Understanding Predictive Maintenance As a Service
Asset maintenance in discreet and process manufacturing is a key priority – one that can have a multiplying effect on plant reliability and bottom line. For industries where even a single hour of production downtime due to asset unavailability can translate into millions of dollars of losses, strategic investments are made to make maintenance more effective. Maintenance management has thus, evolved into a complex field with application of varied digital and analytics solutions to improve efficacy while reducing costs.
Predictive maintenance is a result of this pragmatic evolution and has helped manufacturers save billions of dollars through increased Remaining Useful Life (RUL) of assets and minimized production downtime. By 2025, the predictive maintenance market is expected to grow to USD 25 billion, creating unparalleled value across global industrial value-chains. This article will cover what predictive maintenance is, how it has evolved to suit industry needs, and how it applies in various manufacturing setups.
What is Predictive Maintenance?
Predictive maintenance (PdM) is a maintenance approach which entails use of cloud-enabled technologies to monitor diverse assets involved in production and estimate the maintenance needs on the basis of asset condition and detected anomalies. Condition Based Monitoring (CBM) is the underlying concept, which is used to monitor industrial assets in real time. Technologies such as infrared thermography, vibration analysis, oil analysis, and acoustic monitoring are deployed to collect equipment condition data.
The data thus collected is analysed to detect any deviation in asset performance or irregularities which are otherwise impossible to detect without sophisticated equipment. But predictive maintenance doesn’t just stop at detection of performance issues. With the help of edge diagnostics and predictive analytics, the underlying causes behind poor asset performance are determined. Predictive maintenance solutions can also forecast the time when a monitored equipment will breakdown, if corrective maintenance measures are not taken.
Historically, industrial maintenance has evolved from being reactive to predictive.
This evolution has been driven by an increased focus on Overall Equipment Effectiveness (OEE) and the need to reduce unplanned production downtime. While manufacturing plants no longer rely on a reactive maintenance approach, preventive or scheduled maintenance is the most popular among undigitized setups. Since a preventive approach can deliver 50%-75% OEE on an average, maintenance leaders hesitate to shift away from it towards a predictive maintenance approach. However, understanding the key differences between both practices can help.
How is Predictive Maintenance Different from Preventive Maintenance?
- Preventive maintenance is essentially a periodic and schedule-based maintenance approach that uses asset performance history and routine equipment inspections to plan maintenance events.
- Irrespective of the machine condition and the remaining useful life of the asset, maintenance and replacement activity is performed.
- Preventive maintenance is usually labour-intensive and not very cost-efficient. Maintenance teams have to maintain inventory of equipment spare parts and schedule planned downtimes to execute maintenance.
- The net cost of asset is increased as over the period of its usage, several parts are replaced and consumables are used to keep the asset running optimally.
- Latent problems that may develop in-between inspection schedules could go un-noticed and remain un-addressed. As a result, the dependence on reactive maintenance increases.
- Predictive maintenance relies on continual monitoring of equipment condition through sophisticated technologies that are IoT-enabled.
- Maintenance activity is highly targeted in predictive maintenance and only performed when a fault or anomaly is detected. Routinely organized production downtimes and pre-emptive part replacements are not required.
- PdM is increasingly efficient over time and reduces the labour-intensive nature of maintenance activities. Predictive maintenance is also more cost-efficient, reducing the inventory carrying burden for spare parts.
- Net cost of assets is effectively reduced as the complete useful life of every machine is utilized before any replacement is done.
- The likelihood of unplanned reactive maintenance is reduced by 70-90% with a predictive approach. Since equipment faults are diagnosed well before the actual functional failure, organized maintenance events can help avoid process disruption and reactive efforts.
- Critical equipment problems and underlying causes of diminishing productivity can be accurately identified with predictive maintenance. As a result, a higher network system reliability can be achieved in a cost-efficient manner.
To better understand how both the approaches can have a distinctive impact on the maintenance organization, here’s an example. A single high-pressure roller press can support the production of a 2MTPA cement grinding plant. If it breaks down due to undetected anomalies or fault in its planetary gearbox, it can result in a potential downtime of 24-48 hours. Now, preventive maintenance measures allow plant maintenance staff to schedule regular asset maintenance events, irrespective of the machine condition and abnormalities which may be affecting productivity.
On the other hand, with a predictive maintenance approach, not only can such anomalies be detected in advance, but such critical events can also be avoided with prescriptive measures. (Read the full case here.)
To begin with, all assets available in a production environment must be evaluated to determine how mission critical they are and whether they should be proactively maintained. Once the appropriate machine groups are identified, they can then be outfitted with the cloud-enabled predictive maintenance solutions. These solutions could offer edge-diagnostics with the help of vibration analysis and condition monitoring, but before fault diagnosis can happen, it is important to define performance parameters for assets being monitored.
These performance parameters can include baseline metrics which would be considered as the norm, and continual monitoring would not lead to unnecessary false positive notifications. The assets can then be allowed to operate as per their pre-defined schedule and requirement. Meanwhile, real-time performance data and machine health indicators can be collected and stored on cloud. Advanced predictive analytics can diagnose data anomalies and generate insights regarding potential faults in the monitored assets.
These insights can empower maintenance teams to schedule maintenance activities with minimal production downtime and costs.
In sum, predictive maintenance as a service has evolved with the core objective of making industrial asset management more efficient and less costly. Data accessibility and intelligent analytics have shaped the service into a powerful enabler that allows maintenance and operations teams drive plant reliability objectives. While predictive maintenance adoption has to be strategic and supported by change management, its benefits are quickly felt and undeniable for even the most complex production units.
Infinite Uptime offers responsively designed predictive maintenance solutions in diverse industries such as Cement, Steel, Mining and Metals, Tire, Paper, Automotive, Chemicals, FMCG, Oil and Gas, and more. To understand how predictive maintenance applies to your process plant, explore the digital reliability [PS1] and asset management solutions of Infinite Uptime.