Tankpark scheduling optimisation

Tankpark scheduling optimisation

Baart, Pepijn (TU Delft Mechanical, Maritime and Materials Engineering)

Jiang, X. (mentor)
Schott, D.L. (graduation committee)
Kana, A.A. (graduation committee)
van der Tempel, Wijnand (graduation committee)

Delft University of Technology

Mechanical Engineering | Transport Engineering and Logistics

2020-05-18

This thesis is done in cooperation with Sime Darby Oils Zwijndrecht (SDOZ). SDOZ is a refinery of vegetable based oils. The goal of this thesis is to optimize the scheduling of the Refined Oil Storage (ROS) of the refinery. This is done because within the company it is suspected that the current way of assigning the tanks, that is reactively by hand, is severely limiting the amount of usable free storage volume within the ROS. By increasing the amount of usable volume within the ROS, the production in the refinery, and therefore its profits, could be increased. This thesis starts out by analysing the exact scheduling problem at SDOZ. This information was then used to compare multiple scheduling optimisation methods from the literature in order to find the best applicable method. This was "A novel network-based continuous-time representation for process scheduling". The report continues by verifying this method, by reproducing example cases included in the literature. During these reproductions some issues were found, which were resolved by making modifications to the method. The next step was to improve the method to better suit the method to be used to schedule the SDOZ ROS. This was done by creating additions, which included the use of loading spots, the use of preloaders, and the need for quality checks. The complete method, including all features of the literature, and the additions made in this thesis, was verified after which it was validated. In the validation the developed method was compared to the current situation at SDOZ. This comparison showed that using the method developed in this thesis can improve the amount of usable free storage volume within the ROS significantly. It is therefore recommended for the company to implement the developed method. In this thesis the first steps needed in order to apply the method to the ROS of SDOZ are taken. In future research this could be used to completely optimise the scheduling of the ROS, something that was due to calculation power limitations, not yet possible in this research.

Optimisation
Scheduling
MILP

http://resolver.tudelft.nl/uuid:3892b2df-fa93-4fd9-8491-3afd3907d8d0

2022-05-18

51.8079445, 4.6378353

Student theses

master thesis

© 2020 Pepijn Baart