CN110987996B - 一种判别进口铁矿石产地及品牌的方法 - Google Patents
一种判别进口铁矿石产地及品牌的方法 Download PDFInfo
- Publication number
- CN110987996B CN110987996B CN201911257106.9A CN201911257106A CN110987996B CN 110987996 B CN110987996 B CN 110987996B CN 201911257106 A CN201911257106 A CN 201911257106A CN 110987996 B CN110987996 B CN 110987996B
- Authority
- CN
- China
- Prior art keywords
- powder
- model
- iron ore
- content
- sample
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N23/00—Investigating or analysing materials by the use of wave or particle radiation, e.g. X-rays or neutrons, not covered by groups G01N3/00 – G01N17/00, G01N21/00 or G01N22/00
- G01N23/22—Investigating or analysing materials by the use of wave or particle radiation, e.g. X-rays or neutrons, not covered by groups G01N3/00 – G01N17/00, G01N21/00 or G01N22/00 by measuring secondary emission from the material
- G01N23/223—Investigating or analysing materials by the use of wave or particle radiation, e.g. X-rays or neutrons, not covered by groups G01N3/00 – G01N17/00, G01N21/00 or G01N22/00 by measuring secondary emission from the material by irradiating the sample with X-rays or gamma-rays and by measuring X-ray fluorescence
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N23/00—Investigating or analysing materials by the use of wave or particle radiation, e.g. X-rays or neutrons, not covered by groups G01N3/00 – G01N17/00, G01N21/00 or G01N22/00
- G01N23/22—Investigating or analysing materials by the use of wave or particle radiation, e.g. X-rays or neutrons, not covered by groups G01N3/00 – G01N17/00, G01N21/00 or G01N22/00 by measuring secondary emission from the material
- G01N23/2202—Preparing specimens therefor
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
- Developing Agents For Electrophotography (AREA)
Abstract
本发明公开了一种判别进口铁矿石产地及品牌的方法。其包括如下步骤:S1.取铁矿石中元素含量的数据,建立Fisher判别模型;S2.将待测样品铁矿石的元素含量代入到步骤S1的Fisher判别模型中,确定待测样品铁矿石的产地或品牌。本发明的判别方法具有快速、准确率高的优点。
Description
技术领域
本发明涉及一种判别进口铁矿石产地及品牌的方法。
背景技术
铁矿石是钢铁工业的重要原材料,不同产地来源的铁矿石由于地质成因差异,主次元素含量存在一定区域特征。澳大利亚、巴西、南非等国作为全球铁矿石最主要的出口国,主要铁矿产区相对集中。由于地质成因相似,与产地国别的识别相比,同一国家不同品牌铁矿石的识别将更有难度。
可见光-近红外光谱、微波介电光谱、激光诱导击穿光谱结合化学计量学或机器学习,可实现不同种类铁矿石的识别。武素茹等(中国,申请号:CN102012920A)以67个已知国别铁矿石样本X射线荧光光谱无标样分析数据为基础,采用逐步判别法筛选出CaO、MgO、Al2O3、CuO、V2O5五个特征变量,利用非参数判别方法建立进口国别的判别模型,准确率为74.6%。但至目前为止,尚没有参数判别分析方法在不同铁矿石识别中的报道,亦未有进口铁矿石品牌识别方法的报道。
故建立主要进口国进口铁矿石产地与品牌的快速识别模型,实现“少数”、“异常”铁矿石的快速筛选,对于支撑进口铁矿石的风险监管,保障贸易便利化,具有重要意义。
发明内容
本发明所要解决的技术问题在于解决现有技术中对进口铁矿石产地识别准确率低,且难以对进口铁矿石品牌进行识别的缺陷,而提供了一种判别进口铁矿石产地及品牌的方法。本发明的判别方法具有快速、准确率高的优点。
本发明通过以下技术方案解决上述技术问题。
本发明公开了一种判别铁矿石产地的方法,其包括如下步骤:
S1.取至少3个国别,每个国别至少16个批次的铁矿石中元素含量的数据,建立至少二维以上的Fisher判别模型;S2.将待测样品铁矿石的元素含量代到步骤S1的至少二维以上的Fisher判别模型中,确定待测样品铁矿石的产地;
其中,所述步骤S1和所述步骤S2的元素含量单位均为质量百分含量;
其中,当元素含量能被测试仪器检出,所述元素含量为测试仪器检出元素的含量;当元素含量不能被测试仪器检出,所述元素含量为测试仪器的检出限,所述检出限为0.0015-0.02。
本发明中,本领域技术人员知晓,用于建立模型的国别和批次的数据量为越多越好,因此对于国别和批次的数据量上限不作特别限定,较佳地,国别数为3~6个,批次数为16~298。
本发明中,对于元素含量,本领域技术人员知晓,不同的元素含量测试仪器对于每种元素的仪器检出限均略有偏差,一般可为0.0015-0.02。
本发明中,本领域技术人员知晓,所述步骤S1中的元素含量的测定与所述步骤S2中元素含量的测定为同一台检测仪器。
本发明中,所述元素含量的检测方法为本领域常规元素含量的检测方法,例如,波长色散X射线荧光光谱无标样分析方法或者能量色散X射线荧光光谱无标样分析方法。
较佳地,所述元素含量的检测方法为波长色散X射线荧光光谱中的无标样分析方法。
较佳地,所述元素含量的检测方法为波长色散X射线荧光光谱中的无标样分析方法时,所述步骤S1中的元素含量的测定,与所述步骤S2中的元素含量的测定为同一台检测机器。
本发明中,所述步骤S1中或步骤S2中的元素含量的检测方法,所述铁矿石一般按照本领域常规方法进行前处理,先对铁矿石干燥、之后压片,再进行检测步骤;具体地,可包括如下步骤:将每个铁矿石分装到干燥瓶中于105℃下烘干4h;采用压片机对烘干铁矿石进行压片,压片前用乙醇清洗模具,使用聚乙烯环使粉末铁矿石聚拢,压制所述粉末铁矿石在30t压力下维持30s;检查压制铁矿石样品表面均匀且无裂纹、脱落现象,测量前用洗耳球吹净所述压制铁矿石样品表面。
本发明中,所述至少二维以上的Fisher判别模型本领域技术人员可通过运用商用的软件自带的判别分析模块进行计算,例如,可以是SPSS软件;或者通过本领域人员知晓操作自行编写程序得出判别分析函数。
较佳地,所述至少二维以上的Fisher判别模型可为二维Fisher判别模型;所述元素可为Fe、O、Si、Ca、Al、Mn、Ti、Mg、P、S。
其中,所述二维Fisher判别模型中的2组判别函数为:
F1=0.525X1-0.598X2+1.4X3+32.627X4+0.654X5-3.936X6+37.01X7-29.4X8-58.953X9-24.002X10-16.337;
F2=0.569X1+0.855X2+0.122X3+7.559X4+1.23X5-4.789X6-9.846X7+4.281X8-128.56X9+147.622X10-61.555;
其中,式中X1-X10分别代表Fe、O、Si、Ca、Al、Mn、Ti、Mg、P、S的含量;
其中,所述二维Fisher判别模型中的各产地组质心的坐标为澳大利亚(-1.373,-0.179)、南非(8.003,-0.089)、巴西(-0.611,2.473)。
较佳地,所述至少二维以上的Fisher判别模型可为五维Fisher判别模型;所述元素可为Fe、O、Si、Ca、Al、Mn、Tb、Ti、Mg、P、Na、Cr、K、Sr、Zn、V、Cu、Ba、Ni、Mo、Pb。
其中,所述五维Fisher判别模型中的5组判别函数为:
F1=0.224X1-0.027X2+0.009X3+2.91X4+0.423X5-2.093X6-8.575X7+12.097X8-5.18X9-33.199X10-12.922X11-42.526X12+33.785X13+251.798X14-29.314X15-28.194X16-7.254X17+31.227X18+236.98X19-1591.987X20+518.554X21-15.262
F2=0.174X1-0.057X2+0.032X3+12.148X4+1.047X5-0.885X6+29.473X7-0.112X8-4.821X9-55.242X10+3.792X11-37.251X12-0.76X13+104.144X14+149.363X15+27.059X16+56.533X17-6.595X18-322.156X19-1386.114X20-705.416X21-4.88
F3=0.028X1+0.907X2-0.039X3-3.382X4-0.758X5-0.185X6-31.506X7-4.287X8+2.219X9+60.299X10+15.683X11-23.216X12+15.861X13+96.388X14-24.214X15+27.493X16+24.853X17+2.376X18-75.75X19+1480.018X20-210.037X21-31.304
F4=0.38X1+0.335X2-0.034X3+3.876X4+0.795X5-2.243X6+24.004X7+16.466X8-8.717X9-34.761X10+11.003X11+153.562X12+3.595X13-135.312X14+3.981X15-24.752X16-25.761X17-38.039X18+276.842X19-1076.789X20+257.631X21-36.297
F5=-0.079X1+0.566X2+0.741X3+5.228X4-0.342X5-1.334X6-23.369X7-7.439X8+2.123X9+5.964X10-2.621X11+74.847X12+3.689X13+25.371X14-57.22X15-6.903X16-138.511X17-22.433X18+463.216X19+682.723X20+885.888X21-21.689
其中,式中X1-X21分别代表Fe、O、Si、Ca、Al、Mn、Tb、Ti、Mg、P、Na、Cr、K、Sr、Zn、V、Cu、Ba、Ni、Mo、Pb的含量;
其中,所述五维Fisher判别模型中的各产地组质心的坐标为澳大利亚(-1.608,-1.665,0.476,0.131,0.275)、巴西(-0.883,-0.878,-2.994,0.507,-3.479)、南非(10.092,0.914,0.943,-0.437,-0.07)、哈萨克斯坦(-4.617,14.213,3.55,-0.479,-0.507)、加拿大(-2.326,2.536,-7.937,-5.762,1.015)、印度(2.507,8.331,-8.608,4.423,1.947)。
本发明中,步骤S2,将一待测样品铁矿石的元素含量代入步骤S1的至少二维以上的Fisher判别模型中,确定待测样品铁矿石的产地的确定方式,本领域技术人员知晓为,函数F1得分为横坐标,函数F2得分为纵坐标,根据判别函数和组质心处坐标函数,计算每个样品坐标与质心的距离,与哪个类别的质心最近,该样品就判定为哪个产地类别。
本领域技术人员知晓,本发明的判别铁矿石产地的方法测定待测样品的适用范围为建立至少二维以上Fisher判别模型的国别范围。
本发明公开了一种判别铁矿石品牌的方法,其包括如下步骤:
S1.取至少14个品牌,每个品牌至少6个批次的铁矿石中元素含量的数据,建立至少十维以上的Fisher判别模型;S2.将待测样品铁矿石的元素含量代到步骤S1的至少十维以上的Fisher判别模型中,确定待测样品铁矿石的品牌;
其中,所述步骤S1和所述步骤S2的元素含量单位均为质量百分含量;
其中,当元素含量能被测试仪器检出,所述元素含量为测试仪器检出元素的含量;当元素含量不能被测试仪器检出,所述元素含量为测试仪器的检出限,所述检出限为0.0015-0.02。
本发明中,本领域技术人员知晓,用于建立模型的品牌和批次的数据量为越多越好,因此对于品牌和批次的数据量上限不作特别限定,较佳地,品牌数为14~22个,批次数为6~47。
本发明中,对于元素含量,本领域技术人员知晓,不同的元素含量测试仪器对于每种元素的仪器检出限均略有偏差,一般可为0.0015-0.02。
本发明中,本领域技术人员知晓,所述步骤S1中的元素含量的测定与所述步骤S2中元素含量的测定为同一台检测仪器。
本发明中,所述元素含量的检测方法为本领域常规元素含量的检测方法,例如,波长色散X射线荧光光谱无标样分析方法或者能量色散X射线荧光光谱无标样分析方法。
较佳地,所述元素含量的检测方法为波长色散X射线荧光光谱中的无标样分析方法。
较佳地,所述元素含量的检测方法为波长色散X射线荧光光谱中的无标样分析方法时,所述步骤S1中的元素含量的测定,与所述步骤S2中的元素含量的测定为同一台检测机器。
本发明中,所述步骤S1中或步骤S2中的元素含量的检测方法,所述铁矿石一般按照本领域常规方法进行前处理,先对铁矿石干燥、之后压片,再进行检测步骤;具体地,可包括如下步骤:将每个铁矿石分装到干燥瓶中于105℃下烘干4h;采用压片机对烘干铁矿石进行压片,压片前用乙醇清洗模具,使用聚乙烯环使粉末铁矿石聚拢,压制所述粉末铁矿石在30t压力下维持30s;检查压制铁矿石样品表面均匀且无裂纹、脱落现象,测量前用洗耳球吹净所述压制铁矿石样品表面。
本发明中,所述至少十维以上Fisher判别模型本领域技术人员可通过运用商用的软件自带的判别分析模块进行计算,例如,可以是SPSS软件;或者通过本领域人员知晓操作自行编写程序得出判别分析函数。
较佳地,所述至少十维以上的Fisher判别模型可为十维Fisher判别模型;所述元素为Fe、O、Si、Ca、Al、Mn、Ti、Mg、P、S。
其中,所述十维Fisher判别模型中的10组判别函数为:
F1=0.286X1-0.372X2+0.404X3-9.8X4-0.575X5-2.551X6+54.081X7+18.203X8-95.378X9-17.295X10-12.937
F2=0.726X1-1.244X2+2.718X3-2.715X4-0.808X5-5.169X6-23.321X7+15.246X8-160.116X9-18.187X10-2.608
F3=0.587X1-1.372X2+1.45X3+8.922X4+3.122X5+3.139X6+9.551X7-15.179X8+180.272X9+21.586X10-4.251
F4=0.257X1-0.263X2+0.896X3+33.291X4-1.573X5+1.075X6+37.848X7-34.022X8-62.815X9+25.649X10-6.865
F5=-0.4X1-0.453X2+1.121X3-13.49X4-1.417X5+11.016X6+9.278X7+0.466X8+19.539X9+87.158X10+35.004
F6=-0.012X1+0.213X2+1.171X3+6.748X4+2.843X5-6.586X6+12.703X7-2.634X8-138.53X9-59.512X10-9.303
F7=-0.189X1+0.221X2+2.048X3-4.649X4-3.679X5-0.974X6+10.841X7-7.985X8+221.911X9-69.058X10-0.463
F8=0.337X1+0.494X2+0.457X3+7.023X4-0.625X5-1.537X6-17.619X7+2.087X8+6.88X9+229.064X10-38.458
F9=0.518X1-0.035X2+0.885X3-14.995X4-0.257X5-5.636X6+26.099X7-4.9X8-30.377X9+159.188X10-31.76
F10=1.526X1+1.558X2+1.681X3-1.042X4-0.188X5+4.851X6+6.292X7+0.682X8+7.995X9-75.431X10-145.479
其中,式中X1-X10分别代表Fe、O、Si、Ca、Al、Mn、Ti、Mg、P、S的含量;
其中,所述十维Fisher判别模型中的各品牌组质心的坐标为PB粉(-9.129,-5.309,2.278,-2.045,-0.884,0.199,-0.111,0.337,-0.084,-0.183)、PB块(-8.865,-1.735,-0.529,0.196,-0.914,-1.943,-0.232,0.23,-0.533,0.188)、杨迪粉(-6.897,-0.581,-7.854,0.633,-1.068,0.877,-0.103,-0.226,0.014,-0.119)、纽块(-8.441,0.822,0.838,-0.645,-0.322,-1.095,0.902,0.533,0.658,0.14)、纽粉(-7.648,-2.205,4.196,-3.965,-1.744,3.552,-0.685,-1.19,-0.52,0.917)、津布巴粉(-10.504,-7.276,8.091,-4.139,-0.909,3.428,-0.268,-0.029,0.747,-0.249)、南非粉(-7.329,5.853,6.054,7.442,-2.288,1.773,-0.352,0.422,-0.481,-0.287)、南非块(-6.676,8.807,5.167,6.15,-1.532,0.518,0.25,-0.301,0.646,0.334)、卡拉粉(-6.188,-2.231,4.843,0.443,0.096,-3.793,-1.303,-2.079,0.65,-0.329)、澳精粉(3.146,26.733,-0.494,-4.019,1.429,0.067,-0.61,0.172,-0.033,-0.09)、巴西粉(-6.304,2.881,4.157,-0.639,3.26,0.204,2.768,-1.133,-1.082,-0.266)、国王粉(-6.642,-1.598,-6.442,0.817,-0.226,0.812,1.245,-0.737,0.657,0.168)、混合粉(-7.292,-4.834,-0.377,1.758,6.498,0.57,-0.581,0.249,0.076,0.085)、南非精粉(109.502,-2.359,0.254,0.132,-0.191,-0.026,0.037,0.035,-0.001,0.008)。
较佳地,所述至少十维以上的Fisher判别模型可为二十一维Fisher判别模型;所述元素为Fe、O、Si、Ca、Al、Mn、Ti、Mg、P、Na、Cr、K、Sr、S、Zr、Zn、V、Cu、Ba、Cl、Ni、Mo、Pb。
其中,所述二十一维Fisher判别模型中的21组判别函数为:
F1=-0.053X1-0.099X2+0.048X3-0.253X4-0.402X5+1.104X6+14.021X7+1.777X8+0.356X9+1.422X10-87.361X11+1.833X12+82.245X13-61.82X14-68.402X15-5.214X16+323.547X17+157.553X18+17.288X19+34.918X20+1369.087X21+2182.814X22-748.481X23-13.645
F2=-0.055X1+0.108X2+0.267X3-1.323X4+1.175X5-0.492X6-62.999X7-1.451X8+14.334X9-2.202X10+80.315X11+11.578X12-102.336X13+33.267X14+336.236X15+2.148X16+287.835X17-169.74X18+1.896X19-16.006X20+478.101X21+134.578X22+1.034X23-2.208
F3=0.298X1-0.354X2+0.619X3-2.409X4-0.265X5+0.715X6-14.828X7+0.487X8+10.347X9+2.138X10+1.867X11+48.519X12+80.646X13-48.887X14+113.252X15+70.38X16-0.504X17+139.509X18+71.743X19-9.7932X20+33.683X21+7882.507X22-1789.426X23-17.431
F4=-0.532X1+0.764X2-0.282X3-1.331X4-0.945X5+6.025X6-1.482X7-5.478X8+36.982X9-0.776X10+14.694X11-28.874X12-96.675X13+10.312X14+47.168X15+73.511X16+24.272X17+133.474X18-16.237X19+59.559X20+190.543X21+2321.651X22-1278.666X23+10.576
F5=0.032X1+0.149X2-0.666X3-2.578X4+1.148X5-3.554X6+0.819X7-5.101X8+84.759X9-13.926X10-16.535X11+34.577X12+39.313X13+31.531X14-62.101X15-61.894X16+6.081X17+38.786X18+4.221X19-13.467X20+332.102X21-1020.083X22+428.29X23-11.97
F6=0.42X1-1.19X2+0.299X3+1.469X4+4.142X5+2.331X6-6.509X7-1.344X8+90.751X9-5.005X10+2.877X11-19.373X12-166.907X13+27.71X14+3.087X15+65.946X16+9.687X17+41.817X18-15.122X19-7.81X20-57.365X21-2017.834X22-224.898X23+9.633
F7=-0.065X1+0.66X2+0.071X3+1.662X4+1.355X5+2.717X6+17.37X7-6.992X8-90.46X9+5.554X10+52.493X11+15.791X12+50.956X13-6.753X14+24.939X15+45.909X16+6.518X17-91.513X18-4.024X19+27.505X20-180.137X21-1092.11X22+113.677X23-18.656
F8=0.083X1-0.583X2+0.975X3-3.832X4+0.037X5+7.319X6+6.86X7+0.685X8-35.784X9-10.062X10+49.473X11+2.357X12-38.99X13+60.527X14+86.109X15-105.696X16-35.757X17-134.998X18-2.241X19+100.841X20+321.311X21+128.302X22+549.514X23+5.451
F9=0.021X1+0.268X2+0.401X3-7.766X4+0.964X5+0.896X6-3.487X7+8.214X8+48.753X9+17.048X10+28.419X11+11.511X12-105.516X13+29.067X14+3.746X15-101.435X16+2.703X17-61.216X18-19.516X19+9.189X20-58.156X21+1142.793X22+487.16X23-18.715
F10=0.286X1-0.152X2-0.837X3-1.096X4-1.413X5+10.093X6+0.52X7-5.173X8+62.646X9+8.97X10-129.582X11+3.098X12+16.964X13+41.966X14-73.943X15+34.131X16+13.738X17+117.362X18+3.936X19-10.978X20-226.571X21-1977.339X22-846.084X23-0.723
F11=0.353X1+0.221X2-0.105X3+1.002X4+0.418X5-2.509X6+14.736X7-7.321X8-49.629X9+12.515X10+166.053X11+2.89X12-154.775X13-39.078X14+209.141X15+5.204X16-31.122X17-10.934X18-38.433X19-7.976X20+316.914X21+48.649X22-55.111X23-29.398
F12=0.29X1+0.301X2+0.69X3+4.614X4-1.695X5+1.509X6-4.466X7-3.045X8+75.408X9+2.05X10+59.282X11-2.687X12+94.997X13+187.997X14-128.428X15-41.09X16-2.475X17-147.37X18-36.871X19-48.664X20+371.102X21-4687.471X22+577.282X23-25.107
F13=0.257X1-0.112X2+0.88X3+0.017X4+0.409X5-3.497X6-1.65X7-3.896X8-32.246X9+0.391X10-138.738X11-11.124X12+391.07X13+43.989X14+118.64X15+125.925X16+7.983X17+67.953X18-50.208X19+21.613X20-20.278X21-3933.864X22-540.687X23-2.469
F14=-0.093X1-0.243X2-0.892X3-0.152X4-0.904X5-6.242X6+2.651X7-0.579X8+53.174X9+1.514X10+38.229X11+0.292X12-36.222X13+34.625X14-60.743X15+12.447X16+2.586X17-79.918X18+6.034X19+177.337X20-153.521X21-1149.34X22+215.381X23+15.881
F15=0.248X1+0.377X2-0.251X3+2.776X4+0.659X5+1.686X6-12.487X7+2.43X8-28.118X9+1.007X10-29.367X11-0.693X12+201.635X13-103.05X14-108.573X15-174.727X16+3.728X17+120.573X18-34.519X19+100.36X20+224.851X21+2768.54X22+1534.895X23-40.454
F16=0.026X1-0.052X2+0.516X3-0.632X4-0.176X5-0.523X6+14.344X7-2.21X8-9.7X9+4.072X10+60.055X11-13.256X12+109.5X13+28.814X14-311.137X15-207.363X16+19.933X17+122.527X18+54.953X19-4.244X20-641.366X21+407.307X22+1149.219X23-6.933
F17=0.18X1+0.108X2+0.648X3+2.113X4-0.331X5-0.494X6-14.564X7-3.738X8+21.527X9+4.983X10-159.078X11-8.599X12-93.444X13-42.289X14+557.903X15-68.149X16+15.476X17+193.542X18+79.988X19+41.394X20+263.677X21-454.154X22+398.07X23-19.574
F18=0.48X1+0.324X2+1.307X3+0.854X4-0.298X5+1.029X6+8.675X7-0.564X8+31.132X9-5.279X10-55.278X11-4.933X12-108.48X13-87.987X14-305.093X15+107.737X16+2.717X17+4.216X18+37.751X19+51.161X20-229.377X21+2021.426X22-819.479X23-39.765
F19=0.444X1+0.045X2-0.338X3+0.547X4+1.112X5-1.242X6+10.795X7+2.901X8-91.493X9+0.20X10+621.922X11-11.035X12+57.024X13+75.969X14+153.722X15-1.446X16-24.934X17-221.679X18+37.566X19-21.537X20-90.49X21+1032.202X22-112.522X23-28.029
F20=0.124X1-0.035X2-0.11X3-1.098X4+0.781X5-0.327X6+1.073X7+2.123X8-49.061X9+5.011X10-35.986X11-13.867X12+150.829X13+40.23X14-244.636X15+56.799X16-21.129X17-61.113X18+63.27X19-4.425X20+176.578X21-750.763X22-272.693X23-4.584
F21=0.104X1+0.143X2-0.047X3+1.361X4+0.446X5+0.109X6-17.305X7+1.318X8-29.216X9+2.598X10+81.184X11-10.867X12+146.848X13+10.103X14+21.72X15+81.689X16-1.978X17+66.532X18+47.961X19+19.317X20+486.92X21-1961.691X22-1496.159X23-2.188
其中,式中X1-X23分别代表Fe、O、Si、Ca、Al、Mn、Ti、Mg、P、Na、Cr、K、Sr、S、Zr、Zn、V、Cu、Ba、Cl、Ni、Mo、Pb的含量;
其中,所述二十一维Fisher判别模型中的各品牌组质心的坐标为津布巴粉(-13.457,4.113,-4.925,2.63,4.331,7.5,0.569,-0.793,3.403,-2.302,-0.164,0.017,1.076,-0.79,-0.571,-0.033,0.899,-0.521,0.311,-0.014,0.004)、巴西精粉(-11.852,2.114,-0.785,-2.759,-1.517,3.797,-3.573,2.455,-1.977,0.07,1.254,-0.808,0.328,0.077,-1.034,1.092,-0.607,0.295,0.343,-0.025,-0.007)、PB块(-12.721,1.928,-3.75,1.464,0.556,-0.327,-1.725,-1.283,-0.681,1.615,-0.059,1.013,-0.477,0.248,0.342,-0.246,0.005,-0.000092,0.243,-0.005,-0.03)、纽块(-12.675,2.457,-2.951,0.622,0.462,0.865,-2.689,-0.652,-0.267,0.63,-0.311,0.866,0.398,0.191,0.017,0.15,0.473,0.454,-0.131,0.009,0.017)、国王粉(-12.3,0.362,-4.713,3.919,-1.242,-5.411,0.078,-2.341,-1.758,-0.926,0.732,-0.359,0.721,-0.028,-1.83,-1.075,-0.292,0.223,0.01,0.023,0.002)、PB粉(-12.742,2.104,-4.491,2.504,2.247,2.415,-0.472,-1.311,1.185,-0.287,-0.053,0.392,-0.427,0.269,-0.035,-0.03,-0.379,-0.157,-0.102,-0.058,0.021)、澳大利亚球团矿(167.931,80.586,-4.04,2.091,2.198,-0.503,-2.073,1.832,-0.795,-0.465,0.34,0.017,-0.104,-0.021,0.021,-0.07,0.023,-0.033,-0.018,0,-0.0000016)、杨迪粉(-11.566,1.426,-5.122,3.525,-1.56,-7.06,0.194,-2.217,-1.297,-1.703,0.289,-0.241,0.066,-0.458,0.1,0.409,0.181,-0.128,-0.139,-0.05,-0.022)、哈杨粉(-11.675,1.307,-4.569,2.237,-1.094,-5.382,-0.672,-2.406,-1.701,-0.722,0.684,-0.935,-0.311,0.013,0.762,0.087,-0.018,-0.125,0.292,0.089,0.056)、纽粉(-12.541,2.958,-3.946,0.875,1.778,4.241,-0.24,-0.595,2.646,-3.224,-0.531,-1.718,-0.291,-0.365,0.659,-0.245,-0.228,0.452,-0.029,0.027,-0.022)、南非精粉(87.288,-77.489,-2.119,-1.063,0.445,-0.007,0.054,-0.2,0.317,-0.087,-0.024,0.072,0.015,-0.004,-0.003,0.014,-0.001,0.011,0.008,0,-0.0000344)、澳精粉(-10.919,2.666,8.823,-12.193,-6.046,-2.691,-4.029,2.182,4.786,-0.312,1.478,0.118,-0.236,0.018,-0.013,-0.152,0.084,-0.123,-0.044,0.001,0)、混合粉(-11.55,0.445,-4.604,6.56,-1.641,-1.132,3.715,4.582,1.276,1.858,-0.25,0.421,-0.267,-0.956,-0.027,0.005,-0.104,0.074,-0.005,0.012,0.006)、-卡拉粉(-10.649,-2.193,-1.721,0.961,0.94,4.042,-2.802,-0.876,-2.102,4.902,0.919,-2.245,0.054,-0.134,0.008,-0.096,0.185,-0.277,-0.221,0.021,-0.009)、哈萨克斯坦球团矿(29.37,22.461,3.419,-9.19,-9.964,2.379,8.943,-7.398,2.69,2.48,-1.248,-0.192,0.375,0.129,-0.127,0.236,-0.091,0.12,0.054,-0.001,0)、哈萨克斯坦粉(5.324,-1.676,53.549,16.806,-2.595,1.847,-0.666,-0.926,-0.774,-0.701,0.025,0.13,-0.03,0.013,0.054,0.0000177,-0.004,-0.006,0.006,0.001,0)、加拿大精粉(-9.729,-0.837,-1.004,-6.883,-9.177,2.125,-3.365,3.298,-5.229,-1.649,-4.658,-0.355,-0.276,-0.019,-0.122,-0.201,0.052,-0.244,-0.004,-0.014,0.007)、南非粉(-10.942,3.67,13.141,-11.086,10.607,-3.335,3.506,0.648,-0.612,0.202,-0.887,-0.46,-1.124,0.134,-0.527,0.112,0.26,0.06,0.015,0.006,0.001)、超特粉(-10.997,-0.634,-5.74,7.898,-2.006,-2.31,5.426,5.539,1.253,-0.334,0.217,-1.305,0.673,2.162,0.238,-0.072,0.337,-0.012,0.06,-0.043,-0.005)、麦克粉(-13.041,2.766,-4.999,3.451,1.794,0.67,0.024,-0.052,1.676,-1.23,-0.856,1.174,0.267,0.72,-0.411,0.46,-0.35,-0.397,-0.23,0.183,-0.026)、南非块(-10.75,3.292,8.8,-9.099,5.925,-2.103,0.617,1.238,-1.378,0.607,-0.356,0.366,2.099,-0.265,0.789,-0.146,-0.487,-0.065,-0.002,-0.016,0)、印度球团矿(-12.147,-0.212,-0.848,-7.93,-3.383,7.471,6.083,1.264,-6.712,-2.692,3.374,1.086,-0.421,-0.049,0.275,-0.183,0.146,-0.031,-0.116,0.018,-0.002)。
本发明中,步骤S2,将一待测样品铁矿石的元素含量代入步骤S1的至少十维以上Fisher判别模型中,确定待测样品铁矿石的产地的确定方式,本领域技术人员知晓为,根据判别函数和组质心处坐标函数,计算每个样品坐标与组质心坐标的距离,与哪个类别的质心最近,该样品就判定为哪个品牌类别。
本领域技术人员知晓,本发明的判别铁矿石品牌的方法测定待测样品的适用范围为建立至少十维以上Fisher判别模型的品牌范围。
在符合本领域常识的基础上,上述各优选条件,可任意组合,即得本发明各较佳实例。
本发明所用试剂和原料均市售可得。
本发明的积极进步效果在于:
(1)一种判别进口铁矿石产地及品牌的方法,选取铁矿石中元素含量的数据建立至少二维以上的Fisher判别模型,分别对铁矿石国家和品牌的进行识别;其中,当元素含量能被测试仪器检出,所述元素含量为测试仪器检出元素的含量;当元素含量不能被测试仪器检出,所述元素含量为测试仪器的检出限;
(2)一种判别进口铁矿石产地及品牌的方法,可以快速、准确地进行国家、品牌的识别,适用范围广。
附图说明
图1为实施例1中判别二维函数产地得分散点图。
图2为实施例2中判别三维函数品牌得分三维散点图。
具体实施方式
下面通过实施例的方式进一步说明本发明,但并不因此将本发明限制在所述的实施例范围之中。下列实施例中未注明具体条件的实验方法,按照常规方法和条件,或按照商品说明书选择。
实施例1.1-1.236
本系列实施例为判别铁矿石产地国别的方法。
1、样品收集
根据GB/T 10322.1-2014《铁矿石取样和制样方法》,从我国主要的铁矿石进口口岸采集并制备来自澳大利亚、南非、巴西3个国家的进口铁矿石化学分析样品,包含14个品牌的共计236批次样品。所述14个品牌铁矿石包括皮尔巴拉混合粉、皮尔巴拉混合块、杨迪粉铁矿、纽曼混合块铁矿、纽曼混合粉铁矿、津布巴混合粉铁矿、国王粉、弗特斯克混合粉、澳大利亚铁矿石精粉、昆巴标准粉、昆巴标准块、南非铁矿石精粉、巴西混合粉铁矿、卡拉加斯铁矿石。样品信息如表1所示。
表1铁矿石样品信息
2、样品检测
将样品分装到干燥瓶中于105℃下烘干4h。采用压片机对烘干样品进行压片,压片前用乙醇清洗模具,使用聚乙烯环使粉末样品聚拢,压制样品在30t压力下维持30s。检查压制样品表面均匀且无裂纹、脱落现象,测量前用洗耳球吹净样品表面。
使用德国布鲁克公司S4 Pioneer波长色散-X射线荧光光谱仪中的无标样分析方法检测铁矿石中元素的含量。检测中使用铑靶光管、四个分析仪晶体(LiF200、XS-55、PET和Ge)、流气计数器(FC)、闪烁计数器(SC)等元件。表2列出了仪器的部分测量条件。
表2仪器部分测量条件
针对采集的236个铁矿石样品,采用波长色散X射线荧光光谱无标样分析可以检测到Fe、O、Si、Ca、Al、Mn、Tb、Ti、Mg、P、K、S、Cr、Na、Sr、Zr、Zn、V、Cu、Gd、Ba、Cl、Ni、Co共24种元素的含量,其中K、Cu、Zr、Zn、Na、Cl、V、Sr、Gd、Ni、Ba、Co共12个元素含量存在未检出的情况,未检出比例分别为18.20%、50.00%、51.00%、69.90%、70.30%、73.30%、78.00%、83.90%、84.30%、91.50%、92.80%、97.00%,部分检出元素含量接近方法检测限,考虑到结果带来的误差,分析过程中选取236个样品全部检出的Fe、O、Si、Ca、Al、Mn、Tb、Ti、Mg、P、Cr、S共12种元素含量用于后续分析,元素含量测定结果如表3所示。
表3 12种元素含量测定结果
3、逐步判别分析
采用逐步判别分析对Fe、O、Si、Ca、Al、Mn、Tb、Ti、Mg、P、Cr、S共12个元素含量进行变量筛选,变量能否进入模型主要取决于协方差分析的F检验的显著性水平,当F值大于指定值时保留该变量,而F值小于指定值时,该变量从模型中剔除。选取合适的F值可以用最少的变量达到最佳的判别效果。其中,选取的F值为3.84,经过逐步判别分析,Fe、O、Si、Ca、Al、Mn、Ti、Mg、P、S共10个元素保留在了模型中,Tb与Cr因未通过F检验(F值<3.84)而从模型中剔除,最终10个元素用于建立识别模型。
4、建立二维Fisher模型
使用Fe、O、Si、Ca、Al、Mn、Ti、Mg、P、S共10个元素含量,建立二维Fisher判别模型和相应的组质心处的坐标。
所述二维Fisher判别模型中的2组判别函数为:
F1=0.525X1-0.598X2+1.4X3+32.627X4+0.654X5-3.936X6+37.01X7-29.4X8-58.953X9-24.002X10-16.337;
F2=0.569X1+0.855X2+0.122X3+7.559X4+1.23X5-4.789X6-9.846X7+4.281X8-128.56X9+147.622X10-61.555;式中X1-X10分别代表Fe、O、Si、Ca、Al、Mn、Ti、Mg、P、S的含量;
其中,式中X1-X10分别代表Fe、O、Si、Ca、Al、Mn、Ti、Mg、P、S的含量;所述含量为质量百分含量;
所述二维Fisher判别模型中的各产地组质心的坐标为澳大利亚(-1.373,-0.179)、南非(8.003,-0.089)、巴西(-0.611,2.473)。
将一待测样品铁矿石的元素含量代入所述二维Fisher判别模型中,函数F1得分为横坐标,函数F2得分为纵坐标,根据判别函数和组质心处坐标函数,计算每个样品坐标与质心的距离,与哪个类别的质心最近,该样品就判定为哪个产地类别。
如图1所示的具体分类结果,可以看出模型对南非铁矿石和其它两个国家的铁矿石有明显的区分,在对澳大利亚的铁矿石识别中有少部分样品会落在距离巴西质心更近的位置。
本次判别铁矿石产地国别的方法的实施例包括两类实施例:
第一类为构建模型所用的样品,即建模样品实施例。其中,建模样品实施例分别进行建模样品验证和交叉验证。建模样品验证为将构建模型所用的样品数据回代到模型,进行验证;交叉验证法为建模前每次留出一个作为验证的数据再次代入判别函数,进行验证。经过统计,所建立的二维判别模型对产地的识别正确率如下表4所示,对建模样品分类正确率为97.40%,对南非的铁矿石样品识别正确率为100%,对澳大利亚、巴西铁矿石样品存在识别错误的情况,正确率分别为97.40%、91.70%;交叉验证正确率为95.30%。
第二类为采用未知的待测样品验证的实施例。为了确定二维Fisher判别模型是否可以对未包含在模型中的样品进行识别,选择45个作为测试样品的铁矿石样品。经过统计,所建立的二维判别模型对产地的识别正确率如下表4所示,识别正确率达到95.50%,其中对南非和巴西样品识别正确度都达到100%,说明此模型可以对铁矿石的国别进行很好的识别。
表4二维判别模型识别产地的正确率
实施例编号 | 建模样品验证 | 交叉验证 | 实施例编号 | 测试样品验证 |
1-152 | 97.4% | 95.4% | 192-224 | 93.9% |
153-179 | 100% | 100% | 225-232 | 100% |
180-191 | 91.7% | 83.3% | 233-236 | 100% |
总计 | 97.4% | 95.3% | 总计 | 95.5% |
实施例2.1-2.236
本实施例中的样品收集、样品检测、逐步判别分析均同实施例1.1~1.236。
本系列实施例为判别铁矿石品牌的方法。
使用Fe、O、Si、Ca、Al、Mn、Ti、Mg、P、S共10个元素含量建立十维Fisher判别模型和相应的组质心处的坐标。
所述十维Fisher判别模型中的10组判别函数为:
F1=0.286X1-0.372X2+0.404X3-9.8X4-0.575X5-2.551X6+54.081X7+18.203X8-95.378X9-17.295X10-12.937
F2=0.726X1-1.244X2+2.718X3-2.715X4-0.808X5-5.169X6-23.321X7+15.246X8-160.116X9-18.187X10-2.608
F3=0.587X1-1.372X2+1.45X3+8.922X4+3.122X5+3.139X6+9.551X7-15.179X8+180.272X9+21.586X10-4.251
F4=0.257X1-0.263X2+0.896X3+33.291X4-1.573X5+1.075X6+37.848X7-34.022X8-62.815X9+25.649X10-6.865
F5=-0.4X1-0.453X2+1.121X3-13.49X4-1.417X5+11.016X6+9.278X7+0.466X8+19.539X9+87.158X10+35.004
F6=-0.012X1+0.213X2+1.171X3+6.748X4+2.843X5-6.586X6+12.703X7-2.634X8-138.53X9-59.512X10-9.303
F7=-0.189X1+0.221X2+2.048X3-4.649X4-3.679X5-0.974X6+10.841X7-7.985X8+221.911X9-69.058X10-0.463
F8=0.337X1+0.494X2+0.457X3+7.023X4-0.625X5-1.537X6-17.619X7+2.087X8+6.88X9+229.064X10-38.458
F9=0.518X1-0.035X2+0.885X3-14.995X4-0.257X5-5.636X6+26.099X7-4.9X8-30.377X9+159.188X10-31.76
F10=1.526X1+1.558X2+1.681X3-1.042X4-0.188X5+4.851X6+6.292X7+0.682X8+7.995X9-75.431X10-145.479
其中,式中X1-X10分别代表Fe、O、Si、Ca、Al、Mn、Ti、Mg、P、S的含量;所述含量为质量百分含量;
其中,所述十维Fisher判别模型中的各品牌组质心的坐标为PB粉(-9.129,-5.309,2.278,-2.045,-0.884,0.199,-0.111,0.337,-0.084,-0.183)、PB块(-8.865,-1.735,-0.529,0.196,-0.914,-1.943,-0.232,0.23,-0.533,0.188)、杨迪粉(-6.897,-0.581,-7.854,0.633,-1.068,0.877,-0.103,-0.226,0.014,-0.119)、纽块(-8.441,0.822,0.838,-0.645,-0.322,-1.095,0.902,0.533,0.658,0.14)、纽粉(-7.648,-2.205,4.196,-3.965,-1.744,3.552,-0.685,-1.19,-0.52,0.917)、津布巴粉(-10.504,-7.276,8.091,-4.139,-0.909,3.428,-0.268,-0.029,0.747,-0.249)、南非粉(-7.329,5.853,6.054,7.442,-2.288,1.773,-0.352,0.422,-0.481,-0.287)、南非块(-6.676,8.807,5.167,6.15,-1.532,0.518,0.25,-0.301,0.646,0.334)、-卡拉粉(-6.188,-2.231,4.843,0.443,0.096,-3.793,-1.303,-2.079,0.65,-0.329)、澳精粉(3.146,26.733,-0.494,-4.019,1.429,0.067,-0.61,0.172,-0.033,-0.09)、巴西粉(-6.304,2.881,4.157,-0.639,3.26,0.204,2.768,-1.133,-1.082,-0.266)、国王粉(-6.642,-1.598,-6.442,0.817,-0.226,0.812,1.245,-0.737,0.657,0.168)、混合粉(-7.292,-4.834,-0.377,1.758,6.498,0.57,-0.581,0.249,0.076,0.085)、南非精粉(109.502,-2.359,0.254,0.132,-0.191,-0.026,0.037,0.035,-0.001,0.008)。
将一待测样品铁矿石的元素含量代入所述十维Fisher判别模型中,根据判别函数和组质心处坐标函数,计算每个样品坐标与组质心坐标的距离,与哪个类别的质心最近,该样品就判定为哪个品牌类别。
所建立的十维判别模型对品牌的识别正确率如下表5所示,模型对测试样品识别的准确率有明显提高,正确率达到了100%,所建立识别模型具有很好识别效果。
实施例3.1-3.236
本实施例中的样品收集、样品检测、逐步判别分析均同实施例1.1~1.236。
本系列对比例为判别铁矿石品牌的方法。
使用Fe、O、Si、Ca、Al、Mn、Ti、Mg、P、S共10个元素含量建立三维Fisher判别模型和相应的组质心处的坐标。
所述三维Fisher判别模型中的3组判别函数为:
F1=0.286X1-0.372X2+0.404X3-9.8X4-0.575X5-2.551X6+54.081X7+18.203X8-95.378X9-17.295X10-12.937
F2=0.726X1-1.244X2+2.718X3-2.715X4-0.808X5-3.169X6-23.321X7+15.246X8-160.116X9-18.187X10-2.608
F3=0.587X1-1.372X2+1.45X3+8.922X4+3.122X5+3.139X6+9.551X7-15.179X8+180.272X9+21.586X10-4.251
其中,式中X1-X10分别代表Fe、O、Si、Ca、Al、Mn、Ti、Mg、P、S的含量;所述含量为质量百分含量;
其中,所述三维Fisher判别模型中的各品牌组质心的坐标为PB粉(-9.129,-5.309,2.278)、PB块(-8.865,-1.735,-0.529)、杨迪粉(-6.897,-0.581,-7.854)、纽块(-8.441,0.822,0.838)、纽粉(-7.648,-2.205,4.196)、津布巴粉(-10.504,-7.276,8.091)、南非粉(-7.329,5.853,6.054)、南非块(-6.676,8.807,5.167)、卡拉粉(-6.188,-2.231,4.843)、澳精粉(3.146,26.733,-0.494)、巴西粉(-6.304,2.881,4.157)、国王粉(-6.642,-1.598,-6.442)、混合粉(-7.292,-4.834,-0.377)、南非精粉(109.502,-2.359,0.254)。
将一待测样品铁矿石的元素含量代入所述三维Fisher判别模型中,以函数F1、F2、F3得分为坐标,绘制三维散点图;计算每个样品坐标与组质心坐标的距离,与哪个类别的质心最近,该样品就判定为哪个品牌类别。
如图2所示,三维散点图分布表明,14个品牌的铁矿石可明显地被划分为四个区域,澳精粉和南非精粉与其它类别区分最为明显。从图中还可以看出PB块与纽曼块分类略有重叠,杨迪粉与国王粉的集群非常接近。
所建立的三维判别模型对品牌的识别正确率如下表5所示。结果表明:模型对澳精粉与南非精粉识别完全正确,因为与其他类别的铁矿石相比,这两类的Ti与Mg的含量与其它类别有明显不同。模型对于PB块、杨迪粉、纽块会存在识别错误的情况。这三个品牌的铁矿石都产于澳大利亚皮尔巴拉地区的哈默斯利铁矿带,矿石成因类似、元素含量比较接近,因此相对于其它类别更难以区分。
所建立的三维判别模型虽然对品牌的识别正确率不如实施例2.1-2.236所建立的十维判别模型,但是其识别准确率仍高于现有技术。
表5三维/十维判别模型识别品牌的正确率
实施例4.1-4.434
本系列实施例为判别铁矿石产地国别的方法。
1、样品收集
根据GB/T 10322.1-2014《铁矿石取样和制样方法》,从我国主要的铁矿石进口口岸采集并制备来自澳大利亚、巴西、南非、哈萨克斯坦、加拿大、印度6个国家的进口铁矿石化学分析样品,包含22个品牌的共计434批次样品。所述22个品牌铁矿石包括津布巴混合粉铁矿、巴西铁矿石精粉、皮尔巴拉混合块、纽曼混合块铁矿、国王粉、皮尔巴拉混合粉、澳大利亚球团矿、杨迪粉铁矿、哈杨粉铁矿、纽曼混合粉铁矿、南非铁矿石精粉、澳大利亚铁矿石精粉、弗特斯克混合粉、卡拉加斯铁矿石、哈萨克斯坦球团矿、哈萨克斯坦粉铁矿、加拿大铁矿石精粉、昆巴标准粉、超特粉、麦克粉、昆巴标准块、印度球团矿。样品信息如表6所示。
表6铁矿石样品信息
2、样品检测
将样品分装到干燥瓶中于105℃下烘干4h。采用压片机对烘干样品进行压片,压片前用乙醇清洗模具,使用聚乙烯环使粉末样品聚拢,压制样品在30t压力下维持30s。检查压制样品表面均匀且无裂纹、脱落现象,测量前用洗耳球吹净样品表面。
使用德国布鲁克公司S4 Pioneer波长色散-X射线荧光光谱仪中的无标样分析方法检测铁矿石中元素的含量。检测中使用铑靶光管、四个分析仪晶体(LiF200、XS-55、PET和Ge)、流气计数器(FC)、闪烁计数器(SC)等元件。表7列出了仪器的部分测量条件。
表7仪器部分测量条件
针对采集的434个铁矿石样品,采用波长色散X射线荧光光谱无标样分析可以检测到Fe、O、Si、Ca、Al、Mn、Tb、Ti、Mg、P、Na、Cr、K、Sr、S、Zr、Zn、V、Cu、Gd、Ba、Cl、Ni、Co、Mo、Pb共26种元素的含量,其中Na、Cr、K、Sr、S、Zr、Zn、V、Cu、Gd、Ba、Cl、Ni、Co、Mo、Pb共16个元素含量存在未检出的情况,对于未检出的元素的含量,用检测限代替,元素含量测定结果如表8所示。
3、逐步判别分析
采用逐步判别分析对Fe、O、Si、Ca、Al、Mn、Tb、Ti、Mg、P、Na、Cr、K、Sr、S、Zr、Zn、V、Cu、Gd、Ba、Cl、Ni、Co、Mo、Pb共26个元素含量进行变量筛选,变量能否进入模型主要取决于协方差分析的F检验的显著性水平,当F值大于指定值时保留该变量,而F值小于指定值时,该变量从模型中剔除。选取合适的F值可以用最少的变量达到最佳的判别效果。其中,选取的F值为3.84。经过逐步判别分析Fe、O、Si、Ca、Al、Mn、Tb、Ti、Mg、P、Na、Cr、K、Sr、Zn、V、Cu、Ba、Ni、Mo、Pb共21个元素保留在了模型中,S、Zr、Gd、Cl、Co、因未通过F检验(F值<3.84)而从模型中剔除,最终21个元素用于建立识别模型。
4、建立五维Fisher判别模型
使用Fe、O、Si、Ca、Al、Mn、Tb、Ti、Mg、P、Na、Cr、K、Sr、Zn、V、Cu、Ba、Ni、Mo、Pb共21个元素含量,建立五维Fisher判别模型和相应的组质心处的坐标。
所述五维Fisher判别模型中的5组判别函数为:
F1=0.224X1-0.027X2+0.009X3+2.91X4+0.423X5-2.093X6-8.575X7+12.097X8-5.18X9-33.199X10-12.922X11-42.526X12+33.785X13+251.798X14-29.314X15-28.194X16-7.254X17+31.227X18+236.98X19-1591.987X20+518.554X21-15.262
F2=0.174X1-0.057X2+0.032X3+12.148X4+1.047X5-0.885X6+29.473X7-0.112X8-4.821X9-55.242X10+3.792X11-37.251X12-0.76X13+104.144X14+149.363X15+27.059X16+56.533X17-6.595X18-322.156X19-1386.114X20-705.416X21-4.88
F3=0.028X1+0.907X2-0.039X3-3.382X4-0.758X5-0.185X6-31.506X7-4.287X8+2.219X9+60.299X10+15.683X11-23.216X12+15.861X13+96.388X14-24.214X15+27.493X16+24.853X17+2.376X18-75.75X19+1480.018X20-210.037X21-31.304
F4=0.38X1+0.335X2-0.034X3+3.876X4+0.795X5-2.243X6+24.004X7+16.466X8-8.717X9-34.761X10+11.003X11+153.562X12+3.595X13-135.312X14+3.981X15-24.752X16-25.761X17-38.039X18+276.842X19-1076.789X20+257.631X21-36.297
F5=-0.079X1+0.566X2+0.741X3+5.228X4-0.342X5-1.334X6-23.369X7-7.439X8+2.123X9+5.964X10-2.621X11+74.847X12+3.689X13+25.371X14-57.22X15-6.903X16-138.511X17-22.433X18+463.216X19+682.723X20+885.888X21-21.689
其中,式中X1-X21分别代表Fe、O、Si、Ca、Al、Mn、Tb、Ti、Mg、P、Na、Cr、K、Sr、Zn、V、Cu、Ba、Ni、Mo、Pb的含量;所述含量为质量百分含量;
所述五维Fisher判别模型中的各产地组质心的坐标为澳大利亚(-1.608,-1.665,0.476,0.131,0.275)、巴西(-0.883,-0.878,-2.994,0.507,-3.479)、南非(10.092,0.914,0.943,-0.437,-0.07)、哈萨克斯坦(-4.617,14.213,3.55,-0.479,-0.507)、加拿大(-2.326,2.536,-7.937,-5.762,1.015)、印度(2.507,8.331,-8.608,4.423,1.947)。
将一待测样品铁矿石的元素含量代入所述五维Fisher判别模型中,根据判别函数和组质心处坐标函数,计算每个样品坐标与质心的距离,与哪个类别的质心最近,该样品就判定为哪个产地类别。
本次判别铁矿石产地国别的方法的实施例包括两类实施例:
第一类为构建模型所用的样品,即建模样品实施例。其中,建模样品实施例分别进行建模样品验证和交叉验证。建模样品验证为将构建模型所用的样品数据回代到模型,进行验证;交叉验证法为建模前每次留出一个作为验证的数据再次代入判别函数,进行验证。经过统计,所建立的五维判别模型对产地的识别正确率如下表9所示。
第二类为采用未知的待测样品验证的实施例。为了确定二维Fisher判别模型是否可以对未包含在模型中的样品进行识别,选择104个作为测试样品的铁矿石样品。经过统计,所建立的二维判别模型对产地的识别正确率如下表9所示,识别正确率达到98.10%,说明此模型可以对铁矿石的国别进行很好的识别。
表9判别模型识别国家的正确率
实施例5.1-5.434
本实施例中的样品收集、样品检测、逐步判别分析均同实施例4.1-4.434。
本系列实施例为判别铁矿石品牌的方法。
使用Fe、O、Si、Ca、Al、Mn、Ti、Mg、P、Na、Cr、K、Sr、S、Zr、Zn、V、Cu、Ba、Cl、Ni、Mo、Pb共23个元素,建立二十一维Fisher判别模型和相应的组质心处的坐标。
所述二十一维Fisher判别模型中的21组判别函数为:
F1=-0.053X1-0.099X2+0.048X3-0.253X4-0.402X5+1.104X6+14.021X7+1.777X8+0.356X9+1.422X10-87.361X11+1.833X12+82.245X13-61.82X14-68.402X15-5.214X16+323.547X17+157.553X18+17.288X19+34.918X20+1369.087X21+2182.814X22-748.481X23-13.645
F2=-0.055X1+0.108X2+0.267X3-1.323X4+1.175X5-0.492X6-62.999X7-1.451X8+14.334X9-2.202X10+80.315X11+11.578X12-102.336X13+33.267X14+336.236X15+2.148X16+287.835X17-169.74X18+1.896X19-16.006X20+478.101X21+134.578X22+1.034X23-2.208
F3=0.298X1-0.354X2+0.619X3-2.409X4-0.265X5+0.715X6-14.828X7+0.487X8+10.347X9+2.138X10+1.867X11+48.519X12+80.646X13-48.887X14+113.252X15+70.38X16-0.504X17+139.509X18+71.743X19-9.7932X20+33.683X21+7882.507X22-1789.426X23-17.431
F4=-0.532X1+0.764X2-0.282X3-1.331X4-0.945X5+6.025X6-1.482X7-5.478X8+36.982X9-0.776X10+14.694X11-28.874X12-96.675X13+10.312X14+47.168X15+73.511X16+24.272X17+133.474X18-16.237X19+59.559X20+190.543X21+2321.651X22-1278.666X23+10.576
F5=0.032X1+0.149X2-0.666X3-2.578X4+1.148X5-3.554X6+0.819X7-5.101X8+84.759X9-13.926X10-16.535X11+34.577X12+39.313X13+31.531X14-62.101X15-61.894X16+6.081X17+38.786X18+4.221X19-13.467X20+332.102X21-1020.083X22+428.29X23-11.97
F6=0.42X1-1.19X2+0.299X3+1.469X4+4.142X5+2.331X6-6.509X7-1.344X8+90.751X9-5.005X10+2.877X11-19.373X12-166.907X13+27.71X14+3.087X15+65.946X16+9.687X17+41.817X18-15.122X19-7.81X20-57.365X21-2017.834X22-224.898X23+9.633
F7=-0.065X1+0.66X2+0.071X3+1.662X4+1.355X5+2.717X6+17.37X7-6.992X8-90.46X9+5.554X10+52.493X11+15.791X12+50.956X13-6.753X14+24.939X15+45.909X16+6.518X17-91.513X18-4.024X19+27.505X20-180.137X21-1092.11X22+113.677X23-18.656
F8=0.083X1-0.583X2+0.975X3-3.832X4+0.037X5+7.319X6+6.86X7+0.685X8-35.784X9-10.062X10+49.473X11+2.357X12-38.99X13+60.527X14+86.109X15-105.696X16-35.757X17-134.998X18-2.241X19+100.841X20+321.311X21+128.302X22+549.514X23+5.451
F9=0.021X1+0.268X2+0.401X3-7.766X4+0.964X5+0.896X6-3.487X7+8.214X8+48.753X9+17.048X10+28.419X11+11.511X12-105.516X13+29.067X14+3.746X15-101.435X16+2.703X17-61.216X18-19.516X19+9.189X20-58.156X21+1142.793X22+487.16X23-18.715
F10=0.286X1-0.152X2-0.837X3-1.096X4-1.413X5+10.093X6+0.52X7-5.173X8+62.646X9+8.97X10-129.582X11+3.098X12+16.964X13+41.966X14-73.943X15+34.131X16+13.738X17+117.362X18+3.936X19-10.978X20-226.571X21-1977.339X22-846.084X23-0.723
F11=0.353X1+0.221X2-0.105X3+1.002X4+0.418X5-2.509X6+14.736X7-7.321X8-49.629X9+12.515X10+166.053X11+2.89X12-154.775X13-39.078X14+209.141X15+5.204X16-31.122X17-10.934X18-38.433X19-7.976X20+316.914X21+48.649X22-55.111X23-29.398
F12=0.29X1+0.301X2+0.69X3+4.614X4-1.695X5+1.509X6-4.466X7-3.045X8+75.408X9+2.05X10+59.282X11-2.687X12+94.997X13+187.997X14-128.428X15-41.09X16-2.475X17-147.37X18-36.871X19-48.664X20+371.102X21-4687.471X22+577.282X23-25.107
F13=0.257X1-0.112X2+0.88X3+0.017X4+0.409X5-3.497X6-1.65X7-3.896X8-32.246X9+0.391X10-138.738X11-11.124X12+391.07X13+43.989X14+118.64X15+125.925X16+7.983X17+67.953X18-50.208X19+21.613X20-20.278X21-3933.864X22-540.687X23-2.469
F14=-0.093X1-0.243X2-0.892X3-0.152X4-0.904X5-6.242X6+2.651X7-0.579X8+53.174X9+1.514X10+38.229X11+0.292X12-36.222X13+34.625X14-60.743X15+12.447X16+2.586X17-79.918X18+6.034X19+177.337X20-153.521X21-1149.34X22+215.381X23+15.881
F15=0.248X1+0.377X2-0.251X3+2.776X4+0.659X5+1.686X6-12.487X7+2.43X8-28.118X9+1.007X10-29.367X11-0.693X12+201.635X13-103.05X14-108.573X15-174.727X16+3.728X17+120.573X18-34.519X19+100.36X20+224.851X21+2768.54X22+1534.895X23-40.454
F16=0.026X1-0.052X2+0.516X3-0.632X4-0.176X5-0.523X6+14.344X7-2.21X8-9.7X9+4.072X10+60.055X11-13.256X12+109.5X13+28.814X14-311.137X15-207.363X16+19.933X17+122.527X18+54.953X19-4.244X20-641.366X21+407.307X22+1149.219X23-6.933
F17=0.18X1+0.108X2+0.648X3+2.113X4-0.331X5-0.494X6-14.564X7-3.738X8+21.527X9+4.983X10-159.078X11-8.599X12-93.444X13-42.289X14+557.903X15-68.149X16+15.476X17+193.542X18+79.988X19+41.394X20+263.677X21-454.154X22+398.07X23-19.574
F18=0.48X1+0.324X2+1.307X3+0.854X4-0.298X5+1.029X6+8.675X7-0.564X8+31.132X9-5.279X10-55.278X11-4.933X12-108.48X13-87.987X14-305.093X15+107.737X16+2.717X17+4.216X18+37.751X19+51.161X20-229.377X21+2021.426X22-819.479X23-39.765
F19=0.444X1+0.045X2-0.338X3+0.547X4+1.112X5-1.242X6+10.795X7+2.901X8-91.493X9+0.20X10+621.922X11-11.035X12+57.024X13+75.969X14+153.722X15-1.446X16-24.934X17-221.679X18+37.566X19-21.537X20-90.49X21+1032.202X22-112.522X23-28.029
F20=0.124X1-0.035X2-0.11X3-1.098X4+0.781X5-0.327X6+1.073X7+2.123X8-49.061X9+5.011X10-35.986X11-13.867X12+150.829X13+40.23X14-244.636X15+56.799X16-21.129X17-61.113X18+63.27X19-4.425X20+176.578X21-750.763X22-272.693X23-4.584
F21=0.104X1+0.143X2-0.047X3+1.361X4+0.446X5+0.109X6-17.305X7+1.318X8-29.216X9+2.598X10+81.184X11-10.867X12+146.848X13+10.103X14+21.72X15+81.689X16-1.978X17+66.532X18+47.961X19+19.317X20+486.92X21-1961.691X22-1496.159X23-2.188
其中,式中X1-X23分别代表Fe、O、Si、Ca、Al、Mn、Ti、Mg、P、Na、Cr、K、Sr、S、Zr、Zn、V、Cu、Ba、Cl、Ni、Mo、Pb的含量;所述含量为质量百分含量;
所述十维Fisher判别模型中的各品牌组质心的坐标为津布巴粉(-13.457,4.113,-4.925,2.63,4.331,7.5,0.569,-0.793,3.403,-2.302,-0.164,0.017,1.076,-0.79,-0.571,-0.033,0.899,-0.521,0.311,-0.014,0.004)、巴西精粉(-11.852,2.114,-0.785,-2.759,-1.517,3.797,-3.573,2.455,-1.977,0.07,1.254,-0.808,0.328,0.077,-1.034,1.092,-0.607,0.295,0.343,-0.025,-0.007)、PB块(-12.721,1.928,-3.75,1.464,0.556,-0.327,-1.725,-1.283,-0.681,1.615,-0.059,1.013,-0.477,0.248,0.342,-0.246,0.005,-0.000092,0.243,-0.005,-0.03)、纽块(-12.675,2.457,-2.951,0.622,0.462,0.865,-2.689,-0.652,-0.267,0.63,-0.311,0.866,0.398,0.191,0.017,0.15,0.473,0.454,-0.131,0.009,0.017)、国王粉(-12.3,0.362,-4.713,3.919,-1.242,-5.411,0.078,-2.341,-1.758,-0.926,0.732,-0.359,0.721,-0.028,-1.83,-1.075,-0.292,0.223,0.01,0.023,0.002)、PB粉(-12.742,2.104,-4.491,2.504,2.247,2.415,-0.472,-1.311,1.185,-0.287,-0.053,0.392,-0.427,0.269,-0.035,-0.03,-0.379,-0.157,-0.102,-0.058,0.021)、澳大利亚球团矿(167.931,80.586,-4.04,2.091,2.198,-0.503,-2.073,1.832,-0.795,-0.465,0.34,0.017,-0.104,-0.021,0.021,-0.07,0.023,-0.033,-0.018,0,-0.0000016)、杨迪粉(-11.566,1.426,-5.122,3.525,-1.56,-7.06,0.194,-2.217,-1.297,-1.703,0.289,-0.241,0.066,-0.458,0.1,0.409,0.181,-0.128,-0.139,-0.05,-0.022)、哈杨粉(-11.675,1.307,-4.569,2.237,-1.094,-5.382,-0.672,-2.406,-1.701,-0.722,0.684,-0.935,-0.311,0.013,0.762,0.087,-0.018,-0.125,0.292,0.089,0.056)、纽粉(-12.541,2.958,-3.946,0.875,1.778,4.241,-0.24,-0.595,2.646,-3.224,-0.531,-1.718,-0.291,-0.365,0.659,-0.245,-0.228,0.452,-0.029,0.027,-0.022)、南非精粉(87.288,-77.489,-2.119,-1.063,0.445,-0.007,0.054,-0.2,0.317,-0.087,-0.024,0.072,0.015,-0.004,-0.003,0.014,-0.001,0.011,0.008,0,-0.0000344)、澳精粉(-10.919,2.666,8.823,-12.193,-6.046,-2.691,-4.029,2.182,4.786,-0.312,1.478,0.118,-0.236,0.018,-0.013,-0.152,0.084,-0.123,-0.044,0.001,0)、混合粉(-11.55,0.445,-4.604,6.56,-1.641,-1.132,3.715,4.582,1.276,1.858,-0.25,0.421,-0.267,-0.956,-0.027,0.005,-0.104,0.074,-0.005,0.012,0.006)、卡拉粉(-10.649,-2.193,-1.721,0.961,0.94,4.042,-2.802,-0.876,-2.102,4.902,0.919,-2.245,0.054,-0.134,0.008,-0.096,0.185,-0.277,-0.221,0.021,-0.009)、哈萨克斯坦球团矿(29.37,22.461,3.419,-9.19,-9.964,2.379,8.943,-7.398,2.69,2.48,-1.248,-0.192,0.375,0.129,-0.127,0.236,-0.091,0.12,0.054,-0.001,0)、哈萨克斯坦粉(5.324,-1.676,53.549,16.806,-2.595,1.847,-0.666,-0.926,-0.774,-0.701,0.025,0.13,-0.03,0.013,0.054,0.0000177,-0.004,-0.006,0.006,0.001,0)、加拿大精粉(-9.729,-0.837,-1.004,-6.883,-9.177,2.125,-3.365,3.298,-5.229,-1.649,-4.658,-0.355,-0.276,-0.019,-0.122,-0.201,0.052,-0.244,-0.004,-0.014,0.007)、南非粉(-10.942,3.67,13.141,-11.086,10.607,-3.335,3.506,0.648,-0.612,0.202,-0.887,-0.46,-1.124,0.134,-0.527,0.112,0.26,0.06,0.015,0.006,0.001)、超特粉(-10.997,-0.634,-5.74,7.898,-2.006,-2.31,5.426,5.539,1.253,-0.334,0.217,-1.305,0.673,2.162,0.238,-0.072,0.337,-0.012,0.06,-0.043,-0.005)、麦克粉(-13.041,2.766,-4.999,3.451,1.794,0.67,0.024,-0.052,1.676,-1.23,-0.856,1.174,0.267,0.72,-0.411,0.46,-0.35,-0.397,-0.23,0.183,-0.026)、南非块(-10.75,3.292,8.8,-9.099,5.925,-2.103,0.617,1.238,-1.378,0.607,-0.356,0.366,2.099,-0.265,0.789,-0.146,-0.487,-0.065,-0.002,-0.016,0)、印度球团矿(-12.147,-0.212,-0.848,-7.93,-3.383,7.471,6.083,1.264,-6.712,-2.692,3.374,1.086,-0.421,-0.049,0.275,-0.183,0.146,-0.031,-0.116,0.018,-0.002)。
将一待测样品铁矿石的元素含量代入所述二十一维Fisher判别模型中,根据判别函数和组质心处坐标函数,计算每个样品坐标与组质心坐标的距离,与哪个类别的质心最近,该样品就判定为哪个品牌类别。
所建立的二十一维判别模型对品牌的识别正确率如下表10所示,所建立识别模型具有很好识别效果。
表10二十一维判别模型识别品牌的正确率
Claims (6)
1.一种判别铁矿石产地的方法,其包括如下步骤:
S1.取至少3个国别,每个国别至少16个批次的铁矿石中元素含量的数据,建立至少二维以上的Fisher判别模型;S2.将待测样品铁矿石的元素含量代到步骤S1的至少二维以上的Fisher判别模型中,确定待测样品铁矿石的产地;
所述元素含量的检测方法为波长色散X射线荧光光谱无标样分析方法或者能量色散X射线荧光光谱无标样分析方法;
在所述至少二维以上的Fisher判别模型的建立前,进行逐步判别分析,所述逐步判别分析为识别进入模型的所述元素的含量的协方差分析的F值,当所述F值大于3.84时,保留所述元素的含量,当所述F值小于3.84时,剔除所述元素的含量;
所述至少二维以上的Fisher判别模型为五维Fisher判别模型;所述元素为Fe、O、Si、Ca、Al、Mn、Tb、Ti、Mg、P、Na、Cr、K、Sr、Zn、V、Cu、Ba、Ni、Mo、Pb;
所述五维Fisher判别模型中的5组判别函数为:
F1=0.224X1-0.027X2+0.009X3+2.91X4+0.423X5-2.093X6-8.575X7+12.097X8-5.18X9-33.199X10-12.922X11-42.526X12+33.785X13+251.798X14-29.314X15-28.194X16-7.254X17+31.227X18+236.98X19-1591.987X20+518.554X21-15.262;
F2=0.174X1-0.057X2+0.032X3+12.148X4+1.047X5-0.885X6+29.473X7-0.112X8-4.821X9-55.242X10+3.792X11-37.251X12-0.76X13+104.144X14+149.363
X15+27.059X16+56.533X17-6.595X18-322.156X19-1386.114X20-705.416X21-4.88;
F3=0.028X1+0.907X2-0.039X3-3.382X4-0.758X5-0.185X6-31.506X7-4.287X8+2.219X9+60.299X10+15.683X11-23.216X12+15.861X13+96.388X14-24.214X15+27.493X16+24.853X17+2.376X18-75.75X19+1480.018X20-210.037X21-31.304;
F4=0.38X1+0.335X2-0.034X3+3.876X4+0.795X5-2.243X6+24.004X7+16.466X8-8.717X9-34.761X10+11.003X11+153.562X12+3.595X13-135.312X14+3.981X15-24.752X16-25.761X17-38.039X18+276.842X19-1076.789X20+257.631X21-36.297;
F5=-0.079X1+0.566X2+0.741X3+5.228X4-0.342X5-1.334X6-23.369X7-7.439X8+2.123X9+5.964X10-2.621X11+74.847X12+3.689X13+25.371X14-57.22X15-6.903X16-138.511X17-22.433X18+463.216X19+682.723X20+885.888X21-21.689;
其中,式中X1-X21分别代表Fe、O、Si、Ca、Al、Mn、Tb、Ti、Mg、P、Na、Cr、K、Sr、Zn、V、Cu、Ba、Ni、Mo、Pb的含量;
所述五维Fisher判别模型中的各产地组质心的坐标为澳大利亚(-1.608,-1.665,0.476,0.131,0.275)、巴西(-0.883,-0.878,-2.994,0.507,-3.479)、南非(10.092,0.914,0.943,-0.437,-0.07)、哈萨克斯坦(-4.617,14.213,3.55,-0.479,-0.507)、加拿大(-2.326,2.536,-7.937,-5.762,1.015)、印度(2.507,8.331,-8.608,4.423,1.947);
其中,所述步骤S1和所述步骤S2的元素含量单位均为质量百分含量;
其中,当元素含量能被测试仪器检出,所述元素含量为测试仪器检出元素的含量;当元素含量不能被测试仪器检出,所述元素含量为测试仪器的检出限,所述检出限为0.0015-0.02。
2.如权利要求1所述的判别铁矿石产地的方法,其特征在于,所述元素含量的检测方法为波长色散X射线荧光光谱中的无标样分析方法。
3.如权利要求2所述的判别铁矿石产地的方法,其特征在于,所述步骤S1中的元素含量的测定,与所述步骤S2中的元素含量的测定为同一台检测机器。
4.一种判别铁矿石品牌的方法,其包括如下步骤:
S1.取至少14个品牌,每个品牌至少6个批次的铁矿石中元素含量的数据,建立至少十维以上的Fisher判别模型;S2.将待测样品铁矿石的元素含量代到步骤S1的至少十维以上的Fisher判别模型中,确定待测样品铁矿石的品牌;
所述元素含量的检测方法为波长色散X射线荧光光谱无标样分析方法或者能量色散X射线荧光光谱无标样分析方法;
在所述至少十维以上的Fisher判别模型的建立前,进行逐步判别分析,所述逐步判别分析为识别进入模型的所述元素的含量的协方差分析的F值,当所述F值大于3.84时,保留所述元素的含量,当所述F值小于3.84时,剔除所述元素的含量;
所述至少十维以上的Fisher判别模型为十维Fisher判别模型或二十一维Fisher判别模型;所述元素为Fe、O、Si、Ca、Al、Mn、Ti、Mg、P、S或Fe、O、Si、Ca、Al、Mn、Ti、Mg、P、Na、Cr、K、Sr、S、Zr、Zn、V、Cu、Ba、Cl、Ni、Mo、Pb;
所述十维Fisher判别模型中的10组判别函数为:
F1=0.286X1-0.372X2+0.404X3-9.8X4-0.575X5-2.551X6+54.081X7+18.203X8-95.378X9-17.295X10-12.937;
F2=0.726X1-1.244X2+2.718X3-2.715X4-0.808X5-5.169X6-23.321X7+15.246X8-160.116X9-18.187X10-2.608;
F3=0.587X1-1.372X2+1.45X3+8.922X4+3.122X5+3.139X6+9.551X7-
15.179X8+180.272X9+21.586X10-4.251;
F4=0.257X1-0.263X2+0.896X3+33.291X4-1.573X5+1.075X6+37.848X7-34.022X8-62.815X9+25.649X10-6.865;
F5=-0.4X1-0.453X2+1.121X3-13.49X4-
1.417X5+11.016X6+9.278X7+0.466X8+19.539X9+87.158X10+35.004;
F6=-0.012X1+0.213X2+1.171X3+6.748X4+2.843X5-6.586X6+12.703X7-2.634X8-138.53X9-59.512X10-9.303;
F7=-0.189X1+0.221X2+2.048X3-4.649X4-3.679X5-0.974X6+10.841X7-7.985X8+221.911X9-69.058X10-0.463;
F8=0.337X1+0.494X2+0.457X3+7.023X4-0.625X5-1.537X6-
17.619X7+2.087X8+6.88X9+229.064X10-38.458;
F9=0.518X1-0.035X2+0.885X3-14.995X4-0.257X5-5.636X6+26.099X7-4.9X8-30.377X9+159.188X10-31.76;
F10=1.526X1+1.558X2+1.681X3-1.042X4-
0.188X5+4.851X6+6.292X7+0.682X8+7.995X9-75.431X10-145.479;
其中,式中X1-X10分别代表Fe、O、Si、Ca、Al、Mn、Ti、Mg、P、S的含量;
所述十维Fisher判别模型中的各品牌组质心的坐标为PB粉(-9.129,-5.309,2.278,-2.045,-0.884,0.199,-0.111,0.337,-0.084,-0.183)、PB块(-8.865,-1.735,-0.529,0.196,-0.914,-1.943,-0.232,0.23,-0.533,0.188)、杨迪粉(-6.897,-0.581,-7.854,0.633,-1.068,0.877,-0.103,-0.226,0.014,-0.119)、纽块(-8.441,0.822,0.838,-0.645,-0.322,-1.095,0.902,0.533,0.658,0.14)、纽粉(-7.648,-2.205,4.196,-3.965,-1.744,3.552,-0.685,-1.19,-0.52,0.917)、津布巴粉(-10.504,-7.276,8.091,-4.139,-0.909,3.428,-0.268,-0.029,0.747,-0.249)、南非粉(-7.329,5.853,6.054,7.442,-2.288,1.773,-0.352,0.422,-0.481,-0.287)、南非块(-6.676,8.807,5.167,6.15,-1.532,0.518,0.25,-0.301,0.646,0.334)、卡拉粉(-6.188,-2.231,4.843,0.443,0.096,-3.793,-1.303,-2.079,0.65,-0.329)、澳精粉(3.146,26.733,-0.494,-4.019,1.429,0.067,-0.61,0.172,-0.033,-0.09)、巴西粉(-6.304,2.881,4.157,-0.639,3.26,0.204,2.768,-1.133,-1.082,-0.266)、国王粉(-6.642,-1.598,-6.442,0.817,-0.226,0.812,1.245,-0.737,0.657,0.168)、混合粉(-7.292,-4.834,-0.377,1.758,6.498,0.57,-0.581,0.249,0.076,0.085)、南非精粉(109.502,-2.359,0.254,0.132,-0.191,-0.026,0.037,0.035,-0.001,0.008);
所述二十一维Fisher判别模型中的21组判别函数为:
F1=-0.053X1-0.099X2+0.048X3-0.253X4-0.402X5+1.104X6+14.021X7+1.777X8+0.356X9+1.422X10-87.361X11+1.833X12+82.245X13-61.82X14-68.402X15-5.214X16+323.547X17+157.553X18+17.288X19+34.918X20+1369.087
X21+2182.814X22-748.481X23-13.645;
F2=-0.055X1+0.108X2+0.267X3-1.323X4+1.175X5-0.492X6-62.999X7-1.451X8+14.334X9-2.202X10+80.315X11+11.578X12-102.336X13+33.267X14+
336.236X15+2.148X16+287.835X17-169.74X18+1.896X19-16.006X20+478.101X21+134.578X22+1.034X23-2.208;
F3=0.298X1-0.354X2+0.619X3-2.409X4-0.265X5+0.715X6-14.828X7+0.487X8+10.347X9+2.138X10+1.867X11+48.519X12+80.646X13-48.887X14+113.252X15+70.38X16-0.504X17+139.509X18+71.743X19-9.7932X20+33.683X21+
7882.507X22-1789.426X23-17.431;
F4=-0.532X1+0.764X2-0.282X3-1.331X4-0.945X5+6.025X6-1.482X7-5.478X8+36.982X9-0.776X10+14.694X11-28.874X12-96.675X13+10.312X14+47.168X15+73.511X16+24.272X17+133.474X18-16.237X19+59.559X20+190.543
X21+2321.651X22-1278.666X23+10.576;
F5=0.032X1+0.149X2-0.666X3-2.578X4+1.148X5-3.554X6+0.819X7-5.101X8+84.759X9-13.926X10-16.535X11+34.577X12+39.313X13+31.531X14-62.101X15-61.894X16+6.081X17+38.786X18+4.221X19-13.467X20+332.102X21-1020.083X22+428.29X23-11.97;
F6=0.42X1-1.19X2+0.299X3+1.469X4+4.142X5+2.331X6-6.509X7-1.344X8+
90.751X9-5.005X10+2.877X11-19.373X12-166.907X13+27.71X14+3.087X15+
65.946X16+9.687X17+41.817X18-15.122X19-7.81X20-57.365X21-2017.834X22-224.898X23+9.633;
F7=-0.065X1+0.66X2+0.071X3+1.662X4+1.355X5+2.717X6+17.37X7-6.992X8-90.46X9+5.554X10+52.493X11+15.791X12+50.956X13-6.753X14+24.939
X15+45.909X16+6.518X17-91.513X18-4.024X19+27.505X20-180.137X21-1092.11X22+113.677X23-18.656;
F8=0.083X1-0.583X2+0.975X3-3.832X4+0.037X5+7.319X6+6.86X7+0.685X8-35.784X9-10.062X10+49.473X11+2.357X12-38.99X13+60.527X14+86.109X15-105.696X16-35.757X17-134.998X18-2.241X19+100.841X20+321.311X21+128.302X22+549.514X23+5.451;
F9=0.021X1+0.268X2+0.401X3-7.766X4+0.964X5+0.896X6-3.487X7+8.214X8+48.753X9+17.048X10+28.419X11+11.511X12-105.516X13+29.067X14+3.746X15-101.435X16+2.703X17-61.216X18-19.516X19+9.189X20-58.156X21+1142.793X22+487.16X23-18.715;
F10=0.286X1-0.152X2-0.837X3-1.096X4-1.413X5+10.093X6+0.52X7-5.173X8+62.646X9+8.97X10-129.582X11+3.098X12+16.964X13+41.966X14-73.943X15+
34.131X16+13.738X17+117.362X18+3.936X19-10.978X20-226.571X21-1977.339X22-846.084X23-0.723;
F11=0.353X1+0.221X2-0.105X3+1.002X4+0.418X5-2.509X6+14.736X7-7.321X8-49.629X9+12.515X10+166.053X11+2.89X12-154.775X13-39.078X14+209.14
1X15+5.204X16-31.122X17-10.934X18-38.433X19-7.976X20+316.91
4X21+48.649X22-55.111X23-29.398;
F12=0.29X1+0.301X2+0.69X3+4.614X4-1.695X5+1.509X6-4.466X7-3.045X8+75.408X9+2.05X10+59.282X11-2.687X12+94.997X13+187.997X14-128.428X15-41.09X16-2.475X17-147.37X18-36.871X19-48.664X20+371.102X21-4687.471X22+577.282X23-25.107;
F13=0.257X1-0.112X2+0.88X3+0.017X4+0.409X5-3.497X6-1.65X7-3.896X8-32.246X9+0.391X10-138.738X11-11.124X12+391.07X13+43.989X14+118.64X15+125.925X16+7.983X17+67.953X18-50.208X19+21.613X20-20.278X21-3933.864X22-540.687X23-2.469;
F14=-0.093X1-0.243X2-0.892X3-0.152X4-0.904X5-6.242X6+2.651X7-0.579X8+53.174X9+1.514X10+38.229X11+0.292X12-36.222X13+34.625X14-60.743X15+12.447X16+2.586X17-79.918X18+6.034X19+177.337X20-153.521X21-1149.34X22+215.381X23+15.881;
F15=0.248X1+0.377X2-0.251X3+2.776X4+0.659X5+1.686X6-12.487X7+2.43X8-28.118X9+1.007X10-29.367X11-0.693X12+201.635X13-103.05X14-108.573X15-174.727X16+3.728X17+120.573X18-34.519X19+100.36X20+224.851X21+2768.54X22+1534.895X23-40.454;
F16=0.026X1-0.052X2+0.516X3-0.632X4-0.176X5-0.523X6+14.344X7-2.21X8-9.7X9+4.072X10+60.055X11-13.256X12+109.5X13+28.814X14-311.137X15-207.363X16+19.933X17+122.527X18+54.953X19-4.244X20-641.366X21+407.307X22+1149.219X23-6.933;
F17=0.18X1+0.108X2+0.648X3+2.113X4-0.331X5-0.494X6-14.564X7-3.738X8+21.527X9+4.983X10-159.078X11-8.599X12-93.444X13-42.289X14+557.903X15-68.149X16+15.476X17+193.542X18+79.988X19+41.394X20+263.677X21-454.154X22+398.07X23-19.574;
F18=0.48X1+0.324X2+1.307X3+0.854X4-0.298X5+1.029X6+8.675X7-0.564X8+31.132X9-5.279X10-55.278X11-4.933X12-108.48X13-87.987X14-305.093
X15+107.737X16+2.717X17+4.216X18+37.751X19+51.161X20-229.377
X21+2021.426X22-819.479X23-39.765;
F19=0.444X1+0.045X2-0.338X3+0.547X4+1.112X5-1.242X6+10.795X7+2.901X8-91.493X9+0.20X10+621.922X11-11.035X12+57.024X13+75.969X14+153.722X15-1.446X16-24.934X17-221.679X18+37.566X19-21.537X20-90.49X21+1032.202X22-112.522X23-28.029;
F20=0.124X1-0.035X2-0.11X3-1.098X4+0.781X5-0.327X6+1.073X7+2.123X8-49.061X9+5.011X10-35.986X11-13.867X12+150.829X13+40.23X14-244.636X15+56.799X16-21.129X17-61.113X18+63.27X19-4.425X20+176.578X21-750.763X22-272.693X23-4.584;
F21=0.104X1+0.143X2-0.047X3+1.361X4+0.446X5+0.109X6-17.305X7+1.318X8-29.216X9+2.598X10+81.184X11-10.867X12+146.848X13+10.103X14+21.72X15+81.689X16-1.978X17+66.532X18+47.961X19+19.317X20+486.92X21-1961.691X22-1496.159X23-2.188;
其中,式中X1-X23分别代表Fe、O、Si、Ca、Al、Mn、Ti、Mg、P、Na、Cr、K、Sr、S、Zr、Zn、V、Cu、Ba、Cl、Ni、Mo、Pb的含量;
所述二十一维Fisher判别模型中的各品牌组质心的坐标为津布巴粉(-13.457,4.113,-4.925,2.63,4.331,7.5,0.569,-0.793,3.403,-2.302,-0.164,0.017,1.076,-0.79,-0.571,-0.033,0.899,-0.521,0.311,-0.014,0.004)、巴西精粉(-11.852,2.114,-0.785,-2.759,-1.517,3.797,-3.573,2.455,-1.977,0.07,1.254,-0.808,0.328,0.077,-1.034,1.092,-0.607,0.295,0.343,-0.025,-0.007)、PB块(-12.721,1.928,-3.75,1.464,0.556,-0.327,-1.725,-1.283,-0.681,1.615,-0.059,1.013,-0.477,0.248,0.342,-0.246,0.005,-0.000092,0.243,-0.005,-0.03)、纽块(-12.675,2.457,-2.951,0.622,0.462,0.865,-2.689,-0.652,-0.267,0.63,-0.311,0.866,0.398,0.191,0.017,0.15,0.473,0.454,-0.131,0.009,0.017)、国王粉(-12.3,0.362,-4.713,3.919,-1.242,-5.411,0.078,-2.341,-1.758,-0.926,0.732,-0.359,0.721,-0.028,-1.83,-1.075,-0.292,0.223,0.01,0.023,0.002)、PB粉(-12.742,2.104,-4.491,2.504,2.247,2.415,-0.472,-1.311,1.185,-0.287,-0.053,0.392,-0.427,0.269,-0.035,-0.03,-0.379,-0.157,-0.102,-0.058,0.021)、澳大利亚球团矿(167.931,80.586,-4.04,2.091,2.198,-0.503,-2.073,1.832,-0.795,-0.465,0.34,0.017,-0.104,-0.021,0.021,-0.07,0.023,-0.033,-0.018,0,-0.0000016)、杨迪粉(-11.566,1.426,-5.122,3.525,-1.56,-7.06,0.194,-2.217,-1.297,-1.703,0.289,-0.241,0.066,-0.458,0.1,0.409,0.181,-0.128,-0.139,-0.05,-0.022)、哈杨粉(-11.675,1.307,-4.569,2.237,-1.094,-5.382,-0.672,-2.406,-1.701,-0.722,0.684,-0.935,-0.311,0.013,0.762,0.087,-0.018,-0.125,0.292,0.089,0.056)、纽粉(-12.541,2.958,-3.946,0.875,1.778,4.241,-0.24,-0.595,2.646,-3.224,-0.531,-1.718,-0.291,-0.365,0.659,-0.245,-0.228,0.452,-0.029,0.027,-0.022)、南非精粉(87.288,-77.489,-2.119,-1.063,0.445,-0.007,0.054,-0.2,0.317,-0.087,-0.024,0.072,0.015,-0.004,-0.003,0.014,-0.001,0.011,0.008,0,-0.0000344)、澳精粉(-10.919,2.666,8.823,-12.193,-6.046,-2.691,-4.029,2.182,4.786,-0.312,1.478,0.118,-0.236,0.018,-0.013,-0.152,0.084,-0.123,-0.044,0.001,0)、混合粉(-11.55,0.445,-4.604,6.56,-1.641,-1.132,3.715,4.582,1.276,1.858,-0.25,0.421,-0.267,-0.956,-0.027,0.005,-0.104,0.074,-0.005,0.012,0.006)、卡拉粉(-10.649,-2.193,-1.721,0.961,0.94,4.042,-2.802,-0.876,-2.102,4.902,0.919,-2.245,0.054,-0.134,0.008,-0.096,0.185,-0.277,-0.221,0.021,-0.009)、哈萨克斯坦球团矿(29.37,22.461,3.419,-9.19,-9.964,2.379,8.943,-7.398,2.69,2.48,-1.248,-0.192,0.375,0.129,-0.127,0.236,-0.091,0.12,0.054,-0.001,0)、哈萨克斯坦粉(5.324,-1.676,53.549,16.806,-2.595,1.847,-0.666,-0.926,-0.774,-0.701,0.025,0.13,-0.03,0.013,0.054,0.0000177,-0.004,-0.006,0.006,0.001,0)、加拿大精粉(-9.729,-0.837,-1.004,-6.883,-9.177,2.125,-3.365,3.298,-5.229,-1.649,-4.658,-0.355,-0.276,-0.019,-0.122,-0.201,0.052,-0.244,-0.004,-0.014,0.007)、南非粉(-10.942,3.67,13.141,-11.086,10.607,-3.335,3.506,0.648,-0.612,0.202,-0.887,-0.46,-1.124,0.134,-0.527,0.112,0.26,0.06,0.015,0.006,0.001)、超特粉(-10.997,-0.634,-5.74,7.898,-2.006,-2.31,5.426,5.539,1.253,-0.334,0.217,-1.305,0.673,2.162,0.238,-0.072,0.337,-0.012,0.06,-0.043,-0.005)、麦克粉(-13.041,2.766,-4.999,3.451,1.794,0.67,0.024,-0.052,1.676,-1.23,-0.856,1.174,0.267,0.72,-0.411,0.46,-0.35,-0.397,-0.23,0.183,-0.026)、南非块(-10.75,3.292,8.8,-9.099,5.925,-2.103,0.617,1.238,-1.378,0.607,-0.356,0.366,2.099,-0.265,0.789,-0.146,-0.487,-0.065,-0.002,-0.016,0)、印度球团矿(-12.147,-0.212,-0.848,-7.93,-3.383,7.471,6.083,1.264,-6.712,-2.692,3.374,1.086,-0.421,-0.049,0.275,-0.183,0.146,-0.031,-0.116,0.018,-0.002);
其中,所述步骤S1和所述步骤S2的元素含量单位均为质量百分含量;
其中,当元素含量能被测试仪器检出,所述元素含量为测试仪器检出元素的含量;当元素含量不能被测试仪器检出,所述元素含量为测试仪器的检出限,所述检出限为0.0015-0.02。
5.如权利要求4所述的判别铁矿石品牌的方法,其特征在于,所述元素含量的检测方法为波长色散X射线荧光光谱中的无标样分析方法。
6.如权利要求5所述的判别铁矿石品牌的方法,其特征在于,所述元素含量的检测方法为波长色散X射线荧光光谱中的无标样分析方法时,所述步骤S1中的元素含量的测定,与所述步骤S2中的元素含量的测定为同一台检测机器。
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201911257106.9A CN110987996B (zh) | 2019-12-03 | 2019-12-03 | 一种判别进口铁矿石产地及品牌的方法 |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201911257106.9A CN110987996B (zh) | 2019-12-03 | 2019-12-03 | 一种判别进口铁矿石产地及品牌的方法 |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110987996A CN110987996A (zh) | 2020-04-10 |
CN110987996B true CN110987996B (zh) | 2023-04-07 |
Family
ID=70091795
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201911257106.9A Active CN110987996B (zh) | 2019-12-03 | 2019-12-03 | 一种判别进口铁矿石产地及品牌的方法 |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110987996B (zh) |
Families Citing this family (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112229863A (zh) * | 2020-09-30 | 2021-01-15 | 上海海关工业品与原材料检测技术中心 | 一种铁矿石的原产国或品牌的鉴别方法 |
CN113138178B (zh) * | 2021-04-15 | 2023-07-07 | 上海海关工业品与原材料检测技术中心 | 一种识别进口铁矿石品牌的方法 |
CN115524436B (zh) * | 2022-10-11 | 2024-06-14 | 云南中烟工业有限责任公司 | 一种判别国内外不同产地的丁香的方法 |
Family Cites Families (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
BRPI0500177A (pt) * | 2005-01-21 | 2006-09-05 | Unicamp | espalhamento de raios x e quimiometria para classificação de óleos vegetais, animais, minerais e/ou sintéticos |
AU2013204502B2 (en) * | 2012-09-03 | 2016-01-07 | Source Science And Technology Pty Ltd | Method and system for classifying a foodstuff |
AU2015201156A1 (en) * | 2014-03-05 | 2015-09-24 | Source Science And Technology Pty Ltd | Method and system for determining provenance of a product |
JP6455834B2 (ja) * | 2014-12-24 | 2019-01-23 | 理研ビタミン株式会社 | ワカメの生育海域を判別する方法 |
CN105181907B (zh) * | 2015-10-13 | 2017-01-04 | 中国地质大学(武汉) | 一种定量判别软玉产地的方法 |
-
2019
- 2019-12-03 CN CN201911257106.9A patent/CN110987996B/zh active Active
Also Published As
Publication number | Publication date |
---|---|
CN110987996A (zh) | 2020-04-10 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110987996B (zh) | 一种判别进口铁矿石产地及品牌的方法 | |
TWI573634B (zh) | 光學式粒狀物選別機 | |
CN111007094B (zh) | 判别进口铜精矿产地的方法 | |
CN105181907B (zh) | 一种定量判别软玉产地的方法 | |
CN101701916B (zh) | 一种玉米品种快速鉴定、鉴别方法 | |
TW201442790A (zh) | 光學式粒狀物選別機 | |
CN107335626A (zh) | 自动分拣机及其分拣管理系统和分拣方法 | |
CN111575376A (zh) | 一组用于评估肾透明细胞癌预后的联合基因组及其应用 | |
CN112474387A (zh) | 一种原粮不完善粒的图像识别和自动分拣装置 | |
CN111833296A (zh) | 一种骨髓细胞形态学自动检测审核系统及审核方法 | |
CN105478364B (zh) | 一种不良品检测分类方法及系统 | |
CN114332856A (zh) | 一种血型卡图像的特征提取方法、装置及血型判读系统 | |
CN110569581A (zh) | 一种基于多元素结合随机森林算法对枸杞产地进行判别的方法 | |
CN112229863A (zh) | 一种铁矿石的原产国或品牌的鉴别方法 | |
CN109448842A (zh) | 人体肠道微生态失衡的确定方法、装置及电子设备 | |
CN112798678A (zh) | 基于血清的新型冠状病毒感染快速检测方法 | |
Ordóñez-Calderón et al. | Lithogeochemistry and chemostratigraphy of the Rosemont Cu-Mo-Ag skarn deposit, SE Tucson Arizona: A simplicial geometry approach | |
CN110361412A (zh) | 一种确定不同磁性磁黄铁矿相对含量的方法 | |
CN113355425B (zh) | 一种用于预测肺腺癌总生存时间高低分类的标志物及其应用 | |
CN113960088B (zh) | 一种针对进口散装铁矿卸货前固体废物属性快速筛查方法 | |
KR102330216B1 (ko) | Ai학습 분석을 이용한 골연령 및 건강상태 판독방법 | |
CN102012920B (zh) | 一种建立进口铁矿产地鉴别数据库的方法 | |
CN112666246A (zh) | 一种仿野生岩缝附生栽培铁皮石斛的甄别方法 | |
CN107067418A (zh) | 适用图像处理程序论证分析足迹检验 | |
CN107561036A (zh) | 一种白及品种及真伪的检测方法 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |