The current situation with COVID-19 is forcing production operations to function with minimal staff onsite, so many companies are looking to improve their monitoring capabilities. With fewer eyes actually on the process these days, operators need the earliest possible warning that a process is drifting or a machine is failing. They need guidance — real guidance — on actions to take to mitigate the problem.

 

Safety is always a top priority for these companies, and today’s challenges make it a bigger concern than ever. Operating companies have been shouldering more risk, as they are forced to pare back insurance coverage, and in some cases, the cost of their insurance coverage has doubled. Those increases come on the heels of a dramatic rise in the rate of accidents in refining (a 4x increase over the rates in 2015, with more than 2,000 total incidents recorded in 2019).

 

As a result of all this, operating companies are looking to technology to help mitigate risks. It’s a strategy that could pay double dividends: reduced insurance premiums and improved margins. Predictive analytics for machine failures provides adequate warning to invoke workflows to mitigate the safety and environmental risks. At the same time, the software enables collaboration with planning and scheduling functions to help mitigate the increase in costs.

 

Companies that have adopted this technology are using the longer warning period to change how they respond to potential downtime. They have the time to choose when to do the maintenance based on the overall economic projections, which consider the operating, maintenance and supply chain implications of any potential asset downtime. 

 

Multivariate analysis — specifically, multivariate statistical process control (MSPC) — can be quickly deployed to provide robust monitoring of batch and/or continuous processes for drift or disturbances. The most important benefit of MSPC is it reduces the information contained within all the individual process variables down to two or three composite metrics.

 

Those metrics can be monitored easily in real time to track the efficiency of the process and quickly identify problems. Workflow and procedural automation become strategically important to orchestrate to interoperability between predictive maintenance and other enterprise systems like planning and scheduling.

 

We’re in uncharted waters now, and the only certainty is that life will be different in this volatile, uncertain, complex and ambiguous (VUCA) world. Predictive and prescriptive technologies will play a central role as we do what we do best — come together in times of mutual need.

 

 

Learn more about the potential of machine learning and predictive analytics in our executive brief Maximize Safety, Sustainability and Productivity by Turning Unplanned Downtime Into Planned Downtime.