Back to search

PETROMAKS2-Stort program petroleum

Deteksjon av akutt forurensning uavhengig av sikt, lys og værforhold

Alternative title: Oil spill detection independent of visibility, light and weather conditions

Awarded: NOK 4.1 mill.

Project Number:

235434

Project Period:

2014 - 2017

Organisation:

Regulations that apply to offshore petroleum activities in Norway require that an operator shall establish remote sensing systems to detect and map acute pollution quickly. In this context "remote sensing system" is interpreted as a system that, independent of visibility, light and weather conditions, can detect and map oil spills. A sensor system's operational weather window determines when it is able to detect oil spills. Sensor systems with operational weather window covering all visibility, light and weather conditions do not exist. Thus, a sensor system must always be combined with manual procedures that ensure oil spill discovery in periods when the system is unable to detect oil spills. Unfortunately, it is also the case that sensor systems can indicate that there is an oil spill when there is none. This is typically referred to as false alarms and entails costly follow-up activities that may involve a combination of aircraft, ships, water-sampling and knowledge-intensive evaluations. Our goal is to develop systems that can detect oil spills automatically with low false alarm rate, independent of visibility, light and weather conditions. It is typically infrared camera and X-band radar systems that are installed on exploration and production units for oil spill detection. Infrared camera systems require continuous operator interpretation to detect an oil spill while X-band radar systems can include algorithms for automatic detection. In this project Miros will make use of a new LIDAR based sensor system with automatic detection and combine it with an improved X-band radar system in an integrated system. A new solution with automatic detection based on two independent measurement principles will have a significantly lower false alarm rate and improved operational weather window compared to current solutions. This will enable operators to establish remote sensing systems that satisfy activity regulations with significantly lower resources than current solutions. In 2015, Miros and an operator in the North sea have conducted oil spill tests with a new radar based sensor system and gathered large quantities of data. In the last quartile of 2015 and two first quartiles of 2016 this data has been used to further develop the algorithms for detection of oil spills. Radar based systems with a first version of the new algorithm, designed specifically for a surveillance scenario, have been installed on multiple sites during the first half of 2016. This will, together with data from the 2015 oil spill tests, give us a basis of great value for further development of algorithms and the system as a whole. During the autumn of 2016, the radar based system has been in test operation at 3 installations. We have and will continue to follow up these installations to gather more operative experience in close collaboration with the end user organizations to improve the systems usability. In addition, we have tested the system in a busy harbor entrance with releases of bunker oil and diesel. The test outcome was highly successful. Algorithms are continuously being improved, as new data and experience is gathered from the installed systems. The focus has been gradually shifting from detection capability to suppression of false alarms. We are also working with getting the new LIDAR based sensor system in test operation at one or more test sites, together with an X-band radar system.

Prosjektets overordnede mål er å utvikle et nytt fjernmålingssystem som gir tilstrekkelig informasjon til å sikre at akutt forurensning fra innretningen raskt blir oppdaget og kartlagt uavhengig av sikt, lys og værforhold, med lavest mulig antall falske a larmer. Dette skal oppnås ved å utvikle et integrert system bestående av tre sensorsystemer og et sett anbefalte operative rutiner, som skal kunne tilpasses en gitt operativ kontekst. De ulike virkemåtene av sensorene som brukes til oljedeteksjon gjør at sensor sammensetningen kanskje ikke bør være lik for alle operative forhold som representeres med oljeutvinning i Barentshavet eller ved ekvator. Luftfuktighet, luft- og vanntemperatur og historisk vindstatistikk er alle parametere som påvirker et systems evne til å detektere olje. I prosjektet skal vi dokumentere miljøparameteres effekt på ulike måleprinsipper og utarbeide et underlag som kan brukes til å velge en optimal sensorkombinasjon for et gitt sett med operative forhold. I parallell skal det gjen nomføres forskning og utvikling av en ny sensor og eksisterende teknologi skal videreutvikles med bedre deteksjonsevne. I siste fase av prosjektet skal disse delsystemene med automatisk deteksjon kombineres med et tredje system for manuell verifikasjon av automatiske genererte oljedeteksjoner. Hovedutfordringen i disse aktivitetene er knyttet til å måle små signaler i mye støy og identifisere en målestrategi som gir systemet den ytelsen man trenger for å nå det overordnede målet for et meget bredt spekter av operative forhold.

Funding scheme:

PETROMAKS2-Stort program petroleum