Intelligence in Compressed Air Monitoring, Part 3 – where’s the intelligence?

This article is part 3 of Automate’s Intelligence in Compressed Air Monitoring series.

Read part 1 here
Read part 2 here

The previous two parts emphasized on the need to monitor compressed air systems, and components and practices that lead to compressed air leaks. This article expands that discussion to cover how monitoring activities work and what intelligent solutions bring to the table.

Once the sources of air leaks and pressure drops are known, the simplest tools to detect air leaks are human ears. Any operation/maintenance personnel can pick up major leaks across running systems just by paying attention to the sounds around them as they move about their plants. Ignoring leaks of that magnitude is a fatal mistake no personnel should ever be guilty of.

Audible leaks are a double-edged sword for staff that isn’t trained or equipped for handling compressed air system leaks. Since these leaks are so obviously present, they create a false sense of accomplishment upon rectification as they hide the presence of potentially thousands of smaller leaks that might collectively be causing much more damage than those few large leaks over a longer period.

The traditional way to deal with these leaks then is to perform periodic system audits. These audits can vary from simple walk-the-line inspections where major flaws and leaks are identified and rectified, to full audits with flowmeters, leak detectors and other analyses-ready equipment for a thorough analysis of the entire compressed air system. The best plants have a continuous improvement program through which periodic audits are performed either internally or with help from external subject-matter-experts, these audits are used for a long-term evaluation of plant performance while plant staff are trained and encouraged to perform individual walk-the-line inspections of their systems each time they’re on duty.

While this approach sounds great, any manager who has had to implement a continuous improvement regime can attest to the myriad issues that arise during such large-scale behavioral and business process overhauls. Enter intelligent monitoring: a simple and cost-effective solution to quickly address plant and equipment concerns with minimal dependence on personnel. A simplest monitoring solution is one that just monitors the flow from a compressor. This flow is compared to an ideal value based on a first-principles model of the entire section this compressor feeds. Differences in recorded flow during various processes and load cycles give a very accurate estimate of the amount of compressed air being leaked across each compressor train.

Adding just a few more demand-side flowmeters can enhance the capabilities of this plant model significantly by providing exact locations of leaks and areas to monitor. This monitoring is further enhanced by predictive analytics that estimate the rate of plant performance degradation, potential damage to equipment and components, and even compressor health status. All that with just a few flowmeters and state of the art analytics software.

Perhaps the most useful feature of monitoring and collecting data is the potential to have an AI-enabled learning system to monitor performance signatures in plants and, over time, accurately predict all aspects of installed systems. Imagine having access to information on leak locations and nature of leaks, component health, future leak locations, and eventually overall plant performance including failing components, compressor maintenance and process optimization.

The next part in this series will cover how all this information is utilized to give meaningful insights and actionable data to key decision makers.