Controlling the consistency of aluminium ore concentrate is difficult when there is no proper sampling protocol in place. A proper sampling strategy is necessary to obtain the data for effective quality assurance and quality control. Inputs from several disciplines, including statistics, geostatistics, geology, and sampling, are necessary to establish realistic goals for assessing the accuracy and precision levels regarding the main constituents of aluminum ore, namely available alumina and reactive silica. Selecting an appropriate sampling protocol requires the estimation of inherent heterogeneities and the application of the theory of sampling (Gy 1976) to minimize all sources of bias. The international sampling standards (AS 2806.1 2003; AS 2806.3 2001; AS 2806.4 2004; AS 2806.5 2003; AS 2806.6 2003 and AS 2806.7 2004) describe tests to evaluate precision and accuracy in order to improve data reliability; however, these tests can still be improved upon. This study outlines a proposal for optimizing the sampling protocols used when mining aluminum ore; it employs current standard procedures, Pierre Gy’s Theory of Sampling (Gy 1979), as well as general geostatistical concepts (David 1977).

中文翻译：

---

优化铝矿石的采样方案——一种新方法

如果没有适当的采样方案，则很难控制铝精矿的稠度。为获得有效的质量保证和质量控制数据，必须采用适当的抽样策略。来自多个学科（包括统计学、地质统计学、地质学和采样）的输入对于确定评估铝矿石主要成分（即有效氧化铝和活性二氧化硅）的准确度和精密度水平的现实目标是必要的。选择合适的抽样方案需要估计固有的异质性并应用抽样理论 (Gy 1976) 以尽量减少所有偏差来源。国际抽样标准（AS 2806.1 2003；AS 2806.3 2001；AS 2806.4 2004；AS 2806.5 2003；AS 2806.6 2003和AS 2806。7 2004) 描述评估精确度和准确度以提高数据可靠性的测试；然而，这些测试仍然可以改进。本研究概述了优化铝矿开采时使用的采样协议的建议；它采用当前的标准程序、Pierre Gy 的抽样理论 (Gy 1979) 以及一般地质统计学概念 (David 1977)。