The evolving capabilities of health monitoring systems and the machine learning algorithms that process their data mean more sensor inputs than ever before can be meaningfully analyzed.
However, this does not necessarily mean that engines will pack in ever more sensors, says David Milne, senior director program management, engines, Honeywell Aerospace Technologies, who tells Inside MRO that advances in software can potentially reduce the need for hardware.
“In some cases we have been able to take advantage of sophisticated algorithms and analytics to synthesize parameters, rather than measuring them – thereby eliminating sensors,” he says.
This may help to mitigate any dips in supply, with Milne noting that sensors like many other aircraft systems, have been exposed to supply chain problems.
“Sensors that have integrated electronics exposed to the same challenges regarding availability of foreign-sourced semiconductors,” he says, adding: “The CHIPS act from the US Government is addressing this in a positive way, and has helped Honeywell expand and modernize our electronics manufacturing capability… vertical integration gives us more control of our destiny.”
Software can also improve accuracy Milne says, which is important given that it is difficult to achieve big gains in this respect from the engine sensors themselves, which are a relatively mature and proven technologies.
“With the substantial computing power in our modern FADECs, we can perform sophisticated calibration and linearization to improve accuracy. We can also employ sensor fusion of not only multiple sensors, but also synthetic parameters that are generated by on-board algorithms and models.”
