Properties of an immiscible polymer blend have been proved to be closely related to dispersion uniformity of the minor phase. At present, dispersion uniformity is difficult to evaluate during the blending process, resulting in hysteretic feedback. Aiming at this problem, this work utilized near‐infrared (NIR) spectroscopy to in‐line characterize dispersion uniformity evolution during a twin‐screw extrusion. A multichannel NIR measurement system was set up and applied to evaluate the blending process of polypropylene and polyolefin elastomer (POE). Based on the NIR spectra collected at different positions of the extruder, five prediction models of POE content were established using the light gradient boosting machine algorithm. Dispersion uniformity was characterized through the fluctuation of the predicted content. The evolution of dispersion under such processing parameters was consistent with scanning electron microscopy.

中文翻译：

---

基于近红外光谱的双螺杆共混挤出过程中分散均匀性演变的在线表征

已经证明不混溶的聚合物共混物的性质与次相的分散均匀性密切相关。目前，在混合过程中很难评估分散均匀性，从而导致滞后反馈。针对这个问题，这项工作利用了近红外（NIR）光谱技术来在线表征双螺杆挤出过程中的分散均匀性演变。建立了多通道近红外测量系统，并将其应用于评估聚丙烯和聚烯烃弹性体（POE）的共混过程。基于在挤出机不同位置收集的NIR光谱，使用光梯度增强机算法建立了五个POE含量预测模型。通过预测含量的波动来表征分散均匀性。