Source: EU

Algorithms hold key to help engines stay in good shape Algorithms hold key to help engines stay in good shape

By Aimee Turner in Toulouse

Complex algorithms are being used for the first time as virtual sensors in an effort to advance engine health monitoring techniques.

Speaking at the first user forum for TATEM (technologies and techniques for new maintenance concepts) – a four-year, €40 million ($50 million) project co-funded by the European Commission under its Framework 6 research programme, which aims to cut future generation aircraft maintenance costs by 50% within 20 years – Jean-Remi Masse from Hispano Suiza reported on the progress of latest fusion methods in the field of engine health monitoring.

“Physical methods of diagnostics based upon an engine model have been applied in order to identify the evolution of deterioration of engine components. The first results have been very encouraging and are still being pursued with new real data scenarios. In parallel, non-physical methods are being tested on real data scenarios,” he said.

These non-physical methods include health monitoring on a bi-engine basis with performance and/or vibration parameters adjusted for the two engines before this engine data fusion is subjected to a Baysian belief network.

Belief networks use probability theory to manage uncertainty by explicitly representing the conditional dependencies between different knowledge components. This provides an intuitive graphical visualisation of the knowledge, including the interactions among the various sources of uncertainty.

TATEM’s Jon Dunsdon from Smiths Aerospace says: “You therefore have two different sets of input and start applying a probability-based approach where every decision is not a simple yes or no, but rather a likelihood scenario to give enhanced diagnostics. No-one else is doing this sort of work on aircraft at the moment.”

The TATEM project is also examining virtual techniques for gauging the health of engine accessories such as hydraulic actuators. “When you have an actuator that is not moving much, you should not be doing condition monitoring, but only when it’s working reasonably hard because the variability can be deceptive – the noise from the signal can produce an inaccurate result.”

Adopting a higher level system design, TATEM partners identified that they could calculate those parameters without needing sensors through introducing the concept of virtual sensing. “Using displacement and pressure as the two knowns we can calculate speed and chamber pressure to find out if the seal is degrading,” says Dunsdon.

Masse says that, now these monitoring techniques for engine components have been identified, they will be further developed and validated with tests planned for next year. The TATEM project involves 60 companies from Europe, Israel and Australia.