WNN-EDXRF方法测定地质样品中铁、钛元素含量

地质样品中各元素的定量分析是工矿业生产中重要的一环,为使对地质样品中各元素的定量分析更为精确、简便、行之有效,提出了一种小波神经网络(WNN)结合EDXRF分析技术的一种新的定量分析方法。先对样品进行预处理,进行化学分析,运用EDXRF分析技术得到的X射线强度计数,样品的一部分训练网络,训练过程中进一步研究了小波神经网络中动量因子和小波基函数个数对网络性能的影响。将另一部分样本输入网络进行预测并与化学分析值相比较。最终结果表明:它能够很好地描述各元素X射线强度计数与含量之间的非线性关系,可以得到比较精确的各元素预测值。

WNN-EDXRF Approach to Measure the Contents of Iron and Titanium of Geological Samples

It is an important part of mine industry to measure contents in geological sample.For finding a precise, easy and feasible approach to measure contents in geological sample, this article proposed an advanced prediction technique method by combining wavelet neural network (WNN) and EDXRF.We need to process the geological sample first and make a chemical analysis for obtaining the strength of x-ray by energy-dispersive X-ray fluorescence (EDXRF).A part of samples are used to train the network, during the training process, we studied the affections of factor of momentum and the amount of wavelet functions with the network, and then inputted another part of samples into the network for predicting, compared with the chemical analysis value.The final results show that it is able to express the linearity between the content and during the EDXRF analysis and obtain more precise prediction for elements.