近红外反射光谱快速评定棉粕营养价值的研究

发布时间：2018-05-15 12:55

本文选题：棉粕 + 近红外反射光谱　；参考：《甘肃农业大学》2016年硕士论文

【摘要】：本试验研究近红外反射光谱技术用于棉粕营养价值评定的可行性,选用我国新疆、山东和湖北等不同种植地区的76个色泽和气味正常的棉粕样品作为试验材料,其中56个样品用于建立水分、粗蛋白质、粗脂肪、粗纤维、粗灰分、总能值、蛋公鸡表观代谢能(AME)、真代谢能(TME)和氨基酸的近红外定标模型,对模型进行内部交叉验证,另外20个样品作为外部验证集对模型进行外部验证,探讨了福斯近红外仪(FOSS XDS)和尼尔光电微型近红外仪(Nir Smart Eye 1700)定标与验证模型的稳定性与适用性。试验一应用福斯近红外仪和尼尔光电微型近红外仪,研究近红外反射光谱技术测定棉粕常规养分含量和代谢能的可行性。从全国范围内收集76个不同品种和产地,以及不同加工方式的棉粕样品,测定其常规营养成分(水分、粗蛋白质、粗脂肪、粗纤维、粗灰分及总能)和蛋公鸡代谢能含量,并随机选取定标集(N=56)和外部验证集(N=20)样品,使用改进的偏最小二乘法(MPLS)建立近红外定标模型。试验结果表明:(1)不同来源棉粕的营养成分变异较大,常规营养成分变异系数为2.52%~84.75%,其中水分、粗脂肪和粗纤维的变异系数超过10%;粗蛋白质、粗灰分和总能的变异系数分别为9.58%、9.81%和2.52%;(2)福斯近红外仪棉粕的水分、粗蛋白质、粗脂肪、粗纤维、粗灰分和总能的定标决定系数(RSQcal)为0.924~0.976,交互验证决定系数(1-VR)为0.8247~0.9303,外部验证决定系数(RSQv)为0.879~0.896,定标方程可用于日常分析。(3)尼尔光电微型近红外仪棉粕的水分、粗蛋白质、粗脂肪、粗纤维和总能的RSQcal为0.905~0.951,定标标准差为(SEC)0.169~1.456,RSQv为0.883~0.959,定标方程可用于日常分析,粗灰分的RSQv为0.524,模型不可用。AME和TME的分布范围分别为:4.63 MJ/kg~11.90 MJ/kg和5.39 MJ/kg~13.20 MJ/kg。用福斯近红外仪建模得到的的AME和TME的RSQcal为0.969和0.927,1-VR为0.9170和0.9057,RSQv为0.911和0.892,定标方程可用于日常分析;用尼尔光电微型近红外仪建模得到的AME和TME的RSQcal分别为0.954和0.949,SEC为0.400和0.475,RSQv为0.915和0.907,定标方程可用于日常分析,模型达到了可实用水平。试验二探讨了使用近红外反射光谱法测定棉粕氨基酸含量的可行性。从全国范围内收集76个不同品种和产地,以及不同加工方式的棉粕样品,随机选取定标集(N=56)和外部验证集(N=20)样品,并测定其16种氨基酸相应的含量。结果表明:不同来源棉粕中各氨基酸含量的差异较大;用福斯近红外仪建模得到的天冬氨酸(Asp)、苏氨酸(Thr)、谷氨酸(Glu)、甘氨酸(Gly)、赖氨酸(Lys)、组氨酸(His)、精氨酸(Arg)和色氨酸(Trp)的RSQcal为0.872~0.953,1-VR为0.7813~0.9504,RSQv为0.840~0.887,定标方程可用于日常分析,其它氨基酸的RSQv0.84,尚无法用于实际预测;尼尔光电微型近红外仪建模得到的天冬氨酸(Asp)、苏氨酸(Thr)、丝氨酸(Ser)、谷氨酸(Glu)、甘氨酸(Gly)、亮氨酸(Leu)、苯丙氨酸(Phe)、赖氨酸(Lys)、组氨酸(His)、精氨酸(Arg)和色氨酸(Trp)的RSQcal为0.865~0.970,SEC为0.016~0.537,RSQv为0.845~0.899,定标方程可用于日常分析,其它氨基酸的RSQv0.84,尚无法用于实际预测。在各营养成分、代谢能和氨基酸之间最佳的去散射方法不同。
[Abstract]:The feasibility of using near infrared reflectance spectroscopy to evaluate the nutritional value of cottonseed meal was studied. 76 samples of cotton meal with normal color and smell in Xinjiang, Shandong and Hubei were selected as experimental materials, and 56 of them were used to establish water, crude egg white matter, crude fat, coarse fiber, coarse ash, total energy value, and egg public. Chicken epigenetic metabolic energy (AME), true metabolic energy (TME) and amino acid near infrared calibration model, the internal cross validation of the model, and the external validation of the other 20 samples as external verification set, and the stability of the FOSS XDS and the Neal photoelectric micro type near infrared (Nir Smart Eye 1700) calibration and verification model are discussed. The feasibility of determining the conventional nutrient content and metabolic energy of cottonseed meal by near infrared spectroscopy was studied by using FIR near-infrared instrument and Neal photoelectric miniature near-infrared instrument. The samples of 76 different varieties and producing areas and different processing methods were collected from the whole country. The metabolic energy content of crude protein, crude fat, crude fiber, coarse ash and total energy) and egg rooster, and random selection of N=56 and external validation set (N=20) samples, and using the improved partial least squares (MPLS) method to establish the near infrared calibration model. The results showed: (1) the nutrient composition of the cottonseed meal from different sources varied greatly and the conventional nutrient composition changed. The coefficient of variation is 2.52%~84.75%, in which the coefficient of variation of water, crude fat and crude fiber is more than 10%, and the coefficient of variation of crude protein, crude ash and total energy is 9.58%, 9.81% and 2.52%, respectively. (2) the coefficient of determination of the moisture, crude protein, crude fat, crude fiber, coarse ash and total energy of the fir near infrared instrument is 0.924~0.976, and the interaction between the crude protein and the total energy is 0.924~0.976. The determination coefficient (1-VR) is 0.8247~0.9303, the external verification determination coefficient (RSQv) is 0.879~0.896, and the calibration equation can be used for daily analysis. (3) the moisture of the cottonseed meal, crude protein, crude fat, crude fiber and total energy RSQcal of Neal photoelectric miniature near infrared instrument are 0.905~0.951, the standard deviation is (SEC) 0.169~1.456, RSQv is 0.883~0.959, and the calibration square Cheng Ke is used for daily analysis. The RSQv of coarse ash is 0.524, and the distribution of the model is not available for.AME and TME. The AME and TME RSQcal of 4.63 MJ/kg~11.90 MJ/kg and 5.39 MJ/kg~13.20 MJ/kg. are obtained with AME and TME, 0.969 and 0.927,1-VR are 0.9170 and 0.9057, RSQv is 0.911 and 0.892, and the calibration equation can be used for daily analysis. The RSQcal of AME and TME obtained by Neal photoelectric micro near infrared instrument is 0.954 and 0.949, SEC is 0.400 and 0.475, RSQv is 0.915 and 0.907. The calibration equation can be used for daily analysis and the model reaches the practical level. Experiment two discussed the feasibility of using near infrared reflectance spectroscopy to determine the content of amino acid in cottonseed meal. The samples of 76 different varieties and producing areas and different processing methods were collected, and the N=56 and N=20 samples were randomly selected and the corresponding content of the 16 amino acids was measured. The results showed that the content of amino acids in the cottonseed meal of different sources was different, and the aspartic acid (Asp) was modeled by FTIR. Threonine (Thr), glutamic acid (Glu), glycine (Gly), lysine (Lys), histidine (His), arginine (Arg) and tryptophan (Trp) RSQcal are 0.872~0.953,1-VR 0.7813~0.9504, RSQv is 0.840~0.887, the calibration equation can be used for daily analysis, other amino acids are not used for practical prediction; Neal photoelectric miniature near infrared apparatus is modeled. Asp, serine (Thr), serine (Ser), glutamic acid (Glu), glycine (Gly), leucine (Leu), phenylalanine (Phe), lysine (Lys), histidine (His), arginine (Arg) and tryptophan (Trp) RSQcal are used for daily analysis and other amino acids It can not be used for practical prediction. The best method of scattering is different among nutrients, metabolizable energy and amino acids.

【学位授予单位】：甘肃农业大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：S816.15

【参考文献】