The best solution for solving COVID-19 may be through the assistance of artificial intelligence (AI). This pandemic will have amazing financial implications and a significant loss of human life. COVID-19 is going to have a long-lasting impact on global economic conditions in virtually every country. This pandemic will hopefully be unnoticeable to our young children, but many adults later in life with pre-existing conditions are at risk. AI is one method that may unlock a key to the solution to both vaccines and methods to treat the virus in the seriously ill. Complex mathematics are employed to run millions of equations instantaneously and may help us solve this life-threatening virus.
Yet, how are we not employing AI and machine learning widespread in daily decision-making in the process industry globally?
I suspect the reason is that all of this relatively new and complex. There is a myriad of directions to turn. Should you invest with a supplier that can provide complete plant industrial internet of things (IIoT) solutions, or do you first start to focus on the most critical assets? How large is the budget and how quickly do you want to be there? Where do you want to be and what do you want to accomplish? What are the potential savings and or benefits? The merger of the physical and digital universes is occurring at lightning speed. In 1995, there were approximately 1 million devices connected to the internet of things (IoT). By 2015, there were already more than 10 billion devices connected and by 2025, it is estimated there will be more than 100 billion. This was all made possible by super-cheap and more efficient computer chips.
The advantages of IIoT are agility and efficiency. Products of all types can be equipped with instrumentation that reports back the condition of that asset. Information such as the likely prediction of when that asset will fail can be reported back and that asset can be swapped out before more extensive damage or downtime occurs, and the process can continue without interruption. Both processes and supply chains can be optimized because operators have accurate data about what really is going on.
One of the biggest challenges large manufacturing companies need to overcome when jumping into the digital world, is the hard fact that their key products or equipment were never considered to be digital. They might have a certain level of automation, but this might not suffice the requirements to go on a digitalization journey and, worse, the whole process could become a nightmare with large costs to the company. This is not the only hurdle—we could mention the lack of in-house expertise to develop these kinds of solutions, or the lack of agreement on which road to follow to make a successful transformation. However, understanding this fact is a key first step.
Did COVID-19 Reveal the Need for Digitalization, Machine Learning, AI & IIoT?
I am quite certain that some individuals wish their processes were more visual during this pandemic. Digitalization will continue to grow within the process industry. Pumps & Systems focuses on pumps. So, let’s talk a little about pump monitoring. The natural point on pumps to monitor is bearing condition. There is money that will be saved, processes that will continue to run without interruption and bearing problems will be noticed ahead of failures. But bearing monitoring is only the tip of the iceberg. Adding additional instrumentation to and around the pumps can help generate a more complete picture of pump health and potential problems.
Embedding manufacturer’s pump curves into an IIoT monitoring package and comparing actual performance to the pump curve can help users better manage their pumps. Keeping pumps operating on or close to the manufacturer’s pump curve may save impressively on energy consumption. Also, adding pump best efficiency point (BEP) monitoring can help users run their pumps closer to the BEP and minimize additional unnecessary wear and tear on their valuable assets. IIoT monitoring opens the door to brand-versus-brand analysis on which brand is the best for their process. Energy consumption, efficiency and downtime will be evident to the owner to help them make the best pump selection based on total cost of ownership.
Other inputs that may be beneficial to overall pump performance monitoring is discharge flow rate produced, pressure on the suctions side and discharge of the pump, motor energy consumption, motor and gear box vibration, gear box lubricant temperature and connection to a variable frequency drive (VFD) if one is installed can produce a tremendous amount of valuable data. With additional instrumentation, other valuable information can be captured such as:
- pump impeller wear
- other component wear
- pipe blockages on pump suction and discharge
- conditions not meeting net positive suction head (NPSH)
- poor piping design
- clogged filters
- flow changes
- increased energy consumption
- medium temperature changes
- gear box lubricant temperature escalations
- acoustic emissions
- bearing vibration spectrum
- motor and gear box vibration
- cavitation
Of course, we talked specifically about pumps, but there are massive opportunities for monitoring other critical assets. For example, conveyors, ball mills, thickeners, clarifiers, filters, screening equipment and flotation columns are just a few examples of assets that are natural targets for digitalization. Some of these assets, such as a filter or a thickener, act as a process within the process. Some are already equipped with significant digital instrumentation. These process machines may be ripe for integration of AI. Over time, based on numerous inputs and variables to these process machines, AI may help to produce a likely outcome or result that can be predicted. As process upsets occur, downstream processes can be informed of the upset and likely outcome. Plant personnel can take corrective actions to improve results after the upset occurs.
Digitalization and further automation are going to continue to automate processes globally. Companies that are embracing IIoT strategies are performing in the top 25 percent compared to their peers. Some companies are claiming to be gaining as much as one additional month of production per year due to digitalization. I recommend trying several solutions. There are numerous solutions from many companies—try them and see which one has the best fit for your organization.