高级检索

    分光光度法同时测定石油制品中铁镍钒及钴的含量-小波变换-LSSVM算法的应用

    Simultaneous Spectrophotometric Determination of Iron Nickel Vanadium and Cobalt in Petroleum Products by Applying the Algorithm of Wavelet Transform and LSSVM

    • 摘要: 石油产品样品经燃烧和灼烧除去碳氢化合物后,用稀盐酸溶解残渣,分取部分样品溶液用5-Br-PADAP作显色剂,对其中铁、镍、钒、钴的含量进行分光光度法同时测定.用试剂空白作参比溶液,在540~620 nm波长范围内每隔2 nm测定一次吸光度.所得数据用小波变换法处理以滤去分析信号中的噪声,然后用最小二乘支持向量机(LSSVM)算法解析所得分析数据.结果表明:LSSVM算法具有计算速度快,结果准确及泛化性能好等特点.用模拟样品对LSSVM算法进行预测,测得上述4元素的回收率在96.0%~103.5%之间.应用于分析实样时,所测得4元素结果与原子吸收光谱法的测得结果一致,所得结果的相对标准偏差(n=8)均小于4%.

       

      Abstract: Sample of petroleum products was burned and ignited to eliminate hydrocarbons,and the residue was dissolved in dil.HCl.A aliquot of the sample solution was taken for simultaneous spectrophotometric determination of Fe3+,Ni2+,V(Ⅴ) and Co2+ with 5-Br-PADAP as color reagent.Values of absorbance were taken in the range of wavelengths between 540-620 nm at 2 nm intervals,using reagent blank as reference.Data of absorption spectra were treated by wavelet-transform to filter the noise from the analytical signals.The analytical data were then analyzed by the algorithm of least square support vector machine (LSSVM).As shown by the results,the algorithm of LSSVM showed its special features of speedy calculation,high accuracy and good generalization.In pretesting of LSSVM with some simulated samples,values of recovery for the 4 elements were found in the range of 96.0%-103.5%.In analysis of substantial samples,results obtained by this method were in conformity with those obtained by AAS.Values of RSD′s (n=8) for all the 4 elements were less than 4%.

       

    /

    返回文章
    返回