Artist's vision of the Huygens probe landing on Titan. Image courtesy of NASA's Jet Propulsion Laboratory.By Gawayne Mahboubian-Jones, Director of Engineering, Optimal Industrial AutomationIn mid-January, the Huygens probe completed its long journey towards Titan and transmitted back to exultant waiting scientists a wealth of data â€¦ data which are still being screened, scanned and analyzed. The project aims to extract the last ounce of information, to ensure that our new understanding of one of the strangest places in our solar system is as extensive as practical. It will also allow us to build detailed models and hypotheses about processes occurring on and below the surface, and to allow us to formulate more accurately our next set of questions and to design the associated experiments.The parallels with PAT are really quite striking!Like the surface of Titan, our processes are poorly understood, and often only described by their results. We rely upon trial and error, and upon “experience” to determine our operational conditions.Like the surface of Titan, true understanding has been masked by our inability to measure and observe any of the really important parameters where it matters — real-time and on-site. And, as with Titan, with PAT we are now moving into an era in which, by putting appropriate instruments in close proximity to that which we barely appreciate, we will revolutionize our understanding.But there are other, equally important lessons from Huygens which we have yet to assimilate — lessons which we ignore at our peril:
- New worlds require a root and branch review of our measurement techniques.
The instruments on Huygens were custom-designed to ensure that they could operate reliably after the rigors of atmospheric re-entry, landing, and an environmental temperature 200Â°C colder than an average laboratory.
We persist in trying to use instruments designed for an air-conditioned, vibration-free laboratory in a busy industrial plant.
We persist in accepting laboratory levels of reliability that have no place in a manufacturing environment.
We persist in accepting complex manual set-up procedures when all PAT instruments should be automatically self-testing and self-calibrating.
- We have yet to understand the crucial role of data at the center of PAT.
Huygens was designed and built to squeeze the last once of information from its data, and the last ounce of data from its bandwidth.
We persist in using manual approaches to data collection and analysis; the volumes of data we will soon be generating clearly indicate that anything other than the most transparent, automated collection will rapidly become a bottleneck, crippling our ability to control our processes.
- As Huygens scientists learned by necessity: Automate all possible operations.
When signals take minutes to reach their destination, remote control is both tedious and very risky. Our corollary is that operations reliant upon human intervention, however well-prepared the SOP, are also of high risk. We, too, must minimize them if we are to reap the best fruits of PAT.
- Finally, we need to understand that our instruments are driven by the environment in which they operate, not vice-versa.
For Huygens, this is expressed in terms of the response times needed to make measurements before the batteries fail, and by the power limitations imposed by that same energy source.
This lesson sits equally uncomfortably with many of our current measurement techniques.
We must have instruments which enable us to make measurements in â€˜real-time’ (whatever that is for a particular process).
We must have instruments and methodologies that do not change the process when they are applied â€¦ stopping a process changes it, often irrevocably – stopping a process for a measurement must become a thing of the past.