Introduction: Keeping a complex fleet of assets up and running
When thinking about the steep costs of asset downtime, it quickly becomes apparent how critical artificial intelligence (AI)-enabled asset performance management (APM) systems are for ensuring industrial asset performance. Whether your business is energy, manufacturing, shipping, or defense, keeping expensive assets up and running is key to driving competitiveness and long-term financial performance: assets don’t generate revenue when they’re down.
By integrating AI into daily processes, organizations can take maximum advantage of AI’s breakthrough abilities to extract significantly greater operational value from the deluge of data produced each day by industrial operations using AI-powered solutions like normal behavior modeling (NBM), natural language processing (NLP), and computational knowledge graphs (CKGs).
AI-enabled asset performance management (APM) is predicated on two important capabilities: prediction and prescription. The former provides proactive insights into when failures are about to take place and when, while the latter tells operators what they can do to either prevent the failure or best manage their response to an actual failure to minimize process disruptions along the way. Both of these important capabilities are, however, built upon a critical foundation, i.e., the fact that assets do not operate throughout their lifetimes the same as they did the day they left the factory. Assets age, parts wear out, maintenance takes place, hostile operating environments take their toll. This means that maintenance procedures provided by OEMs for new equipment typically do not ensure long-term performance that maximizes the operating lifetime of assets which have lifespans typically measured in decades. That’s where AI-enabled predictive/prescriptive maintenance enters the picture.
Predicting the future of asset performance
Predictive maintenance uses AI and real-time system performance data to identify the likelihood of an asset failing, the specific nature of the failure, and the estimated time until the failure occurs. This allows maintenance to be conducted precisely when it’s needed, just before a breakdown, avoiding unnecessarily premature servicing and preventing catastrophic failures, with all their expense, productivity, and safety implications. To enable predictive maintenance, AI models are fed a continuous stream of performance data from sensors on critical assets like pumps, turbines, and motors. This data can include parameters like:
- Current/power consumption, operating voltages
- Temperatures, pressures, flow rates
- Vibration (for rotating equipment like motors and pumps)
- Unusual acoustic signatures
- Historical maintenance logs, images/videos
Machine learning (ML) algorithms analyze this real-time data to determine “normal” baseline operating conditions—conditions that vary significantly as equipment grows older and is used more. The ML model then identifies anomalies—subtle drifts, fluctuations, or irregularities—invisible to the human eye but indicative of the early onset of failure.
Finally, advanced AI techniques like Normal Behavior Modeling (NBM) process these patterns and apply prior knowledge from historical failure cases to understand the sequence of events leading up to the forthcoming breakdown and calculate the probability of failure as well as the remaining useful life (RUL) of a system or component. This capability provides an accurate warning of asset failure, typically weeks or even months in advance.
Why does all of this matter?
Every asset-intensive industry—which is to say, nearly every industry—is charged with generating profits from the assets they employ. That means getting the most from each asset while it’s running and, as importantly, keeping these assets in service for as long as possible to forego the profit-sapping downtime associated with failures and the capital costs of asset replacement.
To choose a specific example, approximately 70% of offshore oil and gas (O&G) facilities are over 15 years old. Obviously, older equipment needs more maintenance and attention to ensure efficient performance than newer equipment. There is, though, a secondary problem—a somewhat more subtle one—that’s a factor in this analysis. In the O&G industry, nearly 50% of the workforce is expected to retire in the coming five to ten years. As they move on, there’s concern across the industry about the institutional knowledge and subject matter expertise they will take with them. So, as equipment ages, and experts who know how to keep critical assets operating at peak performance age out of the workforce, it’s even more important than ever that expensive assets operate as reliably and efficiently as possible. This same argument applies with equal validity to utility assets like generation and T&D systems; manufacturing factory equipment; and transportation assets like aircraft, ships, trains, and trucks.
What’s the expected outcome of implementing predictive maintenance?
The adoption of AI-powered predictive/prescriptive maintenance processes delivers significant business benefits:
- Reduced Unplanned Downtime: By anticipating and proactively managing asset failures, organizations can do a better job of scheduling maintenance, driving significant reductions in unplanned downtime.
- Reduced Maintenance Costs: Servicing assets when failures are imminent eliminates the costs (materials, hours, and lost output) of unnecessary, time-based maintenance.
- More Efficient Integration of Operations: Initiating a work order for spare parts procurement (or move if it’s in inventory), crew work order and scheduling, etc. optimizes resource timing and automates work order planning and execution.
- Maximized Asset Lifespans: Proactive/targeted repairs reduce unnecessary strain on assets, extending lifetimes and reducing capital replacement costs.
- Optimized Resources: Maintenance crews and spare parts inventory are managed based on actual, predicted needs rather than fixed schedules or one-off emergency repairs.
- Enhanced Worker Safety: Predicting asset failures minimizes risks to staff members working alongside equipment.
- Better Product Quality: Identifying manufacturing asset failures proactively minimizes the likelihood of defective products making it out the door.
Avathon’s award-winning AI technology, deep domain expertise, and product engineering vision has been developed over 10+ years of real-world engagements serving the largest brands in oil & gas, manufacturing, transportation, and utilities. The patented technology inside Avathon’s Autonomy Platform has yielded successful outcomes in numerous past deployments, including:
- Generating 10x ROI in O&G production efficiency across multibillion-dollar operations
- Averting more than $100K in potential asset downtime costs per predicted failure event for an O&G supermajor
- Delivering a 5% production efficiency gain for a top 50 beverage manufacturer
- Avoiding over $1M in maintenance costs in a single plant deployment for a global food & beverage manufacturer
Conclusion : Predictive Maintenance Drives Optimized Performance
Asset-intensive industries rely on uninterrupted performance to ensure the financial health of their operations and organizations. Failures—particularly unplanned ones—reduce product outputs, compromise quality, put employees at risk, and drive down financial performance. Predictive/prescriptive maintenance helps industry managers and executives keep their assets operating efficiently and economically, maximizing asset lifetimes, and extracting the greatest value from their fleets.
Predictive/prescriptive maintenance technology has advanced rapidly in recent years, and delivery of these performance improvement results is now happening regularly at industry-leading firms, enabling them to identify impending equipment issues and prevent asset downtime. In fact, 36% of industrial asset operators have realized savings on unplanned asset downtime by using predictive maintenance approaches. To put that figure in more concrete financial terms: for a large manufacturing plant where unplanned downtime costs an average of $250,000 per hour, a 36% reduction in such incidents translates to tens of millions of dollars in annual bottom-line savings. A recent “Industry 4.0” analysis by Deloitte indicates that predictive maintenance can increase equipment uptime by 10–20% while reducing overall maintenance costs by 5–10%.
Predictive/prescriptive maintenance is now a widely recognized AI capability that enables industrial operators to squeeze as much life as possible out of the installed equipment that’s producing for them, maximizing ROI on a firm’s asset base while helping to ensure operating efficiency, product quality, and worker safety.
To learn more about Avathon’s predictive/prescriptive maintenance solutions, visit our website.
