Master Thesis Robust Volumetric Flow Control Using Sensor Fusion

Every day, we get opportunities to make a positive impact - on our colleagues, partners, customers and society. Together, we're pioneering the solutions of the future and unlocking the full potential of precious resources. Trusted to act on initiative, we challenge conventional thinking to develop world-leading technologies that inspire progress in vital areas, including energy, food, water and shipping.

As we push forward, the innovative, open spirit that fuels our 140-year-old start-up culture and rapid growth also drives our personal growth. So, as we shape a more resourceful, less wasteful world, we build our careers too.

Robust Volumetric Flow Control Using Sensor Fusion
Background
A high-speed separator separates a liquid into at least two phases according to density. In this application, the relevant output is volume. But due to technical limitations the measurement is by mass.

Currently, volume is estimated from mass measurements and density parameters. This volume estimate is used by PID controllers for volumetric flow control.

Can an improved volume estimate, combined with more sophisticated control algorithms, improve flow control and thereby increase separation efficiency?

In addition, could the enhanced process model provide early warning capabilities that reduce unplanned downtime and improve operational reliability?

Research Questions

• Can a sensor fusion approach that combines the current hardware design with numerical models result in a more accurate and robust volumetric flow estimate than the current method?
• What are the characterized measurement errors and noise properties of the current design under realistic operating conditions?
• Could this improved flow estimate, combined with more sophisticated control algorithms, enable better flow control than the current PID implementation?
• Can this result in improved alarm and early-warning functionality for detecting process deviations and mechanical wear?
• To what extent could these improvements increase separation efficiency?

Workflow suggestion

• Research the current implementation, hardware, and software. What is the untapped potential in the current design?
• Conduct a literature review of methods such as Kalman filters and advanced control algorithms.
• Controller and filter design. Develop new algorithms to be implemented in hardware. This will be done using MATLAB - Simulink / Stateflow.
• Test and develop the algorithms in a simulated environment using the existing simulation environment currently used at Alfa Laval.
• Automatic code generation will be used to deploy the developed algorithms to the target hardware. The control logic implemented in MATLAB Simulink will be translated into code for execution on a Siemens PLC, which serves as the primary controller of the machines installed at customer sites, controlling pumps, valves, and other process actuators. To reduce development risk and enable early validation, subscale hardware test setups may be utilized.

For more information, please contact:
Erik Eketorp, Manager, Smart Separator Systems, erik.eketorp@alfalaval.com

For union information, please contact:
Monica Anderberg, Unionen, at Monica.Anderberg@alfalaval.com

Johan Ranhög, Akademikerna at Johan.Ranhög@alfalaval.com

Stefan Sandell, Ledarna at Stefan.Sandell@alfalaval.com

We review applications continually. If the right candidate appears, the ad will be closed earlier. Applications sent directly via email will be disregarded without notice due to General Data Protection Regulation (GDPR).

We care about diversity, inclusion, and equity in our recruitment processes. We also believe behavioral traits can provide important insights into a candidate's fit to a role. To help us achieve this we apply Pymetrics assessments, and upon application you will be invited to play the assessment games.