Early warning collision avoidance system at INEOS Phenol at Port of Antwerp-Bruges

The area around the INEOS phenol jetty in the Port of Antwerp, Belgium, is busy with a wide variety of vessel movements. A collision with the jetty potentially can have a severe impact on the site’s operations. In time warning of personnel and closing of the valves can prevent severe damage and multi-million EUR operational losses. An early warning collision avoidance system based on AIS and radar data, and the xyzt.ai anomaly detection software prevents such incidents.

The challenge

INEOS operates a jetty at the Port of Antwerp in Belgium, one of the busiest sea ports with a special location far inland on the river Schelde. The area surrounding the phenol jetty is busy with small and large vessel traffic including container vessel traffic from the nearby terminal, and the busy traffic from other vessels passing by along the river Scheldt.

It is not easy to predict the course of each of those vessels, as they have to follow the bend in the Scheldt, watch other vessels and reach their own destination. When a ship accidentally broke away from the quay in the nearby Deurganckdok and drifted towards the INEOS Phenol jetty, it prompted INEOS Phenol to look for an early detection and warning system.

Still today, there is a severe risk of collisions with the jetty. Such incidents need to be anticipated so that the necessary valves can be closed.

The solution

The solution is a real-time early warning system based on anomaly detection algorithms that fuse vessel tracking data with radar data, and audio communication data. Using machine learning, anomalies are detected and a decision engine and visual dashboard is used to alert the operator in advance of a possible collision.

Benny Van Den Heuvel, Smart Manufacturing Excellence Manager of INEOS Phenol Belgium: "Safety is INEOS' first priority, and that includes on and around the jetty. It is used to transport raw materials and end products mainly by water. That's why we were looking for a notification system to be warned as early as possible of potential risks from the Scheldt."

Results

The system is a real-time alerting system based on the xyzt.ai platform and its anomaly detection algorithms. Operators can be warned up to 20 minutes beforehand, depending on the incident building up. In the below test scenarios, 4 abnormal courses are detected successfully using the xyzt.ai anomaly detection algorithms:

A previously moored cargo ship that is adrift, slowly moving to the Ineos jetty
A tanker that is going too fast, unable to take a turn in time and crashing into the jetty
A cargo ship sailing into the wrong shipping lane (e.g. due to a defect rudder)
A terroristic attack with a pleasure craft moving from a nearby yacht club to the jetty at high speed

Drifting vessel on a collision course with the jetty. The path of the vessel becomes a danger and is detected by the xyzt.ai platform.

Early warning collision avoidance system at INEOS Phenol at Port of Antwerp-Bruges

The challenge

The solution

Results

Request a demo