Working in the oil and gas industry, risks surface constantly — much like the commodity these companies are sourcing. As oil, or petroleum, is such a volatile substance, it can pose challenges in the production and refinery processes for machines and engineers.
How did engineers assess risks in the past in order to overcome such pesky difficulties?
The process primarily involved in maintaining an optimal condition after failure has occurred is corrective maintenance. Bear in mind that the traditional method in the past was rather straightforward and simple: fix where it is broken. Assessing what has been affected by the damage, investigating what the root causes were, and how to get the operation to resume again were the key methodology behind this.
We have not taken the fact that the production process was paused into account; however long it was, profit was not being generated. Fixing the broken parts was not going to sustain any business model that wished to grow.
A gradual improvement was discovered and time-based maintenance was an upgraded option over the corrective maintenance. Routine checks performed by engineers helped in the reduction of mechanical failure as they inspect for potential deterioration in mechanical parts. However, should unplanned failure occur out of the blue, time-based maintenance may not be adequate to solve such issues.
Taking numerical readings as variables to calculate the possibility that a failure would arise is another improvement on the prior process. This performance monitoring maintenance takes records of the past to determine, by calculation when an error would occur. This method is contingent on the engineer’s skillset and their accuracy in the assessment. Any miscalculation could see the possibility of failure occur. This form of maintenance counts on humans to be as error-free as possible but humans are not error-free nor are they available around the clock.
On the contrary, what is available twenty-four-seven is computer software. They operate continuously to execute jobs, instructed by humans. And this brings us to the modern methodology where software is heavily relied upon. Although, a small remark to note is that these softwares are specifically designed for specific use with advanced computational functions to support heavy-duty engineering calculation — which means these had to be customized, and result in high investment, not to mention the infrastructure that is required to be installed to facilitate the entire operation.
However, due to the complex nature, the software is capable of running all the time unlike humans who need breaks, annual leaves, mental stress and other factors which do not affect the software. To elevate this point, we now have artificial intelligence which runs almost on its own, executing commands per instructed to detect failure.
The software’s job is to transform data, transferred from physical components, into information which allows it to get ahead of the curve before the slightest sign of failure appears. This predictive form of maintenance is regarded as statistics analytics maintenance.
On top of the previous advantage, it requires less input from humans, to avoid any miscalculation, to perform quicker calculations and deliver faster output, in the form of information which requires experts to break down into reports.
Should any anomalies be detected, engineers are enabled to detect a possible cause for the malfunction before it even happens — eluding catastrophic breakdown that leads to loss and time wasted.
Discover reasons why our predictive statistics analytics maintenance solutions could be suitable for your business.
From past to present, we take a look at how engineers assessed risks based on different models of maintenance from corrective to predictive methods.
Learn more →
Unplanned shutdown in oil production is a costly disaster but it could happen. The question is: how can it be prevented?
Learn more →