Dr. Arne Dankers

I am developing theory and algorithms to estimate the differential equations that govern a component embedded in a larger system from sensor data. One of the reasons for doing this is to monitor the health of that component over time. If you can detect that the physics of the component (the differential equations) are changing, then you can determine whether or not the component is wearing out. If there is a sudden change, then the component may have been damaged. Being able to monitor components of a large system or machine independently is important when the goal is preventative maintenance, or predictive maintenance of the system. 

The methods that I use are rooted in system identification (data driven modelling). Specifically, network system identification which can be seen as an extension of closed loop identification to more complex interconnection structures between modules (or sub-systems, or components)