Close this search box.

Technical News

Ai for optimising terminal efficiency

Written by...

Picture of Meenal Datar

Meenal Datar

Membership specialist

A liquid bulk terminal in the port of Antwerp asked Actemium to optimise its mechanism for tank pressure control. By using artificial intelligence to create an automated planning tool, the incinerator overload could be avoided and a more efficient way of working was guaranteed

Standard tank pressure control

the main function of a tank terminal is, of course, the storage of goods. For liquid chemicals, every tank contains a liquid and gas phase where the tank pressure needs to be monitored in order to avoid the creation of a vacuum or an overpressure. The most common operations are the loading and unloading of ships and trucks. These operations have an impact on the tank pressure and when the pressure exceeds a predetermined level, thermal oxidisers are used to incinerate gas from the tanks.

However, these operations are not the only factor affecting tank pressure. Outside weather conditions also have an impact, especially during spring when nights are still cold, but days can be nice and sunny. The solar radiation increases the temperature inside the tank and consequently, the tank pressure as well. When the predetermined pressure threshold is reached, the incinerator is enabled to reduce the tank pressure to avoid overpressure. Since the temperature affects all tanks at the same time, this results in a huge demand for vapour extraction where the nominal capacity of the incinerator is usually insufficient. Subsequently, this can lead to an overload of the incinerator which will go in shutdown with the risk of shutting down all ongoing terminal operations which entails high costs for the terminal.


Automated planning solution

To solve these issues, Actemium started a co-creation trajectory with their client. Vapour extraction needs to be spread out over time so that the maximum capacity of the incinerator is not exceeded. To moderate this process, an automated planning tool for tank extraction is the solution. This planning tool is based on priorities and on tank pressure prediction. The challenge lies in the latter.

To tackle this challenge, data was needed to be able to develop a machine learning algorithm which can predict the individual tank pressures. In order to obtain such an algorithm, a weather station was installed at the tank terminal to log the local measured weather data. Also, the different process parameters and tank pressures were logged. Next, a team of data analysts combined this input to develop and continuously optimise an artificial intelligence (AI) model which in turn can be applied, with the input of forecasted weather data, to predict the individual tank pressures.

By predicting these tank pressures, taking into account the weather conditions, the incinerator usage planning can be optimised. This will significantly reduce the probability of a complete shutdown and will facilitate overall operations. Furthermore, by anticipating future weather conditions, other saving opportunities arise as well.

On the one hand, the terminal could minimise product loss by avoiding unnecessary incineration when the temperature, and thus pressure, would drop in the near future anyway. On the other hand, this solution could also avoid unneeded heating of the tank if the outside temperature would rise in the next few hours.


More information
and contact details on:


What can Actemium do for you?

Actemium is a brand of VINCI Energies 100% dedicated to industry, among which liquid bulk storage. We optimise your terminal with our Actemium Terminal Management System (ATMS).

Fast changing customers or products, short-term contracts and specific legislations make it important for the tank terminal industry to be flexible and to keep track of everything. ATMS is an independent, completely integrated terminal automation system, covering all necessary elements for a most efficient terminal management from order-to-cash.


Share this article:

Latest technical news