Effect Soft Measurement Model of Steel Slag Powder Repair Heavy Metal Contaminated Soil with Fourier Transform Infrared Spectrum

Effect Soft Measurement Model of Steel Slag Powder Repair Heavy Metal Contaminated Soil with Fourier Transform Infrared Spectrum
复制标题

DOI:
10.3964/j.issn.1000-0593(2017)03-0743-06
复制
发表时间:
2017-03-01
影响因子:
0.7
通讯作者:
Gu, Heng-xing
Gu, Heng-xing
中科院分区:
化学4区
文献类型:
--
作者:
Yang, Gang;Li, Hui;Gu, Heng-xing

文献摘要

被引文献

相似文献

以转炉渣为固化稳定剂,应用毒性特征浸出法(TCLP)和傅里叶变换红外光谱法跟踪检测了重金属在转炉渣粉土壤修复中的效果,以及混合物(转炉渣粉与重金属污染土壤)的微观结构。利用高斯过程回归建立了基于高斯过程回归的矿渣微粉对重金属污染土壤修复效果的软测量模型。结果表明:钢渣微粉对重金属污染土壤具有良好的修复效果,180 d内其修复效果均保持在90%以上;修复过程分为早、中、晚三个阶段,(1)类似于3d)修复模式以离子交换为主的高碱性环境,中期(7类似于42 d)离子交换和凝胶凝固减弱了后期的增强作用(56 d类似于180 d)形成大量CSH凝胶,凝胶固化进一步加强;基于高斯过程回归的钢渣微粉对重金属污染土壤修复效果软测量模型的真实的和预测值与数据吻合较好,绝对误差为-1。35类似于-O。48,相对误差为-1。448%与-0相似。百分之四百九十七
Using converter slag as curing stabilizing agents, applying Toxicity Characteristic Leaching Procedure (TCLP) and Fourier transform infrared spectroscopy method of tracking and detection of heavy metals in soil remediation effect of slag powder, and the micro-structure of mixtures (slag powder and heavy metal contaminated soils). It can establish soft sensor mode which is based on Gaussian process regression slag powder on heavy metal contaminated soil remediation effect, by using Gaussian process regression. The results show that the steel slag powder on heavy metals contaminated soil has good repairing effect, 180 d within its restorative effects are maintained over 90%; the repair process is divided into early, middle and late stages, in which the early (1 similar to 3 d) repair mode is given priority to with ion exchange high alkaline environment, medium-term (7 similar to 42 d) ion exchange and gel setting weaken the enhanced role of the late (56 similar to 180 d) form a large number CSH gel, the gel solidification further strengthened; based on Gaussian process regression Steel slag powder on heavy metal contaminated soil remediation effect soft measurement model of the real and predicted values agree well with the data, the absolute error is -1. 35 similar to-O. 48, relative error of -1. 448% similar to-0. 497%.