The primary objective is to qualify and enhance Embedded Optimization
algorithms for use in model predictive control (MPC) for automatic control in subsea
processing and automated intelligent drilling.
The main secondary objectives are
-
Research
leading to new embedded optimization algorithms, MPC design methods,
implementation and verification tools for embedded MPC suited for
safety-critical applications offshore and subsea;
-
Qualify and
demonstrate embedded optimization for automatic control in subsea processing,
managed pressure drilling and production optimization case studies in
collaboration with Statoil;
-
Train 2 PhD
fellows (Kwame Kufoalor and Benjamin Binder).
Today’s technology for optimization-based control and MPC are essentially
limited to slow processes (update rates in minutes or seconds) that have a
dedicated lower-level control system (such as a decentralized control system)
and/or a dedicated safety-system. Today’s MPC technology is therefore based on
server-type or PC-like computers and software solutions that does not meet the
oil and gas industry’s standard for safety and reliability in stand-alone
operations. In new applications such as subsea processing and automated
intelligent drilling the existing MPC technology has some limitations, and
should be enhanced for computational efficiency and software reliability.
This project’s answer to this challenge in to enable MPC on
ultra-reliable industrial computer system hardware such as microcontrollers and
PLCs, and thereby providing the petroleum industry with automatic control
implementation technology that will enable more advanced functionality to be
more easily built into such control systems. There is a clear trend towards
increased levels of automation, autonomy, built-in intelligence and integrated
software-based functionality in control and monitoring systems that will be
enabled by this project since embedded numeric optimization methods offer the
most potent technology to make real-time choices and automated decisions with
no or little human intervention.