Maintenance in times of COVID and no COVID

Let us start with some imagination. Suppose you go to the plant and everything is normal. People in their day-to-day tasks, no one wears masks, and no one even talks about COVID-19, because it never existed.

As everything is normal, you go back to your office as Head of Maintenance or Plant Manager. You come back because your plant needs you on your usual desk, to receive the usual calls and emails, and then fix the usual problems.

Is this kind of game worth it in the economy proposed by the 21st century? No, and it is not COVID's fault.

In this imaginary world in which COVID did not exist, if there were still other facts that are as verifiable as the virus itself:

• A trade balance for Latin America that has a predominantly declining trend (except for slight skirmishes) whose behavior has been going on since 2011. This, in other words, is why we buy more and more from the world than the world buys from us. In fact, since 2013, this balance has been predominantly negative.
• A participation in the exploration of technologies such as Artificial Intelligence applied to industrial business, which does not manage to be even 1% of the participation that Europe or North America has.
• An integration of robots into our industrial production media that fails to reach 0.1% of the world's total industrial-purpose robot population. Africa, by comparison, already reaches 1.7%.

These many examples outline a clear picture: Latin America's industrial dynamics require less normality, and more disruption. While from a humanitarian point of view, COVID-19 is a catastrophe of incalculable humanitarian costs, perhaps that harsh price we are paying, it will serve as the bucket of cold water that motivates us to think differently.

And not because "COVID", but because we are long ago in arrears of integrating, in a determined and sufficient scale, new ideas in the way we think and decide on our industrial processes and our maintenance.

Maintenance solutions in te 21st century

Today, there are already fully functional applications in various industries where the following scenarios are possible.

• An "intelligent agent", which previously consumed and processed hundreds of work orders and millions of measurements, now receives online signals from a plant's control system. This agent has statistically confirmed his ability to generate, 20 or 30 days in advance, an alarm indicating which failure mode will be presented, which computer, and even for which cause.
• Another agent, who was trained differently, consumed millions of available data regarding how the natural variation of the plant behaved at those historical times where there was no active failure. This agent, unlike the previous one, cannot predict a specific failure, but having been trained with the date of a normal condition plant, it may warn, also days or weeks in advance, when an unidentified or previously unknown failure will occur.

• The intelligent agent system, having detected the possible failure condition at the indicated anticipation, are connected to the RAM Modeling and Simulation Systems. They are informed of when a likely failure is expected. The simulators, on the other hand, determine the likely business losses if nothing is done (base case for the detected event), and run multiple fault management alternatives (lower load, change plant configuration, modified production schedule). As a result of the exploration they deliver quantitative estimates of business value that when compared to the base case, allow maintenance managers to make the best decision for the failure that could be presented in 20 days.

I do not quote this example anecdotally, nor as an introduction to a science fiction novel. I am referencing technological integrations that we have achieved in the real world with real technology, and that enable these benefits, as real as failures can be.

What about the COVID?

The industry's enemy is not COVID, simply because it is not acquired by the pump or compressor. If our industrial process had reliability analyses as a likely resource to manage the plant (resources have been available at least since 1978) if our signals were available on field sensors and then were historized (in the 80's); if we access in a timely manner some of the RAM Modeling and Simulation solutions (also available since the 1980s) and Artificial Intelligence (these are available since 2012); if we had done all this as Latin American industrial companies, recognized throughout the solar system as the companies that know their assets the most and that go to the forefront of industrial knowledge, then the management of a pandemic would already be given.

It would be a matter of sending to the minimum number of people, at the right time (estimated by our A.I. solution) to obtain maximum economic benefit (confirmed by our modeling and simulation systems).

It is clear how 'idealistic' this panorama looks, and of course it must look like this from the perspective of a professional belonging to a technology consumer ecosystem mostly developed elsewhere. That is precisely the perception to challenge in Latin America that we inhabit, because surely this same discourse within SpaceX sounds prehistoric.