Moving from Preventive to Predictive Maintenance in IIoT Projects

Preventive to Predictive Maintenance in IIoT Projects

Over the past few years, a positive effect of IoT on the equipment maintenance has been much praised: businesses from different industries have been striving to benefit from maintenance analytics offered by connected devices. Thus, the notion of “predictive maintenance”, leveraging the IoT potential to the full, has been much discussed as opposed to “preventive maintenance”. So what is predictive maintenance & what new benefits is it bringing to IIoT-driven industries?

Predictive vs. Preventive Maintenance in a Nutshell

Simply put, preventive maintenance (PM) implies performing scheduled maintenance set in accordance with previous statistics of the equipment functioning, while predictive maintenance (PdM) is determined by the condition of equipment monitored during normal operations.

For PM to be efficient, analysts should possess both estimates about the equipment typical functioning, its age and operation hours, and in-depth knowledge and analysis of each individual piece of equipment, to make sound assumptions when and what exactly will fail. Thus, this approach relies heavily on the skills of maintenance engineers and accuracy of their work.

For PdM to be efficient, it needs smart application of advanced prediction models and algorithms to analyze data from various sources: critical equipment sensors, enterprise resource planning (ERP), manufacturing execution system (MES), computerized maintenance management system (CMMS), and others. This makes the approach more autonomous, based mostly on self-learning and self-improvement, and known today as PdM 4.0.

According to ARC Advisory Group’s Enterprise Asset Management and Field Service Management Market Study, preventive maintenance proves ineffective as only 18% of assets have an age related failure pattern, while a full 82% of asset failures occur randomly.

No wonder that with the highlighted figures, the interest of developers, equipment manufacturers and leading providers of digital transformation consulting services in predictive maintenance will only be growing. 

Key Difference: IIoT Technologies Application

Then, if it is the way predictive maintenance treats big data analytics which makes it so effective and differentiates from preventive maintenance, how is the whole system built?

Basically, PdM is divided into 2 types regarding the way it analyzes the collected data:

  • Rule-based PdM. Pre-established rules set a threshold, and once data about assets collected by sensors indicates that this threshold has been reached, an alert is sent.
  • Machine learning PdM. Here, machine learning algorithms process large sets of historical data, thus running different scenarios to predict what will go wrong and when it will happen – and then generate alerts.

In both cases the architecture of this Internet of Things software development solution is the same:

Sensors are implemented in machines or physical products and gather data on their functioning and environment,. Elements chosen for measuring can vary: temperature levels, motor rotation speed, etc.

The central repository works as a data hub, which stores, processes and analyzes incoming data, and can be located either on premise or in the cloud. Business data from ERP, MES and other systems, together with manufacturing process flows, are integrated into the central data repository to provide context to the production asset data.

A flow of data between the monitored assets and the central data store is provided by the chosen communication protocol.

Then it is either the pre-defined rules or machine learning-based algorithms which run all the data multiple times and test them against the real world variables, thus generating insights in the form of dashboards and alerts.

  • If the system shows anything different from the desired behavior, an alert will be sent to manufacturers for further investigation and determination of the corrective action to be performed.
  • A dashboard provided for predictive analytics synthesizes operational data that allows process and maintenance engineers to address actionable insights.
IIoT Applications

This way PdM works at 2 levels of analytics:

  • Predictive analytics, which provides users with recognized patterns and generates insights in the form of dashboards and alerts;
  • Root case analysis, which allows maintenance and process engineers to investigate the insights and determine the corrective action to be performed.

The Value of Predictive Maintenance for Organizations

#1 Reduced Maintenance Time & Increased Efficiency

Installing smart sensors throughout one’s facilities and collecting data on the real state of the equipment allows for performing proactive repairs. Here the statistics come: PdM programs lead to a tenfold increase in ROI, a 25%-30% reduction in maintenance costs, a 70%-75% decrease in breakdowns and a 35%-45% reduction in downtime.

As an example, the largest US automaker General Motors has invested into the adoption of IBM Maximo across its 140 production sites globally, to support broad adoption of predictive maintenance strategies. This is how the tech leader is planning to reduce its total annual operating budget for maintenance, amounting to $1 billion today.

#2 Additional Revenue Streams

With Industrial IoT and data analytics working in tandem, and a predictive maintenance system deployed, organizations can monetize it by offering analytics-driven digital services for their customers. These include: PdM dashboards, optimized maintenance schedules, technician dispatch services performed before parts need replacement. 

One of the examples has been set by the Japanese industrial automation equipment maker FANUC. In cooperation with Cisco and Rockwell Automation, it developed a custom “near-zero downtime” predictive solution, which applies analytics to the operational data generated by FANUC’s manufacturing robots. Once a manufacturer’s customer shares the data with FANUC, the latter analyzes it in the cloud to predict potential problems and allow for its timely remedy.

#3 Operational Cost Optimization

Be it an equipment manufacturer, transportation company or oil & gas unit, each organization strives to minimize person-hours spent performing routine machine inspections, especially when they don’t lead to any key findings, trigger a work order or don’t solve the problems of unplanned downtime.

PdM can help with this onerous task: it automates machine inspection processes and reduce a chance of unpredicted equipment troubles.

#4 Competitive Advantage

Applying PdM can significantly strengthen a company’s branding. A defective device, especially if it is a mission-critical piece of equipment, once going down, can seriously undercut the manufacturer’s reputation. In this case, a smart predictive maintenance strategy can help the manufacturer differentiate himself in a competitive marketplace and reassure customers.

Though the notion of “predictive maintenance” has been around for a while, it is only now that the approach starts getting widespread. The reason is clear: high availability and low cost of digital technologies, coupled with the rise of the digital supply network and a number of specialists working with the technology, are making PdM affordable for facilities and businesses of all sizes.

Alex Makarevich
Alex Makarevich is Content Manager at R-Style Lab – a custom software development company with a business office in San Francisco, CA and dev center in Belarus, Europe. Having worked in the publishing house Éditions Techniques de l'Ingénieur (Paris, France) and got experienced in editing texts both in English and French, she has switched now to topics associated with IoT, web and mobile development.