“And every one of them words rang true
…Pouring off of every page
Like it was written in my soul from me to you”
A Voice from the Past
It’s not every day I get to quote Bob Dylan in a post about process automation, but this is not just any day – we are celebrating AspenTech’s 40th anniversary, a time when the past, the present and the future are mixing together freely as we contemplate our achievements and chart our course forward.
The Dylan lines, from “Tangled Up in Blue,” describe a character’s encounter with “an Italian poet from the thirteenth century” (Dante), and his finding the words so meaningful, urgent and contemporary that he has an epiphany. I’m mentioning them now because I recently had a similar experience with a work from the past crashing into my present.
I was helping my wife clean her father’s study when we happened upon an old book, Petroleum Development and Technology in 1926, that had belonged to her grandfather while a student at the Colorado School of Mines. The engineer in me couldn’t resist flipping through the pages. But what I had thought would be an amusing curiosity instead hit me, repeatedly, with its practicality and prescience.
In fact, I’ve re-read one of the chapters, “Use of Automatic Control in Refining” by Luis de Florez, a few times. It’s uncanny. I’ve found it to be a meaningful touchstone as our team at AspenTech recalls our history, sizes up our accomplishments and considers the challenges and opportunities that lie ahead for us, our customers, and industry.
Ideas Ahead of Their Time
Let me share with you three ideas from the text that really resonated with me.
1. “When you start to analyze every operation that is carried out in a refinery, there isn’t one that cannot be improved if you know the point of control and how to control it.” (p. 446)
Wow, did Mr. de Florez foresee the broad role for control software, or what? But it’s even bigger than that. Nearly 100 years ago, this author, an engineer himself, sensed that behind the enormity of refinery operations lay an efficiency that could be achieved through automation.
At AspenTech, we’ve been on the front lines of this evolution, first targeting discrete processes that can be controlled and optimized to increase efficiencies and yield. Over time, we’ve also played a starring role in elevating and aggregating those individual processes, automating wider and wider swaths of functionality, supplemented by major advances in AI and machine learning, to lay the foundation for, and ultimately enable, the Self-Optimizing Plant.
Of course, our author from 1926 would be amazed by the size, scope, and power of today’s 24x7x365 operating juggernauts known as the modern refinery. But you know what? So would anyone. The first time I visited a refinery, my jaw dropped, and I had started my career in aerospace launching satellites!
2. “It is obvious that the smoothness of operations and the results obtained from the equipment are largely dependent on the individual operator and his state of mind, both uncertain factors. He is influenced by fatigue, weather conditions, personal affairs, and other extraneous circumstances. (p. 431)
“Automatic controls can never be substituted for human intelligence, which is necessary as a guiding factor in all operations. They can, however, carry out mechanical functions required in the maintenance of operations conditions far more accurately and with greater certainty by human agency. (p. 432)
Pretty compelling. I give de Florez credit for not only identifying the macro-opportunity for wide-scale, controls-rooted automation across the refinery, but for also envisioning the interplay, and potential conflict, between humans and machines.
We’re living this moment now. At AspenTech, we clearly see the role of automation and AI as supplemental, something working in concert with, and expanding the capabilities of, a human-run enterprise. But we also know that as knowledge workers in process industries age out, it is vital that their know-how and experience are retained.
Some of that is going to happen in the handoff to the new generation of process-industry workers, but no 25 year-old is going to know how an industrial process will perform in any given situation the way someone who has done it his or her entire working life will. That’s why some of the knowledge transfer is also going to have to be handed down, and to some extent handed over, to the software, control and AI systems that guide refinery and other process operations.
At AspenTech, we see today’s and tomorrow’s process automation solutions being “both/and” propositions, where humans and machines each need to perform vital roles to obtain the synergies that lead to game-changing efficiencies.
3. “As to the delay in adopting automatic controls, it is certainly not lack of intelligence on the part of the designers or operators which is at fault, but more the fact that this intelligence has not been applied in the right manner to the development of automatic controls. Successful development in this line distinctly requires intimate knowledge of both instruments and operations, and there has been and still is a lack of proper cooperation in these fields. (p. 445)
It was true 100 years ago, and it’s just as true today: Domain knowledge matters. In fact, it is essential. Process industries require process experts – not technology generalists – to automate controls.
At AspenTech, we pride ourselves on understanding the process first and automating it, or as our author says, “applying it right,” second. This is an act of hybrid dexterity. On the one hand, we have 40 years of deep, process-industry-specific knowledge and experience. On the other, we know how to marry it with general AI capabilities so that it works together seamlessly – and continues to improve.
A New Era for Process Automation
Petroleum Development and Technology in 1926 punches above its weight even in 2021, nearly 100 years after its publication. I’ve been delighted by that fact, but also curious to know why.
I think the prescience of de Florez’s insights underscores that the tectonic plates that form the process industries are really large. Control automation in the process industry hasn’t been quick or easy, because our industry is complex.
That complexity is a blessing and a curse. It can mean change sometimes comes more slowly than in other industries, but it also rewards deep understanding, and that has been the bedrock of AspenTech’s innovation. It is why it has taken our 40 years in business to develop automated control solutions with a widening envelope, make digital twins more accurate and easier to maintain, implement highly successful predictive maintenance technology, and begin to put the pieces in place for the Self-Optimizing Plant of the future.
But a funny thing happens over time when insights, experience and technological innovation steadily accrue: Breakthroughs begin to happen faster.
I see in the pharmaceutical industry’s rapid development of a COVID vaccine a guiding analogy for control automation in the process industry. When you’ve got the knowledge, the experience and the expertise, and a quickly evolving and improving set of technologies that include AI and machine learning, the ability to accelerate innovation and solve problems dramatically increases.
To me, that bodes very well for the future of our company as we face the dual challenge of scaling to the world’s demands – delivering the outcomes our customers crave – and doing so sustainably. That’s a twist even my new favorite author couldn’t have predicted.
Happy 40th, AspenTech, and here’s to 40 more at fast forward!