MAREMIS – AI-based calculation of air pollution emitted by ships and its dispersion.

Table of Contents

Abstract of MAREMIS

The consortium is developing Big Data and machine learning-based models and a demonstrator to measure, track, and validate emissions-related aspects of maritime transport to reduce emissions (exhaust, no ballast water or waste) from ships and improve local air quality. Validation is required because the basis for the project is data from the Automatic Identification System. These may be partially erroneous or the ship may have disabled its AIS. It is to be investigated here whether, on the basis of the trajectories as a string of position signals in relation to the AIS navigation status, ships can no longer be detected in the AIS or whether they have deactivated their AIS.

Summary of the Development

It will develop and deploy a ship emission model based on real ship movements and sensor-based emission data. The ship emission model will be used to estimate air pollution from maritime traffic in ports. It will reflect spatio-temporal emission dynamics and will track traffic emissions in real time and against scenarios. The impact on regional air quality, i.e. the Northern Germany region as well as Southeast Asia, will be analyzed using a chemical transport model, as these areas dominate the air input to urban areas. Models will continue to be developed to enable emission reduction strategies through changes in port operations and maritime traffic management using Big Data analysis, simulation, and optimization.

Initial situation

Singapore and Hamburg are among the largest ports in the world. The tens of thousands of ships each year are the heartbeat of the city and the economy. Both ports are very close to the city center and therefore have a strong impact on the air quality in these cities. However, the degree of impact has not been fully investigated at this stage. Once this has been determined, recommendations can be made to improve air quality.

Data acquisition and provision

AIS data must be provided for the project. Validation or cleansing is necessary to resolve possible inconsistencies in the reception of AIS telemetry data. This includes, for example, problems with the order of position reports from different sources, as well as possible transmission errors or spoofing. In IT, spoofing refers to various methods of deception in computer networks to conceal one’s own identity. The AIS data to be used must be stored in a highly available memory for the entire duration of the project, including demonstrator operation. This data will be used by both the German and SIN partners. In order to enable the emission calculation, further ship engine data are required. These will be provided in a highly available storage in coordination with the project partners. This Topic is manly handled by JCS and DLR KN.

Modeling of maritime emissions

The modeling of emissions from historical and current traffic helps to determine and assess air pollution from maritime traffic in ports and the port environment. In order to determine immissions, traffic emissions must be determined and coupled via a chemical transport model. On the basis of known or regular ship movements, intervention options are examined. As explained, this will be tested in the sense of a project as methods and technology demonstration at the port of Singapore and Hamburg. If successfully implemented and validated, the results would be transferable to other ports as part of subsequent product development. Modeling emissions from ship traffic based on scenarios enables the assessment and control of air pollution from ships in the port environment. Intervention cases can thus be tested, and traffic emissions can be predicted by coupling them with a chemical transport model and suitable boundary conditions. For the use in a demonstration system the software modules have to be adapted. This will allow a wide variety of scenarios to be tested in the demonstrator. The emission model is the responsibility of DLR.

The chemistry-climate model

Since many atmospheric chemical processes are highly nonlinear, the effect of a specific emission (e.g., from a ship) at a location is highly dependent on the background concentration. In order to evaluate the effect of ship emissions on regional air quality, other emissions, such as land transport emissions or natural emissions, must also be accounted for by the climate chemistry model. While many natural emissions are calculated directly by the climate chemistry model as a function of meteorology, external data sets are required, particularly for other anthropogenic emissions. This includes ship emissions. The effect of emissions from a ship are strongly dependent on the meteorological conditions, which, among other things, influence the chemical processing (chemical production and loss processes, deposition) as well as the transport of trace substances. In addition, the chemical processes strongly depend on the origin of the air masses. In order to analyze the impact of ship emissions on regional air quality in northern Germany and southeast Asia, detailed simulations of the dispersion and chemical transformation of the emissions are necessary. For this purpose, the MECO(n) model system is used, which combines a global climate-chemistry model with a regional climate-chemistry model.

Integration of sensor data

Broad networks of land-based air sensor data exist for routine air quality monitoring and for detecting limit violations. Many of the stations provide the measured data in near real time, although the data are therefore naturally only rudimentarily quality-checked. In addition, it should be noted that not all available stations may be useful for the project. For example, stations that are located very close to other emission sources (traffic, industry) may be strongly influenced by these sources. In addition to land-based air sensors, ship-based air sensors provide important information about the actual pollutant emissions from each ship. The use of existing air sensors on ships, which are also widely installed due to newer regulations, expands the possibilities of the project.

AI Based - Analysis of factors influencing air quality

In order to map the effects of emission changes and the impact of ship emissions on regional air quality in the demonstrator, simplified relationships between ship emissions, background conditions and the meteorological situation are necessary. To achieve this, a prototype of a statistical model will be developed based on the simulation results. The model data will be evaluated with the help of the observed data. In this step of the project, artificial intelligence (AI) and machine learning (ML) will be used to develop response models for the demonstrator. With the incorporation of AI and ML, the response models will have functionality that will allow for continuous improvement of the responses and thus the reliability of the outputs. The technological capability of JCS to compute for each ship feeds into this section of the project. Another major aspect is the evaluation of the results.

The Demonstrator | Outcome

The demonstrator will help port operators understand ship emissions from a situational perspective by also capturing the spatio-temporal dynamics of pollutants and the process of their dispersion into urban areas. The demonstrator will further provide decision support to port operators and policy makers to reduce ship emissions from a port operations and maritime traffic management perspective. Together, the partners from Germany and Singapore will evaluate the application of these newly implemented tools to the world’s busiest transhipment port in Singapore and a major German port. The comparison will take into account the corresponding strategies for maritime traffic management. It is planned to install sensors on ships to validate emission factors in ship emission inventories. Shore stations equipped with air sensors are also envisaged. The project activity will provide Singapore and Germany with insights and options for cooperation in addressing air pollution from shipping. In addition, this project could enable real-time and accurate monitoring of ship and maritime emissions status based on a maritime Big Data approach, even without installing pollution sensors on all ships.

Autonomous Sensor Station

FleetMon has developed an autonomous sensor station in 2020 together with JULIUS Marine as a specialist company for buoys and fairway lighting. Besides being equipped with an AIS receiving antenna, the station is featured with solar panels, a very large battery and power consumption optimized electrical devices. The station is able to host numerous sensors and send the data via GSM (Global System for Mobile Communications). Several of these devices will be equipped with air sensors and placed in the study area.

Visit our blog to read the article How a fully autonomous AIS unit adds value to worldwide vessel tracking.

Table of Contents

German Project Coordinator

Singapore Project Coordinator

IHPC Singapore

Institute of High Performance Computing

Project partner

Bergmann Marine
BM - Bergmann Marine

Maritime consultanting and advisors

Deutsches Zentrum für Luft und Raumfahrt
DLR - PA

Institute for Atmospheric Physics

Deutsches Zentrum für Luft und Raumfahrt
DLR - KN

Institute of Communications and Navigation

ShipFocus Group

A Chemical Shipping Specialist

faurecia | Singapore

Emissions Control Technologies

Associate partner

IAPH
IAPH

International Association of Ports and Harbors

MPA
Singapore Port Authority

Maritime and Port Authority of Singapore

HPA
Hamburg Port Authority

Senate of the Hanseatic City of Hamburg

Hochschule Wismar
HSW - University of Wismar

Department of Maritime Studies, Systems Engineering and Logistics

Equipment Provider

JULIUS Marine GmbH

navigational aids, buoys, marine lanterns and lock signalling systems

Bundesministerium für Bildung und Forschung
GER Funded by

BMBF - Federal Ministry of Education and Research

SIN Funded by

A*STAR | Agency for Science, Technology and Research

FONA
Project Framework

FONA | Research for sustainability, Smart urban mobility 2+2

Deutsches Zentrum für Luft und Raumfahrt
Promoter

DLR - Projekt Management

Official Project Profile

Project Duration

Aug 2021 – Jul 2023

Funding Volume

1,100,000 Euro

Contact

Share This Project

Share on facebook
Share on twitter
Share on linkedin
Share on whatsapp
Share on email