, results showed that:“the color values”of second-class, first-class and superior granulated sugar increased with the increase of pH, and significant and good exponential relationships were shown in the first two: y =5.996e0.4955 x and y = 1.154e0.6528x. used in measuring of the color value without buffer system. Linear regression simulation analysis was made between the color value of superior and refined granulated sugar and absorbance of their 40% (w/w) concentration of sugar solution in the 220, 240, 260, 280, 300nm wavelength, the mathematical model with best correlation was: C=60.334A220+16.657A240-69.062A260-20.220A280+47.662A300-50.207,R=0.965, of which the prediction accuracy and robustness were good.