By Eugene le Roux, FSAIRAC, and Eamonn Ryan
In the complex field of industrial production, where machinery hums and processes flow, the spectre of equipment failure looms large.
Eugene le Roux. © RACA Journal
While the exact moment a machine will falter remains elusive, it’s often possible to quantify the probability of its breakdown over a given period. The application and operating conditions of a machine significantly influence its reliability, making the approach to maintenance a critical strategic decision.
At one end of the maintenance spectrum lies the reactive approach: waiting for a failure to occur before acting.This ‘run-to-failure’ strategy, while seemingly simple, carries significant risks. When a machine breaks down, it often causes consequential damage, leading to more extensive repairs and increased costs. More critically, in production environments where processes are interdependent, the failure of a single machine can halt an entire production line, resulting in costly downtime, missed deadlines and a substantial loss of revenue.The inconvenience of unplanned downtime can far outweigh the perceived savings of delaying maintenance.
Moving beyond reactivity, predictive maintenance emerges as a more strategic alternative. This approach suggests replacing parts on all machines when the predicted probability of failure for those components reaches a predetermined threshold. The underlying philosophy is to proactively address potential failures before they materialise, thus preventing costly disruptions to production.
While replacing parts before they demonstrably fail might seem expensive on the surface, the broader economic context often favours this strategy. In income-sensitive production environments, the luxury of uninterrupted operations can easily justify the cost of preemptive part replacement. By adhering to statistical predictions of component lifespan, businesses aim to minimise unexpected shutdowns, ensuring a smoother, more reliable production flow.This approach shifts the focus from managing breakdowns to managing risk based on anticipated wear and tear.
Condition-based maintenance: monitoring for early warnings
Is there an even smarter way? Enter condition-based maintenance (CBM), an approach that offers a more nuanced and often more cost-effective solution than purely time-based predictive methods. CBM involves continuously monitoring the condition of specific parts or subsystems using various diagnostic techniques. This could include:
- Vibration analysis: Detecting unusual vibrations that indicate bearing wear or imbalance
- Flow measurement: Monitoring resistance to flow in fluid systems, hinting at blockages or pump issues
- Ultrasonic crack testing: Identifying hidden structural flaws before they propagate
- Infrared thermography: Detecting overheating in electrical components or mechanical parts
- Pressure and temperature monitoring: Spotting deviations from normal operating parameters
- Visual inspections: Observing physical wear, leaks, or damage
By collecting and analysing this data, maintenance teams can receive early warnings of impending failures, providing sufficient time to continue production until a convenient window for maintenance opens.This allows for scheduled interventions, minimising sudden disruptions.The acceptable deviation from normal system behaviour for each parameter can be meticulously set by weighing the potential impact of failure against the remaining time before it’s likely to occur. In the absence of formal deviation values, starting with heuristic (rule-of-thumb) values and refining them over time through experience and data analysis is a pragmatic approach.
The monitoring process in CBM can be significantly enhanced through automation, with data observed and analysed at a central point. This integration allows for a holistic view of the plant’s health, combining probabilities from multiple interdependent machines. Such an integrated approach can lead to substantial savings in both cost and time, as maintenance efforts are precisely targeted where and when they are most needed. This intelligent, data-driven strategy is often referred to as preventative maintenance (though some definitions distinguish CBM as a subset of preventative maintenance, or a more advanced form).
Beyond addressing specific component conditions, a truly robust maintenance strategy should also incorporate periodic holistic checks of the entire plant. While individual component monitoring might flag impending failures, systemic inefficiencies can creep in over time without triggering specific failure alerts. Control parameters might drift, or inter- machine dependencies might lead to suboptimal performance without a clear breakdown. Regular, comprehensive assessments ensure that the entire plant remains ‘tuned’ optimally, operating at peak efficiency rather than merely avoiding catastrophic failure. This proactive tuning can uncover creeping inefficiencies, preventing subtle but significant performance degradations that might otherwise go unnoticed, ultimately ensuring sustained productivity and profitability.
In conclusion, effective maintenance is not a one-size-fits-all solution. By understanding the nuances of reactive, predictive and condition-based approaches, and by embracing holistic checks and automation, industrial facilities can move beyond simply reacting to breakdowns. They can instead cultivate a proactive, data-driven maintenance culture that prioritises uptime, optimises costs and sustains peak operational performance.
