Posters

Abstract

Researchers have sought, for some time now, ways to detect and identify pitch system faults. This remains challenging, but some progress has been made. What to do with the resulting information, i.e. the knowledge that a specific fault has happened or is happening, is perhaps well understood in terms of maintenance. However, our focus here is in terms of control: should we stop the turbine, continue operation, de-rate (and, if so, how and by how much)? Note that we are talking about faults that reduce the pitch system's operating envelope, i.e. how fast, hard and/or long it can pitch the blades, without compromising its immediate performance in terms of reference tracking; think a moderate internal leak at one of the hydraulic actuators, or a reduced energy storage capacity (hydraulic or electric). One ought to first consider how the fault affects the pitch system's envelope, then how it, in turn, affects the turbine's, then, finally, how to operate the turbine so that it stays within its reduced envelope. Fortunately, we can draw on a long tradition of Montecarlo simulations being used to fine-tune the turbine control, the pitch system specs and the turbine's load bearing capacity; all we need to do is repeat this for a range of pitch system faults, which we can simulate quite accurately and conveniently if we use a so-called physical model (i.e. a mathematical model based on physical principles) of the pitch system.