以金矿石中的金元素为研究对象, 基于灰色关联度和RSM模型提出原子吸收光谱法分析Au测定条件参数的多目标优化模型。 选取泡塑预处理方式、 振荡时间、 王水浓度和硫脲浓度为优化目标, 确定测量结果相对误差的绝对值为质量指标； 建立基于信噪比的正交设计试验, 分析试验结果的质量指标及对应的信噪比并进行量纲化处理, 计算灰色关联系数和关联度； 确定优化目标的极差分别为0.026、 0.116、 0.176、 0.375, 定性判断泡塑处理方式目标最不显著。 根据RSM模型确定王水浓度、 振荡时间、 硫脲浓度为单因素的Box-Behnken方法试验, 采用三因素三水平曲面设计对测定结果相对误差的绝对值进行分析, 制作显著水平表并完成响应曲面试验； 建立二次多项式回归方程的预测模型并进行显著性分析, 其F=217.24, p<0.000 1表明该模型具有高度的显著性, 模型的相关系数为0.996 9, 校正决定系数为0.992 4, 表明该模型可以解释超过99%的响应值变化； 绘制响应曲面图和等高线图对试验数据进行回归拟合, 通过响应曲面的形状和等高线的陡峭程度进行判定分析, 最终寻找最优化目标参数为王水浓度、 振荡时间、 硫脲浓度分别为11.33%、 27.39 min和0.97%时, 样品测量结果的相对误差最小。 模型验证结果表明, 在最优化目标参数条件组合下选择不同质量浓度的金矿石国家标准物质进行样品制备, 测定结果的正确度和精密度均符合DZ/T 0130.3—2006(地质矿产实验室测试质量管理规范), 表明基于灰色关联度-RSM模型对原子吸收光谱法分析金矿石中金元素的多目标优化参数准确可靠, 验证了该方法的科学性和正确性, 能够被应用于实际的生产中去。 该方法对于定性判断各条件参数间的主次关系, 定量计算各条件参数的最佳组合水平具有独特优势, 有望在寻找多目标优化设计参数领域的平台上发挥作用, 更加有效地确定最佳目标组合。

Based on gray Correlation degree and RSM Model, a multi-objective optimization Model for AAS analysis and determination of gold elements in gold ore is proposed. The foam pretreatment method, oscillation time, aqua regia concentration and thiourea concentration are selected as the optimization objectives, and the absolute value of the relative error of measurement results is determined as the quality index. The orthogonal design experiment based on SNR is established, the test results quality index and corresponding SNR are analyzed and tempered, and the grey correlation coefficient and correlation degree are calculated. The range of determined optimization targets is 0.026, 0.116, 0.176 and 0.375, respectively, and the target of foam treatment is the least significant in qualitative judgment. According to RSM Model, aqua regia concentration, oscillation time and thioureas concentration are determined as single factors in the box-Behnken method test. Three-factor three-level surface design is used to analyze the absolute value of the relative error of the measured results, a significance level table is made, and the response surface test is completed. The prediction model of the quadratic polynomial regression equation is established, and significance analysis is carried out. Its F=217.24, p<0.000 1 indicated that the model had high significance. The correlation coefficient of the model is 0.996 9, and the calibration determination coefficient is 0.992 4, indicating that the model could explain more than 99% of the response value changes. The response surface diagram and contour map are drawn for regression fitting of the test data. The response surfaces shape and the contour lines steepness are determined and analyzed. Finally, the optimal target parameters are found when aqua regia concentration, oscillation time and thiourea concentration are 11.33%, 27.39 min and 0.97% respectively. The relative error of sample measurement results is minimum. The model verification results show that the accuracy and precision of determination results are in line with DZ/T 0130.3—2006 (The Specification of Testing Quality Management for Geological Laboratories) by selecting gold ore national standard substances with different mass concentrations under the combination of optimal target parameters and conditions. The results show that the multi-objective optimization parameters of atomic absorption spectrometry for the analysis of gold elements in gold ore based on gray correlation degree-RSM Model are accurate and reliable, which verifies that the method is scientific and correct and can be applied to practical production and application. This method has unique advantages in the qualitative judgment of the primary and secondary relations among the condition parameters and quantitative calculation of the optimal combination level of the condition parameters, which is expected to play a role in the search for multi-objective optimization design parameter field platform and determine the optimal target combination more effectively.