基于遗传算法的BP神经网络模型在岩心扫描仪测定海洋沉积物多种组分中的应用研究

发布时间：2018-12-18 13:05

【摘要】：海洋沉积物样品成分复杂,由于基体效应的影响,利用岩心扫描仪开展X射线荧光光谱分析只能得到目标元素的强度信息,不利于该方法在成矿机制和古环境等研究领域更好地发挥作用。本文采用岩心扫描仪测定海洋沉积物中的铝硅钾钙钛锰铁钒铬铜锌铷锶钇和铅15种元素,尝试引入BP神经网络模型利用其非线性拟合能力校正基体效应。实验表明,以水系沉积物、海洋沉积物和岩石国家标准物质以及定值海洋沉积物样品为训练样本,采用遗传算法优化BP神经网络的初始权值和偏置,可以有效校正除硅之外的14种元素基体效应的影响,实现了岩心扫描仪XRF测量结果由强度到浓度的转化。本方法的精密度为0.6%~6.8%(RSD,n=11),国家标准物质和海洋沉积物实际样品中15种组分的预测值与参考值的相对偏差在0.5%~17.5%之间,适合于海洋沉积物中多种主次量组分的快速分析,拓展了岩心扫描仪的功能。
[Abstract]:The composition of marine sediment samples is complex. Due to the effect of matrix effect, the intensity information of target elements can only be obtained by using core scanner to carry out X-ray fluorescence spectrum analysis. It is unfavorable for this method to play a better role in metallogenic mechanism and paleoenvironment. In this paper, a core scanner is used to determine 15 elements of aluminum, silicon, potassium, calcium, titanium, ferromanganese, vanadium, chromium, zinc, rubidium, strontium, yttrium and lead in marine sediments. The BP neural network model is introduced to correct the matrix effect by using its nonlinear fitting ability. The experimental results show that the initial weight and bias of BP neural network are optimized by genetic algorithm with the samples of water system sediment, marine sediment and rock national standard and fixed value marine sediment as training samples. The effect of matrix effect of 14 elements other than silicon can be corrected effectively, and the XRF measurement results of core scanner can be transformed from intensity to concentration. The precision of this method is 0.6% (RSD,n=11), and the relative deviation between the predicted value and the reference value of the 15 components in the actual samples of national standard materials and marine sediments is between 0.5% and 17.5%. It is suitable for rapid analysis of various primary and secondary components in marine sediments, and extends the function of core scanner.
【作者单位】：国土资源部海底矿产资源重点实验室;广州海洋地质调查局;
【基金】：国土资源部海底矿产资源重点实验室资助项目(KLMMR-2014-A-01)
【分类号】：P736.4;O657.34

【相似文献】