论文部分内容阅读
工业锅炉水总碱度测量常用指示剂法测定,该方法费时,难于自动化。利用流动注射分析(FIA)技术的快速、准确等优点可以实现对工业锅炉水总碱度的快速测定,将流动注射分析与离子选择性电极联用(FIA-ISE)滴定酸碱溶液可以获取连续变化的电压信号。由于锅炉水中常存在未知物质的干扰,基于局部线性嵌入(Local Linear Embedded,LLE)和偏最小二乘支持向量回归机(Least Square Support Vector Regression Machine,LS-SVR)相结合的电压数据建模方法可以消除未知物质的干扰,同时可以降低电压数据的维数。LLE算法将高维电压信号映射到低维流形空间,实现高维非线性电压数据结构的特征提取,LS-SVR建模方法可以建立总碱度的对数和降维后电压信号之间的非线性回归模型。实际待测数据的计算结果表明:基于LLE-SVR方法获得的锅炉水总碱度回归模型,通过对16组测试样本的分析,得到测量的平均相对误差:NaOH为0.83%,Na_2CO_3为0.66%;对于LS-SVR回归建模,NaOH为0.98%,Na_2CO_3为0.86%。LLE-SVR的运算负担只有LS-SVR的3.3%且测量精度要高于常规的峰面积和总碱度之间的线性拟合方法。
Industrial boiler water alkalinity measurement commonly used indicator method, the method time-consuming and difficult to automate. The rapid determination of total alkalinity of industrial boiler water can be realized by the rapid and accurate flow injection analysis (FIA) technique. The continuous and sequential analysis of FIA-ISE titration acid-base solution with flow injection analysis and ion-selective electrode Varying voltage signal. Due to the common interference of unknown substances in boiler water, the voltage data modeling method based on the combination of Local Linear Embedded (LLE) and Least Square Support Vector Regression Machine (LS-SVR) Can eliminate the interference of unknown substances, and can reduce the dimension of voltage data. LLE algorithm maps high-dimensional voltage signals into low-dimensional manifold space to extract the features of high-dimensional nonlinear voltage data structures. The LS-SVR modeling method can establish the non-linearity between the log of total alkalinity and the voltage signal after dimension reduction Regression model. The calculation results of actual data show that the average relative error of measurement based on the LLE-SVR boiler water total alkalinity regression model is 0.83% for NaOH and 0.66% for Na_2CO_3. For LS-SVR regression modeling, NaOH was 0.98% and Na 2 CO 3 was 0.86%. The computational burden of LLE-SVR is only 3.3% of the LS-SVR and the accuracy of the measurement is higher than the linear fit between the conventional peak area and total alkalinity.