Pipeline cleaning (flushing) is a critical yet often under-optimized process in the lubricant oil packaging industry, performed to eliminate residual oil from pipelines and equipment before introducing a new product. This procedure ensures minimum cross-contamination between different lubricant formulations. However, flushing inevitably generates waste in the form of commingled oil - a mixture of residual and new oil resulting in low value (downgraded) products. Thus, contributing to additional labor demands, production downtime, and material loss. The conventional flushing process depends heavily on the operator’s experience on the required time and volume of new oil used in flushing. This empirical approach, while sometimes effective, leads to inconsistent outcomes and often results in longer flushing time than necessary, thereby increasing commingled oil generation and reducing overall operational efficiency.
This research addresses these challenges by investigating the fundamental dynamics of the flushing process and developing a mathematical model that describe the transition from the residual to the new oil during flushing. Viscosity, being a key parameter between lubricant products, serves as a practical metric for assessing the purity of the transition from the residual to new oil within the system. The primary goal of the model is to provide quantitative guidance for operators with the required time (or volume) necessary to attain desired product specifications. Experimental studies were conducted on a pilot plant designed to replicate the geometry and flow characteristics of an actual industrial packaging plant. Residence time distribution (RTD) studies were used to ensure the pilot plant accurately reflects the fluid dynamic behavior of full-scale operations [1]. With the experimental studies, we are able to assess the performance of our developed models.
The well-mixed model [2] developed initially, provided an 18% reduction in the volume of new oil required for the flushing process at the industry. The well-mixed model, however, is limited in its ability to capture the viscosity profile of the product leaving the system during changeovers. To address this issue, a new two-parameter model has been introduced. In this framework, α (alpha) represents the fraction of the system that behaves as a well-mixed region, while β (beta) quantifies bypass or plug formation phenomena. Both parameters are described by power-law correlations with the physical properties of lubricant oils.
Extensive experimental studies were conducted on a pilot plant system that mimics industrial pipeline geometry and flow behavior. Tests across a wide viscosity range confirmed the reliability of the developed correlations for α and β. The resulting model consistently predicted changeover performance with high precision, enabling more accurate estimation of flushing volumes. Overall, this work provides a scalable modeling approach that enhances decision-making, reduces waste, and improves the sustainability of lubricant oil packaging operations.
References
[1] B. Gao et al., “Improved Design of Flushing Process for Multi-Product Pipelines,” presented at the Foundations of Computer-Aided Process Design, Breckenridge, Colorado, USA, Jul. 2024, pp. 137–144. doi: 10.69997/sct.171679.
[2] S. S. Jerpoth, R. Hesketh, C. S. Slater, M. J. Savelski, and K. M. Yenkie, “Strategic Optimization of the Flushing Operations in Lubricant Manufacturing and Packaging Facilities,” ACS Omega, vol. 8, no. 41, pp. 38288–38300, Oct. 2023, doi: 10.1021/acsomega.3c04668.