Back to search

PETROMAKS2-Stort program petroleum

Sensors and models for improved kick/loss detection in drilling (Semi-kidd)

Alternative title: Sensorer og modeller for forbedret kick/loss deteksjon i olje-boring (Semi-kidd)

Awarded: NOK 16.0 mill.

Project Manager:

Project Number:

255348

Project Period:

2016 - 2020

Project "Semi-kidd" will develop new sensors for oil drilling to improve the management of drill fluid, develop early warning about problems, and reduce non-productive time in an automated drilling process. Norwegian oil and gas production typically takes place offshore. Drilling of production wells is important for efficient production of oil and gas, and Norwegian oil companies are market leaders in this area. Drilling at increasingly greater depths poses challenges to technology, and an increasing complexity and required safety has led to tentatively 30% non-productive time in drilling -- with a steep increase in cost. Good control of the well pressure is mandatory during drilling. Flow of drilling fluid into or out of the well enables pressure control; this fluid also lubricates the drilling crown and transports drilling mud back to the surface. Too high well pressure will press mud into the well formation with resulting loss of fluid; this will simultaneously reduce the productivity of the well. Too low well pressure will allow oil, gas and formation water (or just water) to leak out into the well, and may lead to uncontrolled blowout -- e.g. the Deepwater Horizon accident in the Gulf of Mexico in 2010. Project "Semi-kidd" aims to develop improved monitoring of net addition of fluid to the well during drilling. Positive net fluid addition indicates a high well pressure and damage of the well formation. Negative net fluid addition indicates a low well pressure and leakage ("kick") out of the reservoir and into the well. Early warning of problems requires both knowledge of net fluid addition, and of fluid properties such as density, viscosity, etc. Current sensor instruments are large, heavy and expensive, and give limited information. Project "Semi-kidd" aims to develop new sensor systems based on combining many, cost-effective sensors with computer models to a robust, accurate solution; these sensors will be tested in controlled experiments. So far, 2 PhD candidates have defended their theses, and accepted RnD positions in the industry/academia. Twelve MSc theses have involved problems related to the project, and some tens of MSc students have worked with problems related to the project in courses. Approximately 25 journal and conference papers have been published so far. One PhD candidate has participated in the national Research Grand Prix final, another has presented the project for OG21, and there has been presentations for the industry (SPE, Telemark Offshore conference). Three candidates have been in internships with project partners. At closing of the project, all 4 PhD candidates have defended their theses according to the plan, and accepted RnD positions in the industry/academia. Fifteen MSc theses have been carried out related to the project (USN, NTNU), 8 candidates have done their MSc projects related to Semi-kidd, 40-45 candidates have carried out larger project tasks in MSc courses in instrumentation engineering and Model Predictive Control, and 30 BSc candidates have done laboratory exercises on the Venturi rig of the Semi-kidd project. Twenty papers have been published in international journals, and 19 papers in peer reviewed conference proceedings. One PhD candidate participated in the national finale of the Research Grand Prix, another has presented the project at OG21, and there has been presentations in the industry (SPE, Telemark Offshore conference). All four PhD candidates have had internships with partners. One partner is in the process of commercializing a product based on the ideas in the project.

Nasjonal kunnskapsbase Prosjektet har utviklet: (i) modeller og løsningsalgoritmer for strømning av fluider i åpne kanaler; (ii) metoder for estimering av strømning basert på modeller med nivåmåling; (iii) smart-sensorer ved å fusjonere multiple sensortyper med datadrevne modeller; (iv) observere/estimatorer for fordelte modeller (NTNU). Prosjektet har inkludert 4 ph.d.-studenter, og 95 MSc- og BSc-studenter. Relevans for norsk industri Prosjektet har gitt partnerne førstehånds informasjon om status og videre utvikling, og har pekt på behovet for å forbedre eksisterende målekvalitet. Utviklede metoder har generell nytteverdi. Fordelt strømning er relevant for problemstillinger som olje-/gassproduksjon («DigiWell»), vannkraftproduksjon («HydroCen»), og mange andre felt; estimering i slike systemer er viktig. Maskinlæring og fusjon av multiple sensorer har generell nytteverdi. «Semi-kidd» har gjort USN bedre rustet til å bidra for norsk industri.

Cost-effective drilling is key to increase the recovery on the Norwegian continental shelf. In the last 10 years, well construction costs have increased significantly, while drilling efficiency is reduced. The reason for this is complex, but an important factor is the lack of automation of the drilling operations. Development of improved sensors and models for cost-effective early loss and kick detection is vital to address the industry's need of improving the drilling efficiency and reducing risks of well control incidents with its environmental and costly consequences. The primary objective of the project is to enable cost-effective and automatic kick/loss detection by developing new knowledge on model-based estimation and utilization of new sensor technology for drilling operations. The project is divided into four research tasks. The first is to model and simulate the flow rate and fluid level in an open Venturi channel. The non-Newtonian properties of the drilling fluid is a challenge that has to be taken into account. The second task is to design a smart sensor for estimating the flow rate of drill fluid under steady and transient conditions. Modern variants of Kalman filters will be considered for the estimation algorithm. The third task is to design an improved kick/loss detection (KLD) algorithm by including new measurements, to provide early and robust KLD without false alarms. The fourth task is a feasibility study on utilizing gamma sensors, ultrasonic Doppler and tomographic sensors in the drilling operations. Fusing data from new and existing sensors is essential. Experimental research will be performed on a new test facility particularly designed to carry out realistic tests of flow and density sensors with drilling fluids. The anticipated results of the project are innovation of new models and sensors, which in combination with existing sensors will contribute to improve the kick/loss detection in drilling operations.

Publications from Cristin

No publications found

Funding scheme:

PETROMAKS2-Stort program petroleum